Bank of England Inflation Survey Shows Lower Expectations Amid Provider Change to Savanta

Key Takeaways

• The Bank of England's August 2026 Inflation Attitudes Survey, the first conducted by Savanta in place of Ipsos, shows median one-year inflation expectations falling to 3.2% from 4.0% in May.
• A parallel run shows that part of the decline reflects the provider change rather than a true shift in sentiment.
• Compliance and risk teams should treat quarter-on-quarter comparisons with caution when using this data in models, stress scenarios, or customer-facing materials.

Deep Dive

The Bank of England's August 2026 Inflation Attitudes Survey, the first conducted by Savanta in place of Ipsos, shows median one-year inflation expectations falling to 3.2% from 4.0% in May. A parallel run shows that part of the decline reflects the provider change rather than a true shift in sentiment. Compliance and risk teams should treat quarter-on-quarter comparisons with caution when using this data in models, stress scenarios, or customer-facing materials.

Firms may use it in conduct risk assessments, product governance, customer vulnerability analysis, and ICAAP or stress-testing narratives. The key change is methodological: Savanta took over from Ipsos following a competitive re-tendering, and the sample (quota-based, ages 16-75, surveyed 31 July to 7 August 2026) was collected under the new provider.

The parallel run in May found identical inflation perceptions (5.0%) across both providers. Savanta's expectations were consistently lower: 3.6% vs. 4.0% at one year, 3.1% vs. 3.5% at two years, and 3.3% vs. 3.9% at five years. The headline August results show declines of 0.8, 0.6, and 0.7 percentage points respectively, so a meaningful share of each decline is a provider effect. Roughly 0.4 points of the one-year drop and 0.4 and 0.6 points of the two-year and five-year drops are explained by the provider change alone. Real like-for-like movement is therefore smaller than the headline suggests. The Bank also reiterates an earlier methodological break from May 2020, when the survey moved from face-to-face to online, so long-run series comparisons carry additional caveats.

Other findings are relevant for conduct and market risk. Net satisfaction with the Bank's handling of inflation rose to 5% from -2%. Of respondents, 51% believe rates rose over the past year and 51% expect further rises. Only 19% think rates should go up, while 32% prefer cuts. Also, 35% consider the inflation target too high and 16% too low. These figures suggest continued rate sensitivity among households, which is relevant for lending, mortgage, and savings customers. Firms should not read the apparent fall in expectations as a clear easing of pressure.

Recommended actions: (1) Flag the series break in any models, dashboards, or board papers that use the survey, and avoid presenting May-to-August changes as purely like-for-like. (2) Where practical, use the parallel-run figures to adjust for the provider effect. (3) Review internal scenario assumptions and customer affordability or vulnerability analyses for dependence on this data. (4) Update data lineage and documentation for third-party inputs. (5) Make sure any external communications citing the survey include the methodological caveat to avoid misleading statements.

Overall risk impact is low and mainly concerns data quality and model risk. Misinterpreting the provider effect could lead to underestimating inflation expectations, which could affect pricing, forecasting, and conduct-related judgments. No deadlines or implementation dates apply. The next quarterly survey will be the first with two consecutive Savanta datapoints, which should provide cleaner like-for-like comparisons.

Source: Bank of England