Daniel Albuquerque, Jamie Lenney, Matthew McKernan, David Murakami, Peter Paz and Philip Schnattinger
Wholesale energy prices surged in 2021 and again in 2022, but UK inflation remained high well beyond the direct pass-through of higher energy costs. With energy prices moving sharply again following the recent conflict in the Middle East, could a similar scenario happen again? We use a structural macroeconomic model to show that the answer depends on labour market conditions. When workers are scarce, they are better able to recoup real income losses through higher wages, raising firms’ costs and broadening inflationary pressure. The model implies that the same energy shock is roughly twice as inflationary when vacancies are plentiful relative to job-seekers as when the labour market is loose.
The tight 2022 labour market
In 2022, UK vacancies were near high while unemployment was low. A useful way to summarise this is the Beveridge curve: the empirical relationship between the vacancy rate and the unemployment rate, shown in Chart 1. When the labour market is ‘tight’ (top left), firms find it hard to hire. When vacancies are low and unemployment high, the labour market is ‘loose’ or ‘slack’, workers find it hard to get jobs. During the 2022 episode (red triangle), the UK labour market was off the historical Beveridge-curve relationship: for any given unemployment rate, vacancies were far higher than past experience would predict. This contrasts with the position in 2025 Q3 which is much closer to the historical curve, suggesting the labour market has markedly loosened since the 2022 episode.
Chart 1: The UK Beveridge curve
Notes: The UK vacancy rate against unemployment rate with past episodes of large unemployment increases (1975, 1980, 1991, 2008 and 2020) traced out very different cycles. Arrows trace the labour market’s path over time; the black line is the fitted Beveridge curve.
Standard linear model solutions impose that the economy responds to a shock in the same way regardless of where it starts: a given energy shock adds the same amount to inflation whether unemployment is 3.5% or 8%. However, when the labour market is tight, firms face greater hiring pressures, production is closer to capacity and workers have more bargaining power, so an energy shock is more likely to translate into higher wages, costs and inflation. Our analysis asks whether the state of the labour market matters quantitatively, and finds that it does.
Why labour market tightness matters in our model
We use an open-economy New Keynesian model that integrates the search-and-matching labour market of Broer et al (2025) with the treatment of energy in Chan et al (2024). The broader macroeconomic structure is related to COMPASS, the Bank’s workhorse forecasting model described in Burgess et al (2013).
We solve the model in a way that allows for non-linearity. Specifically, we solve the model using higher-order perturbation rather than linearising it around a long-run average. We then use the time series generated by simulating the non-linear model to compute generalised impulse response functions conditional on the state of the labour market. This preserves three sources of non-linearity already present in our model’s search-and-matching foundations, rather than assuming them away via a linearisation:
- How full is employment? When firms are already running close to full capacity each new hire adds a bit less to output than the last, so a labour market tightening shock pushes costs and prices up more quickly. In the model this is captured by diminishing returns to labour in the production function.
- Matching is harder when slack is scarce. If unfilled jobs outnumber job-seekers, firms end up competing for the same workers. The result is that any change in firms’ demand for labour produces larger moves in vacancies, wages and how quickly people find jobs, that’s the steep part of the Beveridge curve. In the model this is captured by the curvature of the matching function, which makes vacancy-filling rates fall sharply as tightness rises.
- Workers’ bargaining position. When jobs are easy to find, workers are better placed to secure pay rises that keep up with the cost of living because employers competing for scarce staff are more likely to meet them. So an energy-driven jump in prices feeds through more strongly into wages, and from there into firms’ costs. In the model this is captured by the wage bargaining process, in which workers’ effective bargaining position strengthens as the labour market tightens.
All three sources of non-linearity pull in the same direction in our model: a tight labour market amplifies the inflationary effects of an energy shock. Their combined effect can be summarised in a stylised Phillips curve implied by the model:
Here πt is inflation, 𝔼tπt+1 expected future inflation, mctnonlab non-labour marginal costs such as energy, and xt labour market pressure on inflation. The key term is φ(Vt/Ut): the effect of labour market pressure on inflation depends on tightness, and φ′ > 0 means it strengthens as vacancies rise relative to unemployment. In other words, the Phillips curve is steeper when the labour market is tighter.
Previous studies find that the Phillips curve steepens in a tight labour market. These include Schmitt-Grohé and Uribe (2022) and Benigno and Eggertsson (2023) for the US. For the UK, this is consistent with machine-learning evidence of non-linear relationships between inflation and labour market tightness by Buckmann et al (2025) and the evidence in Gaffney et al (2026) using a threshold VAR. These papers establish the state dependence either empirically or in stylised settings. Our exercise is different in kind: because the state dependence arises inside a fully specified structural model solved non-linearly, we can decompose which mechanisms generate it, run controlled experiments that hold monetary policy fixed, and compute how the optimal interest rate response varies with the state of the labour market.
We also examine the effect of real wage resistance: an idea dating back to the 1970s energy crises that workers try to make up lost purchasing power after a burst of unexpected inflation, closely related to the mechanism outlined recently in Lorenzoni and Werning (2023). In response to a squeeze on living standards, people seek pay rises to keep their spending power in line with the cost of living. The economic question is not whether workers want to restore their real wages, but whether prevailing labour market conditions give them the bargaining power to do so. In our model, a worker’s bargaining power depends on labour market tightness, which links real wage resistance directly to the state of the labour market.
Energy shocks bite harder in tighter labour markets via a steeper Phillips curve
We trace the effects of a one-standard-deviation energy-price shock, equivalent to a 13.5% increase based on historical volatility, across tight and loose labour markets. This is smaller than the exceptional increases in 2021–22, and because our model is non-linear, our estimates are therefore conservative for shocks of that magnitude.
Our model simulations confirm that the shock is more inflationary when the labour market is tight: peak wage inflation is roughly twice as high as in the loose regime, and CPI inflation also rises more sharply. Chart 2 summarises this non-linearity through the slope of the price Phillips curve, holding the monetary policy response fixed across regimes. The differences are substantial. In tight states the simulated Phillips curve slope is roughly three times that in loose ones.
Chart 2: Price Philips curve non-linearity
Notes: The chart plots the slope of the reduced form price Phillips curve against the log vacancy to unemployment ratio. Each bin groups Monte Carlo draws by their initial tightness and computes the average price inflation response per unit of output response. The chart shows separate linear fits for the loose half of the distribution and the tight half of the distribution.
What does this mean for monetary policy?
The structural model allows us to go beyond describing the shock’s effects and ask what interest-rate response best stabilises the economy in different labour market states. We compute the model-implied optimal interest-rate path using a stylised loss function that penalises inflation, the output gap and changes in the policy rate. Chart 3 shows the results under alternative policy objectives.
Chart 3: Optimal Rate paths by labour market state in the baseline and wage resistance case
Notes: The chart shows the nominal Bank Rate paths that minimise the policymaker loss function 
following the energy price shock shown in Chart 2, for the tight, normal and loose labour market regimes. Panel (a) reports the baseline specification; Panel (b) reports the case with consumption wage resistance. In each panel, the solid lines are the usual OPP, which targets CPI inflation, and the dashed lines are the look-through variant, which instead targets domestically generated inflation. The counterfactual paths are constructed from local, model-based approximations to monetary transmission that are specific to each regime.
Three key results emerge:
- The model-implied optimal interest-rate response to the energy shock is earlier and larger when labour markets are tight. In panel (a) of Chart 3, the peak rate response starting from a tight labour market is roughly twice that starting from a normal one. In a loose labour market, rates rise only later as activity recovers.
- The case for ‘looking through’ an energy shock is stronger in a loose labour market than in a tight one. In loose states, much of the inflation increase proves temporary and returns to target largely on its own. In tight states, energy prices feed more strongly into wages and other goods inflation, requiring rates to remain high for longer.
- Real wage resistance amplifies inflation persistence in tight labour markets. In panel (b), stronger wage resistance makes the model-implied rate response materially larger and more persistent in a tight market, while making little difference in a loose one. Attempts to recover lost purchasing power translate into wage growth more strongly when workers have greater bargaining power.
Conclusion
The labour market changes how an energy shock affects inflation. In our model, the same shock generates roughly twice the wage inflation and warrants roughly twice the peak interest-rate response when the labour market starts tight rather than normal. The UK labour market has loosened since its 2022 peak, so this amplification mechanism appears less pronounced today than during the 2021–22 energy crisis. A renewed tightening of the labour market would steepen the Phillips curve and raise the inflationary effects of a fresh energy shock. These state-dependent effects are invisible to linearised models, supporting a role for structural models with non-linear labour markets for understanding inflation dynamics.
Daniel Albuquerque, Jamie Lenney and David Murakami work in the Bank’s Macro Model Development Division, Matthew McKernan and Philip Schnattinger work in the Bank’s Structural Economics Division, and Peter Paz works in the Bank’s Economic Outlook Division.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.