Until a vaccine or effective pharmacological treatments for COVID-19 become widely available, individuals can protect themselves and others around them through a combination of social distancing and behavioural changes that reduce transmission when contacts occur. The scientific community has highlighted the importance of limiting social contacts and governments worldwide have either advised or mandated sheltering-in-place orders to promote behavioural changes (Hartl et al. 2020). However, social distancing has major economic and social consequences for both individuals and communities (Glover et al. 2020, Coibion et al. 2020). Therefore, rather than a prolonged period of strict distancing until treatment or immunisation make social gatherings safe, coexisting with the virus is most likely to entail the monitoring of local epidemiological curves and by a sequence of tighter and looser restrictions moving in sync with such curves. It is therefore important to identify what factors are associated with individuals and communities being able to alter their behaviour either voluntarily or in compliance with regulations.
Evidence from the H1N1 pandemic reveals that the likelihood that individuals implemented behavioural changes in 2009 to reduce H1H1 transmission depended, among other things, on their social capital (Chuang et al. 2015; Rönnerstrand 2013, 2014). Social capital reflects the resources and benefits that individuals and groups acquire through connections with others and involves both shared norms and values that promote cooperation as well as actual social relationships (Kawachi et al. 2008).
Evidence on the role of social capital in shaping behavioural responses during the COVID-19 pandemic is beginning to emerge (see Borgonovi and Andrieu 2020, Bargain and Aminjonov 2020, Durante et al. 2020). We contribute to this line of work by examining data from US counties to identify how communities with different levels of social capital responded to the threat posed by COVID-19 by changing one type of behaviour: mobility. We examine mobility patterns between 17 February, when there was little awareness of the public health threat posed by COVID-19 among the general population in the US, and 10 May, by which point as many as 1.3 million individuals had been diagnosed with COVID-19 and about 80 thousand were reported to have died because of COVID-19.
Data on county-level social capital come from the “The geography of social capital” project and mobility data come from the Cuebiq's Mobility Index (CMI). Data from Cuebiq has been used to map movements in Italian provinces prior to and following the implementation of restrictions to movement because of COVID-19.
Figure 1 reveals that both communities with high levels of social capital (counties in top quarter of the social capital index) and communities with low levels of social capital (counties in bottom quarter of the index) changed their mobility patterns between 17 February and 10 May. Mobility was already lower at baseline in areas with higher levels of social capital and decreased markedly in such areas from the week starting on 9 March onwards. Such communities reduced mobility sooner than communities with lower levels of social capital, i.e. when reducing mobility was not yet legally mandated. These communities also reduced mobility to a much greater extent than communities with lower levels of social capital. Mobility returned close to baseline in both high and low social capital communities by 10 May, although observing mobility levels in mid-May that are comparable to those observed in February should be viewed as a form of behavioural change: mobility tends to be greater when the days are longer and weather conditions are more favourable to outdoor activities.
Figure 1 Changes in mobility in US counties in high and low social capital communities
Sources: Cuebiq mobility data and “The geography of social capital” project.
Results presented in Figure 1 do not consider differences across counties with different levels of social capital which may also shape behavioural responses during the pandemic. In Figure 2, we highlight the weekly evolution in mobility differentials across counties that are associated with a one standard deviation difference in social capital after accounting for county-level characteristics (weather conditions, the number of COVID-19 cases diagnosed the week before, if the county was subject to shelter-in-place orders in the week under analysis, the percentage of votes cast that were in favour of Trump in the 2016 presidential elections, the primary economic activity of the county, population density, the share of people over 65, and the share of people living in poverty) as well as state fixed effects.
Figure 2 Social capital differentials in mobility between 17 February and 10 May in the US
Notes: The balanced panel covers 2715 counties which account for 96% of the total US population in 2018. All models are based on Cuebiq Mobility trends data. Dependent variable: mobility index. The figure highlights the change in mobility associated with a one standard deviation difference in social capital while accounting for weather conditions (precipitations), the number of covid-19 cases diagnosed the week before, if the county was subject to shelter-in-place orders in the week, the percentage of votes cast that were in favour of Trump in the 2016 presidential elections, the economic dependence of the county (reference category undifferentiated economic activity), population density, the share of people over 65, and the share of people in poverty.
Figure 2 illustrates the average within state, between county variation in mobility between 17 February and 10 May that is associated with a difference of one standard deviation in social capital. It indicates that the social capital differential in mobility became very large between March and 13 April and decreased from 13 April on but remained statistically significant. This result should be considered in light of Figure 1: at the start of the study period the differential occurred while levels of mobility were declining while at the end of the study period the differential occurred when overall mobility levels were increasing. Crucially, the social capital differential grew before shelter-in-place orders were implemented.
Because behavioural changes in the early stages of COVID-19 are crucial to halt the spread of the virus before cases start to rise exponentially (Dave et al. 2020) these findings are particularly relevant. They also indicate that although sheltering-in-place orders can effectively alter individual behaviours across all communities, they are especially effective in communities with high levels of social capital.
In order to evaluate the likely risk different communities face because of COVID-19 we combine information on community-level social capital and the prevalence of chronic conditions such as diabetes, obesity, high blood pressure, lung disease, and heart disease in the population. Individuals with pre-existing conditions are in fact more likely to face complications if infected with COVID-19 (Jordan 2020). We consider counties with a high prevalence of chronic conditions and low levels of social capital to be very vulnerable while counties with a low prevalence of chronic conditions and high levels of social capital to have low levels of vulnerability. Figure 3 indicate that many counties, particularly in the Southeast, face very high levels of vulnerability, combining high rates of chronic conditions and low levels of social capital.
Figure 3 Geographical disparities in Covid-19 vulnerability
Sources: Author’s calculations based on data from www.policymap.com (risk index) and “The geography of social capital” (social capital index).
Our findings may be important not only to evaluate what happened in the early phase of the COVID-19 pandemic in the US, but also to consider where efforts should be put as legal barriers to the SARS-CoV-2 virus are relaxed. Governments around the world have lifted some of the restrictions to movement that were implemented in February and March 2020 when the number of COVID-19 cases rapidly increased. Our work suggests that the stock of social capital in a community is a crucial factor. Reinforcing the social capital available in a community when this is present and supporting communities when social capital is lacking should be just as much of a priority as sourcing stocks of face masks or testing kits to protect population health.
Bargain, O and U Aminjonov (2020), “Trust and Compliance to Public Health, Policies in Times of COVID-19”, IZA DP No. 13205.
Borgonovi, F and E Andrieu (2020), “Bowling together by bowling alone: Social capital and Covid-19”, Covid Economics: Vetted and Real-Time Papers 17.
Chuang, Y C, Y L Huang, K C Tseng, C H Yen and L H Yang (2015), “Social capital and health-protective behavior intentions in an influenza pandemic”, PLoS ONE 10(4).
Coibion, O, Y Gorodnichenko and M Weber (2020), “Labor markets during the Covid-19 crisis: A preliminary view”, Covid Economics: Vetted and Real-Time Papers 21.
Dave, D M, A I Friedson, K Matsuzawa and J J Sabia (2020), “When do shelter-in-place orders fight COVID-19 best? Policy heterogeneity across states and adoption time”, NBER Working Paper No. 27091.
Durante, R, L Guiso and G Gulino (2020), “Civic capital and social distancing: Evidence from Italians’ response to COVID-19”, VoxEU.org, 16 April.
Glover, A, J Heathcote, D Krueger and J V Ríos-Rull (2020), “Health versus Wealth: On the Distributional Effects of Controlling a Pandemic”, Covid Economics: Vetted and Real-Time Papers 6.
Hartl, T, K Wälde and E Weber (2020), “Measuring the impact of the German public shutdown on the spread of Covid-19”, Covid Economics: Vetted and Real-Time Papers 1.
Kawachi, I, S V Subramanian and D Kim (2008), “Social capital and health: A decade of progress and beyond”, In I Kawachi, S Subramanian and D Kim (eds.), Social Capital and Health (pp. 1–26), New York: Springer.
Jordan, R E (2020), “Covid-19: risk factors for severe disease and death”, BMJ 368:m1198.
Rönnerstrand, B (2014), “Social capital and immunization against the 2009 A(H1N1) pandemic in the American States”, Public Health 128(8): 709–715.
Rönnerstrand, B (2013), “Social capital and immunisation against the 2009 A(H1N1) pandemic in Sweden”, Scandinavian Journal of Public Health 41(8): 853–859.