**Multinomial Logistic Regression to Estimate and Predict the Perceptions of Individuals and Companies in the Face of the COVID-19 Pandemic in the Ñuble Region, Chile**

**Benito Umaña-Hermosilla <sup>1</sup> , Hanns de la Fuente-Mella 2,\* , Claudio Elórtegui-Gómez <sup>3</sup> and Marisela Fonseca-Fuentes <sup>1</sup>**


Received: 16 October 2020; Accepted: 9 November 2020; Published: 17 November 2020

**Abstract:** The Coronavirus Disease 2019 (COVID-19) pandemic is transforming the world we live in, revealing our health, economic, and social weaknesses. In the local economy, the loss of job opportunities, the uncertainty about the future of small and medium-sized companies and the difficulties of families to face the effects of this crisis, invite us to investigate the perception of the local community. Based on a questionnaire applied to 313 citizens and 51 companies, this study explored the perception of these actors on the effects of the pandemic at the local level and determined the main factors that influenced their assessment using a multinomial logistic regression model. The results indicated a systematic concern for issues of employment, job security, and household debt. The variables of age and sex were significant when analyzing the vulnerability of certain groups, especially women and the elderly, to face the effects of the crisis and their role as citizens. At the business level, the focus was on economic policies that support its operational continuity and management capacity to face a changing scenario.

**Keywords:** COVID-19 pandemic; local community; perception analysis; econometric modeling; data science

### **1. Introduction**

Local development studies have strongly encouraged interest to decipher, explain, and understand the role of the citizenry and local governments in promoting development [1,2]. The interaction between the economy, authorities, and the local citizenry is critical for dealing with crisis scenarios and accounts for the integration of the social paradigm in the context of economic policy formulation [3–5].

The cooperation and construction of the community are essential for economic, social, and environmental survival [6]. Likewise, it is considered that the community is the fabric for the collective action of individuals, a mediating structure between individuals/families and the social, economic, political, and environmental spheres, which are typical of local development [7]. As a result, the debate about local development not only involves experts, technocrats, and government authorities, but the assessment of the common citizen and the community also becomes more relevant to define political and social guidelines [8,9].

The vision of local stakeholders is fundamental to detecting gaps and trends in the local economic outlook. The collection of these visions using surveys allows local authorities to undertake efforts for strategic changes that promote local development processes and strengthen communication networks with citizens [10]. As part of a philosophy of community-based environmental monitoring, the survey can provide a viable solution and an attractive means of increasing legitimacy to implement measures that lead to improved quality of life [11,12].

Understanding that the effects of the COVID-19 pandemic have modified the daily dynamics of people, the way they relate to each other and the expectations of the local economy, it is necessary to investigate the social representations that arise in this context of uncertainty, where currently the business and social development of communities around the world is limited.

Following this approach, the socio-economic paradigms require new studies. A sustainable revolution is necessary to tackle climate change and these actions are globally required [13]. In addition, literature shows a positive relationship between sustainability and resilience [14,15].

This study investigates the perception of citizens and local companies about the effects that the COVID-19 pandemic has had on the development of their activities and lifestyle, seeking to understand how the current economic scenario is perceived in terms of employment, economy, and government performance, among others.

Special emphasis has been placed on the factors that influence the assessment of the challenges, risks and opportunities that this new and disruptive scenario offers to the region and its inhabitants, antecedents that allow the general vision of the population to be conceived according to the projection they can make based on their level of family income, supply, level of indebtedness, national and local economic situation.

The research will also allow to deepen the social resilience present in the population, where it has been observed that people and governments organize their efforts to face this global health crisis, thus trying to minimize the loss of human lives, help the population who are more vulnerable and adapt to better assimilate this new reality. In this framework, the study contributes to an emerging literature to assess the effects of the COVID-19 pandemic at the local scale.

This analysis adopts a quantitative approach, limiting the study area to Chile, specifically to the Ñuble region. The following research question is established: What are the appreciations of the inhabitants and businesspeople about their reality in the COVID-19 pandemic? In relation to it, the objective is to identify the perception of citizens on various topics of interest and sensitivity, and their effect on the economic and financial aspects of the country, all marked by the uncertainty of COVID-19.

Regarding the problem, the need to have records related to the effects of the pandemic from social representations is identified, therefore it is established as a research question: What are the appreciations of the inhabitants and business owners about their reality in the COVID-19 pandemic? In relation to it, the objective is to identify the perception of citizens on various topics of interest and sensitivity, and its effect on the economic and financial aspects of the country, all marked by the uncertainty of COVID-19.

This article analyzes the results of the survey "Pulse of the economy in the Ñuble region–Chile," where the perceptions of residents and business owners are evidenced in relation to the effects of the pandemic at the local level. The subject is approached from an analysis that adopts a quantitative approach and the results are presented in the following sections.

### **2. World Context: COVID-19**

The SARS-CoV-2 virus is the cause of the infectious disease known as COVID-19; its most common symptoms are fever, muscle aches, cough, fatigue, and shortness of breath, and it can also cause pneumonia, sepsis, and acute respiratory distress in severe cases. The contagion of this disease is mainly through small drops expelled when talking, coughing or sneezing, which has caused an increase in patients in the world exceeding 30 million infections registered in 190 countries, with the United States, Brazil, and India being among the most affected countries, as their records exceed 100,000 deaths due to COVID-19.

In Chile, the first fatal victim of COVID-19 died on 17 March 2020; to date, according to information provided by the health ministry (until 14 October 2020), 13,415 deaths and 485,372 accumulated confirmed cases have been registered [16]. As in other countries, in Chile social distancing measures have been implemented that include confining families to their homes, in addition to activities such as work and study being carried out at a distance, in an attempt to reduce the number of infected people.

This situation is affecting the individual and social development of people, preventing physical contact, the operation of companies and free movement in the territory, which is generating negative emotional effects due to isolation and constant concern about contracting this disease. Among the most common alterations we find stress, depression, anxiety, insomnia, and dietary changes, among others, which will have a medium and long-term impact, weakening the mental health of the entire world.

Regarding the changes in people's dynamics, the role of the internet and social networks stands out, which currently allows them to carry out daily activities such as working and studying remotely. This new form has kept people connected and occupied from their homes, allowing them to advance and generate resources without exposing themselves to infection, but at the same time it has become a form of exclusion for the population of rural and vulnerable sectors [17,18].

In relation to the effects on economic development, a decrease in the demand for goods and services derived from the confinement of people is identified; among the most affected sectors are tourism and recreation services. Regarding the labor market, although teleworking has been implemented to facilitate activities from home, it has also promoted the exclusion of the most vulnerable sectors derived from rurality, scarcity of resources and all kinds of limits that prevent access to the internet.

### **3. Citizen Perception and Local Economic Development**

Based on different approaches, public opinion studies play a key role in the evaluation of economic scenarios at both the local and national levels [19–22]. Perception and expectations, as an individual and subjective element, are often built on trust in the performance of the economy and support to institutions [23].

The role of public opinion on various aspects of the economy has been the subject of intense debate [24] and growing interest to explore the micro and macro determinants of the social preferences of the citizenry [25–28]. Numerous studies have demonstrated that community participation and citizen empowerment are key instruments for building local capacity, improving quality of life, reducing poverty, and promoting economic development at the local level [29–31].

Empowerment of the citizenry is an increasingly important aspect as a starting point for local development [32] because it provides greater capabilities for communities, especially in the control and information flows of the territorial resources [33,34]. Higher levels of community involvement in the local development processes are associated with the extension of the citizenry and better levels of governance [35].

From the community perspective, citizens participate because they perceive themselves as part of an integrated body that has common goals in the socioeconomic, political, legal, and cultural spheres [36–38]. In the case of opinion polls, the greatest achievements in the participatory processes have been the creation of interactive spaces between civil society organizations and public institutions aimed at implementing development initiatives from a collaborative, dynamic, and multidimensional approach [39–41].

Different authors have emphasized the importance of examining the factors related to the perceptions of local development [42–44]. Some researchers highlight the analysis of the cultural and institutional dimensions to understand the determinants of successful regional processes [45–47]. A better understanding of the factors that influence perceptions is essential to improve communication

strategies and the design of efficient local policies that facilitate development processes, especially in times of crisis, uncertainty, and instability.

To recognize their importance, it is necessary to comprehend that individual perceptions can be transformed into social representations that construct reality through language [48]. Likewise, admitting that cultural identity is established on the basis of the characteristics of a society and its social representations [49,50] provides support for the role that the vision of the citizenry should play in local economic development policies.

Cultural identity can be associated with a sense of belonging inspired by a territory, its social group, and culture [51]. It is a dynamic process of permanent reconstruction that supports social interactions [52,53], and it is directly related to the people's sense of belonging, which is considered fundamental for implementing "social cooperation processes" [54,55]. Furthermore, it invites us to question the notion of territory, so as to review the reality of functional rural-urban territories [56].

"Territory not as an "objectively existing" physical space, but rather it exists as a set of social relations that give rise to, and at the same time express, an identity and a sense of purpose shared by multiple public and private agents (even though such a construction often involves going through processes of conflict and negotiation). It is this identity that gives meaning and content to a development project of a determined space based on the convergence of interests and intentions" [57].

Validating the existence of the territory as "a social construction" that emerges from the local identity [58,59] allows us to explore population dynamics that are present in daily life, and which clearly influence perceptions and representations in the cultural identity of society. For example, special mobility is defined as an "internal, transitory, and frequent territorial movement of people without the need to establish a second residence" [60]; this phenomenon is present in functional territories where people travel to develop activities such as work, study, and have access to public and/or private goods and services. Such mobility constitutes interactions and allows different places to be part of a territory, which have, for example, a common labor market and supply of goods and services [61].

Some changes in the rural sector have been identified as a result of this interaction and influence from the urban sector, including an increase in non-agricultural occupations, higher participation by women in the labor force, and access to education and health, which directly promote economic development [62]. Moreover, the expectations of the rural population have shifted; for example, the educational attainment of the Chilean rural population has improved in recent years [63].

This reality is known as "urban living" [64] and allows people to maintain their residence in rural locations and pursue life goals in urban areas. They can therefore access goods and services not available in their community; this gradually modifies the expectations of new generations and, to some extent, modernizes rural life.

Given the background information and the current situation resulting from the COVID-19 pandemic, it is imperative to investigate the social perception of this "new normality", which has closed borders and paralyzed movement worldwide and activated quarantines and sanitary cordons to protect the population by lockdown and social distancing. This situation has also made the population aware of their responsibility for individual and social well-being [65].

Faced with this contingency that affects how people and organizations function, governments have implemented internal policies to mitigate the effects on economic well-being. Despite all efforts, we are facing a global economic crisis due to the pandemic. The results of the business survey on the impact of the COVID-19 crisis [66] indicate that most companies perceive decreased economic activity due to the health crisis, and smaller companies experience financing and liquidity problems. Another important aspect is how teleworking, production cost adjustment, and tax debt deferral are part of a strategy to continue their activities. As for expectations, there is quite a lot of pessimism as to the recovery of the economy.

Before projecting or comparing this reality, background information must be collected on what people and organizations are experiencing, and prepare representations based on the population's perceptions to identify vulnerability at the local level. It is truly important to know the expectations and emotions related to the effects of the COVID-19 in Chile and to consult the people's evaluation regarding such aspects as the implemented policies and the performance of the local and national governments, which provide an insight into the reality that the country is facing.

### **4. Methodology**

### *4.1. Data and Method*

To understand the perceptions and assessment of the citizens and companies in the face of the pandemic in the Ñuble Region, Chile, a questionnaire was created and administered to residents and business owners of the region. The survey was administered in person and online. The sampling used to perform the measurement on this population was non-probabilistic and of convenience, where those interested in participating were contacted by an invitation. The online survey was circulated in the online version of a local daily newspaper. The survey included the Likert scale, multiple choice, open-ended, and demographic questions; it took approximately ten minutes to complete and respondents were not compensated.

To determine the factors that could predict the perceptions of people and companies as to how to face the pandemic in the Ñuble Region, a multinomial logistic regression model was fitted to the responses, and was used to predict the probabilities of the different possible outcomes [67]. Multinomial logistic regression was used to predict categorical variables or the probability of category membership on a dependent variable based on multiple independent variables [68]. As in binary logistic regression, multinomial logistic regression uses maximum likelihood estimation to evaluate the probability of categorical membership. Thus, this type of model allowed us to characterize the probability of a respondent's decision for a particular multinomial discrete choice, conditional on the values of the explanatory variables [69]. The distribution functions that characterize explanatory variables are often nonlinear. Thus, once the multinomial regression model is created, the parameters are used to make predictions about the probability of an event occurring compared with the reference category.

In this particular case, we wanted to know how changes affected the abovementioned independent variables on the probabilities of the variable (infrastructure choice) in Equation (1) expressed as

$$P(Y = j/X\_1, X\_2, \dots, X\_k) = P(Y = j/K); j = 0, 1, \dots, l \tag{1}$$

In the multinomial case, response probabilities were represented in Equations (2) and (3) as

$$P(Y = j / \mathbf{X}) = \frac{\exp(\mathbf{X}\boldsymbol{\beta}\_j)}{1 + \sum\_{h=1}^{l} \exp(\mathbf{X}\boldsymbol{\beta}\_h)} = p\_j(\mathbf{X}, \boldsymbol{\beta}); j = 1, \dots, l \tag{2}$$

$$P(Y=0/X) = \cfrac{1}{1 + \sum\_{h=1}^{I} \exp(X\beta\_h)} = p\_0(X, \beta) \tag{3}$$

We used maximum likelihood to estimate multinomial logit models in which the logarithm of the likelihood function that usually provides consistent and asymptotically normal estimators is expressed by Equation (4) as

$$I(\beta) = \sum\_{i=1}^{n} \sum\_{j=0}^{I} \mathbf{1}[Y\_i = j] \log \left[ p\_j(\mathbf{X}\_{i\prime}\beta) \right] \tag{4}$$

### *4.2. Study Area*

The study was conducted in the Ñuble Region located in south-central Chile. This region covers an area of 13,178.5 km<sup>2</sup> consisting of a diversity of landscapes ranging from mountainous areas to interior valleys. The region is divided into 3 provinces and 21 communes with a total population of 480,609 inhabitants, and 30.60% live in rural areas, which places the region as the most rural in Chile, well above the national mean of 16.73% [70,71]. As for poverty, the region has the highest national indicators with 16.1% of households living in poverty, which is much higher than the national mean of 9.4% [72]. There have been important demographic changes in recent years and the population has gradually decreased, reflecting the dynamics of the migration of the rural population to urbans spaces and new occupational patterns of the territory through activities linked to tourism in the mountains, nature, and/or agrotourism [73,74].

Historically, the productive and economic base of the region is related to small- and medium-scale agricultural and forestry activity. Thus, the region shows a clear orientation towards agricultural production and this situation is not replicated in other sectors [75]. The seasonality of agricultural activities and the boom in vegetable and fruit export operations have a significant impact on job creation, accounting for 20.7% of jobs in the region. However, rural areas have undergone significant changes in the last 30 years, including the declining importance of agriculture in rural production and labor dynamics [76,77]. On many farms, owners and/or laborers spend part of their time in off-farm work. Rural modernization has changed the traditional ways of living and working. The services sector has become a source of employment and income [78]. In mountain areas, the diversification of activities has been a strategy to ensure the economic viability of farms by promoting various forms of rural tourism and protection and dissemination of both the natural and cultural heritage. Rurality is a strong component of the regional territorial identity [79].

### **5. Results**

### *5.1. Descriptive Analysis*

Within the framework of the complex world and Chilean pandemic scenario set in a historical context of great expectation for public opinion, research was carried out in the newly created Ñuble Region, Chile.

Due to the quarantine as a result of the COVID-19 virus, the present study considered virtual data collection from 13 to 22 April 2020 using a survey in social networks and e-mail, in Chillán and its neighboring communities. A total of 313 valid responses were obtained for the citizenry (individuals) and 51 valid responses for companies (business owners) and two virtual instruments were used:


Therefore, the present work is a quantitative and cross-sectional descriptive study [80] using the survey as its research strategy [81]. The subjects were the owners of regional businesses located in the municipalities of Chillán Viejo, San Carlos, Yungay, Coihueco, Pinto, Bulnes, San Ignacio, and Chillán. For the survey applied to the citizens, study subjects were adults residing in Chillán and other municipalities of the Ñuble Region: Chillan Viejo, San Nicolás, Quillón, El Carmen, Quirihue, Yungay, Coihueco, Coelemu, Pemuco, Bulnes, San Carlos, San Ignacio, Portezuelo, Pinto, San Fabián, and Ranquil.

One of the main results described for individuals was the negative view of the existing economy (47.9%). The surveyed citizens manifested that the economic situation of the country a year ago was better than at the present (71.2%) and that 2021 would be worse than 2020 (65.5%).

Meanwhile, 64.2% of the sample maintained that the economic situation of the region in one more year would be worse than at the present. Some 48.3% of the surveyed individuals pointed out that the existing economic reality of the region was fair or bad.

The lack of employment (33.2%), low salaries (25.9%), and digital connectivity (12.5%) were the three most important problems at the regional level that were identified by the sample in the context of the pandemic.

The majority of the sample indicated that their household income in the last 12 months had decreased (53%), and they were pessimistic and predicted that it would continue to decrease (46.6%). The population admitted to having fear of losing their jobs (74.4%), even though teleworking had been introduced (48.2%). Some 62% of the sample reported some degree of difficulty in their household due to debt. Meanwhile, 43.8% of the citizens indicated they did not feel economically prepared to face the pandemic. In addition, 69.3% of respondents stated that the work of the regional media in addressing the pandemic was "fair to very good" in contrast to the perception of 47.6% of individuals who negatively evaluated the work of the national media.

Regarding the main exploratory results for companies, 54.9% of businesspeople indicated that they were prepared to face the pandemic "only for a while". Some 41.2% positively evaluated the policies developed by the government to provide support to businesses. However, 90.2% were concerned for their levels of debt because they considered that the effects of the coronavirus, in addition to the social upheaval, had affected their commercial activity. Likewise, 64.7% of businesspeople specified that they considered that the economic situation of the country in 2021 would be the same or worse, and 62.7% pointed out that it would be the same or worse than the existing regional situation.

A positive perception was highlighted when associating the creation of the region with the management of future crises. However, there was uncertainty about the region's current preparedness and the responses of regional/community authorities to the COVID-19 crisis.

### *5.2. Multinomial Logistic Regression*

The arrival of COVID-19 in Chile was set in a particular context and probably generated a number of characteristics prior to the emergence of the first case declared on 3 March 2020. The scenario immediately prior to the coronavirus outbreak is a relevant issue due to high sensitivity and concern for the immediate future of the country, both economically and institutionally, among citizens and business people.

This predisposition is very different to how citizens responded to the "subprime crisis" in 2008. At that time, trust in the political institutions and the security in the Chilean economic management during the first presidential term of Michelle Bachelet (2006–2010) and in the Minister of Finance Andrés Velasco were relevant to minimize the effects of such instability. Moreover, Chile was one of the countries in Latin America that most successfully overcame the crisis.

Since October 2019, Chile has experienced a social upheaval of a magnitude that was not on the research agenda, academic studies, or political surveys. In fact, the administration of President Sebastián Piñera was preparing to host two important global forums, APEC and COP25, events that were to strengthen the country's image, public diplomacy, and Chile's international standing vis-à-vis the world.

However, the social crisis became very difficult for the government and the political establishment to manage and those events were therefore cancelled, including visits by Donald Trump and Xi Jinping. This was due to a wave of demonstrations that affected the activities, movements, and routines of the population and which were only interrupted by the arrival of the COVID-19. Before the pandemic context, there was already talk in Chile of a "new normality" due to social indignation or simply the impossibility of returning to "normality."

It is therefore important to point out that the results reported in this study are on a continuum of tensions and concerns about what 2020 would be like for both the citizenry and entrepreneurs or business owners. The complexity of the health, economic, and psychological events related to the coronavirus and its rapid spread in Chile should be measured in the regions that were exposed to the severity of the pandemic.

To respond to the research objective, a multinomial logistic regression model was fitted to know the perceptions of individuals and businesses and quantify the effects of the COVID-19 pandemic in the Ñuble Region. Table 1 summarizes the cases related to the citizens. A linear model was first run on the responses as a function of the predictors to ensure that there were no problems with multicollinearity; only predictors with variance inflation factors (VIF) < 2 were included in these models [82].

Regarding the fitted model information, the chi-squared ratio test had a value of 317.910 (*p* = 0.000), indicating a good model fit. Acceptable values were also obtained for the pseudo R-squared (Cox and Snell: 0.638, Nagelkerke: 0.789). Table 2 shows that the power of our logistic multinomial model was suitable because it correctly classified 88.8% of the known observations and can be expected to project future estimates. Table 3 shows the likelihood ratio tests for the effects of the model and the partials whose low *p*-values show the high significance of the variables in the model.

The perception of the citizenry in the Ñuble Region and the projections related to the health crisis are not encouraging, and it can be seen that there is a systemic concern. Assuming the slump in economic expectations prior to the coronavirus, the preparedness of the country to deal with the pandemic is associated with a number of aspects discussed below.

Age is an important issue in Chile because it has been permeating political and economic discussions over the last decade about the pension system, the fear of aging with low pensions, and the vulnerability of aging. It has also been a generational issue that transpired with the Chilean social upheaval because it established a perspective between the new emerging practices and old styles of leadership. The COVID-19 crisis cannot be excluded from the age-related aspects already incorporated in the perceptions of the population. Therefore, the interpretations of individuals in the face of the coronavirus also emerged from the diversified visions of the citizenry based on age group. The interest in greater social protagonism was found in the under-33 age group with its flexibility, vigor, and resilience to face the virus; this group has opened new spaces for discussion, voices, and empowerment never before observed in a country that shows a trend towards aging.

Gender was also important because the Chilean population has shifted the role of women. The feminine and masculine are perceived as differentiated styles of social and local problem solving. In addition, gender leadership tends to gain a greater presence in the regional social base. This should be considered when addressing such aspects as compliance with sanitary measures, responsibility for self-care to protect the community, or citizen support networks.

Closely related to the above, completion of schooling or educational level was relevant for the perceptions of the country's preparedness for the pandemic. The challenges that the disease poses in the analyzed Chilean reality have required individuals to meet the greatest instructional, technical, professional, or postgraduate demands as an efficient alternative to manage this complex situation. Knowledge, as a value of pandemic preparedness, dispels myths and insecurities and provides information and guidance. Moreover, in the midst of systemic insecurity, education has created new opportunities when faced with the lack or instability of employment generated by the COVID-19.


#### **Table 1.**Summary of citizen responses.


**Table 1.** *Cont.*


**Table 1.** *Cont.*


**Table 2.** Power of classification of citizen model.


**Table 3.** Multinomial logistic regression of citizen model.

From this perspective, household income was a critical measure of the country's preparedness. The cost of living in the regions was not an easy issue for residents: it was directly linked to employment and job security, which was seriously threatened by a standstill of activities caused by physical distancing and sanitary measures. To a large extent, the perceptions and confirmations of residents as to how the country would overcome the pandemic depended on how they felt their income would be affected.

For this reason, household debt was another important variable for the country in the face of the coronavirus. This situation was a matter of concern that has raised awareness, with or without the COVID-19, in both national and regional public discussions. The capacity to take on guaranteed debt in the face of job loss and the direct effects of the crisis or debt that continues to grow at an alarming rate to survive during these difficult circumstances are manifestations of another social aspect that define the preparedness of the country.

At the structural level, the figure of a government in any crisis management is relevant to face such events that impact the world. This was quickly perceived by the residents because the need for the protection of the state was assumed, which was represented by its political and immediate decision making and government response time in an emergency. Leadership, the ability of persuasion of effective and credible communication policies, and clarity in decision making in favor of citizens in the health, social, and economic spheres are essential to manage the risks to the population. Moreover, under the Chilean political presidential system in which the regions must assume that power is centralized, government efficiency is perceived as a matter of collective survival.

However, the conditions pointed out since the Chilean social upheaval in 2019 also highlighted the importance of having local authorities that could provide greater autonomy in their responses to people due to their proximity, knowledge of the territories, and empathy with regional and identity problems. The condition of the prior social crisis, intensified by the pandemic, further differentiates this aspect. Residents have perceived that the country's preparedness to face the coronavirus involves

a more empowered and decentralized regional governance given that Ñuble Region was recently established as a region in 2018.

The economic preparedness of the residents is another characteristic identified by the analysis to face the pandemic. The people of the region perceive what a crisis is and quickly assimilate that it will not only affect life but also material conditions. The protection of the most domestic and intimate limits for basic needs is threatened.

The economic preparation of the citizenry is another characteristic identified in the analysis to face the pandemic. People in the region gauged in their perceptions what constitutes a crisis and quickly assimilated that it would not exclusively affect the valued condition of life but also the material aspect. The protection of the most domestic and intimate boundaries in terms of basic needs has been threatened.

Linked to other previously mentioned aspects, the fragility of the regions could be reverted with stronger companies that have the capacity to protect themselves against these external threats. If these businesses could economically withstand the pandemic and manage to avoid bankruptcy, it would mean that the country was more prepared to face the disease.

From the perspective of the residents, media-related variables emerged that reflected the importance of information and journalism when facing pandemic risks. Given the political centralization of the country, but with a growing need for local empowerment, national media are necessary to provide the population with knowledge about the measures that have an impact on their territory from the decision-making core of power.

However, this information required an adequate local context that was provided by regional media. When the public is informed and interprets the national from the local, it can act more effectively and with greater certainty. This is also important in applying civic behavior, which is necessary in emergency situations.

Finally, the sphere of social networks demonstrated the coexistence with a different informational digital space that was more oriented to the emotions and immediate expectations of the residents along with interactions with many people, accounts, and entities that were important and oriented to the residents. The value of this as a social and supportive resource among residents is undeniably one of preparation in the face of pandemic instability.

As for the analysis of the perceptions of the businesspeople of Ñuble Region, there is a series of other variables that complemented the preparation of the country for the coronavirus and others observed by the residents. Table 4 summarizes the cases related to the business owners. A linear model was first run on the responses as a function of the predictors to ensure that there were no problems with multicollinearity; only predictors with variance inflation factors (VIF) < 2 were included in these models [82,83].




### **Table 4.** *Cont.*

Regarding the fitted model information, the chi-squared ratio test yielded a value of 87.472 (*p* = 0.000), indicating a good model fit. In addition, acceptable values were obtained for the pseudo R-squared (Cox and Snell: 0.820, Nagelkerke: 0.962, McFadden: 0.897). Table 5 shows that the power of the logistic multinomial model was suitable because it correctly classified 94.1% of the known observations and could be expected to project future estimates. Table 6 shows the likelihood ratio tests for the effects of the model and the partials whose low *p*-values show their high significance of the variables in the model.

**Table 5.** Power of classification of business owner model.



**Table 6.** Multinomial logistic regression of business owner model.

The results of the multinomial model applied to both businesspeople and residents of the Ñuble Region highlighted the aspects of sex and age. However, they are associated with characteristics more typical of the elite, whose groups are historically considered to be "prepared" for crisis management; although things are changing, they maintain a symbolic and cultural weight because they are the owners of the companies. They are represented by adult men aged over 40. These characteristics are also close to the classic patterns of the cultural stereotype of political leadership or personalized representation of power in Chile. They are also prominent figures in regional agendas and in local trade associations, and are economically legitimized to act.

A relevant variable to face COVID-19 in Chile was the policies developed by the government in favor of businesses. After months of growing concern due to the social upheaval in 2019, businesses experienced a crisis with effects that were more adverse than expected and caused

by the coronavirus since March 2020. This quickly generated expectations from the owners for a relief package to provide sustainability to their respective businesses.

For the perceptions of the businesspeople, the types of economic policies the government has targeted to companies, as guarantor of the national budgets, are essential to determine the country's immediate economic success or failure in the crisis. Certain regions in Chile do not have a favorable situation for generating employment. In fact, the Ñuble Region was one the regions with the highest unemployment rate before the pandemic, and this vulnerability is of great concern for business stakeholders.

Likewise, corporate debt issues are another sensitive and key point for the preparation of Chile to face the disease in two different situations. First, when the virus was not yet active in Chile and second, once the infectious outbreaks began to subside. The effects of indebtedness can be as devastating as the stopping of productive activities demonstrated under this scenario, although for many businesses they also provided options for survival, leading to the maintenance of regional employment.

The owners are aware that before any crisis arises, they should have a business without major debt problems; this is a necessary basis to face more complex times, which sooner or later come in inevitable cycles to countries open to the world. Indebtedness existing before the virus and the projected post-pandemic scenarios for regional businesses would mark a real level of country preparedness in this crisis, which is necessary when considering its success or failure.

Government measures such as those privileging national companies were appreciated in a context of increased global competition and the presence of international stakeholders in the regions. National companies, and those that were established in the same regions where they operate, tend to have a good relationship with the environments and communities that live there and have a positive impact on local economies. As key stakeholders that add cultural and identity value, they require greater protection from the weight of large foreign groups that have the capacity to protect themselves.

Meanwhile, for any government in times of crisis, confidence is essential in view of the social discipline required for the effectiveness of measures taken by the authorities and the persuasion of the policies they wish to promote in the economic sphere. The political and economic institutionality of the country leads to better preparedness in the face of destabilization by the pandemic.

On this level, business owners are aware that maintaining confidence in government authorities and state entities, especially the Ministry of Finance and the Central Bank which are responsible for overseeing the economic conditions of the country, minimizes uncertainty. This also places the state in its role as an essential coordination center to protect companies during the critical or recovery phases.

Finally, despite the recent creation of the Ñuble Region, a decentralized territory with the capacity to uphold the demands of businesspeople to the national government and efficiently transfer aid measures to the productive activity of local areas is fundamental for an adequate preparedness that speeds up the timeframe and the concerns of the business stakeholders. The need has also been reestablished for a state that is present in the various regional realities to implement systems and early responses to human, material, health, and social catastrophes.

### **6. Discussion and Conclusions**

These are very challenging times for Chile as for all other countries because of the situation caused by the COVID-19 world crisis. The pandemic is rapidly weakening economies and due to social isolation most productive activities have come to a standstill. According to the Economic Commission for Latin America and the Caribbean [84], the sectors strongly affected are tourism services, traditional cultural industry, commerce, repair of goods, hotels and restaurants, transport, fashion, and automobiles.

At the microeconomic level, the Chilean scenario is even more worrisome due to the consequences of the social upheaval that began on 18 October 2019. This is especially reflected in the precariousness of smaller companies, which have high unemployment rates and family debt. This is more complex at the regional level given the high degree of centralism in the country. Based on the results of the

present study, it can be concluded that economic uncertainty and pessimism have increased in the Ñuble Region.

In this context, the present study addressed issues of great interest to the national and international community from a local and decentralized perspective, noting the serious economic and social conditions affecting both families and businesses. It also seeks to highlight the importance of generating local information that is relevant to the population and which can contribute to public policy and private investment decisions. In addition, the study emphasizes the need to strengthen local institutions, mainly universities and research centers.

The new scenario created by the COVID-19 pandemic has not only altered the "normality" of the inhabitants of the Ñuble Region, but it has also increased uncertainty. It also has an impact on household incomes and the economy in general.

There is a high level of economic pessimism because of the health crisis. Results showed that the majority of respondents had a very poor perception of the country's level of preparedness to face the pandemic (only 3.8% considered the country to be well prepared] and that their household income would continue to decrease (46.6%). There is therefore pessimism about the economic situation of the country and the region in the upcoming 12 months. Likewise, it is concluded that there is deep concern about job loss (74.4%), and 43.8% of the respondents indicated that they did not feel economically prepared to face the crisis. There was a negative evaluation (44.1%) as to the efforts of the government to address the situation.

A similar level of pessimism and vulnerability to the effects of the crisis is also observed in the business environment of the Ñuble Region, a territory made up mainly of smaller companies that are protagonists of the local economic development. This defines a complex future scenario generated by the global health crisis of COVID-19. Of the surveyed companies, 52.9% consider that the economic situation of the country in another year would be worse than the current one. Although the companies considered that their indebtedness was not for the most part complicated, they did believe that the situation would worsen significantly as a result of the pandemic.

Furthermore, it can be concluded that the health emergency, together with increased economic uncertainty and pessimism, has raised concerns about pending demands in the country. These are related to the quality of public health, employment, wages, and digital connectivity, which are determinants in the way families can face the challenges caused by this crisis.

Finally, it is necessary to remember that humanity has the ability to overcome and adapt to difficulties, but the pandemic is a global challenge and we do not know how long it will continue to threaten people's health. In this scenario, the concepts of resilience and sustainability are fundamental in the development of future social actions and will be the impulse that allows societies to restructure the lifestyle of families.

From now on, the most important challenge is to reduce the economic gap and for society to generate alternatives to include the most vulnerable in this new social dynamic, where access to technology and internet connection are elementary in "social isolation" and they are changing the way we communicate, learn, and work.

Regarding the limitations, we can mention that when carrying out the data collection through an online survey, an exclusion of the inhabitants who do not have internet access is generated due to the rurality of the town where they live or due to the lack of resources. In relation to this, it was not possible to obtain a proportional participation of all the communes of the Ñuble region and therefore, the response rate was lower than expected. On the other hand, future research will consider other variables and relationships, such as the production linkages and employment effects of the business environment on the rest of the Ñuble economy region [85–88].

**Author Contributions:** Formal analysis, H.d.l.F.-M., B.U.-H., M.F.-F., and C.E.-G.; funding acquisition, B.U.-H. and H.d.l.F.-M.; investigation, H.d.l.F.-M., B.U.-H., M.F.-F. and C.E.-G.; methodology, H.d.l.F.-M., M.F.-F.; project administration, B.U.-H.; software, H.d.l.F.-M.; supervision, B.U.-H. and M.F.-F.; validation, H.d.l.F.-M. and C.E.-G.; writing—original draft, B.U.-H. and M.F.-F.; writing—review & editing, H.d.l.F.-M., C.E.-G. All authors have read and agreed to the published version of the manuscript.

**Funding:** Hanns de la Fuente-Mella and Claudio Elórtegui-Gómez are supported by Grant Nucleo de Investigacion en Data Analytics/VRIEA/PUCV/039.432/2020.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References and note**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **90 Days of COVID-19 Social Distancing and Its Impacts on Air Quality and Health in Sao Paulo, Brazil**

### **Daniela Debone , Mariana V. da Costa and Simone G. E. K. Miraglia \***

Instituto de Ciências Ambientais, Químicas e Farmacêuticas, Universidade Federal de São Paulo (UNIFESP), Diadema 09913-030, Brazil; d.debone@unifesp.br (D.D.); vieiramariana70@gmail.com (M.V.d.C.)

**\*** Correspondence: simone.miraglia@unifesp.br

Received: 8 August 2020; Accepted: 8 September 2020; Published: 10 September 2020

**Abstract:** The COVID-19 pandemic has imposed a unique situation for humanity, reaching up to 5623 deaths in Sao Paulo city during the analyzed period of this study. Due to the measures for social distancing, an improvement of air quality was observed worldwide. In view of this scenario, we investigated the air quality improvement related to PM10, PM2.5, and NO<sup>2</sup> concentrations during 90 days of quarantine compared to an equivalent period in 2019. We found a significant drop in air pollution of 45% of PM10, 46% of PM2.5, and 58% of NO2, and using a relative-risk function, we estimated that this significant air quality improvement avoided, respectively, 78, 337, and 387 premature deaths, respectively, and prevented approximately US\$ 720 million on health costs. Moreover, we estimated that 5,623 deaths by COVID-19 represent an economic health loss of US\$ 10.5 billion. Both health and economic gains associated with air pollution reductions give a positive perspective of the efforts towards keeping air pollution reduced even after the pandemic, highlighting the importance of improving the strategies of air pollution mitigation actions, as well as the crucial role of adopting efficient measures to protect human health both during and after the COVID-19 global health crisis.

**Keywords:** COVID-19; air pollution; air quality; health effects; economic burden

### **1. Introduction**

The coronavirus disease (COVID-19) pandemic caused by rapidly spreading a severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has imposed a unique situation for humanity since the Spanish flu pandemic, transmitted by the H1N<sup>1</sup> influenza A virus, which devastated at least 50 million people in 1918. Thus far, COVID-19 is considered the deadliest pandemic in modern history, exceeding the outbreaks caused by H1N<sup>1</sup> virus, such as the most recent outbreak in 2009, formerly known as swine flu [1–8].

Likewise, in the past two decades, coronaviruses have also shown a continuing potential threat to public health since the emergence of severe acute respiratory syndrome coronavirus (SARS-CoV) and Middle East respiratory syndrome coronavirus (MERS-CoV). SARS-CoV and MERS-CoV can cause severe respiratory distress syndrome in humans, and despite SARS-CoV-2 being highly pathogenic and deadlier to humans, SARS epidemics (2003–2004) and MERS outbreaks (2012) were responsible for the deaths of thousands of people worldwide, as shown in Figure 1 [1–8].

**Figure 1.** Worldwide pandemics' timeline. Estimated numbers of infected people and deaths to Spanish flu, severe acute respiratory syndrome (SARS), swine flu, Middle East respiratory syndrome (MERS), and COVID-19 between 1918 and 2020. \* 25 August, according to WHO Coronavirus Disease (COVID-19) Dashboard [9].

Since its emergence in Wuhan, the capital of Hubei Province at the end of 2019, SARS-CoV-2 has quickly spread worldwide, characterizing a pandemic. On 30 January 2020, the Emergency Committee of the World Health Organization (WHO) declared a global health emergency based on the increasing rates of case reporting on different continents [10].

The first case in Latin America was confirmed in February 2020 in Sao Paulo city, Brazil. The confirmed patient had travelled to Lombardy in northern Italy, and the same origin was confirmed for the virus by genome sequencing [11]. Since then, an exponential increase in confirmed cases was observed in the following weeks. No specific drugs or vaccines are available, and health systems are overburdened in every city, especially in Sao Paulo city, which reported 91,198 confirmed cases of COVID-19 and 5623 deaths up to 14 June (last day of the analyzed period in this study), being considered the epicenter of the pandemic in Brazil and in South America. A follow-up on 25 August recorded 250,171 confirmed cases and 11,141 deaths.

μ In addition to the worrying symptoms of pneumonia, acute respiratory distress syndrome, (ARDS) and multiple organ dysfunction, which can lead to death, mainly in people considered to be at risk groups, such as the elderly and people with chronic noncommunicable diseases and respiratory diseases [12–15], some studies have suggested a higher susceptibility of COVID-19 infection of populations exposed to high concentrations of air pollution [16–18]. A recent study demonstrated that long-term average exposure to fine particulate matter (PM2.5) is associated with an increased risk of COVID-19 death in the United States. They found that an increase of 1 µg/m<sup>3</sup> in PM2.5 is related to a growth of 8% of COVID-19 death rate (95% confidence interval [CI]: 2%, 15%) [19]. Despite these evidences, much remains to be investigated to disentangle the relationship between high levels of air pollution and increasing susceptibility to SARS-CoV-2 infection. However, there is no doubt that in densely populated and polluted localities, such as Sao Paulo, strategies to prevent COVID-19 pandemic progression should be more severe [20].

Sao Paulo, one of the largest urban centers of the world, with a population of approximately 12,252,023 inhabitants in 2019 [21], stands out not only for its economic and industrial performance but also due to its high levels of atmospheric pollutants and greenhouse gas emissions, mainly from motor vehicles, responsible for air quality degradation and negative impacts on public health [21–23]. Considering the fleet of 9.1 million vehicles and 87.83 million km travelled daily by all vehicles in Sao Paulo, automobiles account for 69.5% and are responsible for 72.6% of greenhouse gas emissions per day [24,25]. According to measurements carried out by Sao Paulo's Environmental Agency (CETESB), vehicles are responsible for 97% of carbon monoxide (CO), 74% of hydrocarbons, 62% of nitrogen oxides (NOx), and 40% of particulate matter (PM<sup>10</sup> and PM2.5). Concerning the combustion of biofuels, ethanol does not contribute to the emission of particulate matter and releases pollutants not daily monitored by CETESB, such as formaldehyde and acetaldehyde [22]. Although Sao Paulo is one of the largest producers and consumers of biofuels in the world, emissions from fossil fuels are still prevalent, especially in the transport sector [22,26].

Considering only particulate matter with a diameter less than 2.5 microns (PM2.5), 5012 premature deaths occur per year in Sao Paulo city [27]. In addition, a recent case study reported the strong influence of diesel-fueled heavy-duty vehicles on air quality. Merely during the Brazilian truck-driver strike in 2018, the reduction in the PM<sup>10</sup> concentration resulted in the prevention of between six and eight deaths, which implies between 321 and 442 avoided deaths in a year scenario only in Sao Paulo [28].

Several factors, mainly of anthropogenic origin, are related to increased risk of disease outbreaks of vector-borne and zoonotic diseases, such as mega-trends in human population growth, ecosystem reduction and fragmentation, and land use modification and climate change [29–32]. Therefore, the relationship between a pandemic, such as the current COVID-19, global environmental change, and human health is overly complex and deep.

In this context, this ongoing pandemic has also been a great opportunity to discuss the effects of anthropogenic activities on air quality and its implications on public health, as well as highlight the need for a socioeconomic transformation regarding promoting environmentally friendly transport policies and a low-carbon economy towards a sustainable and resilient society [33].

Therefore, we aimed to investigate the air quality improvement during 90 days of social distancing, as well as to compare the associated avoided deaths to COVID-19 burden deaths, considering the relative risk and the economic outcomes in Sao Paulo megacity. The corresponding analysis was based on the comparison of NO2, PM10, and PM2.5 concentrations during quarantine and equivalent periods in 2019.

The paper is structured as follows: Section 2 introduces the method applied in the present study. Section 3 presents all performed analyses and their results. Section 4 presents the discussion of the outcomes. Finally, in Section 5, we finish the paper with concluding remarks and limitations.

### **2. Materials and Methods**

On 11 March, the World Health Organization declared the new coronavirus a pandemic. In Brazil, on 16 March 2020, some activities started to cease, as universities, schools, and some companies imposed home offices, and the first death was reported in Sao Paulo city on 17 March. Thus, the Sao Paulo government officially decreed quarantine on 22 March.

According to these initial events, we chose to investigate the air quality in the city of Sao Paulo, starting on 16 March and dividing the analysis period into 13 weeks, ending on 14 June. As a reference for the comparison, the days in the equivalent period in 2019 (16 March–14 June 2019) were selected. This set of days was designated as the "control period".

Daily records of precipitation, wind speed, and means of air temperature were obtained from the Sao Paulo Environmental State Agency Air Quality Information System (QUALAR) [34] and from the Brazilian Agrometeorological Monitoring System [35]. These meteorological parameters were used as a criterion to exclude the days of the control period that differed from the range of the meteorological conditions of the quarantine period.

In Sao Paulo, the concentrations of atmospheric pollutants are higher in conditions of low ventilation, reduced relative humidity, and absence of precipitation, which occurs mainly in autumn and winter [36–38].

According to CETESB measurements, during the period between 16 March and 14 June 2020, the autumn season, the quarantine period presented an average daily temperature of approximately 25 ◦C and a wind speed of 2 m/s. This period also presented 65 days with precipitation equal to zero or less than 0.5 mm, equivalent to an average of 1.5 mm in total (Table 1). As a reference for the comparison, the control period, between 16 March and 14 June 2019, showed daily means of air temperature of approximately 26 ◦C, average wind speed of 1.9 m/s, and 52 days with precipitation equal to zero or less than 0.5 mm, equivalent to an average of 1.5 mm in total (Table 1). In addition, both analyzed periods showed similarity in terms of atmospheric pressure (on average, approximately 927 hPa), global solar radiation (on average, approximately 677 W/m<sup>2</sup> ) and relative humidity (mean close to 47%) [34].


**Table 1.** Summary of the daily records of precipitation, wind speed, and mean temperature during quarantine and control periods.

Number of observations (N); Standard deviation (SD).

Therefore, based on these observations, the control period demonstrated meteorological conditions similar to those observed during the quarantine period, justifying the selection of 2019 as a control period, as previously reported in published studies related to the analysis of air pollution during the ongoing COVID-19 pandemic [39,40].

The pollutants selected for analysis were nitrogen dioxide (NO2), particulate matter with less than 10 µm (PM10) and particulate matter with less than 2.5 µm (PM2.5). Hourly atmospheric concentration data were obtained from records of QUALAR [34] for both periods. CETESB has 17 automatic monitoring stations in the municipality of Sao Paulo. Not all stations measure all pollutants. Therefore, PM10, PM2.5 and NO<sup>2</sup> concentrations in both periods were collected, according to the availability of data from 15 monitoring stations (Figure 2), and compiled on their respective week, as previously established, to calculate the pollutants' weekly average. μ μ

**Figure 2.** Location of the analyzed automatic monitoring stations in the municipality of Sao Paulo.

CETESB certifies that all monitoring systems of pollutants and meteorology strictly complied with the quality assurance/quality control (QA/QC) procedure established, as approved by the State Council of the Environment (CONSEMA) of Sao Paulo state [34].

The statistical analyses were performed using GraphPad Prism® 5.0 software for analysis and graphing. The data were analyzed for normality by the D'Agostino-Pearson test. As the pollutants' concentrations behave as nonparametric variables, we used the Mann–Whitney test or Kruskal–Wallis test followed by Dunn's post hoc test to compare the weekly average of quarantine and control periods.

The images of NO<sup>2</sup> were detected through the Tropospheric Monitoring Instruments (TROPOMI) on-board European Space Agency (ESA) Sentinel-5 satellite, based on the Differential Optical Absorption Spectroscopy (DOAS) retrieval method in the 405–465 nm spectral range [41].

To represent the reduction in NO<sup>2</sup> during the quarantine period, two images were captured: 25 May 2020 (4:00 pm UTC) and 21 May 2019 (3:33 pm UTC), with minimal cloud cover. The predominant wind direction data were collected from CETESB and were estimated within the limits of the municipality for both analyzed years. The measurements were performed in the south (23◦40'56.921" S 46◦40'33.884" W) and north (23◦32'57.48" S 46◦36'5.659" W) regions of Sao Paulo.

The extraction and processing of the images were performed by Sentinel Hub EO Browser, Google Earth Pro, and ImageJ. This latter software was used to convert the images from RGB to 8-bit grayscale and to quantify the mean gray value as an approach to estimate the variation of NO<sup>2</sup> concentration on the tropospheric vertical column [42,43].

The social distancing index was made available by the telecommunications service providers through a Big Data platform that is managed by the Brazilian Association of Telecommunications Resources (ABR Telecom). According to the telecommunication service providers, the social distancing index is based on the location obtained by cell phone antennas, which "mark" a reference for the place where the cell phone "slept" between 10 pm and 2 am. During the day, a cell phone that has departed from this reference (which is variable but can reach approximately 200 m in the city of Sao Paulo) is considered out of isolation. All this processing is performed by the operator. The index is updated daily, and the operators aggregate and anonymize the data before the generation of the indexes with respect to privacy of the telephone mobile users. Thus, presenting grouped georeferenced data enables the elaboration of public policies that improve measures of social isolation to confront coronaviruses [44].

Based on air pollution reduction results for the quarantine period, the relative risk and avoided deaths attributed to each pollutant were estimated, adopting regression coefficients from epidemiological studies, a well-established method used to estimate the outcomes of air pollution on health [45–52]. The regression coefficients (β) assumed in this study were 0.0008 for PM10, 0.00405 for PM2.5, and 0.00135 for NO2-related deaths, as previously stated [52–55]. The coefficients were used to estimate the probability of mortality associated with the exposure of the pollutants, named relative risk (RR), obtained by Equation (1)

$$\text{RR} = \exp\left[\,^{\beta}\left(\Delta\mathbf{x}\right)\right] \tag{1}$$

where RR is the relative risk of all-cause mortality due to air pollution; β exposure–response coefficients are related to each pollutant; and∆<sup>X</sup> is the decrease in the pollutant concentrations (µg/m<sup>3</sup> ), the difference in pollutant concentrations between the quarantine period and the equivalent period in 2018.

The RR values were used to calculate the attributable fraction (AF), the fraction of deaths attributable to the risk factor of PM10, PM2.5, or NO<sup>2</sup> variation, defined in Equation (2), to estimate the avoided daily mortality during the quarantine period, based on the mean mortality in the control period.

$$\text{AF} = (\text{RR} - 1) \text{/RR},\tag{2}$$

The daily mean of the all-cause mortality in the control period was calculated and multiplied by the attributable factors to estimate the all-cause mortality avoided per day during the quarantine period.

The latest update of all cause daily mortality between March 2019 and June 2019 was collected from the Brazilian Health System Database [56]. For the effects of short-term exposure to PM2.5, only all-cause mortality in people older than 30 years old was considered. The health effects caused by this pollutant justified the selection of this age group. Exposure to PM2.5 is related to hypertension, lung cancer, type 2 diabetes, and other health conditions, which are mostly observed in this age group [57,58]. For PM<sup>10</sup> and NO2, all ages were considered. The daily data of COVID-19 deaths were obtained from the State System for Data Analysis Foundation database from 16 March 2020 to 14 June 2020 [59].

The economic valuation of health outcomes was performed by the value of statistical life (VSL) estimation, a well-known approach [60–64]. Considering the avoided deaths related to air pollutant emissions reduction, an economic gain is expected. However, due to the deaths caused by COVID-19, a considerable economic loss is expected.

To estimate the economic impact of the quarantine and to calculate the trade-off between mortality and economic costs, data on confirmed deaths caused by COVID-19, collected from governmental

records [59], were multiplied by the VSL established value. The same calculus was used to estimate avoided deaths related to air pollutant emissions.

VSL can be defined as a measure of willingness to pay for reducing the risk of dying and thus consists of the monetary value of postponed deaths [60,64]. For each avoided death and death caused by COVID-19, the VSL assumed was US\$ 1.88 million, previously proposed by the OECD [65].

### **3. Results**

### *3.1. Air Quality Improvement during 90 Days of COVID-19 Social Distancing*

As a first-step approach, we verified the monthly records of PM10, PM2.5, and NO<sup>2</sup> concentrations. Except for March and May, PM<sup>10</sup> showed a significant reduction in 2020. PM2.5 also showed a significant reduction in 2020, except in March. Finally, NO<sup>2</sup> showed a significant decrease in April and June (Figure 3).

**Figure 3.** Monthly comparison of the analyzed pollutants. Bars represent the average ± SEM. The analysis was conducted by Kruskal–Wallis test followed by Dunn's post hoc test (# *p* < 0.05; + *p* < 0.01; \* *p* < 0.001). Until 14 June.

Higher volumes of precipitation could explain the results observed for January, February, and March. In addition, the comparison between incompatible days, in terms of meteorological conditions, can affect the interpretation of the analysis. For instance, NO<sup>2</sup> did not show a significant reduction in March and April (Figure 3). However, several studies have shown its reduction during the pandemic in different regions of the world [66–68].

From this initial analysis, we can conclude that the division of the analysis period on a weekly scale, considering meteorological conditions as an exclusion criterion, could provide more accurate information about the variation of pollutants.

We verified the daily records of precipitation, wind speed, and mean temperature. Comparing both periods (2019 and 2020), we observed that wind speed and mean temperature presented similar ranges. For precipitation, we detected a slight difference in a few days of the control period. Considering that precipitation is an important factor that influences the dilution and dispersion of pollutants [69–72], a 12-day control period with precipitation above 10 mm was not considered in this study. After excluding these outlier days, we observed that the means and ranges of meteorological variables were similar over the weeks of the quarantine and control periods (Table 2).

During 13 weeks of social distancing, we observed that the pollutants showed a gradual increase in concentrations over the weeks, similar to the monthly analysis. For PM10, the average concentrations ranged between 16.9 (week 2) and 32.38 µg/m<sup>3</sup> (week 13). PM2.5 presented average concentrations varying between 9.05 (week 4) and 17.15 µg/m<sup>3</sup> (week 12). For NO2, the average concentrations ranged between 17.21 (week 1) and 40.71 µg/m<sup>3</sup> (week 11) (Figure 4).


**Table 2.** Weekly averages of meteorological variables during quarantine and control periods.

Data are expressed as the mean ± standard deviation (SD). μ

μ

**Figure 4.** Averages of social distancing index and PM10, PM2.5, and NO<sup>2</sup> concentrations during 13 weeks of the quarantine period in Sao Paulo city. Bars represent the average ± SEM.

The average social distancing also fluctuated, presenting the lowest index at week 1 (46.5%) and the highest between weeks 2 and 4 (55.4; 56.9 and 53.3%, respectively). In the remaining weeks, the average social distancing ranged between 51.7 and 48.7%.

The best air quality indexes for all pollutants were observed at weeks 2 and 4, with social distances above 53.3%, and the worst concentrations were detected at weeks 12 and 13, with social distances less than 50%.

From week 4 on, we also found a slackening in the social distancing accompanied by an increase in the average pollutant concentrations, quite evident for PM<sup>10</sup> and NO2. The remaining weeks recorded at least two days, reaching six consecutive days in week 13, with less than 50% social distancing. Some factors have contributed to the reduction in social distancing, such as popular manifestations against this measure that occurred in several days, differences in federal and local governments recommendations concerning COVID-19 prevention actions, and most importantly the poor and

working class who do not have alternatives in this matter, can also explain the reduction in its index and consequently the fluctuation of pollutants' concentrations (Figure 4).

As shown in Figure 5, the comparison between the PM<sup>10</sup> concentrations during quarantine and control periods revealed six weeks with a reduction of at least 25%, the highest reaching –45% in week 3 (*p* < 0.001). For PM2.5, we found eight weeks with a reduction of at least 29%, reaching –46% in week 13 (*p* < 0.001). The air quality improvement in terms of NO<sup>2</sup> showed nine weeks with a reduction of at least 33%, reaching –58% in week 3 (*p* < 0.01).

**Figure 5.** PM10, PM2.5, and NO<sup>2</sup> concentrations and trend-line during 13 weeks of quarantine and control periods. Bars represent the average ± SEM.

Following PM2.5, NO<sup>2</sup> was the best indicator of air quality in the analyzed weeks. This finding could be confirmed from the satellite images detected on 21 May 2019, and 25 May 2020, which showed an abrupt decline of 62.8% in tropospheric NO<sup>2</sup> concentrations, represented by the variation of the red pattern (Figure 6).

### *3.2. Associated Health Economics Outcomes*

Based on the results presented in Figure 5, we estimated the potential health and economic benefits related to pollutant reductions achieved during the analyzed weeks of quarantine. As shown in Table 3, the relative risks (RRs) were estimated to be between 0.993 and 1.013 for PM10, 0.996 and 1.055 for PM2.5, and 0.996 and 1.039 for NO2. Thus, the attributable factors (AF) ranged 0.2–0.8% for PM10, 0.9–3.2% for PM2.5, and 4–6.9% for NO2. As expected, because NO<sup>2</sup> presented the greatest levels of reduction compared to the control period, its weekly RR and AF values were also higher in relation to other analyzed pollutants.

**Figure 6.** Tropospheric vertical column of NO<sup>2</sup> detected on 21 May 2019, and 25 May 2020. Data from Tropospheric Monitoring Instrument (TROPOMI) on board Sentinel-5 Precursor, based on the Differential Optical Absorption (DOAS) retrieval method (405–465 nm spectral range). The north-northwest (NNW) predominant wind direction was estimated within the limits of the municipality for both analyzed years. The measurements were performed in the south (23◦40'56.921" S 46◦40'33.884" W) and north (23◦32'57.48" S 46◦36'5.659" W) regions of Sao Paulo.


**Table 3.** Relative risk (RR) and attributable factor (AF) related to PM10, PM2.5, and NO2 exposures.

Therefore, concomitant with the increase in the number of deaths from COVID-19 infection, which reached 5623 on the last day of this analysis, the decrease in PM10, PM2.5, and NO<sup>2</sup> concentration levels prevented 78,337 and 387 premature deaths, respectively, in Sao Paulo city (Figure 7).

**Figure 7.** COVID-19 deaths, PM10, PM2.5, and NO<sup>2</sup> avoided deaths during the weeks of the quarantine period in Sao Paulo city.

Given the monetary value for the observed health outcomes from the analyzed improvement of pollutant concentrations by the VSL approach, a well-known and widely used method [65,73], we applied the economic valuation of deaths caused by COVID-19 and of the avoided deaths due to the reduction in short-term exposure to PM10, PM2.5, and NO<sup>2</sup> (Table 4).


**Table 4.** Economic valuation of deaths caused by COVID-19 and of avoided deaths due to the reductions of PM10, PM2.5, and NO2.

\* Based on the VSL value.

During the analyzed weeks of the quarantine period in Sao Paulo city, the deaths caused by COVID-19 accounted for an economic health loss of approximately US \$10.5 billion. On the other hand, we can consider a potential economic health benefit in terms of the prevented premature deaths related to the observed improvement in air quality. Thus, we estimated that the expressive number of avoided deaths due to PM10, PM2.5, and NO<sup>2</sup> reductions can represent an economy of approximately US \$720 million, considering the results of NO2, the best air quality indicator observed.

### **4. Discussion**

In a tightly connected and integrated world, the impacts of emerging infectious diseases can be devastating, causing large-scale mortality and morbidity, as we are observing over the days of this global outbreak. However, the adverse effects go beyond, and the size and persistence of the economic impacts of the ongoing COVID-19 pandemic are still uncertain, and our findings raise intriguing questions.

On the one hand, we estimated the economic health loss associated with deaths related to COVID-19 in Sao Paulo city, which was exceedingly raised, even before reaching the important goal of flattening the curve of coronavirus infections, which makes us wonder about the efficiency of the applied public policies and the commitment of the strategies of controlling the pandemic and population's commitment to comply with social distancing. On the other hand, we observed that the epidemic control actions were closely related to the improvement of air quality; due to the major source of air pollution in the region, the emissions from vehicle exhaust were significantly reduced.

The quarantine and cities in shutdown have also caused a supplementary effect, the improvement of air quality, which was observed in different regions of the world, such as India, China, Italy, France, and United States [66,67,74]. In Rio de Janeiro, Brazil, a study showed that the confinement of the population between March and April caused a reduction in road traffic and economic activities and led to a significant decrease in CO (between 37.0 and 43.6%) and NO<sup>2</sup> (between 24.1 and 32.9%) levels and a non-significant reduction in PM<sup>10</sup> concentrations [75].

Regarding NO<sup>2</sup> and particulate matter, our findings showed that, in Sao Paulo, the impacts on its concentrations were more significant. During 16 March and 14 June 2020, NO<sup>2</sup> decreased to 33%, reaching 58%, the decrease in PM<sup>10</sup> ranged between 25 and 45%, and the levels of PM2.5 decreased between 29 and 46% in comparison to the control period in 2019.

Air quality improvement could also be observed by satellite images worldwide. A dramatic reduction in air pollution over China during February 2020 accounted for over 30% less in NO<sup>2</sup> levels compared to early January when power plants were operating at normal levels. In addition, the study also estimated a reduction of CO<sup>2</sup> emissions by 25% [66]. Xu et al. (2020) showed that from January to March 2020, the decrease in NO<sup>2</sup> concentration was of great significance and much higher than other pollutants, such as PM<sup>10</sup> and PM2.5 [68]. Moreover, Anjum (2020) also showed, from satellite images, a remarkable reduction of air pollution in Italy, France, and the USA [67].

Our results corroborate these findings since we also showed a significant reduction in NO<sup>2</sup> concentration in Sao Paulo city, which was greater than the other analyzed pollutants, confirmed by the abrupt decline of 62.8% in tropospheric NO<sup>2</sup> concentrations also observed by satellite images.

Human activities do not change only the atmosphere with the emission of pollutants or only the climate system with the emission of greenhouse gases; the consequences are systemic since they can directly affect human health, socioeconomic, and political stability. Emissions of air pollutants related to the exacerbation of the prevalence of cardiorespiratory diseases, type 2 diabetes, and mental disorders; land use modification associated with deforestation that modifies host/vector interactions, elevates disease risk, and drives wildlife zoonotic disease emergence; and mega-trends in climate accompanied by extreme weather conditions make the population increasingly vulnerable. These conditions can cause several socioeconomic impacts, such as morbidities, mortality, migration, poverty exacerbation, and violent conflict, affecting people of all ages and all nationalities [29,76,77].

In this context, the current global crisis we are experiencing goes beyond showing that environmental degradation is closely related to human resilience. In times of crisis, borders do not matter, showing the importance of cooperation and engagement of governments and all the people to work together for a common purpose and solution.

Foregone income related to laid-off workers or unable to work, disruption of supply chains, restrictions on international travel enforced by governments, and unprecedented consequences on stock markets are examples of already observed economic and social impacts. Individualities and structural differences of each country make forecasts difficult; however, there are some published estimates. According to Fernandes (2020), for Brazil, the GDP growth in 2020, assuming shutdown lasts three months, can reach −4.0% [from −5.9% to −2.1%]. Moreover, according to the OECD, global GDP growth is projected to slow from 2.9% in 2019 to 2.4% in 2020 [78–80]. The COVID-19 pandemic has hit underserved populations and communities of color particularly hard, exacerbating longstanding health disparities in Brazil. This fact can be observed in Sao Paulo municipality, which districts of poorer areas have been suffering most of the cases of COVID-19.

Politicians and some businessmen are against social distancing measures due to the impact on the economy, claiming that the virus is less harmful than presented; they do not value life. Life is more valuable than the economy—there is no need to have good economic indicators if the population dies. Nevertheless, there will be economic effects of the distance measures and the impacts on the mental

health of the populations affected by them. In addition to anxiety due to fear of the disease, there are effects of prolonged quarantine for several weeks faced by a huge number of people around the world.

An impact assessment evaluation can consider the exposure to a multi-pollutant mixture. Adding pollutant-specific effects may be justified when levels of the specific pollutants are clearly not correlated. However, this approach would result in an overestimated analysis due to uncertainties about the contributions of each pollutant to the health effect [81,82].

In this research, we estimated the health effects concerning a reduction in PM10, PM2.5 and NO<sup>2</sup> concentrations. Thus, to avoid a biased analysis, we considered, in the health economic evaluation, the outcomes of NO2, the best air quality indicator observed. Accordingly, our results showed that air quality improvement contributed to a substantial fraction of the avoided deaths associated with significant monetized health co-benefits, signaling that governments have the capability to improve air quality through policy measures.

We found that the decrease in PM10, PM2.5, and NO<sup>2</sup> concentration levels could prevent 78,337 and 387 premature deaths, respectively, corresponding to an economy of approximately US \$720 million and providing evidence for policy making by the quantification and monetization of air pollution-related health effects. This is a tool for environmental planning purposes since the monetized value of the avoided premature mortality typically dominates the calculated benefits of air pollution regulations.

The consideration of exclusively NO<sup>2</sup> abatement in the quarantine period is a conservative simplification. This constraint may have biased our estimation downwards. If we add the effect of PM2.5 and PM10, we double the impact of prevented deaths due to air pollution, meaning a saving of US \$1.5 billion. Considering the maintenance of these mobility patterns in the new "business as usual" (BAU) COVID-19 era, in a 10-year period, this savings would achieve approximately US \$60 billion. Both health and economic gains associated with air pollution reductions give a positive perspective of the efforts towards keeping air pollution reduced even after the pandemic.

Unprecedented challenge demands an unprecedented response. Thus, shifting the BAU approach of life will be the most important and mandatory step to prevent future adverse events. For that, resilience is indispensable for facing new problems and challenges in our increasingly globalized world. The COVID-19 pandemic, forcing a drastic change in the BAU way of life, offers a unique scenario to take the opportunity during the crisis to move toward a more sustainable society and economy, in terms of speeding up the transition of make cities more inclusive, developing urban resilience, improving preventive health measures, and implementing green strategies for economics, especially for transport and energy sectors, since they are major contributors to the degradation of global air quality [20,33,83,84].

In this sense, our analysis presents an estimate of what could be achieved if effective measures would be taken towards air pollution emissions control in an urban center of a developing country, such as Sao Paulo. The implementation of permanent or partial home-office jobs, dynamization of e-commerce, as well as improvement of energy efficiency, support of a low-carbon economy, and investments in cleaner transportation solutions are some recommendations for sustainable procedures that can diminish air pollution emissions in megacities, providing a transition to urban resilience and sustainability.

### **5. Conclusions**

We found that, during 90 days of COVID-19 social distancing in Sao Paulo city, Brazil, 5623 deaths occurred due to this new disease estimated in an economic health loss of US \$10.5 billion. In contrast, we observed a significant air quality improvement. During the analyzed weeks, the decreases in PM2.5, PM10, and NO<sup>2</sup> levels reached 45%, 46%, and 58% reductions, respectively, in comparison to the control period in 2019. The positive impact on air quality in terms of PM2.5, PM10, and NO<sup>2</sup> concentrations avoided 78,337 and 387 premature deaths, respectively, and prevented approximately US \$720 million on health costs, considering the results of NO2, the best air quality indicator observed.

Our study faced some limitations. First, in addition to the reduced number of deaths due to air pollution, we should consider that there are other benefits attributable to the reduction in air pollution itself and could also have positive benefits in reducing preventable noncommunicable diseases [85,86]. Thus, a more comprehensive analysis should contemplate other health outcomes, such as the reductions in hospitalizations due to cardiorespiratory diseases. Second, meteorological data were collected and used as a criterion for selected days of the control period to ensure that the analysis was not carried out on incompatible days in terms of the meteorological similarity of both periods analyzed, minimizing the influence of these factors on pollutant dispersion. In this sense, an air quality dispersion modeling approach could be usefully explored in further future researches.

These findings have significant implications for understanding how important it is to improve the strategies for sustainable practices in urban centers, especially with regard to air pollution mitigation actions. Furthermore, we highlight the crucial role of appropriately allocating investments and adopting efficient measures to protect human health both during and after the COVID-19 global health crisis.

**Author Contributions:** Conceptualization, S.G.E.K.M. and D.D.; formal analysis, D.D., M.V.d.C., and S.G.E.K.M.; writing—original draft preparation, D.D., M.V.d.C., and S.G.E.K.M.; writing—review and editing, D.D., M.V.d.C., and S.G.E.K.M.; supervision, S.G.E.K.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** Supported by Grant Number 1808529 from the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). Supported by Grant Number 2018/26193-3 from the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP).

**Acknowledgments:** The authors are grateful to the Academic Editor Idiano D'Adamo and Etta Shen for all support provided.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
