Gaston Yalonetzky, Associate Professor of Economics at the Leeds University Business School, discusses the vital role data plays when looking at the impact of covid-19 on mental health in the UK.
We can make all sorts of reasonably sounding claims about the economy or society, over a discussion at the pub or the dinner table. Some of them may even have profound policy implications. But it is another thing to find for or against them with empirical evidence. And for that we need appropriate data.
For instance, most people will probably think that the covid-19 pandemic, particularly the first lockdown period, had detrimental mental health effects in the UK. For some, this judgment may even reflect personal experience. But how can we know for sure whether this mental effect was significant and widespread, as opposed to being anecdotal at best, perhaps? Thankfully, the United Kingdom’s Longitudinal Household Survey (UKLHS, aka “Understanding Society”) has been asking people questions on mental health every year for decades now. Moreover, the Institute for Social and Economic Research in charge of the UKLHS, undertook the momentous collection of a special survey during the first months of the pandemic. Thus, we have information on British survey participants before and after the first lockdown was implemented in late March 2020.
A handful of studies (e.g., Serrano-Alarcon et al., 2021; Davillas and Jones, 2021; Bonomi et al., 2021), including ours (Anaya et al., 2023, 2024), benefitted from this data collection effort in order to test and conclude that, indeed, the covid-19 pandemic had a sizable detrimental impact on people’s mental health in the UK; thereby confirming, on this occasion, most people’s intuitive concerns.
For instance, Serrano-Alarcon et al. were interested in identifying the mental health impact of containment measures from other potential sources of distress such as fear of contagion. Cleverly, they took advantage of the nearly two-weeks difference between England and Scotland in the date of lockdown relaxation during May 2020. Comparing the change in mental health during that period between the two populations, they convincingly conclude that lifting the lockdown improved mental health.
Meanwhile, in our study we were worried about potential confounding factors hindering a reliable estimation of the magnitude of the mental health fallout from the first pandemic wave. Specifically, we wanted to neutralise any season effects, as the first wave, and concomitant containment measures, occurred between the end of winter and start of spring. In fact, the first countrywide lockdown was instituted very close to the spring equinox, on March the 23rd, 2020. Hence, with the help of the UKLHS we were able to compare the difference in mental health of a “treatment group” comprising people observed soon before and after the lockdown date, with the respective difference for a “control group” involving people observed soon before and after March the 23rd 2019.
Using the 12-item version of the General Health Questionnaire (GHQ-12), a popular measure of mental health, we found that the first pandemic wave led to a statistically significant average deterioration in mental health. Moreover, the impact was worse among women, migrants, people in BAME groups, and the young. Digging further into other moderating factors, we also found that people reporting financial distress or loneliness, as well as those living in overcrowded homes, suffered a greater mental health fallout. Since the GHQ-12 does not have an intuitive measurement scale, we compared these impacts to those of other life shocks documented in the literature and we noted, for instance, that the average mental health impact among the population is equivalent to a sizable share of the estimated disutility associated with unemployment. Likewise, the average mental health impact is significantly greater than the estimated mental health toll from divorce or widowhood.
We followed up our work by looking into the moderating role of neighbourhood and outdoor dwelling characteristics such as private or shared garden, rooftop, terrace or balcony. This time we merged the UKLHS with two datasets containing information from so-called lower-level super output areas (LSOA) in England and Wales in order to match individuals in the UKLHS with levels of deprivation over several wellbeing dimensions (e.g., income, education, health, etc.) in their place of residence. Likewise, we related these same individuals to average proximities to bodies of water (“blue areas”) and green areas from their place of residence. Among several interesting results, we found that the mental health toll from the first pandemic wave was worse among people living in more socially and economically deprived neighbourhoods. We also found a statistically significant impact of “blue” but not green areas (whereby people nearer bodies of water suffered less mentally).
We also found worse mental health impact among people living without outdoor dwelling space. Perhaps intriguingly, the distribution of outdoor dwelling characteristics partly explains the moderating role of neighbourhood deprivation. By contrast, controlling for socio-demographic traits of respondents (such as sex, age, etc.) does not make a dent on the moderating role of spatial characteristics, which led us to conclude that the mental health burden associated with the pandemic in the UK was significantly shaped by where we live.
We may understand why some people might show fatigue regarding new work on the covid-19 pandemic as it wasn’t particularly a pleasant experience for most and we all wanted to move on with our lives as soon as the emergency subsided. However, sadly, we cannot fully prevent and rule out the onset of new pandemics in the future. Therefore, work documenting the benefits and costs of measures deployed to stem a pandemic (anything from vaccines themselves to lockdown measures, masks and furlough schemes) remains as relevant as it will ever be. And in order to conduct the required research, there’s no proper alternative to top-notch data such as the UKLHS.
About the author
Gaston Yalonetzky is Associate Professor of Economics at the Leeds University Business School, Research Associate at the Oxford Poverty and Human Development Initiative and Visiting Senior Fellow at the LSE’s International Inequalities Institute. He works on distributional analysis and the measurement of wellbeing, with a keen interest in the human and social development processes.