why expectations matter for inflation – Bank Underground

Boromeus Wanengkirtyo, Ivan Yotzov and Mishel Ghassibe

Can tomorrow’s costs affect firm prices today? When a temporary tariff schedule on imported inputs was announced in March 2019, many UK firms adjusted prices in anticipation – despite the potential cost change being in the future. In a recent working paper, we use firm‑level survey data to estimate ‘intertemporal pass‑through’ (IPT): how much expected future marginal costs move current prices. Consistent with modern macroeconomic theory, we find big differences across firms: those that change prices less often, and expect the shock sooner, responded the most. A model shows this variation across firms makes aggregate inflation more forward‑looking, so announcements of future policies can move inflation today.

Methodology

To construct exogenous variation in firms’ expected future marginal costs, we use the announcement in March 2019 of a temporary tariff schedule in the event of a ‘No-Deal’ Brexit. This implied that in the event of no free trade deal with the European Union (EU), UK firms that import inputs from the EU would unilaterally face tariffs that are substantially lower than the likely alternative Most Favoured Nation (MFN) rates. As such, this is a precise and narrow ‘news shock’ – not an evaluation of Brexit’s inflation effects. Since the proposed reductions differ across products, the announcement generates sectoral variation in future effective tariffs, depending on the composition of intermediate inputs. We further account for the indirect effects of the announcement along the supply chain, using an input-output table. Crucially, since the tariffs were lower than the MFN tariffs which would otherwise have taken effect (in the case of a no-deal Brexit), the announcement was a downward news cost shock. In Chart 1 we present the distribution of the news shock across sectors of the economy, after applying the input-output table adjustment. There is substantial heterogeneity, with the largest effects in the accommodation and food and manufacturing sectors. The average news shock across sectors is -4.1%.

Chart 1: Sectoral average effective tariff news shocks

We combine this sectoral variation with firm-level data from the Decision Maker Panel (DMP) Survey on (i) the perceived probability of a ‘No-Deal’ outcome and (ii) the intermediate inputs imported from the EU as a share of total costs. We additionally scale the import cost share by the ratio of total to variable costs at the sectoral level, to get as close as possible to a shock to marginal costs. All combined, this gives us a firm-level tariff news cost shock. The average firm-level tariff news shock is -0.7%, with a standard deviation of 1.3%. We then estimate firm-level IPT by regressing survey-based non-zero price changes after the announcement on the constructed news shocks. In the regressions, we control for the probability of no-deal, the import cost share, the sectoral fixed cost share, exporter status, firm investment, firm employment, and we include time fixed effects. Our main sample period is 2019 Q2 to 2019 Q4.

Main findings

A model of price-setting with firm heterogeneity would predict that IPT varies along at least two key margins. First, firms with longer average price durations (ie ‘stickier’ prices) would have stronger IPT. Intuitively, these firms are more forward-looking because they internalise that they may not get another chance to change prices by the time the shock arrives. Second, firms which expect the shock to arrive sooner would also have higher IPT. Again, the intuition is that if the shock is expected to arrive sooner, the probability that the current reset price is fixed until that period is higher, and therefore firms adjust by more today. In the data, we find evidence consistent with these theoretical predictions. The IPT increases monotonically with average price duration and decreases with the horizon of the shock. Price durations (or the average number of months prices remain unchanged) are estimated using CPI/PPI micro data at the sectoral level, whereas the expected Brexit date is measured using direct survey questions in the DMP. Firms with average price durations of 10–20 months that expected Brexit to occur in 2019 have an IPT of 44% (blue coefficients in Chart 2). This increases to 62% for firms with an average price duration of more than 20 months. These effects are statistically significant. Likewise, expecting the shock to occur sooner leads to a higher IPT. The red coefficients in Chart 2 present the corresponding estimates of IPT for firms that expect Brexit to occur in 2020 instead of 2019. Consistent with the logic above, we do not find evidence of positive and significant IPT for these firms, even for those with stickier prices.

Chart 2: Estimated IPT – interaction with price durations and perceived Brexit horizon

Notes: The chart presents the effect of the tariff import cost shock on firm prices, with interactions for average price duration and expected Brexit date. Standard errors are clustered at the SIC2 level and 90% confidence intervals are reported. The regression results are reported in Column 5 of Table 1 in Ghassibe et al (2025).

We also show that IPT varies depending on a firm’s typical price-setting behaviour. Firms in the DMP are asked whether they typically change prices at regular intervals (ie time-dependent) or in response to changes in demand or costs (ie state-dependent). We find that firms that engage in time-dependent pricing have higher IPT than firms which use more state-dependent pricing. This is consistent with state-dependent firms being more flexible in their price-setting, although crucially the results are robust to controlling for the average price duration.

Does intertemporal pass-through depend on the size of the shock? Using a model solved with fully non-linear methods, we show that IPT increases less than proportionately with shock size. The intuition for this result stems from the fact that these are expected cost shocks. As the shocks grow in size, firms endogenously revise their perceived probability of adjustment in the period when the shock arrives, which in turn lower the IPT. In the limit, as the shock becomes extremely large, firms revise the probably upwards to one, delivering IPT of zero. Importantly, this result contrasts the faster pass-through of large contemporaneous cost shocks which has been documented in the literature (eg Cavallo et al (2024)). In the data, we find similar evidence of a non-linearity in the estimated IPT. Specifically, as the tariff news shock becomes bigger (in absolute value), the impact on price adjustment changes less than proportionately in magnitude. This is shown by the red line in Chart 3, which allows for non-linearity in the estimated IPT. For example, the impact of a -5% tariff news shock on the price level is -3% when allowing for non-linearities, compared with -4.25% under the linear specification (blue line).

Chart 3: Estimated IPT – linear versus non-linear effects

Note: The chart presents the linear and non-linear predicted effect of the tariff import cost shock on firm prices. The results are based on Column 4 from Table 3 in Ghassibe et al (2025). The predicted values are for firms who have a modal expected Brexit year of 2019.

Finally, we use our model to show that the firm-level differences in IPT have important implications for aggregate inflation dynamics. In particular, taking the heterogeneities in perceived shock horizons and adjustment frequencies into account amplifies the response of aggregate inflation to anticipated future shocks, relative to a model with no heterogeneity. This is shown in Chart 4. The solid blue line shows the average IPT with both dimensions of heterogeneity. Turning off heterogeneity in perceived shock horizons (Homogeneous Horizon) implies a slightly lower IPT. However, turning off heterogeneity in price stickiness (Homogeneous Calvo) implies a much lower IPT, closer to the model without any heterogeneity (Homogeneous Calvo and Horizon). Thus, while both dimensions matter for the amplification in the response of inflation to future shocks, we find that heterogeneity in price rigidity is quantitatively more relevant. Importantly, this finding is in contrast to existing results for realised shocks, which suggest that heterogeneity in price rigidity dampens aggregate price movements.

Chart 4: Role of different dimensions of micro heterogeneity for average IPT

Conclusion

Our empirical estimates provide evidence that firm prices respond significantly to expected future cost shocks, in ways predicted by standard models of price-setting. Furthermore, our theoretical results imply that microeconomic heterogeneity in price-setting can amplify the contemporaneous aggregate effects of future policy announcements. For example, announcements about the future path of monetary policy or fiscal policy, to the extent that they affect firms’ expected costs, can have significant impacts on current pricing decisions and therefore inflation dynamics. The non-linearity we find in the pass-through of news shocks can also have potentially far-reaching implications for the optimal size of promises about future policy interventions. If policymakers wish to maximise the contemporaneous price impact of announcing future policies, this suggests they should gradually release information over time (rather than a ‘big bang’ approach), and vice versa.

Boromeus Wanengkirtyo and Ivan Yotzov work in the Bank’s Structural Economics Division and Mishel Ghassibe is an Assistant Professor at the Centre de Recerca en Economia Internacional (CREi).

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.