A Note on the Role of  Cultural,  Institutional, and Urbanization Features in the COVID-19 Pandemic*

Abstract

Countries have taken different approaches to controlling COVID-19. Analyzing the costs and benefits of different policies, Sachs (2020) has concluded that the total cost of using either low-cost epidemic control measures (e.g., hygiene, testing/tracing/isolating, travel restrictions) or high-cost control measures (e.g., economic shutdown) is lower than the cost of not controlling the epidemic at all. Sachs also found that low-cost policies in the Asia-Pacific region are more effective than high-cost ones in the North Atlantic region. Sachs has therefore proposed that the North Atlantic region should learn from the best low-cost practices of the Asia-Pacific region in suppressing the COVID-19 epidemic.This note will supplement the Sachs (2020) analysis in two ways. The first supplement is to identify what we think to be the most important cultural and institutional features that helped generate the different national performances.The second supplement to the Sachs (2020) analysis is to add another dimension to our understanding of the COVID-19 pandemic by presenting evidence that China's spatial and urbanization features exerted powerful influence on the transmission of COVID-19 within China. As the absence of national borders within the Schengen Area1 makes Schengen members akin to provinces within China, our results suggest that these two features were also important in shaping the COVID-19 situation in the Schengen countries. In short, the differences in COVID-19 outcomes among the Schengen countries reflected more than their different public health actions.2The content and effectiveness of a policy depend on many country-specific factors—especially cultural traditions and institutional backgrounds—that could limit the possibility of mutual learning. For example, Sachs (2020) found that “wearing a face mask” could reduce the infection fatality rate, and we would add that the willingness to wear face masks is related to cultural traditions. In North Atlantic regions, people tend to think that healthy people have no need to wear a mask, whereas Asians are willing to do so to protect themselves and others, even if they are healthy. Therefore, Asian Americans are more likely to wear masks in public than Americans,3 which can be said to reflect cultural differences.Social attitudes toward policies such as shutting down schools and workplaces vary greatly across regions, and governments, in turn, consider these public attitudes when issuing policies. These two factors (social attitudes and policies based on such attitudes) jointly determine epidemic control measures in each country. In East Asia, for example, traditional Confucian culture emphasizes the achievement of social stability via individual internalization of social welfare. Therefore, governments throughout much of East Asia tended to adopt comprehensive lockdown policies during the initial outbreak of COVID-19 because they judged their citizens to be willing to comply.By contrast, political traditions in the North Atlantic region tend to emphasize individual freedom. Thus, governments were less inclined to implement lockdown and strict movement control measures in the early phase of the epidemic. Large segments of the public also tended to downplay the threat of individual infection to community health hence did not generally undertake voluntary social distancing. The outcome was the rapid spread of the virus in the North Atlantic countries.It is interesting that the use of economic stimulus packages in response to COVID-19 has also differed across countries. The North Atlantic countries have been notably much more resolute in introducing macro-policies to stimulate the overall economy than in locking down their economies and in conducting comprehensive surveillance of the disease.The Chinese state, on the other hand, has been highly interventionist at the micro-level both in confronting the disease and in propping up the economy. The Chinese state mobilized the population to join the bureaucracy in locking down neighborhoods, and in comprehensive contact tracing of potential spreaders to complement the wide testing of residents. The Chinese state has also enacted micro-targeting economic measures, like direct fiscal-financial support to certain communities (e.g., the poorest groups) and selective enterprises (e.g., high-tech firms), to keep the economy on track in fulfilling the long-term national economic development plan. Without doing injustice to common parlance, China can be called a strong state because of both its institutional capacity and institutional willingness to marshal the bureaucracy and the population quickly to implement highly interventionist social, economic, and public health directives.A pandemic has obvious negative cross-border externalities. An uncontrolled epidemic in one region increases greatly the probability of contagion to the surrounding areas. The ability to implement inter-regional coordination on the spread of the pandemic is therefore fundamental to the success of every region in controlling the disease within its boundaries. The central government of China has been able to perform this inter-regional coordinating role of implementing various policies in different regions to bring the pandemic under control.In Hubei Province, where COVID-19 originated, the government adopted high-cost epidemic control measures (economic shutdown and city lockdown). In other regions, meanwhile, low-cost policies (wearing face masks, maintaining social distance in public places, testing/isolating/tracing) and partial economic shutdown were implemented. By restricting human mobility, the lockdown policies helped to greatly control the number of infections (Fang et al. 2020). With such policies, COVID-19 in China was quickly controlled, and the economy is now recovering. According to the National Bureau of Statistics of China, GDP growth rates in the first and second quarters of 2020 were –6.8 percent and 3.2 percent, respectively. According to truck data provided by Shanghai Pingjia Technology, the scale of traffic flow in 2020 had more or less recovered within two months after the Spring Festival compared with 2019 (see Figure 1), about 600,000 trucks a day.Regional coordination played an important role in the allocation of health care resources during the pandemic. The housing of COVID-19 patients with patients suffering from other sickness together in the same hospital would cause severe cross-infections and exacerbate the pandemic. Therefore, ensuring sufficient medical resources in the worst-hit areas is necessary for controlling an epidemic in the early phase. China's central government was able to mobilize and transfer medical resources and medical workers on a large scale to Hubei Province.Moreover, based on the experience from the severe acute respiratory syndrome (SARS) epidemic in 2003, China quickly built temporary hospitals in Wuhan for severely sick patients. Meanwhile, Hubei conducted centralized isolation rather than home isolation of all suspected infection cases. In all provinces (except Hubei), medical supply enterprises were the first enterprises to resume operations and expand production to meet national demand. In some lockdown cities (e.g., Wuhan, Huanggang), residents were forbidden from leaving their homes. To meet the daily needs of those residents, necessities from neighboring provinces were distributed to them through coordination among the neighborhood committees and the different levels of government.These policies succeeded in containing the spread of the virus in little over a month. Within two months, daily increases in coronavirus cases had fallen to single digits.4To deal with the economic effects of the epidemic, the Chinese government issued various policies, including tax relief for the most-affected industries and temporary low lending rates and deferred loan and interest payments for small- and medium-sized enterprises.China's performance during COVID-19 was, of course, not perfect. First, local governments and residents tended to focus only on their own interests. For example, some local governments blocked roads without authorization to prevent interregional migration. Furthermore, people from Hubei Province were discriminated against in many areas. Such practices have only been partially stopped by the central government. Second, local governments had tended to focus on epidemic control while neglecting other responsibilities. For instance, some local governments narrowed their focus to testing, tracing, and isolating patients with COVID-19 while ignoring the diagnosis and treatment of people with other diseases. Local governments in lockdown cities also tended to neglect low-income people (especially migrant workers), especially in the early days of the epidemic.Third, some interventions by local governments weakened the market, thus undermining entrepreneurs’ incentives to produce medical resources. For example, the government punished those who sold face masks at high prices without considering the impact of the increasing cost. Fourth, although social groups and organizations have played important roles in fighting COVID-19,5 there was low efficiency in collecting and distributing donated goods by some charitable organizations (especially the Red Cross Society of China), resulting in the late delivery of some medical resources.The cross-country analysis of Sachs (2020) is likely to have omitted some important issues on the spread of the epidemic, such as the effects of urban networks and urban population density. Are the routes of disease transmission related to urban economic networks? What role does population density play in transmission? Are large cities more conducive to epidemic prevention, or do they accelerate rates of infection? These issues are closely related to urban governance but are difficult to analyze at the national level.Figure 2 takes a first look at the possible importance of these city variables in determining COVID-19 infection. Figures 2a, 2b, and 2c drop the Wuhan observation, and Figure 2d drops all cities in Hubei. Figure 2a uses the logarithm of resident population size (pop) in 2018 as the horizontal axis; and Figures 2b, 2c, and 2d use the logarithm of population density (popdens) in 2018 as the horizontal axis in the figures. Population density is represented by the ratio of the number of employees in secondary and tertiary industries to the city land area, in the sense of municipal district. The logarithm of the cumulative number of confirmed cases (confirmed) from 20 January 2020 to 20 April 2020 is on the vertical axis of Figures 2a and 2b; and the COVID-19 infection rate (confirmrate), which is measured by the ratio of confirmed to pop, is on the vertical axis of Figures 2c and 2d.Figure 2 shows that cities with larger populations or with higher density suffer more from the pandemic. One might argue, then, that large cities, especially those with high population density, are more prone to spread infectious diseases and thus have negative effects on society. Some even believe that high density itself is to blame for the epidemic and therefore criticize population agglomeration in large cities.Clearly, other factors, such as geographic or economic connections with the epicenter, could also be correlated with the spread of infectious disease. The epidemic in China spread outward from a single center, Wuhan. We took the distance between one city and Wuhan as the measurement of geographic connection with the epicenter. We also built an urban network based on data for bilateral traffic flows of trucks and used the average traffic flow between one city and Wuhan in the first half of 2019 to measure the economic connection with the epicenter. Comparing truck traffic in 2019 and 2020, Figure 3 shows that cities with different distances to Wuhan suffered uneven effects from the epidemic.Table 1 reports the regression model that estimated the effects of the urban network and population density on the COVID-19 infection rate. The dependent variable confirmrate is the daily infection rate (unit: people per million) in each city. The explanatory variables include the logarithm of geographic distance from the city to Wuhan (distwu), the average traffic flow to Wuhan (countwu), and urban population density (popdens). Both of the double and triple interaction terms were added to the regression, and time fixed effects were controlled as well. We did not control for city size, which was used to calculate infection rates. Figure 2.(a) Confirmed cases and population size (b) Confirmed cases and population density (c) Infection rate and population density (d) Infection rate and population density (apart from Hubei). Note:Horizontal axis represents the logarithm of city resident population in Figure 2a and the logarithm of urban population density in Figures 2b, 2c, and 2d. Vertical axis represents the logarithm of a city's confirmed cases in Figures 2a and 2b and the COVID-19 infection rate in each city in Figures 2c and 2d. To ensure positive logarithm values, all variables are added by 1 in advance. Figures 2a, 2b, and 2c drop the Wuhan sample, and Figure 2d drops all cities in Hubei.Source:Dingxiangyuan (DXY); National Bureau of Statistics.Figure 3.Trends of differences in traffic flow between 2019 and 2020 (by distance to Wuhan). Note:0 is the date of the Spring Festival (Chinese New Year); and the vertical line represents the day of the Wuhan lockdown. The vertical axis is traffic flow in 2020 minus traffic flow in 2019 for each distance from Wuhan and on the date that is measured from the Spring Festival. The horizontal axis represents number of days before and after the Spring Festival.Source:Truck data were provided by Shanghai Pingjia Technology Co., Ltd.Table 1.Effects of urban network and population density on COVID-19 infection rate(1)(2)(3)(4)(5)Confirmratedistwu−76.819***−5.432***−14.964***(0.815)(0.827)(3.222)countwu215.321***217.760***716.393***(1.108)(1.496)(42.140)popdens6.580***−9.571***−21.487***(0.558)(0.367)(4.850)distwu_countwu−80.732***(8.651)popdens_distwu3.255***(0.710)popdens_countwu−24.550***(7.632)popdens_distwu_countwu−0.339(1.558)constant529.807***−7.666***−7.080***72.318***103.880***(5.397)(0.420)(2.634)(6.170)(22.268)Time FEYESYESYESYESYESN2418923947235152327323273R20.2700.6130.0060.6240.673Note:Confirmrate is the daily infection rate (unit: people per million) in each city. distwu indicates the logarithm of geographic distance from the city to Wuhan. countwu indicates the average traffic flow to Wuhan. popdens indicates urban population density. Standard errors in parentheses. ****Statistically significant at the 1 percent level.The results in columns (1), (2), and (3) indicate that infection rates were significantly higher in cities that (i) were closer to Wuhan; (ii) had closer economic connections with Wuhan; and (iii) had greater population density. It seems to be consistent with intuition shown in Figure 2 that high density may aggravate the epidemic. However, when these three variables were controlled together (column [4]), the effect of population density became negative. That is, a densely populated city is in fact beneficial for epidemic control after controlling for the geographic and economic link with the epicenter. For example, Yuncheng (population density: 93 people/km2) in Shanxi Province and Xinyu (population density: 55 people/km2) in Jiangxi Province had similar truck traffic flows to Wuhan, but the infection rate in Xinyu (109.8 ppm) was much higher than in Yuncheng (3.55 ppm). Thus, the simple positive correlation between population density and infection rate in column (3) should be decomposed and worth further discussion.In column (5), where all interaction terms are added, the effects of population density, geographic connections, and economic connections to Wuhan can be explained more clearly. The negative coefficient of distwu_countwu shows that for cities with similar population density, the further away they are from Wuhan, the effect of traffic flows to Wuhan on infection rate is smaller. As for variables related to population density, the sign of the interaction term popdens_distwu is positive and that of popdens_countwu is negative. These results imply that high population density may offset both effects from geographic and economic connections with the epicenter on the severity of epidemic to some degree. In other words, connections, rather than density, account more for the epidemic situation.In general, urban population density is negatively correlated with distance from Wuhan and positively correlated with traffic flows to Wuhan, and the latter is more relevant. In fact, large cities have closer economic connections with Wuhan and higher population density as well. Therefore, the incorrect conclusion that population density is bad for epidemic control could be drawn if this economic indicator is omitted from the analysis. Combining these results with those in Table 1, we observe two opposite effects of population density: the risk of epidemic transmission caused by strengthened economic connection and the positive effect of density itself on epidemic control. Although the epidemic has been more serious in large cities, it is inappropriate to criticize agglomeration on that basis.On the one hand, city connections, which can be strengthened by population agglomeration, are highly correlated by the risk of infectious disease as well as economic development. Short-term economic interruption (such as lockdown policy) can efficiently stop the spread of infectious diseases. However, inhibiting long-term economic development because of potential epidemic risk warrants further consideration. On the other hand, because of economies of scale, high-density cities can provide residents with better infrastructure and public services in terms of both quantity and quality, such as sophisticated medical resources, convenient living services, and reasonable isolation policies. Therefore, policymakers should correctly understand the relationship between population density, economic connections, and disease transmission and make appropriate decisions based on objectives for different periods.If governments control population density in large cities for epidemic prevention based only on the apparently positive correlation between density and infection rate, long-term economic growth could be harmed. In fact, it is social distance but not general population density that matters. Therefore, as long as social distance is controlled, the spread of an epidemic can be controlled without losing the positive effect of urban population density on economic activity.These results can also be applied to Schengen countries. We use the traffic flow data of 2019 in the regression. Thus, the estimated coefficient of economic connection is actually more of an early impact. Compared with Schengen countries, China decisively implemented the lockdown policy on the epidemic center, which effectively cut off the channels of disease transmission through economic networks. However, in Schengen countries, due to the lack of coordination by a central Schengen agency, social distancing policies were not fully implemented in time. For example, whereas it is true that some Schengen countries closed their borders after the outbreak of COVID-19 infection, we note that the number that did so is small, that this action was only temporary, and that the large Schengen countries like France, Germany, Italy, and Spain did not do so. Thus, the movements of labor, capital, and commodities remain relatively free, which lead to longer-lasting influence of economic connections and even bring about several new outbreaks.The cultural and institutional differences across countries account for much of the performances in response to fighting COVID-19, especially the emphasis on social stability and the ability to implement inter-regional coordination. The local governments in many countries have to compete rather than to cooperate with each other when facing COVID-19 alone. In addition, global problems exposed by the pandemic—such as a lack of international cooperation in fighting COVID-19—need to be resolved. For example, some countries banned exports of protective medical equipment and withheld these goods on their way to other countries. Ideally, the World Health Organization and other international organizations should play a central role in these issues, but they have limited power to do so.On the other hand, the diversity of cultures and institutions may make it difficult, but not impossible, for countries to learn from each other in terms of policy. In fact, COVID-19 could have long-term effects on social attitudes about public health in the North Atlantic region. For example, many people in North Atlantic countries (e.g., Canada, Germany, Spain) have recognized the importance of wearing face masks in public, and the proportion of people wearing masks in those countries has increased (Sachs 2020). The experience of SARS in the Asia Pacific regions helped governments respond to COVID-19 rapidly while also enabling the public to comply with the control policies. During the epidemic, the tradeoff between public interest (e.g., public health security) and individual rights has also attracted attention in the North Atlantic regions. Such discussions can provide lessons for dealing with similar incidents in the future.Every topic has more than one side, including the performance of countries in response to COVID-19. Focusing on lowering infection fatality rate alone might ignore other problems and deepen divergences across countries. Moreover, although this paper emphasized cultural and institutional differences, evaluating countries based solely on performance during COVID-19 should be avoided. In Asia Pacific regions, the central governments are relatively strong, and their top–down administrative systems do have certain advantages, especially when the objectives are clear and simple. However, such systems can lose their advantages under diverse and personalized objectives. Similarly, in North Atlantic regions, although the performance of local governments in controlling the epidemic has been unsatisfactory, they play an important role in national development. We should therefore take a comprehensive view when evaluating epidemic control measures and performances in each country. Only in this way can international society reduce ideological differences and achieve mutual learning.