Exposure to air pollution in England, 2003–23

1. Introduction

The harmful effects of air pollution on health are increasingly well known. Broadly, air pollution has been shown to damage many aspects of health, particularly for the elderly and young children (see Box 1). The World Health Organisation (WHO) has repeatedly lowered its recommended upper limit for concentrations of air pollutants as understanding develops (World Health Organisation, 2021). Moreover, more attention is being paid to how air pollution can coincide with and exacerbate existing inequalities, such as income and ethnic inequalities. This comes alongside increasing attention on climate policies to reduce greenhouse gas emissions. Such policies often also result in a reduction in local air pollution, making air pollution reduction a potential co-benefit of reaching net zero.

This report documents how local air pollution has changed across England over the last two decades, both geographically and across different income, age and ethnic groups. We study fine particulate matter (PM_2.5), which is the air pollutant most strongly associated with negative health impacts. The three largest sources of PM_2.5 in the UK are domestic burning of wood and other fuels (29.0% of total PM_2.5 emissions in 2022), road transport (17.9% of total PM_2.5 emissions in 2022) – predominantly exhaust pipe fumes – and industrial processes (16.5% of total PM_2.5 emissions in 2022) such as construction and steel manufacturing (Department for Environment, Food and Rural Affairs, 2024).

2. Data and measurement

We use data produced by the Department for Environment, Food and Rural Affairs (DEFRA) on annual PM_2.5 concentrations, measured in µg/m³ (micrograms per cubic metre) at the 1km² grid level for the period 2003 to 2023. This is modelled ground-level air pollution data, meaning that DEFRA uses local air pollution measurements from pollution monitors, information on sources of air pollution (mostly industrial sites and traffic data), local geography and weather patterns, and a model of air pollution dispersion to produce estimates of local air pollution. To measure household characteristics, we use data from the 2001, 2011 and 2021 Censuses for England. We use the lowest level of geography at which key local characteristics can be measured – the lower layer super output area (LSOA). The geographic size of an LSOA varies with population density, but they are relatively small, comprising between 400 and 1,200 households. We restrict our analysis to England only.

We define an individual’s PM_2.5 exposure as the average annual level of PM_2.5 in the LSOA where they live. This is an imperfect measure of an individual’s true exposure for two reasons. First, they may work in or travel through areas with significantly higher or lower levels of air pollution, thus changing their overall exposure level. Second, their home may have particularly high or low levels of PM_2.5 relative to the rest of their LSOA. For example, if they use a wood burner in their home, this may increase their exposure to PM_2.5 in a way our data will not be able to pick up.

The income deprivation rank we use is produced by the Ministry of Housing, Communities and Local Government and includes information about, among other variables, income, employment and education. We use the share of the non-white population in the Census as our measure of the share of ethnic minorities at the LSOA level and also present results for more disaggregated ethnic minority groups.

Most of our results below consider the average exposure to PM_2.5 within groups. We create these averages by taking the average exposure (as defined above) of each individual within the respective group and averaging over these.

Effects of air pollution may be non-linear, with worse effects at high levels of exposure. Health authorities set maximal levels of permissible, or ‘acceptable’, PM_2.5 exposure (discussed further in Box 2). We therefore also consider the share of the population whose exposures is above 10µg/m³.

The threshold of 10µg/m³ is the current annual mean concentration target for England. This threshold was introduced in 2023 and the target is to meet it in all areas of England by 2040. This level also used to be the WHO’s recommended limit before it was lowered to 5µg/m³ in 2021. The limits for the different nations in the UK, the EU and WHO, as well as an overview of air pollution policy since 2000, are discussed in Box 2.

3. Air pollution over time

The average level of air pollution exposure faced by an English resident has fallen significantly since 2003. Figure 1 displays the annual average PM_2.5 exposure, defined as a population-weighted average across small areas’ (LSOAs’) annual average. The dashed line indicates the 10µg/m³ threshold (England’s 2040 target). Figure 2 shows the share of the English population exposed to pollution levels above that threshold.

Both graphs show a large decrease in exposure to fine particulate pollution over the period, from a mean of roughly 15µg/m³ in 2003 to under 7µg/m³ in 2023. The share of the population facing PM_2.5 levels above 10µg/m³ goes from close to 100% to 0%.1 Progress is non-linear, with a sharp decrease in the 2003–07 period, followed by a period of slight increase until 2014, and a significant decrease in 2020.

During the COVID-19 pandemic, air pollution fell dramatically and has remained at these lower levels since. In 2019, 35% of the population still faced PM_2.5 levels above England’s target of 10µg/m³. Average levels of PM_2.5 have fallen by 27% since 2019, and very few individuals remain exposed to more than 10µg/m³. Two-fifths of the progress made on air pollution exposure over the last two decades occurred in the year 2020 alone.

Despite this substantial fall in levels of air pollution in recent years, the vast majority (96%) of England is above the World Health Organisation’s recommended limit of 5µg/m³. Levels of PM_2.5 above this level are associated with significant public health risks according to the WHO. This recommended limit was lowered from 10µg/m³ in 2021 in response to growing evidence of damaging effects of PM_2.5 even at relatively low levels.

The fact that PM_2.5 did not rebound as England came out of the lockdown period of 2020 and 2021 is hard to explain. There are a few potential reasons that could have played a part in this persistence. Since 2020, several of the largest cities in England have introduced clean air zones, designed to dissuade people from driving the most polluting cars into cities. There has also been sluggish growth in the manufacturing sector, with the steel sector (one of the largest industrial sources of PM_2.5) not recovering to pre-pandemic production levels. Further work will look into the extent to which these changes explain the persistent drop in air pollution after the pandemic.

Figure 3 shows how emissions of PM_2.5 have changed over time, broken down by source. Emissions of PM_2.5 – i.e. how much fine particulate matter a certain sector is putting into the air – differ from the measures of ground-level PM_2.5 exposure used in Figures 1 and 2 earlier in three important ways.

First, the emissions may be dispersed by the wind or brought to the ground by rain, meaning a large emission of PM_2.5 may not translate into significantly higher amounts of PM_2.5 in the air we breathe.

Second, if the emissions take place far away from people (e.g. shipping emissions in the English Channel) then these emissions will not expose many people to higher levels of PM_2.5.

Finally, Figure 3 only measures so-called primary PM_2.5 – fine particulate matter that is directly emitted. PM_2.5 concentrations in the air include a lot of so-called secondary PM_2.5. Secondary PM_2.5 is fine particulate matter that is formed from other air pollutants, such as NO₂ and SO₂, emitted from various sources (including combustion in energy industries, industrial combustion and transport) that turn into PM_2.5 a few hours to a few days after being emitted. Considering only primary PM_2.5 tends to overestimate the importance of woodburning as opposed to transport and energy combustion.

The three largest sources of primary PM_2.5 in the UK are domestic burning of wood and other fuels (29.0% of total PM_2.5 emissions in 2022), road transport (17.9% of total PM_2.5 emissions in 2022), and industrial processes and product use (16.5% of total PM_2.5 emissions in 2022) such as construction and steel manufacturing. The only source of PM_2.5 emissions that has increased over the period is domestic combustion. Three-quarters of the domestic combustion emissions of PM_2.5 came from woodburning in 2022.

Air pollution over time by age group

Young children and older individuals are particularly vulnerable to the health effects of air pollution. Are they exposed to worse levels than the rest of the population? Figure 4 plots the average PM_2.5 exposure of children under 5 years of age and individuals over 65 over time. The overall pictures are very similar for the two groups and for the average population (Figure 1), because these age groups are fairly evenly distributed across England.

Air pollution over time by region

Figure 5a plots the evolution of average PM_2.5 exposure for each region of England separately. Figures 5b and 5c present average air pollution levels in 2003–04 and 2023 at a more disaggregated regional level, and the map in the dropdown shows average air pollution levels for every year from 2003 to 2023 by middle layer super output area (MSOA).

The most striking takeaway is the difference between London and the other regions: air pollution exposure in London is consistently 15–30% higher than the average throughout the period. This can partly be explained by the fact that London is an overwhelmingly urban region, as air pollution is systematically higher in urban areas. However, since 2019, London has experienced a rapid fall in average levels of PM_2.5, falling by 30% between 2019 and 2023.

We see limited convergence across regions over time, although the absolute gap between the most polluted and least polluted regions has fallen, mostly since 2019. The regions with the biggest falls in PM_2.5 are the South West and the South East, with the North East seeing the smallest fall.

Air pollution over time by income deprivation level

We rank LSOAs using their income deprivation rank in 2019 (latest data available) and consider the average exposure to PM_2.5 in each quintile of income deprivation over time in Figure 6.2

The most deprived quintile consistently has higher PM_2.5 pollution levels than the least deprived quintile, and this gap has widened since 2017. Individuals in the most deprived quintile are exposed to average PM_2.5 concentration levels that are 8% higher than those in the least deprived quintile in 2023.

The relationship between income deprivation and air pollution is not simple: in several periods, the most deprived quintile faces lower air pollution levels than the second-most deprived quintile (see also Figures 9 and 10 later), and the least deprived quintile often had higher levels than the second-least deprived quintile. This is partly a result of several of the richest areas, such as Kensington & Chelsea, being in urban areas and some of the poorest areas being relatively rural.

Air pollution over time by ethnicity

Ethnic minorities were exposed to levels of air pollution 13% higher than average levels for white populations in 2003. This ‘ethnic pollution gap’ shrank to 6% by 2023. Figures 7 and 8 plot the average exposure to air pollution over time and the share of people facing levels above 10µg/m³ (England’s 2040 target) respectively for different ethnic groups. Table 1 later presents average exposure by group in 2003–04 and 2023, as well as the ethnic pollution gap: the difference between average exposure to PM_2.5 faced by ethnic minorities and that faced by white people.

The gap between ethnic minorities and white people was substantial in the early 2000s: ethnic minorities were exposed to on average 1.8µg/m³ more PM_2.5 than white people – a 13% gap. Both the levels of exposure and the gap are similar to those observed in the US by Currie, Voorheis and Walker (2023). Average exposure was 12.3µg/m³amongst the white population in the US in 2001, compared with 14.2µg/m³in England in 2003–04, and the ‘racial pollution gap’ (comparing black and white populations only) in the US was 1.5µg/m³ in 2001.

Looking at Figure 7, we see that average exposure fell substantially for all groups, but more for ethnic minorities. This leads to the ethnic pollution gap being divided by four over the last two decades: by 2023, the gap is 0.4µg/m³ with ethnic minorities exposed to on average 6% higher PM_2.5 levels than white people. This again is similar to the evolution in the US: Currie, Voorheis and Walker (2023) find that the racial gap is divided by three, though only over a 13-year period.

Pollution exposure in England is particularly high for the black population: the black–white pollution gap was 2.6µg/m³in 2004 and fell to 0.5µg/m³ by 2023. In general, however, there is a lot of similarity between different minority ethnic groups’ exposure despite significant differences in ethnic groups’ economic status (Mirza and Warwick, 2022).

Figure 8 presents the share of the population exposed to high levels (above 10µg/m³). We see a particularly dramatic fall amongst ethnic minorities since the pandemic: in 2019, 50% of those individuals were exposed to more than 10µg/m³; by 2023, this number had fallen to close to zero.3

Finally, Figures 9 and 10 plot the correlation between air pollution and the ethnic minority share or the income deprivation rank at the LSOA level, in the first and last periods. We see a clear positive correlation between the ethnic minority share and air pollution concentration. In 2023, the relationship (as measured by the correlation coefficient) is a third of what it was in 2003–04: a one standard deviation increase in the ethnic minority share was associated with an increase in PM_2.5 of 0.95µg/m³in 2003–04 and of 0.31µg/m³in 2023. In contrast, the correlation between the income deprivation rank and air pollution is weaker and getting slightly stronger over time.

Explaining the fall in the ethnic pollution gap

The ethnic pollution gap has fallen in part due to ethnic minorities moving from more polluted areas to less polluted areas (i.e. moving away from London) and in part due to the areas where they live becoming less polluted. In the last two columns of Table 1, we perform a decomposition exercise to investigate the role played by these two factors. The first decomposition (in column 3) considers exposure to PM_2.5 in 2023 had the pollution level in each LSOA remained as it was in 2003–04. Any difference between columns 2 and 3 is therefore due to ethnic minorities and white populations being distributed differently across LSOAs in the two periods, and not pollution changes – a pure ‘sorting’ effect. The second decomposition (in column 4) considers exposure to PM_2.5 in 2023 holding constant the distribution of population by ethnic group across LSOAs as it was in 2003–04. This number tells us the average exposure in 2023 if the population and ethnic mix of each LSOA had not changed.4

Table 1. Observed and counterfactual PM2.5 exposure by ethnic group

Note: Ethnic minorities are defined as those who report an ethnicity other than ‘white’. Columns 1 and 2 present the observed pollution exposures for white and ethnic minority populations averaged across the two respective periods. Column 3 is the counterfactual pollution exposure for the two groups if the pollution level in each LSOA in 2023 remained as it was in 2003–04. Column 4 is the counterfactual pollution exposure for the two groups if the two populations’ geographic distribution in 2023 remained as it was in 2001.

Source: Authors’ calculations using DEFRA PM_2.5 data and English Census, various years.

Table 1 shows that the movement of people of ethnic minorities away from polluted areas was more important in lowering their exposure to air pollution than falls in air pollution. In other words, even if air pollution had not fallen at all, the gap in air pollution exposure between white people and ethnic minorities would still have halved (as shown by comparing columns 1 and 3 of the table) because people of ethnic minorities have moved to less polluted areas. Since the pandemic, however, falling air pollution has been a more important channel than ethnic minorities moving to different areas.

All of the decrease in PM_2.5 exposure faced by white people can be explained by pollution levels falling: the average exposure in the counterfactual in 2023 in which air pollution is kept constant is the same as the observed average exposure in 2003–04. In contrast, part of the decrease in exposure faced by ethnic minorities can be explained by ethnic minority populations growing more in less polluted areas. Even in the absence of any change in air pollution in England, the average exposure to air pollution faced by ethnic minorities would have fallen by 1µg/m³on average. This implies that changes in sorting behaviours by ethnic group (such as ethnic minorities moving to less polluted areas) alone led to a reduction of the ethnic pollution gap over the period equal to 71% of the observed reduction in the gap. Changes in pollution levels while keeping the distribution of ethnic groups constant across LSOAs lead to a smaller decrease in the ethnic pollution gap, of 57% of the observed decrease.5

Table 2 shows the share of ethnic minorities in the population of each region in each Census year. We see that at the start of the period, ethnic minorities represented a small share of the population in all regions except London, the region with by far the highest air pollution levels (and, to a lesser extent, the West Midlands with roughly average air pollution levels). Looking across columns, we then observe ethnic minorities ‘moving out’ of more polluted regions over time, relative to the white population. The share of ethnic minorities increases in all regions, and doubles in England overall, between 2001 and 2021, but the smallest increases take place in more polluted regions (London, East Midlands) whilst the largest increases by far can be found in the North East and South West, two regions with much lower air pollution levels than average. This is consistent with findings from recent research documenting that in England, residential mobility is associated with air quality improvement on average, with a stronger effect for households with a migration background than for British natives (Rüttenauer et al., 2023). Overall, ethnic minorities are initially highly concentrated in one high-pollution region (London) but then spread out across English regions over the period. This plays a large role in explaining why the ethnic pollution gap falls over time.

Table 2. Share of ethnic minorities in each region over time

Note: Ethnic minorities are defined as those who report an ethnicity other than ‘white’.

Source: Authors’ calculations using English Census, various years.

4. Conclusion

Remarkable progress has been made in lowering average levels of PM_2.5 exposure across England. Almost everywhere in England is now below the target the UK government set England for 2040, well ahead of expectations. This progress has coincided with significant policy activity in this area – from tighter regulations on industrial plants to the proliferation of clean air zones in many of England’s cities.

In this report, we have found a large drop in PM_2.5 over the last two decades, with a substantial fall during the COVID-19 pandemic. Between 2003 and 2023, the average annual level of PM_2.5 fell by 54% in England from around 14µg/m³ to under 7µg/m³, with half of that fall occurring since 2019. The lower levels of PM_2.5 in 2020 have persisted through to 2023 (latest data available).

Every region of England has seen its average level of PM_2.5 roughly halved over the period. Some regions, such as the South East and the South West, have seen slightly larger falls, while others, such as the North East, have seen slightly smaller falls. London consistently has the highest levels of PM_2.5, around 15–30% higher than the national average.

We have provided evidence that air pollution in England has been consistently higher for ethnic minorities over the last 20 years.6 On average, ethnic minorities were exposed to levels of PM_2.5 6% higher than those for white people in 2023, down from 13% in 2003. Others have shown that there is a ‘race pollution gap’ in the US – black individuals face substantially more air pollution than white individuals (Currie, Voorheis and Walker, 2023; Colmer et al., 2024). Campaigners have long raised concerns about ethnic minorities experiencing higher levels of air pollution (ClientEarth, 2021). There is a case that environmental policy, including that related to air pollution, should be part of policy efforts to narrow racial health gaps.

We have also documented a gap in levels of air pollution between the most and least deprived areas. Using the Ministry of Housing, Communities and Local Government’s income deprivation rank, we show a gap of, on average, 5% between the most deprived 20% of areas (lower layer super output areas, LSOAs) and the least deprived 20%. Unlike the gap between ethnic minorities and white people, this gap shows no clear trend.

There has been both significant policy action and substantial progress in lowering levels of PM_2.5 across England. At the same time, our understanding of the harms of PM_2.5, even at relatively low levels, has progressed rapidly. This report has shown that there are still sizeable inequalities in exposure to PM_2.5 across regions, by ethnicity and by income deprivation and that average levels of PM_2.5 are still above the WHO’s recommended limit for 96% of England’s population.

Appendix

References

Alexander, D. and Schwandt, H., 2022. The impact of car pollution on infant and child health: evidence from emissions cheating. Review of Economic Studies, 89(6), 2872–910, https://doi.org/10.1093/restud/rdac007.

Bishop, K. C., Ketcham, J. D. and Kuminoff, N. V., 2023. Hazed and confused: the effect of air pollution on dementia. Review of Economic Studies, 90(5), 2188–214, https://doi.org/10.1093/restud/rdac078.

Brook, R., Rajagopalan, S., Pope, C., Brook, J., Bhatnagar, A., … and Kaufman, J., 2010. Particulate matter air pollution and cardiovascular disease: an update to the scientific statement from the American Heart Association. Circulation, 1 June. Retrieved from https://www.semanticscholar.org/paper/Particulate-Matter-Air-Pollution-and-Cardiovascular-Brook-Rajagopalan/be8bb10cc06bb61c8447ed09d33418cde36a8eeb.

Brook, R., Zhang, H. and Sammut, J., 2023. Greater London Authority air quality exposure and inequalities study. https://www.london.gov.uk/sites/default/files/2023-06/Air%20quality%20exposure%20and%20inequalities%20study%20-%20part%20one%20-%20London%20analysis.pdf.

ClientEarth, 2021. What is environmental justice? https://www.clientearth.org/latest/news/what-is-environmental-justice/.

Colmer, J. M., Qin, S., Voorheis, J. L. and Walker, R., 2024. The changing nature of pollution, income, and environmental inequality in the United States. National Bureau of Economic Research, Working Paper 32060.

Costa, L. G., Cole, T. B., Dao, K., Chang, Y-C. and Garrick, J. M., 2019. Developmental impact of air pollution on brain function. Neurochemistry International, 31626830, https://pubmed.ncbi.nlm.nih.gov/31626830.

Currie, J., Voorheis, J. and Walker, R., 2023. What caused racial disparities in particulate exposure to fall? New evidence from the Clean Air Act and satellite-based measures of air quality. American Economic Review, 113(1), 71–97, https://doi.org/10.1257/aer.20191957.

Department for Environment, Food and Rural Affairs, 2024. Emissions of air pollutants in the UK – particulate matter (PM10 and PM2.5). https://www.gov.uk/government/statistics/emissions-of-air-pollutants/emissions-of-air-pollutants-in-the-uk-particulate-matter-pm10-and-pm25#:~:text=Domestic%20combustion%20is%20a%20major,closed%20stoves%20and%20open%20fires (accessed 15 November 2024).

Deryugina, T., Heutel, G., Miller, N. H., Molitor, D. and Reif, J., 2019. The mortality and medical costs of air pollution: evidence from changes in wind direction. American Economic Review, 109(12), 4178–219, https://doi.org/10.1257/aer.20180279.

European Environment Agency, 2021. Air pollution in Europe. https://www.eea.europa.eu/publications/air-quality-in-europe-2021.

Health Foundation, 2024. Inequality in likelihood of living in polluted neighbourhoods. https://www.health.org.uk/evidence-hub/our-surroundings/air-pollution/inequalities-in-likelihood-of-living-in-polluted#:~:text=Less%20than%201%25%20of%20people,with%2011%25%20of%20white%20people.

Institute for Health Metrics and Evaluation, 2021. GBD results. https://vizhub.healthdata.org/gbd-results/.

Klauber, H., Holub, F., Koch, N., Pestel, N., Ritter, N. and Rohlf, A., 2024. Killing prescriptions softly: low emission zones and child health from birth to school. American Economic Journal: Economic Policy, 16(2), 220–48, https://doi.org/10.1257/pol.20210729.

Liu, Q., Xu, C., Ji, G., Liu, H., Shao, W., … and Zhao, P., 2017. Effect of exposure to ambient PM2.5 pollution on the risk of respiratory tract diseases: a meta-analysis of cohort studies. Journal of Biomedical Research, 31(2), 130–42, https://doi.org/10.7555/JBR.31.20160071.

Mirza, H. and Warwick, R., 2022. Race and ethnicity. IFS Deaton Review of Inequalities, https://ifs.org.uk/publications/race-and-ethnicity.

Rüttenauer, T., Bader, F., Ehler, I. and Best, H. 2023. Breathing unequal air: environmental disadvantage and residential sorting of immigrant minorities in England and Germany. SocArXiv Papers, https://doi.org/10.31235/osf.io/snzbx.

World Health Organisation, 2021. WHO global air quality guidelines: particulate matter (‎PM2.5 and PM10)‎, ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide. https://www.who.int/publications/i/item/9789240034228.