Individualism and the fight against COVID-19

The study investigates how societal culture—specifically the individualism–collectivism dimension—affects the containment of COVID-19. The authors argue that in more individualistic societies, governments may be more hesitant to impose restrictions that curtail personal freedom, and citizens may be less likely to comply with such measures. This combination can hinder effective pandemic control, leading to more severe health outcomes. Using global data from 2020–2021 and a within-country analysis of Germany, the paper examines whether individualism is associated with higher COVID-19 infections and deaths and explores mechanisms through policy adoption and adherence.

Theoretical foundations draw on Hofstede (1980, 2001) and Schwartz (1990) to define individualism/collectivism at societal levels, highlighting tensions between personal freedoms and social obligations. Triandis (1996) and others conceptualize individualism/collectivism as psychological syndromes that shape adherence to norms and authority. Prior work links collectivism with acceptance of authority and restrictions (e.g., Hofstede 1980; Gelfand et al. 1996; Kemmelmeier et al. 2003; Conway et al. 2006). In the COVID-19 context, earlier research documents the efficacy of NPIs (e.g., Tian et al. 2020; Chinazzi et al. 2020; Mitze et al. 2020) and suggests cultural influences on behaviors (masking, distancing) and policy timing/diffusion. The German case literature shows persistent cultural differences post-reunification affecting attitudes toward redistribution and authority (Alesina & Fuchs-Schündeln 2007; Schwartz & Bardi 1997; Schmelz 2021), motivating a natural experiment on East vs. West districts.

International analysis (2020–2021):

• Outcomes (country-day): Confirmed Cases = log(1 + cumulative cases per 10,000); Deaths = log(1 + cumulative deaths per 10,000); Death Ratio = deaths/confirmed cases (%). Data: Our World in Data.
• Key regressor: Individualism index from World Values Survey (emancipative values, education, social movement activity).
• Controls: Land_area; Population_age65; GDP_growth; Export; Consumption; Alcohol; Arrival (tourists/population); Refugee; Internet; Broadband; UHC service coverage; Healthsecurity (2019 GHS Index); Directflight_WH; SARS_experience (2003). Pandemic controls: Test (log(1 + cumulative tests per 10,000)) with Test_missing; Vaccination (log(1 + cumulative vaccinations per 10,000)) with Vaccination_missing; Relativeday (days since first case) and Relativeday². Continent fixed effects.
• Instrumental variables: 2SLS using Nobel_Prize (cumulative Nobel Prize winners to 2019) as instrument for Individualism; alternative instrument uses Nobel_Prize_LP (Literature and Peace prizes only) to avoid potential endogeneity from science/medicine/economics prizes. Analyses conducted on four reference dates (Apr 15, Oct 15 in 2020 and 2021) and daily IV regressions covering Jan 1, 2020–Dec 31, 2021.
• Policy adoption: Ordered logit regressions for eight NPIs (school closing; workplace closing; cancelling public events; restrictions on gatherings; transport closure; stay-at-home requirement; restrictions on internal movement; international travel controls), regressed on Individualism, Confirmed Cases, Vaccination, testing variables, national controls, and continent fixed effects, over 2020–2021.
• Policy effectiveness: Country fixed-effects panel regressions of outcomes on Policy intensity, LowIndividualism (indicator for below-median Individualism), their interaction (LowIndividualism·Policy), vaccination/testing controls, Relativeday terms; Jan 1, 2020–Dec 31, 2021.

Within-Germany natural experiment (399 districts, Berlin excluded):

• Outcomes (district-day): same definitions for Confirmed Cases, Deaths, Death Ratio. Data: Robert Koch Institute (RKI)/Kaggle.
• Key regressor: East (1 if former East Germany district, 0 otherwise). Subsample/border analyses use distance to former border and bands (on border; within 50/100/150 km) from Federal Agency for Cartography and Geodesy.
• Controls: Relativeday and Relativeday² (district-specific first case timing); Vaccination (fully vaccinated per 100 people; RKI/GitHub; 2021); district/state characteristics from Statistisches Bundesamt: Land_area (log km²), Population_age65, GRP_growth, Export (% GRP), Healthcare (per capita public assistance), Migration (arrivals per 100), Serviceperson (employment in services per 100), Student (students per 100), Educationexp (higher education expenditure per capita), Cinema (visits per capita), Residentialbuilding (per capita), Passenger (bus/rail passengers per capita), Accommodation (arrivals per capita), Rail (track length per 10,000 people), Police (per 100). Cross-sectional regressions at reference dates and daily regressions with controls over 2020–2021.

• International IV results: Instrument relevance holds (Nobel_Prize positively predicts Individualism; first-stage F around 3.3–5.6). Second stage shows Individualism positively and significantly associated with Confirmed Cases and Deaths across Apr 15, 2020; Oct 15, 2020; Apr 15, 2021; Oct 15, 2021. Example coefficients (z/t-stats):
  • Confirmed Cases: 0.145 (2.51), 0.323 (2.14), 0.177 (2.72), 0.123 (2.17).
  • Deaths: 0.082 (2.59), 0.173 (2.05), similar significance across dates.
  • Death Ratio: coefficients mostly positive but statistically insignificant across dates.
  • Daily IV coefficients (2020–2021) are predominantly positive and significant for cases and deaths; largely insignificant (but positive) for death ratios.
• Policy adoption: Across 2020–2021, Individualism does not systematically predict the timing/stringency of the eight NPIs (daily ordered logit coefficients largely insignificant; brief early-2020 slowdown in workplace closures in more individualistic countries).
• Policy effectiveness: In high-individualism countries, Policy coefficients are often positive (NPIs less effective). Interaction terms LowIndividualism·Policy are negative and significant for Confirmed Cases across all 8 policies and for Deaths across most policies (e.g., school closing −0.176 (−3.87), workplace closing −0.310 (−5.12), cancelling public events −0.298 (−4.15); for Deaths: workplace closing −0.123 (−3.28), cancelling public events −0.148 (−3.09), stay-at-home −0.157 (−4.35)). Death Ratio interactions mostly negative, with some significant (e.g., school closing −0.490 (−2.21), restrictions on internal movement −0.403 (−1.99)). This indicates NPIs are effective primarily in low-individualism contexts.
• Germany natural experiment: Without controls, East districts show less severe COVID-19 in 2020 but more severe in 2021. After controlling for vaccination and extensive district characteristics, East coefficients are negative and significant on all four reference dates for Confirmed Cases, Deaths, and largely for Death Ratio, indicating former East German districts had less severe outcomes throughout 2020–2021. Daily controlled regressions show consistently negative and significant effects of East on cases and deaths. East districts also exhibited lower vaccination rates in 2021, highlighting the importance of controlling for vaccination when assessing cultural effects.

Findings support the hypothesis that individualism hampers COVID-19 containment primarily by reducing adherence to government-initiated non-pharmaceutical interventions rather than by preventing their adoption. Internationally, higher individualism predicts more cases and deaths across the pandemic trajectory, and NPIs are less effective where individualism is higher. Within Germany, leveraging historical cultural divergence, former East (more collectivist) districts experienced less severe COVID-19 outcomes once demographic, economic, mobility, and vaccination differences were adjusted for. The results underscore how cultural norms influence collective action and public health externalities during pandemics, clarifying that cultural orientation affects the equilibrium efficacy of policies via societal compliance.

The study shows that individualism is associated with more severe COVID-19 outcomes internationally and reduces the effectiveness of NPIs. Instrumental variable evidence suggests a causal path from individualism to pandemic severity. A complementary within-country analysis exploiting East–West German cultural differences confirms that, controlling for district characteristics and vaccination, collectivist-leaning districts experienced less severe outcomes. The work contributes by documenting persistent cultural effects across the full 2020–2021 period and by distinguishing between policy adoption and adherence mechanisms.

• Instrument strength is marginal in some specifications; potential endogeneity concerns with Nobel science/medicine/economics prizes are addressed by using Literature and Peace prizes as alternative instruments with similar results.
• Policy adoption and effectiveness analyses may suffer from endogeneity due to policy responses to local outbreaks and policy diffusion; results focus on heterogeneity by cultural context rather than absolute causal effects of policies.
• Data limitations include missing testing and vaccination data for some country-days (handled with indicators and zeros), and cross-country heterogeneity in reporting quality.
• German setting: Cultural convergence over time and internal migration may attenuate East–West differences, biasing against detecting cultural effects; Berlin excluded due to internal mobility complicating East/West assignment.
• Observational design limits definitive causal claims beyond IV strategy; unobserved confounders may remain.