Research on the long-term effects of pollution on mortality has been held back by lack of historical pollution data. In a new working paper, Walker Hanlon of NYU works around these limitations using details of London’s famous fog events.



London’s The Shard under heavy fog. Photo by By mattbuck [CC BY-SA 2.0], from Wikimedia Commons

For more than a century, London struggled with some of the worst air pollution on earth. Starting as early as the 17th century, there are reports of air pollution problems stemming from the burning of coal in the city. These problems worsened substantially with the onset of the Industrial Revolution, as industrialization, population growth, and rising income increased the amount of coal being used in the city. Only in the late 1950s, after the Great London Fog of 1952 killed several thousand Londoners, did the city take substantial steps to combat air pollution.


Today, cities around the developing world—from Beijing to Delhi and Mexico City to Jakarta—continue to struggle with high levels of air pollution. Yet most of what we know about the costs of pollution come from the United States and Europe, where air pollution levels today are relatively modest and health care relatively advanced. As policymakers consider the costs of urban air pollution, and how these costs evolve as cities develop, there are lessons to be learned from London’s long experience.


In my paper “London Fog: A Century of Pollution and Mortality, 1866-1965,” I study the impact of air pollution in London across a full century. This is the first study to look at the impact of air pollution on health across such a long span. Previously this has not been possible, mainly because consistent direct measures of air pollution are not available before the middle of the 20th century. To get around this, I develop a novel approach that involves tracking the occurrence of London’s famous fog events, which trapped pollution in the city, dramatically increasing exposure levels.


The formation of fog requires calm conditions, with little wind, often accompanied by temperature inversions (a layer of colder air trapped below a layer of warmer air). Under these conditions, the pollution constantly being generated in the city became trapped there, dramatically increasing pollution exposure. We can see this association clearly in the period after 1951, when direct measures of pollution become available.


In figure 1, I plot maximum and mean weekly levels of total suspended particulates (TSP), a measure of one important type of air pollution, against the number of fog days in each week. The graph covers 1951–52, the first two years for which direct pollution measures are available. It is clear that the fog events were associated with spikes in pollution levels, and that the highest levels of pollution seen in the city always corresponded to the occurrence of fogs. Because the timing of fog events depends on a complex interaction of climate conditions, including temperature, humidity, cloud cover, and wind speed, I argue that the occurrence of fog provides variation in pollution levels that is as good as random after controlling for underlying weather conditions.


Figure 1: Fog Days and Pollution Levels, 1951–52


To study how the elevated pollution exposure caused by fog events affected mortality patterns, I gathered weekly mortality data covering 1866 through 1965. This is a really unique data set in terms of the frequency of observations, the length of time covered, and the richness of detail (over 350,000 observations), which includes information about age and cause of death.


One result to emerge from my analysis is that the effect of elevated pollution exposure associated with fog events accounted for more than one out of every 200 deaths in London across the century from 1866–1965. This is a surprisingly large figure that puts the acute effects of air pollution on par with causes of death like suicide, smallpox, or diphtheria. Moreover, these deaths appear across all age groups. While today we often think of infants and the elderly as the main populations at risk of dying from pollution exposure, I find that many older children and prime-aged adults were also dying.


The effect of elevated pollution exposure associated with fog events accounted for more than one out of every 200 deaths in London across the century from 1866–1965.

Digging deeper into the data, I found that the pattern of effects across age groups can be explained by understanding how pollution was interacting with the infectious disease environment, particularly in the late 19th and early 20th century. Specifically, pollution exposure increased mortality from two infectious diseases of the respiratory system: tuberculosis (TB) and measles. The interaction of pollution with these diseases can explain why children and prime-aged adults were affected by pollution exposure in the setting that I study, and also why these groups may be less affected by the same level of exposure in a modern developed country. For example, measles is a highly contagious infectious disease of the respiratory system that mainly affects children after infancy.


The interaction between measles and pollution exposure drove the overall effect of pollution exposure on older children that I observe. Take measles away, and the effect of fog events on older children falls by more than half. Similarly, TB is a disease that impairs pulmonary function. My results suggest that exposure to air pollution increased mortality among those with TB. Since this disease was the main killer of prime-aged adults during the period I study, this interaction drove the overall effect of pollution on adults that I observe.


The interaction of acute pollution exposure with these infectious diseases has interesting implications. In particular, this interaction suggests that the costs of pollution are dependent on the underlying disease environment. As the prevalence of these diseases in London fell, this lowered the health costs of air pollution. For example, the reduction in the prevalence of measles in the interwar period, relative to the period from 1851 to 1911, led to a 18 percent reduction in the mortality associated with fog events.


These results tell us that pollution exposure will be more deadly in developing countries with higher infectious disease burdens. This also means that progress in improving underlying health conditions may be an effective way to reduce the health costs of pollution exposure, particularly given the continued prevalence of TB and measles in developing countries. The WHO reports that there were 1.3 million deaths due to TB in 2016 and just under 90,000 children died of measles. Reducing the prevalence of these diseases through, for example, expanding measles vaccinations or improving the treatment of TB may be as effective at curbing the health costs of pollution in some locations as direct emissions restrictions.


Walker Hanlon is an Assistant Professor at NYU Stern School of Business. He is an economic historian with a research focus on urbanization and the environment.


Disclaimer: The ProMarket blog is dedicated to discussing how competition tends to be subverted by special interests. The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty. For more information, please visit ProMarket Blog Policy.