Homophones – Remarks by Huw Pill

Thanks to our hosts at the Edinburgh Chamber of Commerce for the opportunity to speak at the roundtable this evening.^[1]

As a member of the Bank of England’s Monetary Policy Committee, it is always a pleasure to engage with the business community here in Scotland: not only to celebrate your opportunities and successes, but also to inform our policy discussions through a better understanding of your experiences and concerns.^[2]

As we heard reverberate across the cities of North America this summer: “No Scotland, no party!”

In my remarks today, I will survey the impact of recent events in the Middle East on the outlook for monetary policy here in the UK. Drawing on staff analysis prepared for the MPC’s deliberations, I will explore the issue through three distinct (though homophonic^[3]) lenses.

Wait and see

The outbreak of hostilities in the Gulf a few months ago has added considerable additional uncertainty to an already uncertain outlook for the UK economy.

In our discussions within the MPC, we have distinguished two aspects of this uncertainty: (a) that directly associated with events in the Middle East and their implications for international energy prices (“sizing the shock”); and (b) that associated with the propagation of this external shock to UK consumer price inflation (in particular, the potential for “second-round effects” in price and cost dynamics).

As regards the former dimension of uncertainty, experience suggests that many of the crucial issues (which, in principle, we would like to understand better before calibrating the monetary policy stance) may simply be ‘unknowable’ in the current environment.

While a combination of political and technological constraints are likely to impose a ceiling and a floor on the level of global energy prices over time, within the wide range defined by those limits and at horizons relevant for monetary policy, the predictability of geo-political events and their implications for energy prices has proved to be low.

On this dimension, we are facing a more profound form of uncertainty – something closer to ‘Knightian uncertainty’ – than traditional risk.^[4] For me, it is hard to see this Knightian uncertainty disappearing in the foreseeable future. And it is not just the amplitude of the risk at any point that is uncertain, but also the duration of this uncertainty through time. Even if a new ceasefire were announced tomorrow, experience suggests we would be hard-pressed to assess its effectiveness, how long it might last and what would follow its expiry. Likewise for any re-escalation of the conflict.

Living with such Knightian (or ‘radical’) uncertainty is a normal part of monetary policy formulation.^[5] Of itself, it is neither a justification for inaction nor a reason for giving up on the monitoring of geo-political and energy price developments.

But recognising the Knightian character of these uncertainties both warns against conditioning monetary policy on the higher frequency movements in energy prices (which are likely to be volatile and noisy, rather than a meaningful monetary policy-relevant signal) and cautions against believing (still less signalling) that our understanding of oil prices will improve significantly at a meeting-to-meeting frequency. By nature, something that is ‘unknowable’ cannot be ‘learnable’.

The latter dimension of uncertainty considered by the MPC surrounds the propagation of energy price shocks through the UK economy in general, and to UK inflation in particular. As I suggested above, our discussion rightly focuses on the potentially persistent role of second-round effects in cost and price dynamics, recognising that the impact of direct and indirect effects is more mechanical, mostly unavoidable and largely beyond the influence of monetary policy.^[6]

By contrast with the geo-political dynamics driving events in the Middle East, these propagation mechanisms are – at least in principle – ‘learnable’. And we are learning about them: from the experience of 2022-23; from our forecast evaluation exercises;^[7] and from the insightful analytical work prepared by the Bank’s staff on structural and empirical modelling of inflation and the labour market that has been published in recent Monetary Policy Reports.^[8]

The challenge for the MPC on this dimension is that different members of the MPC are learning different things. This is neither a new story, nor a problem for the formulation of monetary policy.

My own position on this issue is well known and I think well understood. For some time, I have shown a higher-than-average-for-a-current-MPC-member concern about the potential for second-round effects – and thus greater intrinsic inflation persistence – in UK consumer price inflation.^[9]

I recognise this view does not command universal agreement, either within or outwith the MPC. But that is a feature rather than a flaw of our system: discussion across the diverse views held by individual MPC members should improve the robustness of the final collective decision; and making those differences of opinion transparent helps external stakeholders understand the character of MPC discussions and, on that basis, the outlook for Bank Rate.

Inside the MPC, a key issue has been the ‘state-contingency’ of second-round effects. I am broadly convinced by modelling work suggesting that second-round effects are likely to be more modest in an environment of labour market slack (such as we are facing now) than when the labour market is tight (as experienced in 2022-23).^[10] At the same time, I am not convinced that the existence of some labour market slack implies there will be no second-round effects at all: after all, our baseline models estimated on data from the ‘halcyon days of inflation targeting’ at the turn of the millennium themselves incorporate some modest second-round dynamics.

And there are also reasons to believe that second-round effects now will be stronger than estimated in those ‘halcyon days’. As I have argued in the past, there may have been permanent shifts in price setting behaviour in recent years (e.g., stemming from reduced contestability of UK goods and factor markets following Brexit). And even with regard to the state contingent character of second-round effects, arguments of increased attentiveness to nominal dynamics after the 2022-23 inflationary episode^[11] or greater salience of prospective rises in energy and food prices on household inflation expectations^[12] both suggest the risk of second-round effects on this occasion may be greater than on average over the inflation targeting regime.

Further complicating matters, the two dimensions of uncertainty recently considered by the MPC interact with one another: they are distinct but related phenomena. On my reading, the MPC broadly agrees that, other things equal, a stronger direct inflationary impulse from energy prices is likely to strengthen second-round effects. This is intuitive (and accounts for the correlation between the two forms of uncertainty traced out by scenarios A, B and C published in the MPC’s April Monetary Policy Report, as reflected in the diagonal shown in Chart 1).

But there is no such consensus within the MPC about whether second-rounds effects stronger than already embodied in our current forecasting models may emerge, even if the direct effect from energy is modest (a view I, for one, entertain and which would imply a greater likelihood of outturns in the more darkly-shaded ‘north-east corner’ of Chart 1).

I will return to how the MPC has used scenario analysis to explore and communicate the implications of the uncertainties set out in Chart 1 in a moment. But first I want to explore the response of monetary policy to these uncertainties.

In the face of such substantial and complex uncertainty, it might seem natural to adopt a ‘wait-and-see’ approach to setting Bank Rate.^[13] Let the uncertainties resolve themselves so that the MPC can form a reliable view about the level and forward path of energy prices and the strength of the direct, indirect and second-round effects on our target variable (UK CPI inflation), and only then set Bank Rate at a level that best serves the lasting achievement of the 2% inflation target.

But the problem with such a wait-and-see approach is that these uncertainties may not resolve themselves as quickly or definitively as we would hope, leading to a status quo bias in the setting of Bank Rate. In turn, such a bias could lead monetary policy to fall ‘behind-the-curve’ in addressing emerging inflationary risks.

After all (and I said during a talk in March in Washington, DC): if you follow a ‘wait-and-see’ approach and then do not ‘see’, all you have done is waited. And in that case, you may have waited too long.

There is ample reason to doubt that we will we see a definitive resolution of the multiple and profound uncertainties we currently face any time soon.

As regards events in the Middle East (and without wanting to repeat myself), the vagaries and complexities of geo-political and geo-economic decision making by all parties to the current conflict render the energy price outlook difficult to forecast – not just at near-horizons, but also for a period of uncertain and thus potentially prolonged duration.^[14] While we may now be able to rule out the extreme scenarios that were entertained by some at the onset of the conflict in early March, the ‘bounded Knightian’ – and thus profound and ‘radical’ – uncertainty on this dimension that I have already sketched out seems unlikely to dissipate any time soon. Events over the past few days bear this out.

Turning to second-round effects, we should also be wary of believing uncertainties surrounding their strength will be resolved quickly.

On the positive side, I am convinced that events in the Gulf have not triggered a de-anchoring of longer-term inflation expectations that would threaten to feed back into pricing behaviour today. Any such de-anchoring would have been seen quickly (say in market pricing of long-dated indexed swaps or longer-term survey measures of household and firm inflation expectations). The absence of such de-anchoring is reassuring – but this has never been my main concern. If we were to see suggestions of any such de-anchoring, the MPC would certainly act promptly and decisively to address it.^[15]

Rather my concerns around second-round effects are driven by the threat of more insidious ‘catch-up’ nominal dynamics in the UK, as different sectors (firms, households and government) respond to the substantial relative price shock associated with the increase in commodity prices (and its real impact, including the distributional implications). On my reading (and of course with the considerable benefit of hindsight), this behaviour is what was under-estimated in the 2022-23 inflation episode. And, thus far, such dynamics are still to be wholly embodied in the Bank’s modelling and forecasting machinery.^[16]

As I have already suggested, I remain concerned that changes to the structure of the UK economy stemming from Brexit (lower contestability of UK markets), the global financial crisis (lower trend productivity) and Covid (lower participation and productivity of young people in the labour market) render the UK more vulnerable to these effects than our standard models (estimated over a more benign period, the ‘halcyon-days-of-inflation-targeting’ in the early 2000s) suggest.

By their nature, these catch-up effects emerge more slowly than a one-off expectational shift, but – once established – are more likely to drive the self-sustaining intrinsic persistence of domestically-generated inflation that we should most fear – and which is likely to prove more costly to halt and reverse.^[17] In my view, it remains much too early to tell whether or not we are seeing a revival of these catch-up dynamics. In consequence, I do not draw much comfort from the relatively benign developments in prices, costs and wages (and shorter-term expectations thereof) thus far. This is a story that will play out over a longer horizon, as headline inflation rises, wages are re-negotiated and contracts are re-priced. We will not have definitive evidence any time soon.

And, of course, new uncertainties may emerge. There is a danger that our focus on challenges stemming from events in the Middle East may be to the detriment (or even exclusion) of analysis of other risks that ultimately prove bigger threats to the MPC’s price stability mandate: the potentially inflationary impact of El Niño on food and energy prices; the monetary and financial consequences of fiscal developments; and the implications of artificial intelligence (AI) on investment and productivity all spring to mind. How these risks play out will only be revealed with the passage of time.

Further complicating matters, in assessing the overall risk environment, interactions among these new uncertainties (and with those existing uncertainties originating from the Middle East) will need to be considered.^[18] Again, this not only adds complexity to our assessment, but also prolongs the wait for resolution of overall uncertainty.

In my view, in this environment we cannot wait for uncertainties to resolve themselves before acting. It is now six months since the onset of conflict in the Middle East. How or when the conflict will be resolved – and, more importantly, the magnitude of its implications for UK inflation – remain unclear: essentially as unclear as they were six months ago.

Given all this, I am uncomfortable with a ‘wait-and-see’ framing of the MPC’s current decisions over Bank Rate.

In line with its rhetoric and decisions thus far, the Committee should be making an active choice about Bank Rate; one that establishes a monetary policy stance appropriate to achieve the 2% inflation target in the face of all the challenges it currently faces, including the uncertainties over the size and propagation of the inflationary shock emanating from events in the Gulf. This last point is crucial: the appropriate policy response needs to reflect the magnitude and character of the uncertainty the MPC is confronting, not just our ‘best guess’ of how events may turn out.^[19]

That leaves open the question of how to act. Since March, the MPC has decided to keep Bank Rate unchanged at 3.75%. The majority of the Committee have judged that level of Bank Rate (together with the tightening in broader financial conditions, including the emergence of an upward sloping money market yield curve over the next year) as sufficient to contain inflationary risks coming from higher energy prices.

As is reflected in my votes at recent MPC meetings (and for the reasons that I have already discussed), my preference has been for a modestly higher level of Bank Rate. But my point here is not to argue for 25 basis points higher or lower. Such differences reflect healthy diversity in the balance of risks to achievement of the inflation target.

Rather my intention is to emphasise that at each of its meetings the MPC will need to make an active choice about the stance of monetary policy.

A danger of keeping Bank Rate unchanged is that it will be seen as reflecting a status quo bias in decision making, which itself influences market expectations and – given the self-referential, ‘hall of mirrors’ relationship between monetary policy makers and financial markets – then feeds back into influencing policy decisions themselves. Worse, a decision to keep Bank Rate unchanged can lapse into the perception there is a preference for holding rates unchanged in the hope that events will work out in a relatively benign way that justifies such inaction. Such perceptions may delay pre-emptive action that reduces the likelihood that self-sustaining and persistent inflationary dynamics emerge.

At each of its meetings, the MPC must strive to argue that its choice of Bank Rate (and any associated communications) best serves lasting achievement of the 2% inflation target. In doing so, the Committee needs to explain why that choice addresses the challenges to its price stability market thrown up by circumstances, including the impact of uncertainty. This is the essence of making an ‘active choice’ as distinct from adopting a ‘wait-and-see’ approach. Emphasising the active nature of that policy decision guards against the risk that holding rates unchanged is viewed as signalling a bias to the status quo.

This is what the MPC needs to do. And I am sure the Committee will deliver.

Weight and see

If the MPC is to take its Bank rate decisions in this ‘active’ and forward-looking manner, it needs tools to assess the flow of data so as to draw implications for the evolution of risks to lasting achievement of the 2% inflation target.

Historically, the MPC has placed significant weight on its inflation forecast (and the associated fan chart) in assessing and communicating risks to the inflation target. This approach worked well in the more benign global environment seen at the outset of the UK’s inflation targeting regime. But in recent years – as the magnitude of external shocks has increased and their character has shifted to the supply side, creating the attendant more difficult trade-off environment for monetary policy makers^[20] – this framework has been found wanting. Both experience and its codification in Prof. Ben Bernanke’s review of MPC forecasts and communication^[21] suggest that a more robust approach – drawing on a richer and more diverse set of tools, models and scenarios beyond the MPC’s central forecast – would help support better monetary policy decisions.

By nature, a diverse set of tools is more voluminous than a single forecast, so – in the interests of time – I won’t attempt to summarise all staff initiatives in pursuit of the Bernanke recommendations here. (Many are illustrated in the Boxes published in recent Monetary Policy Reports.) Rather I will focus on one new empirical model that seeks to draw a signal about the underlying inflation dynamics most relevant for monetary policy decisions by re-weighting the components of the UK’s consumer price index.^[22]

Re-weighting CPI components is not a new approach to analysing inflation. Core inflation measures exclude components that historically have been seen as noisy (e.g. CPI excluding energy and/or food prices) or those that have moved significantly of late (e.g. trimmed means or inflation measures excluding outliers). Statistical filters can assign higher weight to those components that tend to be more associated with the overall trend of inflation. And the MPC has in recent years focused on components (e.g. services prices) that are seen as more associated with domestic inflation dynamics rather than external shocks.^[23]

In the novel approach discussed here, higher weight is given to CPI components that either have proved to be ‘stickier’ in the past and/or derive from sectors that are most central to the economy’s production network. Reweighting the CPI in this way yields a measure of inflation that should better reflect real marginal cost pressure on prices, thus capturing the underlying dynamics of the inflation process as envisaged by a standard new Keynesian macroeconomic model.^[24]

As regards placing higher weight on stickier prices,^[25] this model recognises that inflation in stickier-price sectors brings welfare losses, as firms in those sectors are constrained in their ability to adjust prices in response to shocks. By contrast, inflation in flexible-price sectors is more likely to reflect efficient changes in relative prices, which it would be natural for monetary policy to ‘look through’. Weighting the former more than the latter creates a measure of underlying inflation that better captures the real costs and distortions associated with inflation, and thereby gives content to those aspects of the MPC’s remit that demand efforts to avoid undesirable volatility in output and employment.^[26]

Turning to placing higher weights on ‘network-central’ prices, this approach recognises that some prices take on a particularly important role because the associated goods or services are important inputs across sectors and / or are consumed by both other businesses and consumers. Energy prices are a prominent example. Network-central prices tend to constitute input costs to many other (often downstream) firms. Consequently, whenever network central prices rise, marginal costs tend to rise across multiple sectors.

Bank staff have constructed a measure of underlying UK consumer price inflation on the basis of these considerations (which has been labelled ‘sticky-central inflation’). This is shown against the conventional headline measure in Chart 2.^[27]

The underlying measure broadly tracks the headline measure, but – in particular, in the face of the significant rise in international energy prices following the Russian invasion of Ukraine in early 2022 – does not show the same amplitude of variation. The difference between the two series can be thought of as a measure of how much of the rise in headline inflation the MPC can safely ‘look through’.

Reassuringly, the underlying inflation measure shown in Chart 2 demonstrates disinflation towards the 2% target following the 2022-23 inflationary episode, albeit with some stalling from mid-2024 onwards. Moreover, the most recent data show a healthy step down towards target, even as the headline measure remains above^[28]. This is consistent with a large portion of the rise in headline inflation stemming from events in the Gulf seen thus far can be ‘looked through’ by the MPC.

Admittedly more speculative forward-looking analysis unfortunately makes for less comfortable reading. Chart 3 shows the BVAR median projection for underlying ‘sticky-central’ inflation under the assumption Bank Rate remains unchanged at 3.75%, comparing this with an equivalent forecast for headline CPI inflation. While the latter falls back towards target in 2027, the underlying measure remains stuck meaningfully above 2%, with the risks around this projection being clearly to the upside.^[29] Such analysis gives reason for caution in sounding the all clear as regards lasting achievement of the inflation target, in line with my recent policy votes.

Moreover, Chart 3 offers a timely reminder the MPC should not be seduced by falling headline inflation rates as the impact of external energy price shocks unwinds. As we saw in the first half of 2024, in this context substantial falls in headline inflation may coincide with a stalling (or even rise) in underlying inflation.

More generally, this analysis suggests that the MPC cannot afford to wait-and-see whether underlying inflationary pressures will moderate. Rather it needs to analyse the data flow carefully – perhaps by adopting a novel weighting of CPI components in its internal analysis and discussion – in order to see how those underlying dynamics are evolving in real time, and then respond with Bank Rate to the resulting assessment.

Weight and C

If the previous discussion has given some insight into new tools that are being used in real time to assess the underlying dynamics of inflation internally in the MPC, the presentation of the rationale for monetary policy to external stakeholders nonetheless remains important.

Monetary policy communication cannot be an independent instrument of policy design: if words do not match deeds, they will cease to be credible. But effective communication can shape private expectations in ways that strengthen monetary policy transmission and support the achievement of the 2% inflation target. Given the environment at present, a key question for the MPC is how it should communicate about and in the face of significant uncertainty.

Scenarios are one way to explore the implications of the many current uncertainties for monetary policy. The Bank staff’s response to the Bernanke review has created a foundation upon which to develop and present scenarios, building on a long history involving the use of scenarios for internal analysis. Nevertheless, using scenario analysis comes with pros and cons – the MPC needs to magnify the former while managing the latter.

One thing scenario analysis does permit (in a way that fan charts around a baseline do not) is the ability to distinguish across different types of uncertainty. This may be particularly important in the current environment. As I have argued above, Gulf-related uncertainties have a ‘bounded Knightian’ character at present. It is difficult (if not impossible) to assign statistical probabilities to various outcomes.

This is something that scenario analysis can help to reveal and assess, in a way that being placed in the upper half of a fan chart (which derives from statistical probabilities) only obscures. In developing scenarios that span both ‘Knightian’ and ‘probabilistic’ uncertainty, the MPC is thus adopting a novel and ambitious approach.

Scenarios can also address financial market participants’ understandable demand for more information about how the MPC is likely to respond to energy price shocks. This represents a ‘what if’ approach, giving guidance as to how the inflation outlook (and, conditionally, monetary policy decisions) would differ if energy prices entered an adverse or severe trajectory. Clarifying this by publishing a scenario ahead of the event implies that markets will price the implications of new shocks into interest rates and asset prices in a way that supports the MPC’s actions in response to the shock, thereby reducing noise and volatility, and strengthening policy transmission.

But there are also risks with scenario analysis.

The danger of favouring centrifugal over centripetal forces in external communication has been much discussed of late.^[30] In the MPC context, this danger is amplified by the individual accountability for votes over Bank Rate embodied in the legislation governing monetary policy. For me, the answer here is for the MPC to put greater emphasis and weight on its collective communications (such as the Monetary Policy Summary published with the MPC minutes), recognising more explicitly that these collective vehicles represent the view of the Committee as an entity rather than an average of individual views (whether of the whole of the Committee or any subset). Credibly establishing this role for collective communication would help external stakeholders understand the core message offered by the MPC as a whole.

Against this background (and even with the advent of individual paragraphs in the MPC minutes summarising the view of each member), I do not want to have ‘my own scenario’ published in the Monetary Policy Report or elsewhere. Rather I see staff-designed scenarios a defining a common MPC space within which I can express my personal views in a way consistent and comparable with those of other Committee members, allowing the differences and similarities to be identified and evaluated by our various external stakeholders. I have found the current matrix framing risks from geo-politics / energy and second-round effects (captured in Chart 1) helpful in that regard.

The choice of scenarios will also shape market perceptions of where the Committee stands. Of itself, this is neither a pro nor a con: the issue at hand is whether market expectations are credibly directed in a way the serves the transmission of MPC decisions in pursuit of the inflation target.

That issue is particularly important at present. One aspect of recent MPC discussions has been whether the Committee, rather than raising Bank Rate, can rely on developments in a wide set of asset prices in general – and the ‘premia’ in the front-end of the OIS curve in particular – to tighten financial conditions sufficiently to address the inflationary consequences of conflict in the Middle East. Behind this lies two further questions: (1) whether the ‘premia’^[31] creating the upward slope of the money market yield curve simply represent market expectations of higher Bank Rate or something else; and (2) whether the cause of that upward slope matters for monetary policy transmission.^[32]

These are difficult issues, which I won’t resolve today. I have already mentioned the complex ‘hall-of-mirrors’ character of interactions between monetary policy makers and financial markets that governs how asset prices respond to decisions on Bank Rate and communication on how the MPC views the potential state and evolution of the economy.

But there is one important point I will emphasise: in the face of such complexity, the MPC cannot simply treat any premia in the as a ‘fact-of-life’ with which it has to deal. At least in part, such premia are creatures created by the MPC’s communication itself.

To explore this assertion, it is useful to revisit MPC communication around its April decision to leave Bank Rate unchanged. In the accompanying Monetary Policy Report, the MPC chose to publish three scenarios (labelled – without much creativity – A,B and C), while not identifying a baseline or central case. The main features of the three scenarios are described in Chart 1: scenario A involved energy prices following the path implied by futures prices, with no second-round effects; scenario B had a similar peak for energy prices as implied by futures, but greater longevity of that peak and stronger second round effects; while scenario C had much stronger energy price rises and a temporary de-anchoring of longer-term inflation expectations.

In a recent Bank Insights note,^[33] staff have used ‘scenario synthesis’ techniques^[34] to explore how the potential Bank Rate paths implied by these three scenarios compare with a financial market-implied reference distribution for Bank Rate. More specifically, the article explored how well the Bank Rate paths derived from model-based optimal policy projections associated with each of the three scenarios, spanned the one-year-ahead distribution of Bank Rate embodied in options prices. The results are illustrated in Chart 4.

Two results emerge from this exercise.

First, in this ‘implied Bank Rate one-year ahead’ space, scenarios A and B are essentially indistinguishable. Both assign a high weight to leaving Bank Rate unchanged.

Second, in order to span the forward Bank Rate distribution implied by markets, the more ‘hawkish’ scenario C is required. It needs to be given a meaningful (although not modal) role. More importantly, the resulting scenario synthesis distribution embodying all three of the published scenarios is bi-modal – it has two peaks: one implying little need to raise rates from 3.75%; the other suggesting that over the next year Bank Rate will need to rise substantially towards the around 5% level last seen in the aftermath of the Russian invasion of Ukraine and consequent rise of headline inflation into double digits.

How the latter result is interpreted by markets becomes an important aspect of the MPC’s overall communication. Whether implicitly (via their internal thinking or analysis) or explicitly (via the material published in the Bank Insights article), it is natural for market participants to interpret this set of scenarios as suggesting the MPC is seeking to keep rates on hold but would raise rates aggressively if the inflation outlook were to deteriorate significantly. In other words, the bi-modal distribution points to two distinct ‘parallel universes’: one where Bank Rate is on hold in the hope that inflationary pressures and risks diminish largely of their own accord; and another where inflation jumps substantially prompting the MPC to raise Bank Rate quickly.

Such an interpretation naturally points to market pricing a base case with Bank Rate unchanged,^[35] while any response to the inflationary impulse is pushed into a ‘premium’ that steepens the money market curve. Such an interpretation might bolster any perception of status quo bias in Bank Rate, such as that which is inherent in a wait-and-see strategy as I argued above.

Recognise that the configuration of modal Bank Rate expectations and slope of the forward OIS curve priced by the market are – at least in part – as much a creation of the MPC’s communication as something that the MPC has to respond to. Choosing one set of scenarios rather than another will influence market expectations and thus the transmission of monetary policy. Scenario choices should not be made lightly: in current circumstances, the design of an upside scenario like C will influence the weight placed upon it and thus how markets price forward Bank Rate.

Repeating this exercise using the scenarios published in the MPC’s July Monetary Policy Report gives a quite different picture (as shown in Chart 5). On this occasion, a baseline central projection was re-introduced, with both upside and downside inflation scenarios drawn around it. Using the same scenario synthesis techniques, matching the forward Bank Rate distribution embodied in market option prices with the model-based projections for Bank Rate in each of the three published scenarios results in a single-peaked distribution of forward Bank Rate in the scenario synthesis, with a higher mean and mode. Arguably, this is a distribution that would support an upward shift in Bank Rate expectations and a lower risk premium being embodied in the money market curve: it is more consistent with achieving some tightening of financial conditions by raising Bank Rate modestly, rather than relying on the steepness of the money market curve.

One might argue that if the same overall broad financial conditions are engineered by the MPC’s choices across Bank rate and communication, then the balance between the two drivers is not relevant. But there may be reasons to favour one approach relative to the other.

While we can continue for a while under the assumption that markets ‘have done our job for us’ by repricing the short-end following the energy price shock, when those market expectations depart from an unchanged path for Bank Rate (as they do at the time of writing), at some point we will unavoidably face the choice of validating the forward rate curve (by hiking Bank Rate) or acquiescing in the curve shifting downward (by continuing to hold Bank Rate).

There is no free lunch. And the danger exists that the market may lose confidence that expected Bank Rate increases will be implemented just at the time as it starts to entertain doubts that the MPC ability or willingness to bring inflation down to target. Indeed, one would expect that to be the case. So, there is potential ‘wrong-way risk’ here: in this setting, the market will ease financial conditions just when the MPC needs them to tighten.

Just as we the MPC was wary of allowing markets to ‘get-ahead-of-themselves’ with respect to prospective Bank Rate hikes in the immediate aftermath of hostilities breaking out in the Gulf, we should also recognise that markets may have already ‘gotten-ahead-of-themselves’ with respect to prospective Bank Rate cuts before the energy price shock.

At a time of profound uncertainty, it is important that the MPC offers an anchor for the short end of the market curve, rather than gives the impression that it is chasing market expectations. As reflected in my vote in recent policy rounds, I see benefit in acting clearly, promptly and decisively with Bank Rate. This would cut through the noise inherent in the current uncertain environment in a way that bolsters the clarity and effectiveness of our policy choices.

This is preferable to relying on the evolution of risk premia in the money market curve – over which our control is certainly incomplete – to contain inflationary pressures. Communicating that the MPC is prepared to act in relatively extreme circumstances (such as April’s scenario C) runs the risk that the market eventually concludes that it will only act in such circumstances: in that context, we hazard that tough conditional messages on Bank Rate end up being treated as ‘cheap talk’.

Yes sir, I can boogie … but I need a certain song

My policy assessment starts with the question:^[36] Recognising the profound geo-political and economic uncertainty we currently face, what level of Bank Rate should the MPC establish now to put itself in the best position to manage the significant and (on balance) self-evidently upside risks to price stability as they unfold in the future?

As reflected in my vote at the July MPC meeting (and indeed prior meetings), my own response to this question has pointed to a need to raise Bank Rate to 4% on the grounds:

(a) this represents a clear and unambiguous signal of the MPC’s willingness and ability to address upside risks stemming from events in the Middle East, cutting through noise in commodity and asset price developments and better anchoring our reaction function;

(b) it recognises that ‘fine-tuning’ the response in the face of ‘bounded Knightian’ uncertainty about energy prices is problematic (especially if that fine tuning extends to addressing volatility in activity and/or employment in a ‘trade-off management’ setting); and

(c) it reflects my ‘greater-than-MPC-median-voter’ concerns about the UK’s vulnerability to ‘catch-up’ second round effects, in turn reflecting a deterioration in the supply side over the past two decades.

Raising Bank Rate on this basis need not be the start of a prolonged and aggressive series of increases. Indeed, implemented and communicated effectively, a prompt increase in Bank Rate may serve to head-off some of the potential insidious ‘catch up’ nominal dynamics that threaten to make temporary departures of inflation from target more persistent.

Communication is key to managing how a decision to hike influences household and firm behaviour, as well as market expectations of future policy moves. In spelling out its analytical framework, MPC communication influences how markets will expect future policy decision to be shaped by changes in circumstances as the economy evolves and the current substantial uncertainties slowly resolve themselves – albeit possibly only to be replaced by new uncertainties.

In this context, MPC communication must ensure that our published scenarios bolster the signals we wish to offer about the policy outlook. The Committee should be cautious about using relatively extreme ‘what if’ scenarios to explain its analytical framework for policy making, for fear of these being dismissed as ‘cheap talk’.

While the MPC should not be steering forward Bank Rate through (now somewhat discredited, early 2010s-style) ‘forward guidance’,^[37] the Committee should anchor the money market yield curve around a systematic approach to monetary policy (‘reaction function’) that (1) accords high weight to forward underlying inflation developments^[38] and (2) recognises the uncertainties surrounding our assessment of the output gap and equilibrium real interest rate.

Acting clearly, promptly and decisively with Bank Rate while pointing towards a ‘simple robust policy rule’ of the type sketched out above would cut through the noise inherent in our uncertain environment in a way that bolsters the clarity and effectiveness of monetary policy choices.

References

Abiry, R., J. Hurley, P. Labonne, D. Latto, H. Li, A. Moreira, J. Oyegoke and S. Singh (2026). “Learning from forecast errors: The Bank’s enhanced approach to forecast evaluation,” Bank of England macro technical paper no. 6.

Adrian, T., D. Giannone, M. Luciani and M. West (2025). “Scenario synthesis and macroeconomic risk,” Federal Reserve economics and finance discussion paper no. 2025-036.

Aikman, D. (2026). “Insights from the recent monetary policy roundtable,” NIESR blog, 12 January.

Anesti, N., V. Esady and M. Naylor (2025). “Food prices matter most: Sensitive household inflation expectations,” Bank of England staff working paper no. 1125.

Aoki, K. (2001). “Optimal monetary policy responses to relative-price changes,” Journal of Monetary Economics 48(1), pp. 55-80.

Bailey, A. (2026). “Remaining anchored: Monetary Policy in an unpredictable world,” speech at the Reykjavík Economic Conference, 29 May.

Bank of England (2024). “Response of the Bank of England to the Bernanke review,” Bank of England, 12 April.

Bernanke, B.S. (2024). “Forecasting for monetary policy making and communication at the Bank of England: A review,” Bank of England, 12 April.

Blinder, A.S. (1998). Central banking in theory and practice, Lionel Robbins lectures, MIT Press.

Brainard, W.C. (1967). “Uncertainty and the effectiveness of policy,” American Economic Review 57(2), pp. 411-425.

Brignone, D., S. Goel, S. Lloyd, G. Mantoan, N. Raviraj and A. Renzetti (2026). “Making scenarios add up: spanning risks with scenario synthesis,” Bank of England Insights, 16 July.

Broadbent, B. (2023). “Signal versus noise,” speech at the London Business School, 18 December.

Dhingra, S. and J. Page (2023). “Accounting for imported and domestically generated inflation: Supply chains, monetary policy, and the UK’s cost of living crisis,” VoxEU, 25 May.

Gaffney, A., K. Petrova, G. Potjagailo and A. Sisko (2026). “Inflation thresholds and oil shock transmission in the UK: A self-exciting threshold VAR approach,”. Bank of England macro technical paper no. 8.

Greene, M. (2025). “The supply side demands more attention,” speech at the Adam Smith Business School, University of Glasgow, 24 September.

Kanngiesser, D. and T. Willems (2024). “Forecast accuracy and efficiency at the Bank of England – and how errors can be leveraged to do better,” Bank of England staff working paper no. 1078.

Kay, J. and M.A. King (2020). Radical uncertainty: Decision-making for an unknowable future, Bridge Street Press.

Knight, F.H. (1921). Risk, uncertainty, and profit, Houghton Mifflin.

Lengyel, A., and D. Walker (2026). “Bank Rate expectations in the UK curve following the war in Iran,” Bank of England Insights, 17 July.

Lorenzoni, G., and I. Werning (2023). “Inflation is conflict,” NBER working paper no. 31099.

Pill, H. (2023). “Inflation persistence and monetary policy,” International Centre for Money and Banking Lecture, Graduate Institute, Geneva, 4 April.

Pill, H. (2025). “The courage not to act,” remarks on the monetary policy outlook at a briefing hosted by Barclays, 20 May.

Pill, H. (2026). “Robustness,” speech at the National Bank of the Republic of North Macedonia and SUERF conference – Central Banking Amid Persistent Global Shifts: Fostering Stability, Innovation and Resilience, 24 March.

Rubbo, E. (2023). “Networks, Phillips curves and monetary policy,” Econometrica 91, pp. 1417-1455

Söderström, U. (2002). “Monetary policy with uncertain parameters,” Scandinavian Journal of Economics 104(1), pp. 125–145.

Yotzov, I., N. Bloom, P. Bunn, P. Mizen and G. Thwaites (2025). “The speed of firm response to inflation,” Bank of England staff working paper no. 1085.

Notes

1. The views expressed in these remarks are not necessarily those of the Bank of England or its Monetary Policy Committee. I would particularly like to thank Kavya Saxena, Lou Everett and Claire Willis for their help in the preparation of these remarks. We draw heavily on the analysis done by Davide Brignone, Satyam Goel, Simon Lloyd, Giulia Mantoan, Tim Munday, Joseph Oyegoke, Adrian Paul, Nades Raviraj, Andrea Renzetti and Tim Willems. I also thank Nick Bate, Alan Castle, Becky Maule, Alice Pugh, Amar Radia, and Ryland Thomas for their comments on earlier drafts. Opinions (and all remaining errors, whether of commission or omission) are my own. ↑

2. Thanks to the Bank of England’s Agents based in Scotland – Will Dowson and team – for their facilitation of the meetings during this regional visit. ↑

3. Homophones are words that sound the same when you say them, but they have different meanings. They might be spelled the same way or spelled differently. ↑

4. In other words, we are confronting a situation where we cannot assign statistical probabilities to future outcomes. ↑

5. See Kay and King (2025) who discuss the role played by ‘radical uncertainty’ in shaping economic outcomes and framing economic policy choices. ↑

6. This interpretation takes as it starting point that the transmission lags in the monetary policy transmission from Bank rate changes to impact on consumer prices are “long and variable”: specifically, that they are longer than the lags from the incidence of external energy price shocks to inflation. This has been challenged by recent empirical results which point to an expectational channel of monetary policy transmission (including effects through exchange rates). Such expectational channels may operate more quickly, making the decision to focus on second-round effects on inflation rather than direct and indirect effect more one of choice rather than necessity. ↑

7. See Abiry, et al. (2026) and Kanngiesser and Willems (2024). ↑

8. For example, see Box A: A framework for monitoring second-round effects of higher energy prices and Box B: Evidence on second-round inflation effects from the recent energy price rises in the July Monetary Policy Report or Box B: How might firms respond to higher energy prices? and Box G: How should monetary policy respond to an energy price shock? in the April Monetary Policy Report. ↑

9. In the interests of time, I will not rehearse all the arguments behind this view again here. For an exposition of this view, see Pill (2025). ↑

10. See the staff analysis presented in Box C: How will prevailing economic conditions affect the impact of the energy shock on inflation? in the April Monetary Policy Report. ↑

11. See Yotzov, et al. (2024) and Gaffney, et al. (2026). ↑

12. See Anesti, et al. (2025). ↑

13. Indeed, the well-known Brainard (1967) uncertainty principle would point in this direction. Former Federal Reserve Vice-Chair Alan Blinder (1998) famously stated: “The Brainard result was never far from my mind when I occupied the Vice Chairman's office at the Federal Reserve. In my view…a little stodginess at the central bank is entirely appropriate.” However, more recent research (e.g. Soderstrom, 2002) has demonstrated that the nature of uncertainty matters when assessing the policy implications (something I took up in Pill (2026)). The Brainard result is not as generalisable as the Blinder quotation suggests. ↑

14. Of course, market prices can give some guide to the outlook for energy prices. Indeed, the MPC’s baseline macroeconomic projection is conditioned on the path implied by energy price futures contracts. But market participants face the same uncertainties as we do. In this context, we should be wary of over-interpreting an energy futures curve that reflects financial conjectures at least as much as physical reality. By the same token, we should be concerned about (re-)pricing (in either direction) of the sterling OIS curve driven by short-term volatility (‘noise’) in oil or gas prices in this environment. As a result, in my view we should be ‘looking-through’ the considerable day-to-day volatility (in energy price futures and OIS curves) to focus on underlying fundamentals – including the character of the uncertainty surrounding energy prices. Cutting through the noise would improve the clarity of our policy signal and the effectiveness of policy transmission. ↑

15. While the experience of the 1970s shows that monetary policy makers cannot take their credibility for granted, the main success of our inflation targeting regime has been to convince markets that ultimately inflation will return to the target level. That is a substantial achievement, to be jealously guarded. ↑

16. See Lorenzoni and Werning (2023) for a discussion of such dynamics. ↑

17. See Pill (2023) for a definition of ‘intrinsic’ inflation persistence and the importance of its distinction from ‘extrinsic’ inflation persistence in determining monetary policy responses. ↑

18. In other words – and to the extent that it is possible to assign statistical probabilities to uncertain outcomes in an environment of such profound uncertainty – it is the joint distribution of all risks rather than the marginal distribution of each risk that matters for the formulation of monetary policy decisions. ↑

19. This conclusion is at odds with the ‘certainty equivalence’ of the optimal monetary rule in many academic treatments. This principle states that the optimal response of policy does not depend on the magnitude of ex ante variance of economic shocks. But – as discussed in Pill (2026) – this principle only holds in linear quadratic model specifications and is not robust to alternative (and more realistic) stylised characterisations of the monetary policy problem. ↑

20. See Greene (2025). ↑

21. See Bernanke (2024) and Bank of England (2024). ↑

22. Just to be clear from the start, this is not to suggest a re-definition of CPI inflation or a change to the design of the current 2% inflation target (which are determined by the ONS and the Government respectively). Rather it is an exploration of a novel tool that can support policy decisions in pursuit of the existing target. Note also that this measure is not intended to be a pure measure of domestically generated inflation. ↑

23. By the same token, the MPC has placed emphasis on domestic nominal indicators beyond CPI components, such as PPI components (Dhingra and Page, 2023) or wages and pay (Broadbent, 2023). ↑

24. See Rubbo (2023). ↑

25. And in the spirit of Aoki (2001). ↑

26. By implication, it avoids reliance on imperfect measures of the output gap, an empirically elusive concept. ↑

27. The staff exercise uses the COICOP 4 classification, which distinguishes between 85 different subcomponents of CPI. To calculate which subcomponents of headline CPI are the stickiest, we use the microdata. For every item in the CPI for which data are available, we filter individual price quotes to ensure that the same item has been recorded in consecutive months in the same retail outlet. We then calculate the probability that a given item changes price over the course of a year, and we aggregate those probabilities up at the COICOP 4 class level. To determine which subcomponents of headline CPI are the most central in the UK’s production network, we use the ONS input-output tables. We map CPA sectors to COICOP definitions to construct appropriate weights for each COICOP 4 class. These so-called ‘Domar weights’ represent the sales share of each sector in overall GDP. For further discussion, see Box C: “A granular measure of underlying inflationary pressures,” Monetary Policy Report, July 2026. ↑

28. The ‘sticky-central’ measure may be relatively more affected by certain CPI components than headline measure. ↑

29. The risk analysis represented by the asymmetric 90% confidence intervals around the median forecasts shown in Chart 3 is derived from simulations of the BVAR model. Note also that the ‘sticky-central’ measure of underlying inflation is normalised such that a rate of 2% is consistent with headline CPI inflation at the 2% target. (This normalisation was not implemented in the material published in the July Monetary Policy Report.) ↑

30. See, for example, some of the comments reported in Aikman (2026). ↑

31. Here I am using the word ‘premia’ in a loose sense, namely as a label for the gap between survey-based measures of the modal Bank Rate outlook (from the Banks’ MaPS survey, which at present point to levels unchanged at 3.75%) and the pricing of forward Bank Rate in the OIS curve. As such, the ‘premia’ in this context incorporate both first and second moment effects of the Bank Rate outlook. ↑

32. See Lengyel and Walker (2026) for an initial empirical discussion of these issues. ↑

33. See Brignone, et al. (2026). ↑

34. See Adrian, et al. (2025). ↑

35. As, for example, reflected in recent MaPS surveys, which report that market participants have foreseen unchanged Bank Rate at 3.75% as the most likely outcome even while the money market curve is upward sloping. ↑

36. Note that this is a different question to asking what level of Bank Rate is appropriate in the ‘worst-of-all-worlds’ that we might envisage as of today, i.e. a so-called minmax approach often associated with the application of robust control techniques to the monetary policy problem, as discussed in Pill (2026). Nor is it the question of what level and path of Bank Rate is ‘optimal’ (on the basis of some welfare criterion) in a specific model and setting, tested to ensure it does not lead to problematic results in alternative settings (which is one characterisation of how scenarios might be used around the MPC’s central projection). ↑

37. Which may have been appropriate when easing monetary policy via cuts in Bank Rate was constrained by its effective lower bound, but which has proved difficult to implement effectively over the past decade. ↑

38. In other words, an approach with a low ‘state-contingent lambda’ in the framework employed by the Governor in his recent Reykjavík speech (Bailey, 2026). ↑