Euro area inflation nearcast model Report Interpretation

Overview

Goldman Sachs introduces a high-frequency, bottom-up model for forecasting Euro area headline and core HICP inflation. The model suggests inflation pressures have rebuilt since August, with upside risk concentrated in core goods even as recent spot core inflation has remained relatively benign.

Core views

Goldman Sachs argues that interpreting Euro area inflation has become difficult because energy prices are highly volatile, national and area-wide releases arrive at different times, and surveys, producer prices, import prices and wholesale prices can send conflicting signals. While August HICP inflation excluding food, energy, alcohol and tobacco was 2.40% year-on-year—broadly unchanged from February—official, consensus and ECB projections have continued to point to potential upside in core inflation. The institution therefore introduces a daily-updated nearcast based on more than 600 domestic and global price-related variables. The first stage is a dynamic factor model that condenses non-HICP indicators into common upstream-cost factors. Its inputs include sectoral producer and import prices, wholesale prices, wage measures, supply-chain indicators, commodity prices, refined-product prices, surveys and the ECB trade-weighted euro index. The model is designed both to extract common signals from noisy sector data and to project those pressures through economically restricted cross-sector links. For example, chemical-cost pressures can be influenced by earlier oil-price moves, while wage pressures are not allowed to respond mechanically to paper-sector costs. Second, Goldman Sachs uses input-output tables to translate estimated labour and non-labour cost factors into product-specific cost indices for 23 exhaustive area-wide HICP components. These weights reflect the inputs required to supply each consumer category: restaurant services combine labour and food inputs, whereas clothing retail has greater exposure to textiles. The institution says these component-level indices are generally better leading indicators of component inflation than less detailed, noisier measures such as raw producer-price indices. Third, the component cost indices are combined with supplementary domestic indicators in mostly vector autoregressions and linear regressions. Specifications vary by component. Fuel inflation is forecast from oil futures and refined-product prices, assuming product spreads gradually narrow toward historical norms; gas and electricity incorporate slow wholesale-gas pass-through to reflect utility contracts. For core items, survey price expectations, wholesale prices, foreign exchange moves or lagged headline inflation may supplement cost indices. Component forecasts are then aggregated into core and headline inflation and can use early national detail from EMU4 economies before the final Euro area release. In pseudo-real-time tests using data available with average release lags, the model delivered large out-of-sample gains over a bottom-up AR(1) benchmark for major components and outperformed that benchmark for core inflation at 3-, 6- and 12-month horizons. It would also have outperformed ECB projections and the Survey of Professional Forecasters median during the 2022 inflation cycle: by mid-2022, its nearcast would have been around 9% for headline and 5% for core inflation, close to the inflation rates eventually realized in December. Since 2023, however, performance has been broadly in line with other forecasts. The current signal is firmer inflation pressure. The model-implied end-2026 headline rate has risen from below 2% year-on-year in February to 3.85% currently, broadly following commodity prices and tracking market-implied inflation reasonably well. End-2026 core inflation is projected at 2.62%, up from around 2% expected before the war. Goldman Sachs attributes the increase first to higher energy prices and more hawkish manufacturing price surveys, supply-chain indicators and producer prices in the first half; pressures eased slightly early in summer but strengthened again from August as energy commodities rose and non-HICP indicators turned more hawkish. Goldman Sachs says its standing forecast is now close to the nearcast, implying headline inflation just under 4% year-on-year at year-end and core inflation peaking at 2.8% early next year. Relative to its own and the ECB's latest forecasts, the nearcast's upside risks arise mainly from core goods—particularly vehicles, appliances and ICT equipment. Conversely, it sees a more modest food-inflation increase, peaking at just over 3% year-on-year. The model will run daily alongside the institution's judgment-based official forecast, which uses its commodity strategists' forecasts, to interpret incoming inflation data and express risks around that forecast.

Analysis framework

Goldman Sachs first compresses a broad set of upstream price and cost indicators into sectoral factors, then applies input-output cost shares to map those pressures to 23 HICP components. It forecasts each component with tailored statistical models, aggregates the results into core and headline inflation, and evaluates the framework through pseudo-real-time out-of-sample comparisons with simple and professional-forecast benchmarks.

Methodology notes

• Other

  Dynamic factor model (DFM) of upstream cost pressures

  The model reduces hundreds of noisy price-related indicators into a smaller set of common sectoral cost factors and projects them using economically restricted spillover relationships.

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  Input-output-based component cost indices

  Direct input-cost shares are used to pass labour and non-labour upstream pressures through to product-specific consumer-price components.

• Other

  Component-level VARs and linear regressions

  Tailored statistical models combine cost indices with indicators such as price expectations, wholesale prices, foreign exchange and lagged inflation to forecast each HICP component.

Key data

• Model input coverageOver 600 variablesDomestic and global price-related indicators used in the daily nearcast.
• HICP target components23Exhaustive area-wide basket components forecast individually before aggregation.
• August core HICP inflation2.40% yoyHICP excluding food, energy, alcohol and tobacco; broadly unchanged from February.
• End-2026 headline HICP nearcast3.85% yoyUp from below 2% in February and close to Goldman Sachs' official forecast.
• End-2026 core HICP nearcast2.62% yoyUp from around 2% expected before the war.
• Core inflation peak in standing forecast2.8%Expected early next year.
• Food inflation peak in nearcastJust over 3% yoyMore modest than Goldman Sachs' and the ECB's existing forecasts.

Impact & implications

The report indicates that apparently benign current core inflation data may understate pipeline pressures. Goldman Sachs sees its current official inflation path as broadly validated by the nearcast, while highlighting additional upside risk from core goods and comparatively less food-price pressure.

Risks

• The strength of indirect energy-price pass-through into core inflation remains highly uncertain.
• Energy and commodity prices remain highly volatile, increasing uncertainty around the inflation path.

What to watch

• Daily nearcast updates as new price-relevant data are released.
• Energy commodity prices, refined-product prices and market-implied inflation.
• Manufacturing price surveys, supply-chain indicators and producer-price data.
• Core-goods inflation in vehicles, appliances and ICT equipment.