Have democracies achieved better control over social interactions?
For now, we have reached the half-way point in our attempt to answer the question of whether democracies are proving more or less successful in the fight against the pandemic. In the previous article1 we have presented empirical evidence that at least they have not been weak. On the contrary, they have been more «aggressive» in imposing more stringent lockdowns and have been more ambitious in terms of testing. However, what we are really interested in is not stringency, but effectiveness. It could be that democracies have been more aggressive in their anti-COVID health policies because they are rather ineffective in achieving the desired social behaviour, namely the reduction in mobility (which, as we know, leads to the decline in social contacts and the breakdown of the chains of contagion). This is a rather more complex question than that addressed in the previous article, since more factors potentially intervene. So let us take it step by step.
In specific terms, the question we are interested in answering is whether, with the same intensity of lockdown restrictions or testing, two countries with different degrees of democracy have achieved different reductions in mobility among their population.2 Again, the result of our empirical analysis is favourable towards democratic countries. The desired behaviour (a reduction in social interactions and mobility) has occurred to a greater extent3 in more democratic countries with an equal application of the tools at hand.4
- 1. See the article «Have democracies been more lax with lockdowns and testing?» in this same Dossier.
- 2. An initial, preliminary analysis has been performed to determine whether the increase in the stringency of lockdowns corresponds to a reduction in mobility using the following panel regression with fixed effects: \(Mobility_{i,t}=\alpha_0+\alpha_1Stringency_{i,t}+u_i+u_t+\mu_{i,t}\) where \(Mobility_{i,t}\) is a measure, developed by Google, of the reduction in mobility in country i and on day t compared to the previous year and \(Stringency_{i,t}\) is a measure, developed by Oxford University, of the severity of the lockdown measures. The regression analysis confirms the expected relationship and, therefore, that the more stringent the lockdown measures, the greater the reduction in mobility. In short, the tool chosen by countries «works» in achieving the desired social outcome.
- 3. The difference in favour of democratic countries is of a relatively small magnitude (see first chart), but the results are statistically significant.
- 4. To perform this analysis, the regression described in note 1 above is expanded by adding an interaction between \(D_i\), a measure of the political system in the country on a scale ranging from 0 to 100 based on data from Freedom House, and \(Stringency_{i,t}\). In this regression, \(Mobility_{i,t}=\alpha_0+\alpha_1Stringency_{i,t}+\alpha_2\;\left[Stringency_{i,t}\times D_1\right]+u_i+u_t+\mu_{i,t}\), the coefficient \(\alpha_2\) can be interpreted as the differential of the reduction in mobility, for the same degree of lockdown stringency, in democracies compared to more authoritarian countries. A significant coefficient is obtained, with an \(R^2\) of 0.645.
However, this relationship is only a top-level one. In the first article of this Dossier we have encountered the notion that a democracy’s success depends on three different aspects: the political one, the capacity of the state and the cultural one. Now it is time to find out whether what we have interpreted as the good performance of the more democratic countries in reducing mobility could in fact be due to the degree of efficiency of public policies or to the prevalent type of sociocultural behaviour – that is, the importance of individualism or collectivism (working on the basis that, in the former case, stricter tools are required since the degree of social «discipline» is foreseeably lower).
With these sophistications, the results are maintained. Firstly, when we try to explain mobility in terms of the degree of democracy and state capacity, we find, as expected, that the greater the administrative capacity to implement public policies, the greater the reduction in mobility. For instance, according to our results, if the lockdown stringency index (measured on a scale between 0 and 100) were to increase by 10 points in two identical countries, but where one had the administrative capacity of Spain and the other that of Colombia (which is significantly lower), then the mobility index would fall by around 5.6 points in the country with the greater administrative capacity, while it would fall by only 5.1 points in the second country, assuming the same conditions apply in both of them. In spite of the control arising from the capacity of the state, being more democratic continues to have a positive correlation with the desired social behaviour, that is, a reduction in mobility.5
Something similar occurs with sociocultural elements and, in particular, with the importance of individualism in a country.6 Generally speaking, the greater the degree of individualism – and always taking the stringency of lockdown measures into account – the worse the result in terms of control over mobility. Repeating the previous exercise, mobility would fall by around 6.7 points in a country with the degree of individualism of Spain (51/100), while it would fall by only around 6.1 points in a country with the degree of individualism of the US (91/100). This makes sense, since in societies in which the collective element or the importance of the group are greater, the expected level of discipline is also greater than in more individualistic societies. However, when we incorporate the degree of democracy into the analysis, we once again see that democracies are more effective in achieving the reduction in mobility.
- 5. An interaction between \(Stringency_{i,t}\) and \(Eff_{i}\), a measure of state capacity developed by the World Bank, is added to the regression described in the footnote above, such that the final regression is: \(Mobility_{i,t}=\alpha_0+\alpha_1Stringency_{i,t}+\alpha_2\;\lbrack Stringency_{i,t}\times D_1\rbrack+\alpha_3\;Stringency_{i,t}\times Eff_i\rbrack+u_i+u_t+\mu_{i,t}\). Significant coefficients are obtained both for \(\alpha_2\) and for \(\alpha_3\), with an \(R^2\) of 0.736.
- 6. The impact of culture on the relationship between the stringency of lockdown measures and the reduction in mobility in each country is analysed using the following panel regression with fixed effects: \(Mobility_{i,t}=\alpha_0+\alpha_1Stringency_{i,t}+\alpha_2\;\lbrack Stringency_{i,t}\times Individualism_i\;\rbrack+u_i+u_t+\mu_{i,t}\), where \(Individualism_{i}\) is a measure of the individualistic/collectivist culture in the country, according to an analysis by Hofstede (2001). A significant coefficient is obtained for \(\alpha_2\), with an \(R^2\) of 0.779.
In other words, although states with a greater capacity to effectively implement policies and countries with more collectivist tendencies have been better at controlling mobility, even when we incorporate these factors democracies continue to do so better than autocracies.
Do these results mean that democracies are saved and can rest easy? Clearly not. For starters, the human drama we are living through does not allow us the luxury of such complacency. Moreover, being better in relative terms may not be enough for citizens. Exercises like the one presented in this article, well-publicised, should help us to make a less critical reading of the true state of democracy. But let us not be fooled, citizens will express their dissatisfaction with the response from democracies based on far more transcendental elements than the obligatory and necessary effort of economists and technicians to interpret the world properly. What history will ultimately decide is what role this health crisis has in the broader context of the political crisis. In other words, it will help to shed some light on the question of whether or not the pandemic can act as a spark to ignite the regeneration of democracy. This judgement is still a few years away, but we can attempt to explore the matter now – a tentative exercise which we conduct in the next article.