Changes in Social Capital during the COVID-19 Pandemic: Empirical Evidence from Singapore and Switzerland

Social capital (SC), i.e., social ties that enable connections, trust, mutual aid, or collective action (Fraser and Aldrich 2021), has been identified widely as a dominant and crucial dimension of social resilience (Kwok et al. 2016; Saja et al. 2018). Both horizontal SC (i.e., ties between individuals of similar status and power within a community) and vertical SC (i.e., ties between individuals and decision makers or authorities) (Aida et al. 2009; Bhandari and Yasunobu 2009) at local and national levels are important for effective coping, response, and recovery during and after crises (Aldrich and Meyer 2015; Kwok et al. 2016).

The Coronavirus disease 2019 (COVID-19) led to the declaration of a global pandemic in March 2020, and almost all countries worldwide implemented social distancing measures such as travel restrictions, citywide lockdowns, extensive quarantines, school closures, and soft public health messaging (Onyeaka et al. 2021). The unprecedented levels and protracted duration of social distancing measures during the COVID-19 pandemic, which limited physical interactions and traditional channels of social support, raise the question of whether these measures may have triggered a decrease in SC (Pitas and Ehmer 2020). Understanding changes in SC due to social distancing measures is crucial for assessing the consequences of the COVID-19 pandemic and gaining insights into the potential consequences of future pandemics (Sun et al. 2022). Such insights are important because a decrease in SC might decrease a country’s social resilience by reducing compliance with mandates (Barrios et al. 2021), disrupting information dissemination, increasing frequent risky behaviors, or diminishing capacities for recovery or response in future crises (Kwok et al. 2016).

However, few studies have analyzed changes in SC over time while social distancing measures were in place. Although SC changes in the contexts of natural disasters (Saja et al. 2018; Layla et al. 2020) or financial crises (Lindström and Giordano 2016) have been well studied, there is limited applicability of these studies to the COVID-19 pandemic context (Busic and Schubert 2022) because they typically do not consider limitations to social interactions between individuals. An established framework is lacking for assessing SC in pandemic contexts (Layla et al. 2020) with prolonged limits on physical interactions. Existing studies of SC during the COVID-19 pandemic largely have been limited to single SC indicators (Borkowska and Laurence 2021; Yu et al. 2021; Tatarko et al. 2022) or specific segments of populations (e.g., only students, workers, or the elderly) (Yu et al. 2021; Saraniemi et al. 2022; Sato et al. 2022). Some studies have shown that societies with higher collectivistic orientation (CO) had higher compliances with social-distancing rules and better health outcomes during the COVID-19 pandemic (Biddlestone et al. 2020; Shekriladze et al. 2021). However, the relationship between SC changes during a pandemic crisis and the levels of collectivistic orientation (Beilmann and Realo 2012) rarely have been evaluated.

In summary, there is an evidence gap on changes in both horizontal and vertical SC over time during a pandemic for representative population samples as well as for groups with differing levels of collectivistic orientation. This study closes these gaps. Changes in SC were assessed using two waves of cross-sectional surveys during the COVID-19 pandemic, administered to adult representative samples in two countries, i.e., Singapore and Switzerland. These two chosen are characterized by similar population size, income conditions, high quality of education, and similar economic and political stability (Guo and Woo 2016), but different pandemic-related interventions. The analyses for each country comprised three parts. First, changes in horizontal SC (between residents) were assessed; second, changes in vertical SC (between residents and leaders) were assessed; and third, horizontal and vertical SC changes for groups with different levels of collectivistic orientations were investigated.

We first analyzed changes in overall SC from Wave 1 to Wave 2. Fig. 1 shows a significant increase in the overall horizontal and vertical SC for Singapore ($p<0.001$ for all) during the pandemic (between June 2020 and July 2021), with social distancing measures in place. In Switzerland, the horizontal SC mean score increased significantly ($p<0.001$), but the vertical SC had no significant changes ($p=0.44$). However, the magnitudes of the significant SC increases in Singapore, and of the significant horizontal SC increase in Switzerland, were modest (Table S1).

Second, we analyzed changes in the main SC indicators. For Singapore, the horizontal SC main indicators, including social cohesion, social network, and social support, increased significantly during the pandemic (between June 2020 and July 2021), despite social distancing measures being in place ($p<0.001$ for all) (Fig. 2). Similarly, the vertical SC main indicators, including ties with government leaders and ties with local leaders, increased significantly in Singapore during the pandemic ($p<0.001$ for all), in spite of the prevalent social distancing measures (Table S1).

In Switzerland, the horizontal SC main indicators, including social cohesion, social network, and social support, increased significantly during the pandemic (between June 2020 and July 2021), in parallel with social distancing measures in place ($p<0.001$ for all) (Fig. 3). However, similar to Singapore, the significant changes in horizontal SC main indicators were modest (Table S1). The vertical SC main indicators, including ties with government leaders ($p=0.55$) and ties with local leaders ($p=0.47$), did not change significantly during the pandemic in Switzerland.

Thirdly, we analyzed the effects of citizens’ collectivistic orientation on SC changes. Both Singapore’s and Switzerland’s respondents with higher collectivistic orientation had a significantly higher mean score for overall horizontal and vertical SC compared with those with lower collectivistic orientation ($p<0.05$ for all) (Fig. 4). Singaporean respondents with either low or high collectivistic orientation had significant positive changes (i.e., increases) in both horizontal SC and vertical SC ($p<0.05$ for all) in the presence of social distancing measures (Table S2). Swiss respondents with either low or high collectivistic orientation also had significant positive changes (i.e., increases) in horizontal SC ($p<0.001$ for all), but there were no significant changes in vertical SC in both collectivistic orientation groups ($p\ge 0.05$ for all).

Furthermore, we found for Singapore a significantly larger increase in vertical SC among respondents with low collectivistic orientation compared with those with high collectivistic orientation ($p\text{-}\mathrm{interaction}<0.001$) [Fig. 4(b)]. However, the extent of changes in horizontal SC in Singapore did not vary by the levels of collectivistic orientation ($p\text{-}\mathrm{interaction}=0.85$). In Switzerland, the changes in overall horizontal SC and vertical SC did not differ significantly with the levels of collectivistic orientation ($p\text{-}\mathrm{interaction}\ge 0.05$ for all) (Table S2).

We further assessed the effects of sociodemographic variables on horizontal and vertical SC changes. In Singapore, changes in horizontal SC and vertical SC did not differ between gender and ethnicity groups ($p\text{-}\mathrm{interaction}\ge 0.05$ for all). However, there were differences, although modest, between age groups for horizontal SC ($p\text{-}\mathrm{interaction}=0.0011$) and vertical SC changes ($p\text{-}\mathrm{interaction}=0.012$). All age groups had significant increases in horizontal SC and vertical SC, except the age group 50–59 years with respect to horizontal SC changes and the age group 40–49 years with respect to vertical SC changes (Table S3). In Switzerland, horizontal and vertical SC changes did not differ significantly between gender or age groups ($p\text{-}\mathrm{interaction}\ge 0.05$ for all). Horizontal SC changes also did not differ between language groups ($p\text{-}\mathrm{interaction}=0.51$), but there were modest differences between language groups with respect to vertical SC changes ($p\text{-}\mathrm{interaction}=0.011$). Italian speakers had a small but significant increase in vertical SC, but no significant changes were observed for other language groups ($p\ge 0.05$ for all) (Table S3).

Key findings from this study are that horizontal and vertical SC remained stable in Singapore and Switzerland during the pandemic crisis (between June 2020 and July 2021), even while social distancing measures were in place. In both countries, overall horizontal SC and all horizontal SC main indicators had small but significant increases. Overall vertical SC and all vertical SC main indicators increased significantly and modestly in Singapore, but did not undergo significant changes in Switzerland.

The finding that overall there were no losses in SC in the presence of social distancing interventions in both countries is an encouraging result. The ability of a country to retain stable SC levels may be seen as an important insurance in the event of similar future pandemics, which might again require social distancing interventions (Bartscher et al. 2021). Although the findings of this study do not provide causal evidence, we offer a few plausible explanations for these findings, which may be explored further in future studies.

First, the increase in horizontal SC and horizontal SC main indicators, including social cohesion, social support, and social networks may have resulted partly from experiences of isolation and stress due to social distancing mandates during the pandemic. Individuals may have become more aware of the importance of mutual support, more worried about people in their networks, or more supportive of one another (e.g., checking on sick neighbors or making phone calls to friends to minimize their feelings of isolation) (Bartscher et al. 2021; Tatarko et al. 2022). Furthermore, it has been suggested that communities comprising individuals who are innovative in creating new ways or modes of collaboration to stay well connected could better share information or resources (Aldrich 2010), and hence provide more-reliable or -adequate assistance to group members (Fraser and Aldrich 2021). Social media platforms have been reported to maintain social support systems when physical contact is challenged, through the expansion or activation of individuals’ social networks beyond existing networks and the linkage of resources from local and central governments to community members (Page-Tan 2021; Samutachak et al. 2023). The significant increases in overall horizontal SC in both Singapore and Switzerland thus may reflect the abilities of individuals to invent and adapt to new ways of expediting collective action, connection, and interaction with their group members or leaders during a pandemic, when social distancing rules limit physical contact. The stable horizontal SC in Singapore and Switzerland found in this study concurs with findings from a cross-sectional study of 10,540 Chinese youths (Yu et al. 2021). Conversely, a cross-sectional study of 500 Russian adult respondents reported a decrease in social cohesion with distancing measures (Tatarko et al. 2022). However, the explanatory power of both the Chinese and Russian studies may be limited because the assessment of SC was implemented only at a single point in time.

Second, consistent across Singapore and Switzerland, vertical SC remained stable when social distancing measures were in place during the COVID-19 pandemic. Whereas vertical SC remained unchanged in Switzerland, it increased significantly (although modestly) in Singapore. The stability of vertical SC in the two countries may be attributed to context-specific reasons. The findings in Singapore may be attributed partly to the consistent efforts of local and central government leaders to maintain regular contact with residents to convey relevant, accurate information on the latest pandemic insights and to debunk misinformation. Communication strategies for social distancing interventions, focusing on social responsibility (Tan and Chua 2022) further sought to align with the populations’ values (Yip et al. 2021). Thus, measures in Singapore were designed not only to be congruent with recommendations of recognized international health organizations or scientific evidence (Chua et al. 2020; Wong and Kohler 2020), but also with the country’s value system or social norms (Yip et al. 2021). Such efforts may have supported the pre-empting of policy shifts, and restored vertical trust in the legitimacy of national or local responses (Wong and Kohler 2020; Lim et al. 2021). In addition, respect for government leaders and regulations is high in Singapore (Chua et al. 2020), and residents hold expectations that the government will make decisive and adequate decisions to protect the citizens (Abdullah and Kim 2020). Thus, despite more stringent measures such as strict rules and tough penalties, a balance between legitimacy, authority, and trust could be maintained (Woo 2021), and hence may have preserved the social contract underlying the relationship between citizens and government. Taken together, these factors may have prevented a loss of vertical SC in Singapore during the pandemic.

In Switzerland, the federal government assigned a high degree of autonomy to cantonal leaders, allowing for responses tailored to specific local conditions (Pleninger et al. 2022). Communication of distancing measures largely focused on recommendations, and less on prohibitive mandates (Chen et al. 2021a; Pleninger et al. 2022). Such measures placed greater emphasis on more-persuasive instruments and voluntary compliance than on formal interdictions, which seem to be aligned with Switzerland’s value system and social contract (Sager and Mavrot 2020; Zimmermann et al. 2023). The stable vertical trust levels in both countries may further reflect beliefs in leaders to provide adequate protection, unbiased information, and scientific recommendations (SteelFisher et al. 2023). Thus, social distancing measures that were evidence-based (SteelFisher et al. 2023) and considered the respective countries’ value systems (Siegrist and Bearth 2021) or social contract framework conditions (Chin et al. 2023) may have supported the stability of vertical SC in a pandemic context.

Third, the results in this study point to higher horizontal and vertical SC levels among citizens with higher collectivistic orientation, compared with those with lower collectivistic orientation. This finding relates to studies suggesting that features underlying a higher collectivistic orientation, such as residents’ interdependencies (including shared goals) or in-group solidarity (including respect for authority) are linked to higher levels of horizontal trust (Shin and Park 2004; Pathak and Muralidharan 2016) and vertical trust (Jakubanecs et al. 2018; Leonhardt et al. 2020). Because higher trust levels reflect higher SC, higher collectivistic orientations that are expressed through shared goals, and support for and trust in each other, may result in higher levels of horizontal and vertical SC.

Furthermore, changes in SC were found to vary between groups with different levels of collectivistic orientation. In Singapore, individuals with lower collectivistic orientation had a significantly larger increase in vertical SC, compared with those with higher collectivistic orientation. However, a similar effect was not found in Switzerland, where changes in SC did not differ between groups with different levels of collectivistic orientation. It may be reasonable to speculate that in Singapore, the pandemic crisis and social distancing measures may have amplified or caused individuals with lower collectivistic orientation and possibly less community support to rely more on authorities for support and protection (Stevens and Haines 2020). This amplified reliance may have enhanced the exposure of individuals with low collectivistic orientation to government agencies or effective local services, thereby promoting a large and significant increase in vertical SC.

Our findings indicate that the links between collectivistic orientation and SC changes are multifaceted, differ by nations and contexts, and should be investigated in further detail. Modest differences in SC changes between age groups in Singapore, and in vertical SC changes between language groups in Switzerland, further suggest that changes in SC may depend on different sociodemographic characteristics across nations and contexts. The use of social archetypes to identify patterns of value configurations among residents in specific contexts may be explored to uncover further endogenous heterogeneity between population groups, even beyond single standard sociodemographic variables. Based on the respective insights, tailored interventions to cushion potential negative impacts of future social distancing measures on SC may be conceived.

There are limitations to this study. Changes in SC were assessed with repeated cross-sectional surveys during the pandemic (in June 2020 and July 2021) in Singapore and Switzerland. Thus, the findings are correlational. Reverse causation cannot be ruled out, and dynamic trends in SC within the study period cannot be captured. Longitudinal (panel) studies would be needed to confirm our results and to assess long-term trends in SC changes.

The use of online surveys may have introduced a sampling bias into the study. For example, it may be assumed that respondents with high digital literacy were overrepresented. However, this survey method offered efficient access and the possibility to collect demographically representative survey data on a national level. This was particularly valuable because our surveys were conducted in the pandemic context with social distancing measures in place. Hence, the use of other traditional survey methods was hardly feasible.

Furthermore, an evaluation of causal factors contributing to SC stability during pandemic contexts is needed. Although measures of SC in this study focused on the most widely cited indicators (Saja et al. 2018), we cannot exclude the possibility that other structural-level factors (e.g., the effectiveness of specific civic organizations, the existence of social welfare programs, or the economic framework conditions) may have influenced the stability of SC indicators. Future studies should analyze these factors to arrive at a more comprehensive understanding of changes in SC in the presence of social distancing rules.

The finding that horizontal and vertical SC remained stable in two different countries, in spite of the presence of social distancing rules, is encouraging, and potentially may suggest sustained social resilience during a pandemic. Evidence-based social distancing interventions tailored to the value system of a country and the innovative, adaptive capacities of citizens may have contributed to these positive results. Policymakers and relevant agencies working in social resilience potentially could benefit from increasing tailored communication and messaging efforts in future pandemic contexts. Further studies are warranted to determine the causal pathways responsible for retaining SC in pandemic settings.

All data, models, or codes that support the findings of this study are available from the corresponding author upon reasonable request.

This research was funded by the National Research Foundation, Prime Minister’s Office, Singapore under its Campus for Research Excellence and Technological Enterprise (CREATE) program (Project No. 2019-N-162). All authors were involved in the production and writing of the manuscript. The authors thank the Home Team Behavioural Sciences Centre Singapore for their contributions to the study design, and Natalia Borzino and Ante Busic-Sontic, who contributed to the study design and data collection efforts.