In this piece I am going to look at a really interesting piece of longitudinal research which looks into the impact of a significant episode of air pollution that occurred in 1952 in London, UK. The study “Air pollution, foetal mortality, and long-term health: Evidence from the Great London Smog” was written by Alastair Ball in 2014 and was recently presented at the Royal Economic Society’s annual conference (1).


We know about some of the ill effects that exposure to air pollution has in childhood for example causing increases in respiratory illnesses. However the effects of air pollution on foetuses is much less well known. It is understandably pretty difficult to perform experimental trials on this population. Furthermore the outcome of any shock on the body at this important time developmentally could have impact throughout life rather than just an immediate effect. For these reasons this study gives a fascinating insight into the long-term effects of severe air pollution.

London Smog of 1952

In December 1952 a persistent weather system affected London for five days, this prevented the removal of atmospheric pollution. Particulate matter from traffic and coal burning, normally dispersed by the wind, was trapped. Thick ground level smog settled over the city, with visibility reduced to just a few metres. The weather continued for 5 days. There is good information on the pollution levels during the smog, with measurements of sulpur dioxide and particulate matter. Levels of pollution recorded increased to 300% of normal and were close to those experienced in modern day Beijing or Mumbai

The official advice was for people to stay at home, however people were used to high levels of smog at the time so initially there was no particular concern. However official data on deaths and hospitalisations that soon came to light revealed that something serious had happened. The ministry of health attributed 4000 excess deaths to the smog and it led to the passing of the Clean Air Act in 1956. This policy was drafted with the aim of reducing any further smogs.

Reason for the study

WHO estimated pollution caused an extra 3.7million premature deaths worldwide in 2012, with 88% of these in low and middle income countries (2). However there is little data on the effects of pollution in utero nor on long-lasting economic effects of pollution exposure. The goal of this study was to gather evidence about the long-term health effects of high pollution such as those currently found in low- and middle-income countries and relate the health effects to the effects on economic productivity. There is a known link between levels of ill-health and earnings.

The good quality data collected at the time of the London smog and data from the Office of National Statistics (ONS) allowed the authors to follow the exposed population over 60 years. This enables us to look at possible future effects of the high levels of pollution currently seen in many parts of the world both in terms of health and economic outcome.

Study design

The study split areas in London affected by the smog into areas of highest pollution and lower pollution (still very high). The most affected areas experienced an increase in combination of sulphur dioxide and particulate matter on average of 1800ug/m3, the less affected areas experienced increases of 800ug/m3. These areas were then compared to unaffected areas of England and Wales.

In addition two populations, 7000 people in total, were compared. The 1952 cohort, conceived in 1952 and exposed to the smog during pregnancy; and the 1953 cohort, conceived in 1953 and hence not exposed to the smog. Apart from the smog the two cohorts are exposed to other aggregate shocks at similar ages e.g all these people would have been exposed to post-war rationing of food. The average levels of pollution that these cohorts were exposed to during pregnancy are broadly similar as was the weather for the two years so the two cohorts are considered to be similar enough to be directly comparable.

Foetal deaths

This study used data from the historical record to estimate the number of stillbirths caused by the severe episode of air pollution in 1952. Foetal deaths often go unreported, they may not be noticed by mother if early in pregnancy or are just never reported to authorities. Therefore the study authors used a complementary identification process that uses geographic variation in pollution combined with expected data for numbers of live births to identify any “missing” children in the 9 months after the smog.


Survivors were identified by district and quarter of birth, and their health outcomes were observed at 50 and 60 years old. Health was measured by a self-reported questionnaire and the outcome between the two cohorts was compared. It is worth noting that although self reported health is a widely used metric in health economics, it may not capture all health affects so has to be relied on only with caution.

Education and employment attainment

Education level achieved and current employment status were also included in the questionnaire at age 50 and 60 years. These act as surrogate markers for economic effects of the pollution. People who have had negative health or developmental effects due to the pollution are likely to do less well at school and earn less through their working lives. Employment can be a good way of looking at health effects that do not show up on hospital or mortality records, it can show low level or chronic ill health that has had a negative effect on an individuals ability to work but not been significant enough to cause hospital admissions or death. Average population levels of unemployment were very similar for the two cohorts.


The study used data from the Office of National Statistics (ONS). This data connects information on individuals from multiple rounds of the census which occurs every 10 years in the UK. It is a survey based on a 1% sample of the population in England and Wales and is linked to data on births, migrations and deaths. The survey holds data on about 1 million people with linked data to 2001 and 2011 censuses for more than 400 000 individuals. Inclusion in the survey depends on being born on one of 4 dates in the year. The author used this information to identify and follow survivors and question them about their health, their educational achievement and their current employment status both 50 years and 60 years after the event.


This study draws two main conclusions from the findings. First, foetal exposure to pollution can lead to stillbirth, with only the stronger children likely to survive to be observed. Second, foetal exposure to pollution can cause damage to the child’s health, meaning that all individuals are likely to be less healthy than they would have been.

The author also summarizes his conclusions as showing that survivors are in general less healthy, less likely to have a formal qualifications and less likely to be employed than those unaffected by the smog.

Reduced foetal survival

A quarterly historical register of stillbirths collected between 1948 and 1964 showed a 2% increase against expected national trends in the 9 months after the smog. When this was verified by examining actual numbers of live births compared to expected numbers, to counteract the affect of under reporting of stillbirths, there was a reduction of approximately 4000 or 3% of live births for the affected areas.

Surprisingly those affected by the smog were 2% less likely to have died by ages 50 and 60 than comparative groups not affected by the smog. This was considered by the author to be a possible result of the survival of the fittest theory. “what doesn’t kill you makes you stronger”.

Worse health outcome

The second conclusion that the author draws becomes less intuitive for a variety of reasons which I will try to summarize. So we know that fewer children were born than we expected and we are attributing that to an increase in foetal deaths due to the pollution. Those who survived into adult life were more likely to be alive than comparative groups. However those observed at age 50 years were 2% more likely to report themselves to be in poor health than non-affected individuals. On face value this supports the conclusion that pollution leads to long-term ill health. The difference is lost by age 60 where there is no difference in the self-reported health of those exposed and not exposed. As these are the same individuals this can be explained by the unexposed “catching up” in terms of ill health by age 60. We all become less healthy as we get older. But why doesn’t the same “what doesn’t kill you makes you stronger” apply to all health affects rather than just death?

Education and employment

It gets more complicated when we also start looking at the results relating to education and employment status. Those affected by the smog were 3% less likely to have A level qualifications and 1% less likely to be in employment at 50y than unaffected peers. The study also found that affected people are 5% less likely to hold a degree, and men aged over 50 were 4% less likely to be in any kind of employment.

There was significant heterogeneity in these results, and again interestingly there is some counterintuitive data. Those from the low exposed areas group were more significantly affected by poor health at age 50 than those exposed in high areas. They were 3% more likely to be in ill health at age 50 and 4% more likely at age 60. They were also far less likely to hold qualifications than those from the ‘high” exposed areas. This affect is again attributed to the survival of the fittest theory. Those who were exposed to highest levels of pollution as foetuses and survived showed less marked affects of the pollution on their long-term outcome because they were much stronger to start with.

So we appear to have two groups of survivors, those who were exposed to high levels of pollution but if they survived were not as significantly affected as those who were exposed to lower levels of pollution, it is likely that more of whom survived to birth, but then showed more evidence of long-term health and labour market effects.

What we now need to know is whether there were higher percentages of foetal deaths in the areas with higher pollution. If this is the case then the results do seem to be explained by a “what doesn’t kill you makes you stronger effect”. Unfortunately there would be no way to work out which boroughs the “missing” infants would have been born in to be recorded and followed up.

So again what it looks like is that affects on foetal survival were less marked for those in “low pollution districts” of London. This led to a lower strong survivor response e.g lower still births, more babies born who were exposed and affected and hence more ill health affects in later life. Among this specific group they were 6% more likely to report themselves in poor health, 10% less likely to have an A level qualification, 4% less likely to be in work than non-affected individuals. Conversely in the “high pollution districts” there was a 3% reduction in reported ill health compared to unaffected population, they actually reported themselves as healthier. Consequently foetal mortality must have had a strong effect on the surviving population.


The conclusion I draw from this paper is that there does appear from the evidence presented to be an affect of exposure to severe level of pollution in utero. That being exposed to air pollution in utero not only causes an increase in stillbirths and childhood ill health but also causes significant problems later in life, and that these affects could have an impact on the economic productivity of the affected region. This is a really important finding given the levels of air pollution worldwide and particularly in certain cities with frequent extremely high levels of pollution. When you add in the potential economic effects of the health effects of pollution on the working population it is another stick to be used to lobby for better air pollution policy and control.


longitudinal research — A longitudinal study is an observational research method in which data is gathered for the same subjects repeatedly over a period of time. Longitudinal research projects can extend over years or even decades. In a longitudinal cohort study, the same individuals are observed over the study period.

shock — a sudden, unexpected event that potentially has physiological effects on the developing foetus

WHO — the World Health Organisation — International organisation whose primary role is to direct and coordinate international health within the United Nations’ system.

Office of National Statistics — The UK’s largest independent producer of official statistics and the recognised national statistical institute of the UK.

Cohort — a group of people with a shared characteristic.

A level qualifications — Advanced level — The later of two standardized tests in a secondary school subject, used as a qualification for entrance into a university.

heterogeneity — the quality or state of being heterogeneous; composition from dissimilar parts; disparateness.


1. Air pollution, foetal mortality, and long-term health: Evidence from the Great London Smog” was written by Alastair Ball in 2014. Accessed via URL: Date accessed 28th March 2016.

2. Burden of disease from Ambient Air Pollution for 2012. Accessed via URL: Date accessed 6th April 2016