Impact of hot days on productivity in Great Britain methodology

• The Office for National Statistics (ONS) has developed an experimental method to estimate the potential scale of lost output, in terms of gross value added (GVA) lost from hot days.

• Initial model-based estimates suggest that between 1998 and 2021, the biggest impacts were in 2020 where hot days could have reduced productivity in Great Britain by £5.3 billion if we assume some adaptation measures, such as air conditioning, were in place.

• Impacts from hot days on productivity were lowest in 2007 and 2008 at under £100 million each year.

• The average loss between 1998 and 2021 was estimated at £1.2 billion per year with adaptation.

We produce the UK natural capital accounts, which include estimates of the costs of environmental damage. Such costs include estimates for the number of human life years lost from air pollution, and the benefits of outdoor recreation.

However, there are no consistent, detailed estimates available for the current economic impacts of climate change. These costs of climate change are difficult to observe, and they are not captured by current economic monitoring systems.

Estimates of the future monetary impacts of climate change are also contentious among environmental economists (as explained in the Annual Review of Environment and Resources journal). Some suggest that long-term impacts of significant atmospheric warming might only amount to a few per cent of global gross domestic product (GDP), but this is hard to assess without comprehensive data. More information can be found in the presentation Climate change: the ultimate challenge for economics (PDF, 1.11MB) on the Nobel Prize website. Further details can be found in the research article Economic impacts of tipping points on the climate system, (which can be found on the Proceedings of the National Academy of Sciences of the United States of America [PNAS] journal website).

The UK natural capital accounts include estimates for urban heat regulating services, measuring nature's ability to cool down urban environments on hot days, which helps to avoid productivity losses.

We complement that analysis by estimating the costs of hot days in terms of labour productivity reflected in output changes, using the most recent and most spatially granular gross value added (GVA) data available to estimate the scale of the costs across Great Britain. This method update also improves our estimates of the benefits of urban cooling in the UK natural capital accounts.

We set out to provide estimates of how hot days in Great Britain – those where the maximum daily temperature reaches at least 28 degrees Celsius – might affect gross value added (GVA).

We use GVA as it is a measure of industrial production. Hot days are assumed to particularly affect industries that involve outdoor or manual labour, where heat is likely to have a greater impact on the ability to undertake such labour, as explained in the Economics for the Environment Consultancy (eftec) report on the Department for Environment, Food and Rural Affairs (Defra) website.

The Met Office states that, as greenhouse gas concentrations increase, heatwaves are expected to occur more frequently:

"A scientific study by the Met Office into the summer 2018 heatwave in the UK showed that the likelihood of the UK experiencing a summer as hot or hotter than 2018 is a little over 1 in 10. It is 30 times more likely to occur now than before the industrial revolution because of the higher concentration of carbon dioxide (a greenhouse gas) in the atmosphere. As greenhouse gas concentrations increase, heatwaves of similar intensity are projected to become even more frequent, perhaps occurring as regularly as every other year by the 2050s." Met Office – What is a Heatwave? article.

This work shows the potential scale of one economic effect of that change. We also sought to use up-to-date weather data from the Met Office, update our methods to use more recent GVA data at the more granular local authority level, and improve and adjust existing methods to determine the impact of hot days on labour productivity in local authorities in Great Britain.

This work also improves on methods to estimate the reduction in productivity losses gained from urban heat regulating (cooling effects of nature), which we have previously published in our natural capital accounts.

Our previous urban heat work was developed as part of the Economics for the Environment Consultancy (eftec) report for the Department for Environment, Food and Rural Affairs (Defra). More information can be found in the UK Scoping Urban Natural Capital Accounts – Extension to develop temperature regulation estimates – NR0172 methodology.

Because of data constraints at the time, our previous urban heat work looked at 11 city regions, with coverage across England, Wales and Scotland. Updates in local area data availability allow us to improve coverage to all local authorities with more densely populated areas in Great Britain. Our current methods excluded Isles of Scilly as the population of this local authority is under 5,000.

The current methodology uses up-to-date weather data from the Met Office, recent gross value added (GVA) data, and more granular local authority-level GVA data than what was previously used for urban heat regulating estimates in Office for National Statistics (ONS) natural capital accounts.

Spatial granularity and GVA data

We were able to further breakdown the regional boundaries for weather data and GVA data by local authorities in Great Britain (England, Scotland and Wales), improving the representation of urban heat impacts.

We looked at more densely populated areas within local authorities, using the ONS definition of built-up areas. Find more information in our Towns and cities, characteristics of built-up areas, England and Wales: Census 2021 article.

We refined this to focus on those sections of built-up areas with populations of 5,000 or above in England and Wales, and as a proxy for population size, areas with 2,500 or more houses in Scotland. This was done using the QGIS mapping tool. Mapping for Northern Ireland uses different definitions for urban areas leading to a data discontinuity. Future work could address this, but in this experimental project, we have limited the analysis to Great Britain.

Within a local authority, rural areas may have cooler temperatures than urban areas, where much of the economic activity is associated. In addition, the impact of heatwaves on agriculture will be more complicated than the labour productivity impact alone as heatwaves could directly affect crop growth for instance.

A different treatment is likely necessary for rural areas compared with urban. Therefore, focusing on the built-up areas within local authorities helps to avoid complicating the impact of heat. We therefore only include built-up areas. Updated GVA data (data time series up to 2021) was collected from regional accounts at the ONS.

Weather data

The Met Office provides maximum daily temperatures for 1 kilometre grid squares across the country. We used QGIS mapping software and Python to overlay weather data grid squares over our built-up area local authority map (described in the previous section).

Temperature bands (in Celsius) were: 28 to 29 degrees, 29 to 30 degrees, 30 to 31 degrees, 31 to 32 degrees, 32 to 33 degrees, 34 to 100 degrees.

We then average the number of hot days reported per year – by temperature band – in each 1 kilometre grid square within the densely populated parts of each local authority.

Percentage of productivity loss

Our productivity loss percentage value was based on the work Climate change, heat stress and labour productivity: A cost methodology for city economies (PDF, 1.08MB), from Costa and others, and Scoping UK Urban Natural Capital Accounts - Extension to develop temperature regulation estimates, from eftec (see previous) that identified productivity loss estimates (percentage of productiveness compared with "fully productive work") for each industry affected under the different hot day temperature bands being used.

The Costa study assumes that industries with lower elasticity between labour and capital are most affected by increased heat. For example, construction is an industry highly affected by heat. This is because the manual labour involved cannot easily be replaced by a substitution, such as with machinery, during a heatwave or in response to heatwaves becoming more common. Costa and others based their estimates on International Organization for Standardization (ISO) standards for recommended hourly work rates at different levels of "wet bulb" temperature. This is a measure that takes humidity and energy requirements into account to perform different activities.

Productivity lost from "hot days"

Our method to calculate the productivity loss from hot days builds on the previous studies and our own previous work on urban heat regulating services.

We show productivity lost from hot days as the output lost in GVA in pounds from days where temperatures reach over 28 degrees Celsius, taking into consideration how hot days will affect industries differently. These impacts are modelled to increase through temperature bands up to 35 degrees Celsius, above which we assume for simplicity that the impacts are the same as at 35 degrees Celsius.

The potential productivity loss by temperature band and industry is first produced as a percentage loss from the studies mentioned previously. The annual GVA by industry and local authority can then be divided by 365 (366 for leap years) to estimate daily GVA, assuming that output is evenly distributed across the year.  Multiplying the potential loss for each temperature band by each daily GVA produces the loss expected from a "hot day". This value is then multiplied by the actual number of hot days in a year by temperature band. The output loss per year from hot days in GVA is the sum of the number of hot days in a year (by temperature band), multiplied by the expected productivity daily GVA losses (deflated to 2022 prices) by industry and local authority.

Adaptation measures

To account for the potential of adaptation to heat, we also used an adjusted loss multiplier. Such adaptation measures could include changing working hours to avoid early afternoon work, increasing ventilation, adding solar blinds and using air conditioning.

These measures would reduce productivity losses, but this would vary by industry. It is easier to avoid the impacts of heat in offices with air conditioning, while for labour-intensive outdoor work such as construction, air conditioning is not an option.

Our approach follows Costa and others' paper (PDF, 1.08MB) and eftec's report because industry sectors that are more affected by heat because of being outdoors or more labour intensive are able to avoid an estimated 40% of productivity losses because of adaptation with behavioural change, and industry sectors that are more often indoors may avoid an estimated 85% of productivity loss from adaptation with air conditioning.

Our results are modelled based on expected effectiveness of labour at different temperatures, and therefore should be considered as an indication of what could be happening. These are not actual losses in productivity based on observed data. All GVA figures are in 2022 prices. 

We estimate that hot days in Great Britain resulted in an estimated total of £28.6 billion (in 2022 prices) of lost GVA over the 24-year period 1998 to 2021, with adaptation measures (Table 1 and Figure 1). The average estimated annual GVA loss was £1.2 billion assuming some level of adaptation.

There was considerable variation between years (Table 1). Modelled losses were highest in 2020 with a GVA loss of £5.3 billion with adaptation measures, approximately 0.2% of GDP that year. While there were more hot days in 2018 than in 2020, temperatures in 2020 were higher and therefore had a greater impact on GVA.

Note that local authority estimates are in the supplementary tables.

Our modelling has limitations. Air conditioning is now common in offices and may be common in more industry sectors than are captured in our adaptation models. We also do not know with certainty how much less work people can complete in the heat or how they may make up for a loss of output on a hot day on the cooler days that follow.  

This method is not based on observed impacts, but is a model based on assumptions about human energy use and efficiency at different temperatures. 

However, while this work is based on assumptions, the scale of the impact of hot days on productivity in these results indicates that it may be useful to attempt to observe impacts empirically. 

It would also be useful to examine correlations between "hot days" and gross value added (GVA). The challenge of observing a correlation between hot days and reduced output seems clear in 2007, the year of the financial crisis, and in those following it, as those years coincided with fewer hot days (Figure 2). The effects of hot days in most years are likely to be much smaller than prevailing economic drivers (such as geopolitical events or the coronavirus (COVID-19) pandemic) or even potentially other weather events, such as heavy rain each year. Monthly or quarterly data might make it easier to disentangle any confounding factors, since the small number of hot days would be more concentrated into shorter time periods.  

In the short term, the improved mapping and GVA data from this work will improve our UK natural capital accounts estimates. We use this approach in those accounts to estimate how the cooling of cities by trees and rivers reduces the impacts of hot days on productivity. We are now able to do this for a wider area of the country with up-to-date GVA data. These data will be included in our next UK natural capital accounts publication due later this year.

Official statistics in development 

These statistics are labelled as "official statistics in development". Until September 2023, these were called "experimental statistics". Read more about the change in the Guide to official statistics in development.  

As a working paper this output is particularly exploratory. We are developing how we collect and produce the data to improve the quality of these statistics and considering whether to continue development. If developments are continued to completion, we will review the statistics with the Statistics Head of Profession. We will decide whether the statistics are of sufficient quality and value to be published as official statistics, or whether further development is needed. Production may be stopped if they are not of sufficient quality or value. Users will be informed of the outcome and any changes. 

We value your feedback on these statistics. Contact us at natural.capital.team@ons.gov.uk

Gross value added (GVA) 

The value generated by any unit engaged in production and the contributions of individual sectors or industries to gross domestic product. It is measured at basic prices, excluding taxes, less subsidies on products.  

Gross domestic product (GDP) 

Gross domestic product (GDP) measures UK economic growth based on the value of goods and services produced during a given period. GDP at market prices is equal to the GVA at basic prices, plus taxes on products, less subsidies on products.