Current and projected regional economic impacts of heatwaves in Europe

The study investigates how extreme heat, particularly heatwaves, reduces labour productivity and thereby economic output across Europe. With rising temperatures and an observed doubling of days above the 90th percentile threshold since 1960, heatwaves are becoming more frequent and prolonged, posing threats to occupational health and economic performance. Prior work often assessed impacts using average temperatures rather than well-defined extreme events, leaving gaps in understanding the spatially and sectorally heterogeneous effects of heatwaves. This paper aims to quantify current and future economy-wide damages from heatwave-induced productivity losses with high spatial, temporal, and sectoral resolution, using a heat stress index (WBGT) and a regionalised general equilibrium model, and to project impacts under a high-emission scenario for mid-21st century.

Previous studies have shown heat’s adverse effects on labour productivity and time allocation but largely focused on average temperatures rather than extreme heat events. Orlov et al. analysed past European heatwave effects without characterizing extent and duration of episodes; Knittel et al. (Germany) and Orlov et al. (global) projected productivity impacts based on average workplace temperature conditions. Emerging literature indicates increasing European heatwave frequency and severity linked to anthropogenic climate change. The current study extends the literature by explicitly identifying heatwaves (TX90p), using hourly heat stress (WBGT) for indoor/outdoor contexts, and integrating sector-specific productivity responses into a sub-national CGE to capture direct and indirect effects.

Study area comprised 274 regions across EU-27, UK, and EFTA. Heatwaves were identified using TX90p (regional daily max temperature above the 90th percentile for ≥3 consecutive days) with 1981–2010 as baseline. Past events: four anomalously hot years (2003, 2010, 2015, 2018) and the full historical period were analysed. Climate data: ERA5-Land hourly meteorological variables (air temperature, dew point, wind speed, solar radiation) were used to compute hourly Wet Bulb Globe Temperature (WBGT) for outdoor (sun) and indoor (shade) environments via the R HeatStress package. A Wet Bulb Degree-Day (WBDD) metric (WBGT above 26 °C during working hours) summarized cumulative heat severity. For future projections (2035–2064), two EURO-CORDEX RCM simulations under RCP8.5 (MPICSC-REMO2 driven by MPIESM; KNMI-RACMO driven by HADGEM) were bias-corrected using empirical quantile mapping against WFDEI observations; daily mean and max WBGT were downscaled to hourly using the 4+4+4 method. Population weighting: Gridded UN WPP-adjusted population data (SEDAC GPWv4) were used to weight productivity impacts; for future, SSP5 population grids for 2030–2070 were interpolated. Seasonal adjustment: Quarterly National Accounts (Eurostat) informed seasonal patterns; economic shocks were weighted by heatwave days per quarter. Heat exposure functions: Workability (0–100%) was derived from hourly WBGT using ISO 7243 (benchmark), NIOSH, and Hothaps functions, with sector workload classes (low, moderate, high; M=200/300/400 W). ISO/NIOSH workability used thresholds dependent on metabolic rate; Hothaps used a logistic curve with a 10% lower bound. Workday was assumed 9:00–17:00; daily averages computed for heatwave days, then aggregated to annual equivalents. Economic modelling: A sub-national computable general equilibrium (CGE) model based on GTAP (calibrated to GTAP8, year 2007) with regional SAMs constructed from Eurostat and national sources. Production uses Leontief aggregation of intermediates and CES for value added (capital, labour). Trade uses Armington CES for domestic vs aggregate imports and CRESH to split imports by source, with intra-national elasticities increased 20% (border effect). Labour and capital fully employed, mobile across sectors within regions. Coupling: Annual-equivalent sector-region labour productivity shocks τ_rs = (1 − ω_rs) were applied to the model to obtain economy-wide effects. Historical losses over 1981–2010 were recomputed by identifying heatwaves and deriving regional-sector shocks from ERA5-Land WBGT time series.

- Spatial heterogeneity: Heatwave frequency, duration, and severity vary widely; summer events are longer and more severe. Average heatwave affects 27–38% of Europe annually, ~49% during summer, with maxima >95% coverage in large episodes. Southern regions consistently show higher cumulative heat stress (WBDD). - Current economic impacts: In 2003, 2010, 2015, 2018, heatwaves reduced European GDP by 0.3–0.5% (vs historical average ~0.2% over 1981–2010). Regional losses show a strong north–south gradient; many southern regions exceed 1% of regional GDP losses, some >1.5–2%. - Sectoral propagation: Direct productivity losses occur mainly in outdoor sectors; indirect effects spread via intermediate inputs to services and other indoor sectors. Trade partially buffers shocks but is outweighed by intermediate input linkages. Under current climate, indoor work is minimally affected directly (WBGT_shade seldom exceeds damaging thresholds), while outdoor ambient-exposed work shows pronounced impacts. - Sensitivity to exposure functions: Relative to ISO, damages are ~11% lower using NIOSH and ~30% lower using Hothaps. - Projections (RCP8.5, 2035–2064): Europe-wide annual GDP losses from heatwaves increase steadily: from 0.21% (1981–2010 avg) to 0.77% (2035–2045, s.d. ±0.16%), 0.96% (2045–2055, s.d. ±0.26%), and >1.14% in the 2060s (s.d. ±0.25%), i.e., ~five-fold increase vs historical average, with rising interannual variability. Indoor workers are projected to face more direct heat stress in southern/central regions if current conditions persist. - Country-level: Southern Europe is most affected. Cyprus has the highest relative losses. Portugal, Spain, and Croatia rise from ~2% losses around 2040 to ~3% by 2060. Balkans, Italy, Greece exceed 2% (2055–2064). Central/northern Europe experiences smaller but significant losses (e.g., Germany ~0.5% by 2050). UK, Iceland, Scandinavia: mild losses (0–0.2%). Differences across climate models are modest through ~2050; scenario divergence increases thereafter.

The findings demonstrate that heatwaves already impose measurable macroeconomic costs in Europe, disproportionately burdening southern regions due to higher thermal exposure and larger shares of outdoor economic activity. The CGE analysis reveals that direct outdoor labour productivity shocks propagate broadly via intermediate input linkages, affecting indoor and service sectors, while intra- and international trade only partly mitigates impacts. The projection exercise indicates steadily increasing damages through the mid-21st century—even under stringent mitigation pathways, given limited scenario divergence before ~2050—implying that adaptation in workplaces (especially for outdoor sectors) will be crucial. The results address the research question by quantifying both current and future economy-wide impacts of extreme heat with regional and sectoral detail, highlighting hotspots for vulnerability and the mechanisms of impact transmission within the economy.

This study quantifies the current and projected regional economic impacts of European heatwaves using hourly WBGT-based productivity losses coupled to a sub-national CGE model. It shows 0.3–0.5% EU GDP losses in recent extreme years versus ~0.2% historically, with strong south–north gradients and significant within-country heterogeneity. Under a high-emission pathway, losses are projected to rise to ~1% of GDP by the 2060s, with southern countries experiencing the largest damages. The approach offers a tool for assessing occupational heat risks and informing local adaptation policies. Future research should: (1) integrate heat-induced occupational injuries and health costs; (2) resolve urban heat island effects and apply human heat balance metrics (e.g., MRT, UTCI) at city scale; (3) incorporate adaptation measures (e.g., air conditioning, shading, scheduling) and their sectoral penetration; (4) refine exposure–response functions with sector- and task-specific field evidence, including impacts below 26 °C WBGT; (5) combine climate and socioeconomic pathways (RCP–SSP) with hourly, spatially downscaled heat stress projections and regionally consistent economic structures to assess uncertainty and adaptation pathways.

- Reported costs are likely a lower bound: occupational injuries and related public health costs are excluded. - Heat exposure functions are calibrated from limited empirical studies; subsector heterogeneity in workloads is simplified. - Adaptation and insulation measures (e.g., air conditioning) are not explicitly modeled; while limited in outdoor sectors now, they could alter future impacts, especially for indoor work. - Economic model calibration uses 2007 SAMs, which may embed some historical heat impacts and reflect that year’s economic structure. - Climate projection uncertainties (model structure, initial conditions) remain; only two RCMs under RCP8.5 were used, albeit bias-corrected. - Assumptions on working hours (9–17), constant intra-day schedules, and the 4+4+4 reconstruction for hourly WBGT may not reflect all occupational practices. - Historical population data gaps (pre-2000) required using year-2000 weights; trade elasticities for intra-national flows are conservatively adjusted but not fully validated via sensitivity analysis.