POLITICAL DETERMINANTS OF COVID-19 MORTALITY; FACTIONAL POLITICS IN VIETNAM; RISING MEDIA ENCLAVE EXTREMISM by THUY THI NGUYEN A DISSERTATION Presented to the Department of Political Science and the Division of Graduate Studies of the University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy June 2022 DISSERTATION APPROVAL PAGE Student: Thuy Thi Nguyen Title: Political Determinants of Covid-19 Mortality; Factional Politics in Vietnam; Rising Media Enclave Extremism This dissertation has been accepted and approved in partial fulfillment of the requirements for the Doctor of Philosophy degree in the Department of Political Science by: Tuong Vu Chairperson Yeling Tan Core Member Clare Evans Core Member Michael Dreiling Institutional Representative and Krista Chronister Vice Provost for Graduate Studies Original approval signatures are on file with the University of Oregon Division of Graduate Studies. Degree awarded June 2022 2 © 2022 Thuy Thi Nguyen 3 DISSERTATION ABSTRACT Thuy Thi Nguyen Doctor of Philosophy Department of Political Science June 2022 Title: Political Determinants of Covid-19 Mortality; Factional Politics in Vietnam; Rising Media Enclave Extremism My dissertation consists of three independent projects. The first one investigates possible relationships between certain macro-systemic political variables and Covid mortality rates. Using multilevel modeling, I analyze countries’ trajectories of Covid mortality rates between March 2020 and January 2022. I found that countries with federal system, on average, tend to have higher death rates than those in unitary system. Democracy is found to be negatively associated with Covid mortality overall, but democracy’s effects on the trajectory of Covid rates depends on what subgroup of countries are considered. Government effectiveness persists as a significant factor that is negatively associated with Covid deaths. In countries where people have higher trust in government the curves of death tolls tend to be flatter. The second project addresses the debate on authoritarian resilience with evidence from Vietnam. The dominant view in the debate focuses on political institutions and argues that institutions help dictators to resolve “the problem of authoritarian power-sharing”. I test two main claims of this dominant views: institutions facilitate power access and rule-based power succession. I found that these processes are rather superficial in the case of Vietnam and show how the persistence of factional politics, coupled with the historical context of the country, are 4 embedded in current politics. In contrast to the expectation of institutionalist scholars, top leaders were not bound by retirement age limits. Furthermore, evidence indicates that hometown ties constitute rigid political factions. The third project examines the rise of media enclave extremism, showing how it successfully mobilized a historically inactive ethnic population into the far-right circle. My research provides insight into the production side of media enclave extremism in an ethnic minority news outlet on the far- right, The Epoch Times. I conduct a discourse-historical analysis of its political news articles and identify two main discursive strategies. First, the outlet links their traditional enemy, the Chinese Communist Party, with US political entities such as the Democratic Party, Liberals, the mainstream media, and “bad” immigrants. Second, The Epoch Times groups itself with US right-wing media, “good” immigrants, naturalized citizens, and conservatives. 5 CURRICULUM VITAE NAME OF AUTHOR: Thuy Thi Nguyen GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon, Eugene The University of Tokyo, Japan National University, Vietnam DEGREES AWARDED: Doctor of Philosophy, 2022, University of Oregon Master of Science, 2020, University of Oregon Master of Arts, 2013, The University of Tokyo Bachelor of Arts, 2006, Vietnam’s National University AREAS OF SPECIAL INTEREST: Data Science Quantitative Research Methods PROFESSIONAL EXPERIENCE: Data Scientist, Reason Foundation, Aug. 2021 - present Project Officer, Vitens-Evides International, 2014 - 2018 GRANTS, AWARDS, AND HONORS: Oregon Sylff Fellowships for International Research, University of Oregon, 2021 Klonoski Teaching Assistant Award, University of Oregon, 2020 Southeast Asia Studies Award, University of Oregon, 2019 Promising Scholars Award, University of Oregon, 2019 - 2020 6 PUBLICATIONS: Thuy Nguyen. Review of Vietnam's Strategic Thinking during the Third Indochina War. The Journal of Military History, vol. 85 no. 1, 2021, p. 285–86 Thuy Nguyen (with Tuong Vu). Book chapter. Vietnam in the Reform Era, Cambridge History of the Vietnam War. The Cambridge History of the Vietnam War, Volume III: Endings (Cambridge University Press, forthcoming [2022]) 7 ACKNOWLEDGMENTS I wish to express sincere appreciation to Professor Tuong Vu for his guidance throughout this project. I also thank the members of my committee, Professor Yeling Tan, Professor Clare Evans, and Professor Michael Dreiling for their valuable input. The data collection in this project was supported in part by the US-Vietnam Research Center, University of Oregon. 8 I dedicate this dissertation to my daughter, An Nguyen, who gives me all the strength to overcome many challenges in both private life and academic life. I also dedicate this work to my husband, Khoi Nguyen, who has been a great source of support and encouragement. For my beloved grandmother, who gave me my best childhood memories. A special feeling of gratitude to my advisor, Professor Tuong Vu whose excellent mentorship has raised me up during the time in graduate school. 9 TABLE OF CONTENTS Page CHAPTER 1: MACRO-SYSTEMTIC POLITICAL VARIABLES AND PUBLIC HEALTH OUTCOMES............................................................................................... 15 I. Political Epidemiology ............................................................................................. 15 II. Politics as Determinants of Public Health Outcome ............................................... 19 Democracy ............................................................................................................. 19 Systems of Internal Governance (Formal State Organization) .............................. 20 Government Effectiveness ..................................................................................... 22 Trust in Government .............................................................................................. 23 III. Data ........................................................................................................................ 24 Outcome Variable: Covid Mortality Rates ............................................................ 24 Time-invariant Predictors of Covid Mortality Rates ............................................. 25 IV. Analysis ................................................................................................................ 27 Level 1 Variables ................................................................................................... 27 Level 2 Variables ................................................................................................... 28 V. Result ...................................................................................................................... 29 Average Observed Growth of Covid Mortality by Country .................................. 29 Partitioning Within and Between-country Components ........................................ 30 Linear and Quadratic Growth Curve Model .......................................................... 30 VI. Conclusion ............................................................................................................. 36 Reference ..................................................................................................................... 38 10 Page CHAPTER 2: FACTIONAL POLITICS, INSTITUTIONS, AND AUTHORITARIAN RESILIENCE: EVIDENCE FROM VIETNAM ......................................................... 43 I. Authoritarian Resilience ........................................................................................... 43 II. Communist Regime ................................................................................................. 44 III. Vietnam Case: Factional Politics and Historical Legacy ....................................... 47 Factional Politics .................................................................................................... 47 Factional Politics and Regionalism in Vietnam ..................................................... 48 IV. Data ........................................................................................................................ 50 V. Retirement Age Limit Rule ..................................................................................... 52 Rules and Norm of Retirement Age Limit ............................................................. 53 Violation of Age Limit Rule .................................................................................. 54 VI. Factions: Analysis of Hometown Networks .......................................................... 55 Transforming the Networks ................................................................................... 56 Person-to-working Term Network ......................................................................... 56 Person-to-person Network ..................................................................................... 57 Term-to-term Network ........................................................................................... 58 Identification of Hometown Networks .................................................................. 59 Regionalism and Hometown Factions ................................................................... 59 Degree of Centralization Over Time...................................................................... 63 Flow of Power Across Time and Branches............................................................ 65 VII. Conclusion ............................................................................................................ 67 Reference ..................................................................................................................... 68 11 Page CHAPTER 3: RISING MEDIA ENCLAVE EXTREMISM....................................... 74 I. The Rise of Media Enclave Extremism .................................................................... 74 II. Radical Media and Democratic Values ................................................................... 76 Radical Media ........................................................................................................ 76 How Media Enclave Extremism Rise .................................................................... 77 Enclave Extremism in a Far-right Minority Ethnic News Media .......................... 80 III. Data and Method .................................................................................................... 83 Data Collection ...................................................................................................... 83 Data Partition ......................................................................................................... 83 Discourse-historical Approach............................................................................... 84 IV. Analysis ................................................................................................................. 84 Bundling Enemies: The CCP, Democrats, “Bad” Immigrant, Antifa, BLMs ....... 85 Bundling Identities: “Our” Knowledge, “Our” Values, “Our” People .................. 90 V. Conclusion ............................................................................................................. 97 Reference .................................................................................................................... 98 Appendices .................................................................................................................. 101 A. List of Countries Coded as Federal States ........................................................ 101 B. Special Cases in Coding Hometown ................................................................. 101 12 LIST OF FIGURES Figure Page 1. Top 5 countries started out and ended with highest mortality rates ...................... 30 2. Intercepts and slopes of the mortality trajectories ................................................. 32 3. Mortality at different level of democracy .............................................................. 33 4. Number of working years left ................................................................................ 54 5. Average age ........................................................................................................... 55 6. Person-to-working term network ........................................................................... 57 7. Term-to-term network ............................................................................................ 58 8. The core and peripheral provinces ......................................................................... 60 9. Change in number of seats ..................................................................................... 62 10. Degree of centralization ......................................................................................... 63 11. Seats at the hometown tables ................................................................................. 64 12. Flow of power across time and branches ............................................................... 66 13. Number of articles published 2013 - 2021 ............................................................. 82 14. Analytical framework of two main discursive strategies deployed in ET ............. 85 13 LIST OF TABLES Table Page 1. Descriptive statistics .............................................................................................. 24 2. Analysis result ........................................................................................................ 35 3. The study’s hypotheses and summary of findings ................................................. 36 4. Count of seats in the Politburo and the government .............................................. 51 5. Description of variables ......................................................................................... 52 6. Number of seats held by the 5 provinces across 6 terms ....................................... 61 7. Examples of “Experts” interviewed or quoted in ET political news ..................... 92 14 CHAPTER 1: MACRO-SYSTEMTIC POLITICAL VARIABLES AND PUBLIC HEALTH OUTCOMES I. Political Epidemiology “Medicine is a social science, and politics nothing but medicine at a larger scale” - the founding father of social medicine and modern cellular pathology, Rudolf Virchow, once famously stated. His approach has inspired a field of studies that views politics as determinants of health. The statement is interpreted as “human health and disease are the embodiment of the successes and failures of society as a whole, and the only way to improve health and reduce disease is by changing society by, therefore, political action” (Mackenbach 2009, 181). Facing multiple global diseases in recent decades, scholars call for more explicit embrace of “health political science”, maintaining that understanding policy processes is pivotal for the potential to influence policy change (de Leeuw, Clavier, and Breton 2014). “Political epidemiology” becomes a more urgent interdisciplinary quest of scholars in political science and epidemiology. While conventional research often focuses on social factors such as living conditions, education, healthcare access, and community as key determinants of health, Dawes (2020b) contends that political determinants are the causes of the social determinant of health - the determinants of the determinants – which are far more influential and more insidious drivers of health and health equity. That is, they “create the social drivers that affect all other dynamics of health and create structural barriers to equity for population groups that lack power and privilege” (Dawes 2020a, 80). In other words, political determinants of health are “the instigators of the causes of inequities”, which then have a cascading effect on public health and social life (Dawes 2020b). Bambra (2005) argues health is political because 1) it is unequally distributed among groups just like any other resource or commodity, 2) its social determinants depend on political action, and 3) it is an important aspect of human rights and citizenship. Kickbusch (2015) further elaborates that examine health through the lens of political determinants means “analyzing how different power constellations, institutions, and processes, interests, and ideological positions affect health within different political systems and cultures and at different levels of governance”. Vu (2011) shows how politics matters greatly in dealing with epidemics through a framework of political 15 processes that involves four phases: pre-political, announcement, mitigation, and rebuilding. For example, the lack of democracy in Thailand and Vietnam accounted for delays in making public announcement of the outbreaks, while a low degree of power centralization was responsible for the slow and infective responses in Malaysia and Vietnam. Viewing politics as “the process of making and executing collective decisions”, Mackenbach (2014) identifies three main political variables that impact public health: structures (states, legislative and executive agencies), processes (elections, lobbying, law-making), outputs and outcomes (the laws, taxes, social security benefits, public services). Dawes (2020b) further develops a model that integrates three elements of politics that affect health in America: the systematic processes of structuring relationships, distributing resources, and administering power. In this paper, I expand the above framework to examine the extent to which macro-systemic political variables relate to health outcomes using Covid mortality data. As the Covid-19 pandemic swept through the world, countries were hit with different level of severity, resulting in mortality rates ranging from hundreds to thousands of deaths per million population. I ask: do macro-systemic political factors have any bearing with the outcome, in particular the mortality rates? By macro-systemic political factors, I refer to the following: 1) political regime (level of democracy), 2) institutions (systems of internal governance), 3) state capacity (government effectiveness), and 4) trust in government. These factors speak to countries’ preparedness in the face of a global health crisis like Covid-19 pandemic. In other words, I seek to examine the degree to which countries’ political features can account for the variance in Covid mortality rates at the beginning phase and throughout the period between March 2020 and January 2022. Do countries in different level of these political characteristics experience different level of Covid deaths? Is there a certain kind of macro- systemic political features, which had been built up for years, that appear to equip countries with better ability to handle this health crisis? With the political and governance resources that were already in place before the pandemic hit, which type of countries tend to be able to reduce Covid deaths better than others? Putting in the framework of multi-level modeling method, the above 16 questions can be translated as: Is there a role of country political characteristics that can explain the variability of starting point and trajectory of death rates from country to country? What type of countries started out with high death rates in the first time point? What type of countries increased their death rates more rapidly over time? This project differs from the existing literature on politics and public health outcomes in two ways. First, while the scholarship often focuses on macro-systemic factors in their relationships with long-term public health outcomes in a non-crisis condition, my research establishes relationships between these factors and a short-term health outcome in the context of a global crisis. The most often asked question is the role of democracy in delivering public health results. Others consider the degree of power centralization in a country, the efficiency of public policy implementation, or the ability of states to ensure compliance to improve public health outcome. In these research projects, outcomes are often measured by mortality rates of children under five years old, life expectancy at birth, or death rates of some preventable diseases. Clearly, there is no perfect match for the challenge we are currently facing – understanding the relationships of political factors and millions of deaths due to Covid-19. Second, recent research projects that address the Covid-19 mortality as a dependent variable often focus on social policies such as travel ban, lock-down, social distancing, masking, surveillance and schools closing (Haug et al. 2020; Liang et al. 2020; Aleta and Moreno 2020; Pung et al. 2021; Migone 2020a). My work looks deeper into the macro-systemic factors that might influence those social policies. Recent studies found seemingly contradictory results with respect to the relationship between some governance indicators and pandemic control. While Haug et al. (2020) found no significant positive correlation between the Worldwide Governance Indicators and the level of virus control effectiveness, Liang et al. (2020) found that Covid-19 mortality is negatively associated with government effectiveness. As time progresses, countries have diverged in their responses to the pandemic and witnessed different consequences. This calls for asking deeper questions on political determinants that underlie the social determinants, the causes of the causes, and their relationships with mortality in health crises. 17 This paper establishes an association between common macro-systemic political variables and a newly observed outcome. That is, once controlled for the level of wealth and healthcare infrastructure, how much of the variation in the Covid mortality rates could be explained by national factors such as democracy, internal governance system, level of government effectiveness, and trust in government? More specifically, I ask: Are democracies better at saving lives? Does it matter what type of systems of internal governance when it comes to millions of lives at stake in a global health crisis? Can we expect those with higher score of government effectiveness, an indicator often positively associated with economic development, to be also better at limiting mortality? Do governments who have higher level of citizens’ trust reduce more death rates due to Covid? I found that democracy has negative main effect on Covid mortality overall, but effects on the trajectory of Covid rate increase over time depends on what subgroup of countries are considered. Types of internal governance matters to the extent that countries with federal system, on average, tend to have higher death rates than those in unitary system. Government effectiveness persists as a significant factor that is negatively associated with Covid deaths. In countries where people reported to have higher trust in government, the curve of death rates tends to be flatter. My research contributes a timely analysis of a special case to the interdisciplinary literature that examines the relationships between major national characteristics of political system and public health outcomes. Unlike any other extreme outcomes that we can dismiss to stay focus on the common trend, an extreme outcome in public health such as Covid mortality is particularly important. In the most recent evaluation of countries’ preparedness for the next global health crisis, a panel of respectful researchers concludes that “all countries remain dangerously unprepared for meeting future epidemic and pandemic threats (The 2021 Global Health Security GHS report).”1 That is, rich and poor countries alike need to reconsider their approaches to protect public health. My analysis provides a direct implication which suggests that we need to account for macro-systemic variables when thinking about future crises management. 1 https://www.ghsindex.org/report-model/ 18 II. Politics as Determinants of Public Health Outcome In this section, I identify some macro-systemic political determinants of public health that exist in the literature. Also, I explore the mechanisms by which these factors impact the health outcomes that scholars have been studied. Democracy Democracy is argued to be more beneficial for public health than other political regimes thanks to two mechanisms. First, democratic regime empowers more people to have their voice in political decision making and get politicians to be more responsiveness (Franco, Álvarez-Dardet, and Ruiz 2004; Ncayiyana 2004). Miller (2008) finds that the enfranchisement of women in democracy helps reducing child mortality in the US in the early twentieth century. Kudamatsu (2012) shows infant mortality drops by 1.2 percentage points following democratization in sub- Saharan Africa since 1990. In both studies, the authors argue that democracy does improve health by improving political responsiveness. Second, democracy can indirectly contribute to better public health by improving economic and living conditions. A study conducted on a panel data of 170 countries spanning 46 years reveals that democracies are more likely than autocracies to bring about health gains in some cause- specific mortality such as cardiovascular diseases and injuries. That is, increases in democratic experience had direct and indirect effects on increases in government health spending, hence declines in mortality from some diseases (Bollyky et al. 2019). Democracy may contribute to better population health indirectly by improving socioeconomic factors such as income, income inequality, education, and social access to health care services (Safaei 2006). Democracy is also found to have clear and strong impact on population health, particularly reducing child mortality rates and improving life expectancy at birth. A specific analysis of 33 countries that underwent democratic transition concluded that, on average, democratization reduced child mortality (Pieters et al. 2016). On the other hand, there are studies that find mixed effect or raise questions about the non-effect of democracy on health or even cases where a non-democracy could do better. A cross-national 19 study of 168 states between 1970 – 2000 by Ross (2006) shows that although democracies spend more resource on education and health than nondemocracies, these benefits often accrue to middle and upper income groups, leaving the low income group little or no benefit, which renders democracy’s overall non-effect on infant and child mortality rates. Lin (2015)’s examination of the relationship between the state and disaster vulnerability in 150 countries between 1995 and 2009 shows that state capacity, regime type, and their interactions determine the magnitude of human losses. More specifically, democracies are more effective in preventing human losses due to predicable disasters, but not when the causes are unpredictable. Even more, there are studies that demonstrates how a non-democracy can be more effective in preventing public from diseases, as how China handled the Severe Acute Respiratory Syndrome SARS epidemic in 2003. Although the lack of democratic freedom, i.e., government’s censorship of news, silencing people, is said to dramatically impacted Chinese population during the early phase of SARS epidemic (Ruger 2005), the authoritarian Chinese regime is argued to be effective in imposing draconian quarantine measures, closing public space, screening a large part of the population, hence successfully contained and ended SARS (Hesketh and Zhu 2004). Also, according to these authors, China has historically been successful in eradicating some diseases by aggressive closure of brothels and opium dens by using dictatorship power. As the world is facing the global pandemic Covid-19 that has claimed millions of lives, the relationships of democracy and public health is, once again, urgently contested (Diamond 2020).2 However, the magnitude and direction of such effect is not yet clear. This paper is an effort to contributes to the large empirical literature that tries to identify the effect of democracy on health. I first hypothesize: Hypothesis 1: Countries with higher level of democracy tend to have lower Covid mortality rates. Systems of Internal Governance (Formal State Organization) 2 https://www.theatlantic.com/ideas/archive/2020/04/americas-covid-19-disaster-setback-democracy/610102/ 20 System of internal governance, or formal state organization, is the way a state organizes its political power over its territory. Political scientists generally dichotomize system of internal governance into federal or unitary. A federal polity is characteristically noncentralized where “the powers of government within them are diffused among many centers, whose existence and authority are guaranteed by the general constitution, rather than being concentrated in a single center” (Elazar 1987, 34). About forty percent of the world’s population live in federal and quasi-federal countries today. In contrast, unitary states have one center of power at the national level and multiple subordinate bodies at lower levels. The debate on the degree to which power concentration is more efficient in public management and resolving crises gives rise to both proponents and opponents of federalism model. On the one hand, federalism can induce good governance thanks to its mechanism of including multiple interests (Palermo and Alber 2015). Some economists argue that creating space for initiatives from grassroot, empowering local people to take care of their own wellbeing is the best way to recover from a disaster (Storr 2015). Austrian school of thoughts embraces the idea that local people have the local knowledge that would best help them overcome challenges during crises. That is, a less centralized power system, or a more polycentric approach would benefit communities in many ways (McGinnis 1999). On the other hand, federal systems can be thought to be less effective than unitary systems to handle significant national crises because “with their structures of shared rule and self-rule, they cannot always respond to crises in the rapid, focused, complementary, and coordinated ways said to be more characteristic of unitary systems” (Kincaid, Tarr, and Wälti 2010). Canada, for example, experienced mixed effects of its federalism on emergency management during Covid- 19 that “enables specific responses to local realities, which represent an important failsafe in a federal system, but at the same time which can reduce cooperation and misalign critical policy steps” (Migone 2020b). Also during the pandemic, the case of Italy is “emblematic of a low- capacity response” in a decentralized system that responded slowly and incrementally to a fast- moving crisis (Capano 2020). 21 However, not all federal states responded to the pandemic in the same way. Hegele and Schnabel (2021) find a great variation among federal states along two dimensions of federal decision making: the degree to which power was centralized and the mechanism by which decision- making process was conducted, unilateral or coordinated. They found that while Austria and Switzerland adopted a centralized approach in introducing containment measures, Germany used decentralized decision-making approach. With respect to coordination, Austria and Germany utilized coordinated schemes, while unilateralism prevailed in Switzerland. To further test the relationships between the systems of internal governance, or the form of state organization and health outcome in the Covid-19 context, I hypothesize that: Hypothesis 2: Federal states tend to have lower Covid mortality rates than unitary states. Government Effectiveness Crisis management, even in such diverse events as disasters, riots, or terrorism, can be said to share some similarities because it demands decision makers to confront with situations of great danger of people lives, uncertainty, and urgency (Rosenthal, Charles, and Hart 1989). Policymakers generally face four critical tasks in a crisis like Covid-19: sense-making (understanding a crisis); making critical decisions and solving emerging problems; crisis communication; and terminating a crisis through “exit strategies” (Boin, Lodge, and Luesink 2020). The policymakers’ ability of crafting effective responses to an existential threat such as Covid-19 arguably depends a large part on the governance infrastructure and decision-making processes that are pre-installed before such crisis taking place. It is well perceived that governance effectiveness and corruption control has strong impact on economic development and social equity (Arora and Chong 2018). These factors might manifest themselves in many ways that could save lives during a public health crisis like Covid-19. These might be distribution of urgent healthcare resources, coordinating human resources, or other issues related to policies and spending to control Covid. It is found that in crises such as refugee migration and early phase of Covid-19, administrations that were “structurally prepared, learned during preceding crises, and displayed a high quality in their network cooperation with other 22 administrations and with the civil society” performed significantly better, on average (Schomaker and Bauer 2020). I thus hypothesize: Hypothesis 3: Countries with higher score of governance effectiveness have lower Covid mortality rates. Trust in Government The notion of citizens’ trust in government has been discussed extensively in political science, especially in relations to development and prosperity. Fukuyama (1995) proposes a thesis of trust and economics, which posits that social trust can influence a country’ economic development and hence brings a more prosperous economy by lowering transaction costs and enhancing market efficiency. An overall trust in government that is built prior to a crisis has lingering effect on forming more positive view about the efficiency of public management during a time of crises (Keele 2007; Khan 2016). Several studies focus on the relationships between trust in government and resolving social and health problems, but during routine times rather than in crises (He and Ma 2021; Van de Walle and Bouckaert 2003). In the Covid-19 context, trust in government could mean that citizens would believe that the government is doing their best to protect the public from death. Furthermore, trust influences citizens’ behavior in the form of protective actions and their support for intervention politics intended to combat the pandemic (Robinson et al. 2021). That is, the higher level of trust in government, the more legitimacy people grant their government, the more willing they are to cooperate with it in handling the crisis (Mizrahi, Vigoda‐Gadot, and Cohen 2021). Schraff (2020) found that the rising number of Covid-19 cases was associated with increased political trust, while two standard determinants of political trust, economic evaluations and social trust, lost explanatory power. In the study, the author defines political trust as “diffuse support for political institutions, referring to the perceived competence and legitimacy of public institutions”, measured by a household survey in the Netherlands in March 2020. I seek to continue examine the linkage between the level of trust in government and Covid deaths across countries, thus I propose the fourth hypothesis: Hypothesis 4: Countries with higher level of trust in government had lower Covid mortality rates. 23 III. Data The data used in this analysis is complied from multiple public sources: Our World in Data, Economist Intelligence Unit Democracy Index, The World Bank’s World Development Indicators, OECD Trust in Government Indicator, Global Health Security Index. This is a repeated measurement data, where only time points and the corresponding Covid mortality rates calculated at each time point are varied. All other covariates are considered time-invariant. Except Covid mortality data, all other variables are for year 2019. I purposefully chose data year 2019 rather than the earlier or later years to fit the purpose of the analysis, which is to understand how already established political features impact countries’ ability to handle Covid-19 pandemic. Table 1. Descriptive statistics Statistic N Mean St. Dev. total_deaths_per_million 3,384 480.46 769.77 gdp_percap 3,384 15,251 20,893 democracy 3,384 56.04 21.75 gov_eff 3,339 0.02 0.98 health_index 3,315 44.01 14.18 trust_gov 873 49.89 16.22 unitary_federal 3,384 0.14 0.35 unitary_semi_federal 3,384 0.22 0.41 Notes: Number of observations (N) is the sum of the product of the number of the time point measurements each country has and the number of countries. For three variables gdp_percap, democracy, unitary_federal, the sample consists of 151 countries, of which 115 countries have 23 time point measurements, 20 countries have 22, 7 countries have 21, 6 have 19, the last 3 countries each has 14, 13, and 11 points, respectively, hence 115*23 + 20*22 + 7*21 + 6*19 + 14 + 13 + 11 = 3384 total points of measurements. The same method of calculation is applied for other variables. 24 Outcome Variable: Covid Mortality Rates The data of Covid mortality rates was retrieved from Our World in Data in February 2022.3 The data consists of 3384 observations of monthly Covid deaths per million population, spanning from April 1, 2020, to February 1, 2022. Although the original data has measurements of January and February 2020, I do not include these two months because there are too few observations during this time. The data is partitioned in 23 monthly time points, where month 0 starts on 2020-04-01 and month 22 ends on 2022-02-01. Time-invariant Predictors of Covid Mortality Democracy Index The EIU Democracy Index measures the quality of democracy in 167 countries on a continuous scores 1 to 100. The Democracy Index is calculated based on five categories: electoral process and pluralism, political participation, the functioning of government, civil liberties, and political culture. The value used in this analysis is “Democracy index (EIU)” for year 2019.4 I use this EIU measurement rather than Polity V or other method of dichotomous regime type because EIU’s continuous scale provides more nuanced differences between countries. Systems of Internal Governance I accounted for two ways of categorizing systems of internal government. The first method records that there are 26 federal states in the world (see appendix 1), which results in 22 federal states in my dataset. The second method records additional 11 states that have devolved governance power structure, regionalized unitary, and federacy (see appendix 2).5 In these semi- federal states, but central governments delegate a part of governance power to regional and local level governments. This approach recognizes a total of 37 federal states in the world, which results in 32 federal states in my dataset. China is special case. Although two semi-autonomous regions in China, Hong Kong and Macau, have some economic autonomy, I code China as unitary state in this study to reflect the fact that China has highly centralized power of pandemic 3 https://github.com/owid/covid-19-data/tree/master/public/data/excess_mortality 4 https://www.gapminder.org/data/documentation/democracy-index/ https://www.eiu.com/topic/democracy-index 5 https://cs.mcgill.ca/~rwest/wikispeedia/wpcd/wp/l/List_of_countries_by_system_of_government.htm 25 control. The two approaches of coding systems of internal governance led to different outcomes, which are presented in the analysis section below. Government Effectiveness Government effectiveness is an indicator in The World Bank’s World Development Indicators, ranging from -2.5 (less effective) to 2.5 (more effective). The indicator measures the quality of public services, the quality of the civil service, the quality of policy formulation policy implementation, and the credibility of a government's commitment to their policies. I retrieved the data from the World Bank’s data package “data360r”. The indicator used in this analysis is “Government Effectiveness, Estimate” for year 2019, which includes data of 190 countries. Trust in Government Trust in government data is collected by OECD in 40 advanced industrial countries.6 There is no equivalent measurement of trust in government for other countries that I am aware of. This limitation inevitably shrinks the whole sample size in the estimating the effect of trust in government on Covid deaths. Notably, countries in this subset tend to have higher scores in all other covariates such as democratic index score, GDP per capita, and Health Security Index. However, I argue that the sample size of 38 data clusters (countries) and 23 repeated measurements still provides sufficient statistical power to estimate the effect of each covariate. Controlling Factors To single out the impact of the political factors, I control for factors that might confound the relationships between those explanatory factors and the death tolls outcome. Two controlling variables in this analysis are public health infrastructure (Global Health Security Index) and the level of wealth of nation (GDP per capita). Global Health Security Index 6 OECD (2022), Trust in government (indicator). doi: 10.1787/1de9675e-en (Accessed on 14 February 2022) https://data.oecd.org/gga/trust-in- government.htm 26 Global Health Security Index is a “comprehensive assessment and benchmarking of health security and related capabilities across the 195 countries that make up the States Parties to the International Health Regulations (IHR [2005])”.7 The index score is ranging from 1 to 100. The data used in this analysis is “Global Health Security Index” for year 2019. GDP per capita I use GDP per capita as a proxy for the wealth level of countries. Controlling for the wealth level would help us better understand the effect of political determinants on Covid deaths that is separated from other material factors. This data is collected by the World Bank, retrieved from R package ‘data360r’ for year 2019. IV. Analysis I use multilevel method in this analysis, which is more appropriate than a traditional general linear model (GLM) for two reasons. First, the repeated measurement data, where each country unit forms a data cluster, violates the key assumption of GLM models of residual independent and identically distributed. Second, the research question of interest is inherently multilevel. That is, I examine how predictors at level-2, i.e., countries’ characteristics, affect the outcome at level- 1, i.e., measurement at different time points. A multilevel approach has clear advantages in dealing with clustering data and this type of research question. Multilevel modeling is flexible in representation of time, making it a great advantage in longitudinal data analysis. The model can accommodate the differences in the number of measurement occasions between countries (Raudenbush 2002). Level 1 Variables Variables at level 1 are: unique countries ID, monthly Covid mortality rates per million population, time points. Covariates at level 2 are the countries’ political characteristics including level of democracy, form of internal governance system, level governance effectiveness, and level of trust in government. Internal governance systems is a binary covariate, coded 1 for 7 https://www.ghsindex.org/about/ 27 federal states and 0 for unitary states. Three other covariates, namely democracy, government effectiveness, and trust in government, are continuous measurements with different scales. For the ease of interpretation, I standardized these three measurements when fitting the model. The analysis controls for two level-2 potential confounders, gdp per capita and health security index. I use R packages lmer, lmer4 to fit the models. Plots were generated using R package ggplot2. Data management was handled using dplyr, tidyverse, purr, lubridate, etc. I first apply an unconditional growth model to serve as a baseline for comparison to the succeeding models. In the second modeling step, I apply conditional multilevel growth models to examine model results. The last step is to add country characteristics as additional predictors of Covid mortality. The conditional models are: 1) a level-1 model that specified Covid mortality rates predicted by a quadratic function of time of measurement, 2) and a level-2 model composed of the prediction of level-1 model parameters as a function of country characteristics. The initial level-1 model can be described as follows: 𝑀𝑜𝑟𝑡𝑎𝑙𝑖𝑡𝑦 𝑡𝑖 = 𝛽0𝑖 + 𝛽1𝑖 (𝑡𝑖𝑚𝑒) + 𝛽2𝑖 (𝑡𝑖𝑚𝑒 𝑠𝑞𝑢𝑎𝑟𝑒𝑑) + 𝑟𝑡𝑖 Where 𝑀𝑜𝑟𝑡𝑎𝑙𝑖𝑡𝑦 is the mortality rates for country 𝑖 at time 𝑡 and is the initial status or intercept for country 𝑖 at time 0, 𝛽1𝑖 is the linear rate of change, 𝛽2𝑖 is the curvature that represents the acceleration or deceleration in each country’s Covid deaths growth trajectory, 𝑟𝑖𝑗 is the residual for each country. I used a quadratic model based on inspection of the data and statistical testing of alternative growth functions. Compared to linear model, quadratic function improved the model fit significantly, Chisq = 267.12, p < 0.001 (see appendix 4). Level 2 Variables At level-2, the level-1 parameters are modeled using countries’ characteristics. The dichotomous predictor (unitary_federal) is uncentered, all other continuous variables are standardized. That means, coefficient for dichotomous predictor represented the effect for the group coded 1, countries with federal form of state organization. Coefficient for continuous predictors (democracy, government effectiveness, trust in government) represented the effect for a country 28 with average score in that predictor. The level-2 equations for the mortality initial status and growth rate parameters are as follows: Initial Status, 𝛽0𝑖 = 𝛾00 + ∑𝛾0𝑘(𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑜𝑟 𝑘) + 𝑢0𝑖 Linear Rate of Change, 𝛽1𝑖 = 𝛾10 + ∑𝛾1𝑘(𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑜𝑟 𝑘) + 𝑢1𝑖 Curvature, 𝛽2𝑖 = 𝛾10 + ∑𝛾2𝑘(𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑜𝑟 𝑘) + 𝑢2𝑖 Where 𝛾00 is the Covid mortality rates intercept at the first time point in the dataset, i.e., April 1, 2020, for all countries, each 𝛾0𝑘 represents the average partial regression coefficient relating the predictor of interest to country’s initial status, and 𝑢0𝑖 is the residual between the fitted predictor value for each county and the country’s observed Covid mortality rates at the first time point. For each rate of change parameter (i.e., 𝛽1𝑖 and 𝛽2𝑖), each country’s rate of change, 𝛽p𝑖, is modeled as a function of average Covid mortality rates of change, 𝛾pk. Each 𝛾pk represents the average partial regression coefficient relating the predictor of interest to country’s rate of change, and 𝑢p𝑖 is the residual between the fitted predictor value for each country’s rate of change and the observed rate of change. To detect multicollinearity, the problem in which a predictor is already explained by other predictors, I calculate variance inflation factor (VIF). All VIFs are small than 10, the limit value often recommended as a good limit. V. Result Average Observed Growth of Covid Mortality by Country First, I conduct an observational analysis (Figure. 1), which shows that most typical trajectories are curvilinear in form, with accelerating growth across time point. All countries show increasing variability in Covid mortality over time. Notably, five countries that have highest mortality rates at initial status in 2020 (Belgium, Spain, Italy, Netherlands, and France) are those that have relatively low slopes. Top five countries that have highest mortality rates at the ending point 29 (Peru, Bulgaria, Bosnia and Herzegovina, Hungary, and Montenegro) are those that have relatively higher slopes than other countries. All countries have positive growth over time, although there are apparent differences in both linear rate of change and rate of curvature. I test these observations using the multilevel statistical models in the next section. Figure 1: Top 5 countries started out and ended with highest mortality rates Partitioning Within- and Between-country Components Next, I partition the variance in the outcome variable Covid mortality per million population into its within- and between- country components. The sample consisted of 151 countries, making a total of 3384 points of measurement. There is some degree of imbalance on group sizes. The number of monthly data point measurements for each country ranges from 11 to 23. There are 115 countries (76.2% of all countries) that have full 23 time point measurements, 20 countries have 22 measurements, 7 countries have 21 points, 6 countries have 19 points, one has 14, one has 13, and the last one has 11 time point measurements. This is not a huge problem in multilevel model, where the maximum likelihood estimator can account for that imbalance in group size. 30 The fixed effect for the intercept means that the grand mean of Covid mortality rates per million population is 472.48 for the pooled data (all time points in all countries). The level-1 residual variability is 299012, the level-2 random intercept variance is 291718. The intra-class correlation ICC is 291718/(291718 + 299012) = .494, which means the model implies that 49.4% of the variance in Covid mortality is associated with between-country differences. Equivalently, on average, within a given country, Covid mortality rates has a .494 correlation. This means there is a meaningful amount of dependence in the Covid mortality per million population rates that needs explanation from countries’ characteristics. Linear and Quadratic Growth Curve Model The estimated random effects show that variance among the intercepts is 40216. Compared to the unconditional model above, the variance associated with countries has been reduced vastly (70%, from 291718 to 40218) when we account for time factor. Variance among the slopes is 3297, and level-1 residual variance is 20943. The results indicate that there is variability among the country intercepts and country slopes. The random effects of the intercepts and slopes are negatively correlated (-0.53), meaning that those countries who have high death tolls at the early phase tend to increase at slower rates, while countries who have lower death tolls tend to increase at faster rates. There remains significant variability at level 1 after accounting for time. The fixed effects indicate that there is a significant mean linear slope (𝛾 = 52.02, sd = 4.68, p < 0.001). The model implies that on average, there’s significant increase of 52 deaths per month. The intercept with negative value is not meaningful in this context. Result from quadratic growth curve model indicates a strong support for a quadratic component for this data. The fixed quadratic effect, 0.78, is statistically significant (sd = 0.06, p < 0.001). Furthermore, the formal test result also indicates a great improvement in model fit with the inclusion of the quadratic component (Chisq = 148.4, p < 0.001). For this reason, I will include quadratic term in the succeeding models. Systems of Internal Governance 31 Starting from this step, I expand the model to include level-2 predictors. The first predictor to be included is system of internal governance of a country, i.e., a federal or unitary system, controlling for gpd per capita (log) and health security index (standardized). Two ways of categorizing federal states mentioned in section 3.2.2 result in two estimate outputs. Using the first method that counts 22 states in the dataset as federals results in no significant effect of systems of internal governance on the Covid mortality. When using the second method that counts 32 states in the dataset as federal, systems of internal governance become a significant factor that is associated with higher Covid deaths, controlling for GDP per capita (log) and health security index. Considering the fixed effect results, the main effect of internal governance system (federal coded 1) is significant. The model implies that federal state has intercept higher than a unitary state by 92.8 Covid deaths per million population (t = 2.42, SE = 38.2, p = 0.016). When including the cross-level interaction between time and federal, the regression of trajectory slopes on federal is not significant. This means that the model-implied slopes do not differ in magnitude of slopes between federal and unitary states. The figure below demonstrates that the intercepts of the trajectory of Covid mortality are significantly higher for federal states compared to unitary states, but the slopes of the trajectories do not significantly differ from each other. Figure 2: Intercepts and slopes of the mortality trajectories Democracy 32 In this section I added democracy as an additional predictor at level 2 to examine its effect on Covid mortality, controlled for GDP per capita and Health Security Index. The original scale of democracy index is a continuous with 1 point incremental, I standardized democracy scale to have more meaningful interpretation. M𝑜𝑟𝑡𝑎𝑙𝑖𝑡𝑦 𝑡𝑖 = (𝛾00 + 𝛾01𝑡𝑖𝑚𝑒 + 𝛾0Democracy) + (𝛾01∗𝑡𝑖𝑚𝑒 + 𝛾0 ∗ Democracy) + (𝑢0𝑖 + 𝑢1𝑖 Democracy + 𝑟𝑡𝑖) The fixed effects indicate significance in both main effect and interaction effect of democracy. Countries with higher democracy score have lower intercept and steeper regression slopes. One standard deviation on democracy score is associated with 63.89 less Covid deaths per million population (p = .004). The significant interaction between time and democracy (𝛾^ = 21.78) shows that the slope of trajectory of Covid mortality increases by approximately 21.78 units for each one standard deviation increases in the predictor democracy (p < 0.001). Figure 3: Mortality at different level of democracy The plot above further examines the relationships between democracy and Covid mortality. Three lines are democracy value at the mean, one standard deviation above the mean and one standard deviation below the mean. One standard deviation of democracy index in the dataset is equivalent to 21.75 points. Accordingly, countries with higher democracy have lower number of Covid deaths per million population at initial time point, but its trend increases quicker. Countries with lower democracy index score started out with higher death rates but its trend 33 increases slower. In other words, higher democratic countries have lower death rates early in the pandemic, but as time progresses their mortality rates increases faster than those who are less democratic. Starting at around time point 3, i.e., July 2020, they suffered more from Covid deaths compared to lower democratic countries. Government effectiveness Next, I included government effectiveness to the model. This predictor is also standardized for an ease of interpretation. The model’s results indicate that government effectiveness matters greatly when it comes to Covid mortality. On standard deviation increases in government effectiveness resulted in -68.77 lower Covid deaths (p = 0.043). I also run a model with a cross- level interaction term between the slope (time) and government effectiveness, which showed no significant effect of this predictor. Trust in Government Trust in government have negative effect on the slope of Covid mortality. For each standard deviation increase in trust in government, the slope was 1.78 unit lower, p = 0.026. Note that the model now has 873 observations from 38 countries rather than 147 countries due to the limited availability of the indicator trust in government. Because of the difference in sample size, the variance estimates in the random effects can only be compared to the previous models with caution. Both the between-country (31278) and within country variances (1874) get smaller. The correlation between time in month and the intercept, -0.46, is also smaller than previous models. Interestingly, in this smaller sample size democracy show significant negative effect on the slope of Covid trajectory, which is an opposite direction compared to the previous model. That is, one standard deviation increase in democracy is associated with 2.78 lower unit in the rate of change of Covid mortality. The differences in the effect of democracy on Covid mortality is probably due to the distinctiveness in the scores in the two groups of countries. In the smaller sample of 38 countries, democracy is much less varied, ranging from 31.1 (Russia) to 98.7 (Norway), mean = 78.18, sd = 13.61, compared to the score in the bigger sample size of 148 countries which ranges 34 from 13.2 to 98.7, mean = 56.04, sd = 21.75 in. Essentially, the subgroup of democracy in this model is those in much higher level of democracy than those in large sample. In this join model using a smaller sample size, government effectiveness factor persists its negative main effect on Covid deaths. That is, one standard deviation increases in government effectiveness is associated with 2.67 less death, p = 0.025. The table below summarizes the results of hypotheses testing. Table 2: Analysis result Variables Model 1 Model 2 Model 3 Model 4 time_month 34.644***(5.01) 34.800***(4.69) 34.912***(4.714) 109.285***(15.06) I(time_month^2) 0.793***(0.065) 0.794***(0.065) 0.797***(0.065) 0.014(0.146) GDP per capita 2.683(13.345) 5.907(14.172) 31.251(18.718) 228.876**(80.442) (log) Health Security 2.682(1.494) 3.106(1.611) 4.680**(1.753) 5.753(3.922) Index Federal system 92.857*(38.272) 93.612*(38.349) 87.149*(38.956) 159.318**(61.088) (Yes = 1) scale(democracy) -63.891**(22.2) -50.582*(23.645) 153.128*(77.777) time_month:scale 21.777***(4.46) 21.581***(4.482) -27.869*(12.216) (democracy) scale(gov_eff) -68.772*(33.759) -266.969*(113.229) scale(trust_gov) -9.196 (37.384) time_month:scale -17.894*(7.640) (trust_gov) Num.Obs. 3315 3315 3293 873 RMSE 136.50 136.50 136.94 161.43 * p < 0.05, ** p < 0.01, *** p < 0.001 Notes: Standard errors shown in parentheses. 35 Table 3: The study's hypotheses and summary of findings H1 Countries with higher level of Partially supported by the data. democracy tend to have lower In the sample of 148 countries, countries with higher Covid mortality rates. level of democracy have lower death rates at the beginning, but the death tolls tend to increase faster than those with lower level of democracy. In the group of 38 OECD countries, those in higher democracy experience slower rate of Covid deaths. H2 Federal states tend to have Data shows an opposite effect. lower Covid mortality rates Federal states tend to have higher rate of Covid than unitary states. deaths compared to those in unitary system. H3 Countries with higher score of Supported by the data. governance effectiveness have Government effectiveness significantly associated lower Covid mortality rates. with lower rate of Covid deaths. H4 Countries with higher level of Supported by the data. trust in government had lower In the groups of 38 OECD countries, those with Covid mortality rates. higher level of trust in government experience significantly slower rate of Covid deaths. VI. Conclusion The result from analysis above both supports and challenges some findings in the literature on the political determinants of public health. On the contrary with the argument that a more decentralized power structure would enable local entrepreneurship, hence helping community revival after disaster, I found an opposite effect of power decentralization in relations to Covid deaths. That is, states under federal system tend to have higher Covid deaths. 36 Democracy has negative main effect on Covid mortality overall. That is, higher democracies experienced less Covid deaths tolls on average. However, democracy effects on the trajectory of Covid rate increase over time are different depending on what subgroup of countries are considered. In the group of 148 countries, democracy shows positive relationship with the trend of Covid deaths, meaning higher democracies also tend to have higher increase in death rates. In the smaller group of 38 OECD countries, democracy shows negative relationship with the slope of Covid deaths, which mean those with in advanced democracy witnessed slower rates of Covid deaths overtime. Two other factors, government effectiveness and trust in government, are statistically significant predictors of Covid mortality. Government effectiveness persists as a significant factor that is negatively associated with Covid deaths. My result is aligned with Liang et.al. (2020) at the early phase of the pandemic that shows Covid-19 mortality is negatively associated with government effectiveness. Trust in government is an important factor that negatively associated with the trajectory of Covid deaths. In countries that people reported to have higher trust in government, the curve of death tend to be flatter. The research contributes to the emergent studies that examines the relationships between major political and governance features and public health outcomes. This paper establishes a correlation, not causal relationship, between macro-systemic factors and a specific public health outcome, Covid-19 mortality rates. Nevertheless, by showing how countries’ political characteristics relate to the consequences of Covid deaths, the paper strongly suggests that politics matters a great deal in crises management. It also supports the notion of politics being “the determinants of the determinants” developed by Dawes (2020b) that is introduced at the beginning of this paper. A major part of political science works focuses directly on the impact of political factors on the economy, which can be indirectly inferred to population health. A better economy, arguably, can lead to better public health outcomes. But as we have witnessed in the current pandemic, richer countries are not necessarily better at saving more lives. As this paper points out, political and 37 governance factors have a large role in limiting deaths. My work is an effort to establish an association between macro-system political factors and Covid mortality. Further examinations in terms of causal inferences will be needed to deepen our understanding of politics and public health. Reference Aleta, Alberto, and Yamir Moreno. 2020. "Evaluation of the potential incidence of COVID-19 and effectiveness of containment measures in Spain: a data-driven approach." BMC medicine 18 (1):157-157. doi: 10.1186/s12916-020-01619-5. Arora, Puneet, and Alberto Chong. 2018. "Government effectiveness in the provision of public goods: the role of institutional quality." Journal of applied economics 21 (1):175-196. doi: 10.1080/15140326.2018.1550593. Bambra, Clare, Debbie Fox, and Alex Scott-Samuel. 2005. "Towards a politics of health." Health promotion international 20 (2):187-193. doi: 10.1093/heapro/dah608. Boin, Arjen, Martin Lodge, and Marte Luesink. 2020. "Learning from the COVID-19 crisis: an initial analysis of national responses." Policy design and practice 3 (3):189-204. doi: 10.1080/25741292.2020.1823670. Bollyky, Thomas J., Tara Templin, Matthew Cohen, Diana Schoder, Joseph L. Dieleman, and Simon Wigley. 2019. "The relationships between democratic experience, adult health, and cause-specific mortality in 170 countries between 1980 and 2016: an observational analysis." The Lancet (British edition) 393 (10181):1628-1640. doi: 10.1016/S0140- 6736(19)30235-1. Capano, Giliberto. 2020. "Policy design and state capacity in the COVID-19 emergency in Italy: if you are not prepared for the (un)expected, you can be only what you already are." Policy & society 39 (3):326-344. doi: 10.1080/14494035.2020.1783790. Dawes, Daniel E. 2020a. "Health Inequities: A Look at the Political Determinants of Health During the COVID-19 Pandemic." American journal of health studies 35 (2):77. doi: 10.47779/ajhs.2020.191. Dawes, Daniel E. 2020b. The political determinants of health. Baltimore, Maryland: Baltimore, Maryland : Johns Hopkins University Press. 38 de Leeuw, Evelyne, Carole Clavier, and Eric Breton. 2014. "Health policy - why research it and how: health political science." Health research policy and systems 12 (1):55-55. doi: 10.1186/1478-4505-12-55. Elazar, Daniel J. 1987. Exploring federalism. Tuscaloosa, AL: Tuscaloosa, AL : University of Alabama Press. Franco, Álvaro, Carlos Álvarez-Dardet, and Maria Teresa Ruiz. 2004. "Effect of democracy on health: ecological study." BMJ 329 (7480):1421-1423. doi: 10.1136/bmj.329.7480.1421. Fukuyama, Francis. 1995. Trust : the social virtues and the creation of prosperity, Social virtues and the creation of prosperity. New York: New York : Free Press. Haug, Nils, Lukas Geyrhofer, Alessandro Londei, Elma Dervic, Amelie Desvars-Larrive, Vittorio Loreto, Beate Pinior, Stefan Thurner, and Peter Klimek. 2020. "Ranking the effectiveness of worldwide COVID-19 government interventions." Nature human behaviour 4 (12):1303-1312. doi: 10.1038/s41562-020-01009-0. He, Alex Jingwei, and Liang Ma. 2021. "Citizen Participation, Perceived Public Service Performance, and Trust in Government: Evidence from Health Policy Reforms in Hong Kong." Public performance & management review 44 (3):471-493. doi: 10.1080/15309576.2020.1780138. Hegele, Yvonne, and Johanna Schnabel. 2021. "Federalism and the management of the COVID- 19 crisis: centralisation, decentralisation and (non-)coordination." West European politics:1-20. doi: 10.1080/01402382.2021.1873529. Hesketh, Therese, and Wei Xing Zhu. 2004. "Effect of restricted freedom on health in China." BMJ 329 (7480):1427-1427. doi: 10.1136/bmj.329.7480.1427. Keele, Luke. 2007. "Social Capital and the Dynamics of Trust in Government." American journal of political science 51 (2):241-254. doi: 10.1111/j.1540-5907.2007.00248.x. Khan, Haroon A. 2016. The idea of good governance and the politics of the global South : an analysis of its effects. New York: New York : Routledge, is an imprint of the Taylor & Francis Group, an Informa business. Kickbusch, Ilona. 2015. "The political determinants of health—10 years on." BMJ (Online) 350 (jan08 2):h81-h81. doi: 10.1136/bmj.h81. 39 Kincaid, John, G. Alan Tarr, and Sonja Wälti. 2010. "Federalism and the Global Financial Crisis: Impacts and Responses: Introduction." Europe en formation n° 358 (4):3-14. doi: 10.3917/eufor.358.0003. Kudamatsu, Masayuki. 2012. "HAS DEMOCRATIZATION REDUCED INFANT MORTALITY IN SUB-SAHARAN AFRICA? EVIDENCE FROM MICRO DATA." Journal of the European Economic Association 10 (6):1294-1317. doi: 10.1111/j.1542- 4774.2012.01092.x. Liang, Li-Lin, Ching-Hung Tseng, Hsiu J. Ho, and Chun-Ying Wu. 2020. "Covid-19 mortality is negatively associated with test number and government effectiveness." Scientific reports 10 (1):12567-12567. doi: 10.1038/s41598-020-68862-x. Lin, Thung-Hong. 2015. "Governing Natural Disasters: State Capacity, Democracy, and Human Vulnerability." Social forces 93 (3):1267-1300. doi: 10.1093/sf/sou104. Mackenbach, J. P. 2009. "Politics is nothing but medicine at a larger scale: reflections on public health’s biggest idea." Journal of epidemiology and community health (1979) 63 (3):181- 184. doi: 10.1136/jech.2008.077032. Mackenbach, Johan P. 2014. "Political determinants of health." European journal of public health 24 (1):2-2. doi: 10.1093/eurpub/ckt183. McGinnis, Michael D. 1999. Polycentric governance and development : readings from the Workshop in Political Theory and Policy Analysis. Ann Arbor, MI: Ann Arbor, MI : University of Michigan Press. Migone, Andrea Riccardo. 2020a. "The influence of national policy characteristics on COVID- 19 containment policies: a comparative analysis." Policy design and practice 3 (3):259- 276. doi: 10.1080/25741292.2020.1804660. Migone, Andrea Riccardo. 2020b. "Trust, but customize: federalism's impact on the Canadian COVID-19 response." Policy & society 39 (3):382-402. doi: 10.1080/14494035.2020.1783788. Miller, Grant. 2008. "Women's Suffrage, Political Responsiveness, and Child Survival in American History." The Quarterly journal of economics 123 (3):1287-1327. doi: 10.1162/qjec.2008.123.3.1287. 40 Mizrahi, Shlomo, Eran Vigoda‐Gadot, and Nissim Cohen. 2021. "How Well Do They Manage a Crisis? The Government's Effectiveness During the COVID‐19 Pandemic." Public administration review 81 (6):1120-1130. doi: 10.1111/puar.13370. Ncayiyana, Dan J. 2004. "Is democracy good for people's health? A South African perspective." BMJ 329 (7480):1425-1426. doi: 10.1136/bmj.329.7480.1425. Palermo, Francesco, and Elisabeth Alber. 2015. Federalism as decision-making : changes in structures, procedures and policies. Leiden, Netherlands ; Boston, [Massachusetts]: Leiden, Netherlands ; Boston, Massachusetts : Brill Nijhoff. Pieters, Hannah Dr, Daniele PhD Curzi, Alessandro PhD Olper, and Johan PhD Swinnen. 2016. "Effect of democratic reforms on child mortality: a synthetic control analysis." The Lancet global health 4 (9):e627-e632. doi: 10.1016/S2214-109X(16)30104-8. Pung, Rachael, Alex R. Cook, Calvin J. Chiew, Hannah E. Clapham, Yinxiaohe Sun, Zongbin Li, Borame L. Dickens, Stefan Ma, Kenneth Mak, Chorh Chuan Tan, Derrick Heng, Mark I. Cheng Chen, and Vernon J. Lee. 2021. "Effectiveness of Containment Measures Against COVID-19 in Singapore: Implications for Other National Containment Efforts." Epidemiology (Cambridge, Mass.) 32 (1):79-86. doi: 10.1097/EDE.0000000000001257. Raudenbush, Stephen W. 2002. Hierarchical linear models : applications and data analysis methods. 2nd ed. ed. Thousand Oaks: Thousand Oaks : Sage Publications. Robinson, Scott E., Joseph T. Ripberger, Kuhika Gupta, Jennifer A. Ross, Andrew S. Fox, Hank C. Jenkins‐Smith, and Carol L. Silva. 2021. "The Relevance and Operations of Political Trust in the COVID‐19 Pandemic." Public administration review 81 (6):1110-1119. doi: 10.1111/puar.13333. Rosenthal, Uriel, Michael T Charles, and Paul 't Hart. 1989. Coping with crises : the management of disasters, riots, and terrorism. Springfield, Ill., U.S.A.: Springfield, Ill., U.S.A. : C.C. Thomas. Ross, Michael. 2006. "Is Democracy Good for the Poor." American journal of political science 50 (4):860-874. doi: 10.1111/j.1540-5907.2006.00220.x. Ruger, J. P. 2005. "Democracy and health." QJM : monthly journal of the Association of Physicians 98 (4):299-304. doi: 10.1093/qjmed/hci042. Safaei, Jalil. 2006. "IS DEMOCRACY GOOD FOR HEALTH?" International journal of health services 36 (4):767-786. doi: 10.2190/6V5W-0N36-AQNF-GPD1. 41 Schomaker, Rahel M., and Michael W. Bauer. 2020. "What Drives Successful Administrative Performance during Crises? Lessons from Refugee Migration and the Covid‐19 Pandemic." Public administration review 80 (5):845-850. doi: 10.1111/puar.13280. Schraff, Dominik. 2020. "Political trust during the Covid-19 pandemic: Rally around the flag or lockdown effects?" European journal of political research. doi: 10.1111/1475- 6765.12425. Storr, Virgil Henry. 2015. Community revival in the wake of disaster : lessons in local entrepreneurship. Houndmills, Basingstoke, Hampshire ; New York, NY: Houndmills, Basingstoke, Hampshire ; New York, NY : Palgrave Macmillan. Van de Walle, Steven, and Geert Bouckaert. 2003. "Public Service Performance and Trust in Government: The Problem of Causality." International journal of public administration 26 (8-9):891-913. doi: 10.1081/PAD-120019352. Vu, Tuong. 2011. "Epidemics as Politics with Case Studies from Malaysia, Thailand, and Vietnam." Global Health Governance IV (2). 42 CHAPTER 2: FACTIONAL POLITICS, INSTITUTIONS, AND AUTHORITARIAN RESILIENCE: EVIDENCE FROM VIETNAM Authoritarian Resilience The durability of authoritarian regimes depends largely on rulers’ ability to resolve two fundamental conflicts: “the problem of authoritarian control”, in which dictators face the threats from the masses, and “the problem of authoritarian power-sharing”, in which they encounter the challenges from the elites (Svolik 2012). Authoritarian scholarship in recent decades show that dictators are increasingly engaging democratic-like political institutions such as parties, legislature, governance committees, courts to resolve both problems. On the problem of authoritarian control, dictators often utilize institutions to limit mass discontent and boosting legitimacy. Common strategies are: distributing patronage to some selected groups using mass organizations or electoral clientelism (Handlin 2016; Blaydes 2011), boosting political ideology and nationalism through media (Tismaneanu 2013; Zhao 1998), collecting and mediating popular grievances via legislatures (Salomon in Balme and Sidel 2007; Malesky 2009), helping the public resolve garden-variety disputes through judicial system (Solomon 2007; Anthony 2007). On the problem of authoritarian power-sharing, the main interest of my paper, political institutions are also believed to be useful. Some scholars maintain that institutions are established to reduce the threat of rebellion by opposition (Gandhi 2008; Gandhi and Przeworski 2006), and to co-opt with the “selectorate” (Bueno De Mesquita 2003). Others emphasize that institutions facilitate power-sharing among the ruling elites by alleviating commitment and monitoring problems (Geddes 1999; Brownlee 2007; Svolik 2012; Boix and Svolik 2013; Magaloni 2008). Svolik (2012) argue that formal, deliberative and decision-making institutions in dictatorship can ease relations among elites by helping the ruling coalition monitor the ruler's commitment to share power. Boix and Svolik (2013) further develop this notion by showing the complex interaction between collective action, commitment, and monitoring problems. The authors identify two ways through which institutionalized interaction between the dictator and the allies can help stabilize authoritarian power-sharing and therefore enhance the survival of authoritarian regime. First, such institutionalized interaction increases transparency among those in power, reducing the potential for misperceptions about the dictator’s compliance with a power-sharing 43 commitment, thus avert unnecessary rebellions. Second, when power-sharing is institutionalized, formal rules embody the power-sharing compromises, making the dictator’s compliance a publicly observable signal. Studies indeed shows that dictators who successfully manage to institutionalize power-sharing and the transition of power often last longer (Geddes 1999; Brownlee 2007; Svolik 2012; Boix and Svolik 2013). Among the authoritarian regimes, single-party regimes tend to be more durable than military and personalist types thanks to its inclination to negotiation. In fact, single- party regimes outside the Soviet system were observed to “have tried to negotiate institutional changes that allow the opposition some participation and satisfy international donors and lenders, while not actually giving up control of the government and the resources attached to it” (Geddes 1999, 141). Communist Regime While the core pillars of the communist system collapsed in 1989-1991, five surviving regimes namely China, Vietnam, Laos, Cuba, and North Korea, resist the waves of regime transition. As a subgroup of dictators, communist rulers also face similar threats from the masses and the ruling coalition as other types of autocracies. Finding from studies on this special regime type resonates with scholarship on general authoritarian regime. That is, leaders of communist party do deploy adaptive political institutions for the purpose of stabilizing the regime (Bernstein 2013; Dimitrov 2013). To control the mass, they make use of ideology (Chen 2016; Amstrong 2013), limit the exposure of revolutionary contagion (Koesel and Bunce 2013; Bunce and Wolchik 2013), develop a proxy of accountability for citizen complains (Dimitrov 2013). To control the elite coalition, they allow for changes in the composition of the selectorate (Gallagher and Hanson 2013, 203) , co-opt private entrepreneurs (Tsai 2013). In short, scholars contend that institutional design plays an important role in maintaining communist regime’s resilience. Dimitrov (2013) concludes that the fate of authoritarian resilience is “predicated on sustained experimentation in the economic and ideological realms, as well as on effort to increase inclusion and to create meaningful institutions of accountability” 44 (304). Abrami, Malesky and Zheng (2013) assert that “the survival or collapse of a regime is ultimately a matter of the ability of those in power to obey their own rules” (239). Despite this expectation, we can observe several cases where rulers blatantly disobey their own rule. A major case in the authoritarian regime literature is China, whose resiliency was largely explained by the regime’s tendency to become more institutionalized and rule-based regarding its power succession (Nathan 2003). Yet, the noncompliance on rules and well-established norms among the Chinese leadership in recent years signals a different story. Xi Jinping, part of the third generation of Chinese Communist leaders after Deng, is not expected to be bound by the precedent of a two-term limit. Similarly, Russia, in its fall back to authoritarianism, witnessed Putin defy the norm of limited terms to hold on to the top power position. Vietnam, the case of central interest in this paper, seems to follow the same suit. Nguyen Phu Trong was expected to step down in 2021 at the age of 77 after two terms as the Party Secretary General and did not. What do the institutionalists miss? Why do we see authoritarian rulers violate their own rule and at the same time maintain the regime’s stability? Was there other driving force other than institutions that prolongs the longevity of the regime? In other words, was institutionalization a real and meaningful process that contributes to the durability of in communist regime? I seek to answer these questions through the case of communist Vietnam. More specifically, I ask to what extent does Vietnam embrace institutionalization to resolve the authoritarian power- sharing conflict? I further ask what is distinctive in Vietnam’s context that contributes to the longevity of the regime in power? To do so, I evaluate two features of institutionalization that are often viewed to be important for power-sharing between the dictator and his ruling coalition in authoritarian regimes, in turn contributing to regime resilience. One is formal institutional rules such as term limits and retirement ages, which holds that leaders relinquish power upon reaching the age of retirement. Such rule-based power succession mechanism allows elites to detect noncompliance with agreements by dictators (Svolik 2012). Second is institutionalized power access in party autocracies, which implies that rulers delegate the power-sharing deals to institutions rather than making it a private transfer (Magaloni 2008), and the co-optation of economic actors into primary circles of power (Tsai 2013). In the context of Vietnam, where the 45 main driving force of the national economy is in the South, I argue that the second tenets, institutionalized power access and co-optation, could be translated into balancing regional politics. That is, if power-sharing was institutionalized meaningfully in Vietnam, we should expect to see meaningful share of seats for the Southerners in the important political bodies. Using social network analysis, I show that over the last 25 years, institutionalization has been a rather superficial process in Vietnam. The retirement age limits rule was frequently violated. More importantly, I find the persistence of hometown-based factional politics, with the northern faction consistently holds on to power-controlled seats. This means that despite all the rules that the Communist Party has set up, power remains concentrated in the few hubs that have existed for a long time. The data used is a novel dataset of 159 individuals who held seats in two most powerful political entities, the Politburo and the government cabinet. The Politburo is the decision-making body of the Communist Party of Vietnam, normally consists of 15 to 18 individuals, but could size up to 23 in some special circumstances. The government cabinet, in this specific project, refers to the following positions: president, prime minister, vice prime ministers, and ministers or ministerial-level chairs. Those who hold key positions in the government cabinet such as president, prime minister, minister of security, minister of finance are often also members of the Politburo. My contribution is twofold. First, the paper provides empirical evidence of factional politics in Vietnam through an analysis of hometown-based networks in the Politburo and the government cabinet. By highlighting the superficiality of the institutionalization process, the findings in this paper invites scholars to revisit theories on the true driving force behind maintaining political order in single-party regimes like Vietnam. Second, the paper brings a theoretical contribution to scholarship on institutions by paying attention to the history and geographic contexts that can potentially shape the circulation of power. I echo Pye’s (1988) view of political power from a cultural perspective. That is, institutionalization is not a universal process that takes place in democracies and non-democracies alike. We rather need a mid-range analytical approach to specify the processes by which regimes institute alongside the historical timeline context of each country. How elites reproduce power around regimes is not only a matter of rational calculation 46 to maximize their likelihood of survival, but also a function of historical legacy that predates the current regime. Vietnam Case: Factional Politics and Historical Legacy Factional Politics In an authoritarian regime, one can expect that factional politics always exists underlying the formal democratic-like institutional arrangement. Nathan (1973) defines political factions base on clientelism personal ties, as opposed to authority relations, that “does not become corporatized after recruitment but remains structured along the lines of the original ties which formed the bases of recruitment (42). Domes (1984) considers factions as coherent groups that “are based on alternative platform and exclusive claims for political power and overall control”, which can be either structural groups that are based on common demographic characteristics (shared hometown, educational background, or organizational experiences), or functional groups that are based on interest of functional subsystem (central or regional civilian party, central military machine) (27-29). Similarly, Dittmer and Wu (1995) conceptualize two types of relationships in informal politics. Value-rational relationships are those that are valued as an end in itself, including shared kinship ties, common geographic origin, former classmates, teachers, or students. Purpose-rational relationships, on the other hand, are instrumental to achieve other ends, forming with those colleagues, subordinates, and superiors in occupational contacts (470). Observing recent development of factional politics in China, Bo (2017) argues that political elites are better characterized as “factional groups” rather than “factions” because of the factional overlapping - some politicians have multiple identities which can be sort into different factional groups. Studies that applied factionalism model reveal important insights in political relations and processes. Huang (2000) shows that factionalism is the primary explanatory factor to understand the outcomes of Chinese leadership relations and decision making between the late 1930s and 1987. Accordingly, a leader’s power is based on the strength of his factional networks; decisions are made in accordance with “the vision of those who have prevailed in the power struggles rather than through the due process in which everyone is entitle to present his preferences” (5). The extant literature using social network analysis to understand politicians’ relation and 47 factionalism often analyses common characteristics of political leaders such as hometown origins, family connections, alumni, and military units to draw conclusions about their factions (Bo 2007; Shi 1997; Shih, Adolph, and Liu 2012; Li 2001; Choi 2012). Keller (2016), for example, finds that coworker networks best capture the underlying process of alliance formation among the contemporary Chinese elites, while provincial origin and alumni ties are not reliable proxies for their informal ties. In contrast, hometown origin is a significant factor of emergence of some group of officials, which helps identify the Shaanxi Gang or Mishu cluster (Li 2014, 2015)1. Comparative studies on elite networks in Togo and Ghana reveals systematic differences between elite configurations depending on the level of democracy (Anja 2018). Factional Politics and Regionalism in Vietnam Some scholars have noted that Vietnamese elites are considered less divisive, or that factional clarity in Vietnamese politics is weaker than their Chinese counterparts (Ben, Anita, and Jonathan 1998; Trinh 2020). Those who do study Vietnamese factional politics stratify elite individuals into factions with prominent and distinctive characteristics, interests, and strategies. Vuving (2010) suggests that post-Cold War Vietnamese politics can be seen as a game of four key players: regime conservatives, modernizers, rent-seekers, and China proponents. The author later characterizes the four top positions after leadership transition in 2016 as follows: “the chief of the party is a conservative with moderate tendencies, the head of the government is a modernizer, the country’s collective leadership is dominated by moderates, but these leaders have to rely on a structure heavily influenced by rent-seekers for policy recommendations and implementation” (Vuving 2017). An indicator of factional politics is also found in the legislature. Previously, the fact that seats in the parliament are distributed to different geographic areas, was interpreted as an institutional means for the regime leader to co-opt alternative voices into the policy-making process (Malesky and Schuler 2010, 483). However, Schuler (2021) shows that legislative institutions in single-party regimes do not work as a means of providing information about the population or as a “check” on autocrat power. Rather, they signal dominance and provide legitimacy for the regime. That is, the Vietnamese National Assembly, with its minimal 1 “Xi Jinping’s Inner Circle (Part 1: The Shaanxi Gang).” China Leadership Monitor 43 and “Xi Jinping’s Inner Circle (Part 4: The Mishu Cluster I).” China Leadership Monitor 46. 48 political impact, is entirely subsumed under the Communist Party and is nothing more than a representation of the prevailing political faction. I explore a geographic dimension of factions in Vietnamese politics. As Gainsborough (2010) observes, political networks in Vietnam are “fundamentally about personal relationships – blood or marital ties, shared hometown” (178). The norm is that “everyone owes their position to someone, which puts people in a hierarchical relationship, which in turn comes with further debts and obligations”. This is the main driving force behind being appointed to public office, where “the strong connections in people’s minds between public office, making money, and other forms of personal advancement, lie(s) at the heart” (Gainsborough et.al, 2009).2 My paper clarifies one type of personal ties that connect networks of politicians – hometown network. I will show that geographic origin of birth network constitutes political factions, which itself reflects the long historical legacy of regionalism. Regionalism in Vietnam is a topic of great interest to many people, especially around the time of electing a Politburo board or a government cabinet every five years. However, the extent to which a region sees their number of seats expand or contract remains a matter of anecdotal observation. The party secretary general Nguyen Phu Trong once said in the 12th Party Congress: “A party secretary general must be a northerner [who is] equipped with [communist] theory”. The statement points to underlying strength of the CPV, openly asserting that the major power players in Vietnam have been and must be from the North. Despite the war’s end, the 17th parallel that used to divide the country into two parts for decades during the Vietnam War continues to mark a division of leadership between the North and South. There is a mixed finding in regionalism in Vietnam. Trinh (2020) argues that Vietnamese political factions are less based on geographic as other factors given that “military casualties during the revolutionary period discourage(d) background sorting by erasing the salience of North-South differences”. Some recent studies find evidence of more balancing regionalism. That is, the allocation of the top 3 positions -Secretary General, Prime Minister, and President- have generally been divided equally among the three regions of the North, Center, and South 2 https://www.undp.org/content/dam/vietnam/docs/Publications/25524_6_Anti-Corruption.pdf 49 (Giang 2020). On the other hand, some scholars find a persistence of regional differences and regionalism in Vietnam in its past development (Dell, Lane, and Querubin 2018; Beresford and McFarlane 1995). Historian Taylor sees a deep division of factional groups resulted from centuries of warfare (Taylor 1998). Regional discrepancy is clearly revealed in the level of economic development with the South being the main engine. Although high foreign direct investment, most likely in the South, appears to have significant impact on a province’s autonomy, by which provincial leaders could spearhead their local economic policies (Malesky 2008), such local power does not necessarily translate into seats in the national political bodies. With an ever-growing economy reliant on the South, should we expect to see a regional balance of power in the country? If not, what is the structural makeup of hometown preferences in each case? Sharing a birthplace is an important network tie to understand politician relations, among other measurements of a faction’s ties - including broad ties, complete work ties, early work ties, and restrictive work ties (Meyer, Shih, and Lee 2016; Domes 1984; Dittmer and Wu 1995). One concern other scholars have raised about using hometown ties is the use of such an expansive definition of political ties that may generate a higher number of false positives (Keller 2016). Huang argues that in Chinese politics, the professionalization of the cadre body has rendered factions based more on work experience than birthplace ties (Huang 2008). However, Meyer, Shih and Lee (2016) posit that when a party’s general secretary pursues a broad strategy of factional recruitment, the broadest definition of a faction can perform just as well as the most restrictive definition in predicting promotions (57). Indeed, Chinese leaders in the reform era “continued to recruit faction members on a broad basis, including those with whom they shared birth province and school ties” (51). In the case of Vietnam, as detailed below, hometown network is particularly meaningful because it carries a historical legacy that goes back to centuries ago. Data 50 I collected a dataset of 322 seats in the two most powerful political bodies in the country, the Politburo and the government cabinet, between 1996 and 2022. These 322 seats were occupied by 159 individuals.3 Out of that total, 104 seats in the Politburo were held by 64 individuals, other 218 seats in government were held by 123 individuals. I collected information on individuals including their membership in each branch, working term, age, and hometown. There were only 10 females who covered 16 seats (5%) in the two institutions during the 25-year frame.4 Year of the term Politburo seats Government seats Total (Politburo/ government) 1996 – 2001 / 1997 - 2002 235 36 59 2001 – 2006 / 2002 - 2007 15 39 54 2006 – 2011 / 2007 – 2011 15 29 44 2011 – 2016 16 52 68 2016 – 2021 17 34 51 2021 – 2026 18 28 46 Total 104 218 322 Table 4: Count of seats in the Politburo and the government. Note: Before 2011, working terms in the government was lagged 1 year behind working terms in the politburo. For example, the politburo terms 1996 – 2001, 2001 – 20016, 2006 – 2011 were paralleled to the government terms 1997 – 2002, 2002 – 2007, 2007 - 2011. After 2011, the terms started and ended in the same year. The data was coded entirely by hand using official national and provincial websites. The main sources are: dangcongsan.vn and chinhphu.vn. In several cases, information about a seat was not found on a national website in which case I relied on news media sources or other non-official local sites.6 Exceptional cases to coding hometown are noted in the appendix. 3 http://caicachhanhchinh.gov.vn/tin-noi-bat/co-cau-thanh-vien-chinh-phu-gom-27-thanh-vien-co-4-pho-thu-tuong-18-bo-truong-4-thu-truong-co- quan-ngang-bo- 8907.html#:~:text=18%20th%C3%A0nh%20vi%C3%AAn%20Ch%C3%ADnh%20ph%E1%BB%A7,Th%E1%BB%83%20thao%20v%C3%A0 %20Du%20l%E1%BB%8Bch%3B 4 https://chinhphu.vn/thanh-vien-chinh-phu-qua-cac-thoi-ky-68352 5 Four seats were added in Dec 1997: Pham Thanh Ngan, Nguyen Minh Triet, Phan Dien, Nguyen Phu Trong. https://tulieuvankien.dangcongsan.vn/ban-chap-hanh-trung-uong-dang/bo-chinh-tri/khoa-viii/danh-sach-bo-chinh-tri-khoa-viii-1996-2001-393 6 Nguyen Danh Thai: https://vnexpress.net/ong-nguyen-van-an-trung-cu-chu-tich-quoc-hoi-1968644.html Do Que Luong: https://tinbds.com/hai-phong 51 Table 5: Description of variables Variable Description of variable seatID unique ID of each seat. A seat in the politburo starts with “pol”. A seat in the government cabinet starts with “gov”. year_branch_term_serve 5-year term, prefixed by “pol” for politburo and “gov” for government position. For example: pol96_01 is the politburo term of from 1996-2001 term_serve 5-year term of the position, without prefix person_name name of person in the job position gender Female/Male year_birth year of birth of each person hometown name of hometown province of each person branch Politburo or Government. Terms with “gov” denotes government, terms with “pol” denotes politburo, followed by last 2 digit of the year. For example, gov02_07 refers to government cabinet 2002 – 2007; pol06_01 refers to politburo 1996 – 2001. year_starting_term starting year of each term year_working_left number of working years a person has left by the beginning year of the term. The retirement age is 62 for men and 60 for women.7 Retirement Age Limit Rule Absent of free and fair elections, power transition is a major vulnerable spot in authoritarian regime. Institutionalists observe one measure that is often deployed as a “check” on the power of authoritarian leaders is using the age limit to control access to certain positions. I examine this rule in practice in the Politburo and the government cabinet between 1996 – 2022 to assess the degree to which it holds true. Although it is claimed that the age limit is one factor influencing the possible end to a candidate’s career, is the retirement age a real institution with teeth? In this 7 https://vietnamnet.vn/vn/thoi-su/chinh-tri/tuoi-bo-nhiem-gioi-thieu-nhan-su-ung-cu-lan-dau-khi-tang-tuoi-huu-752163.html 52 particular case, to what extent was it respected or violated? Are there varying levels of compliance between the Politburo and the government? Rules and Norm of Retirement Age Limit Vietnam has a hard rule on a certain age limit for public officials. Since the early 2000s, the retirement age was 60 for men and 55 for women. Two documents issued by the Communist party, Directive 35-CT/TW and Conclusion 60-KL/TW, specify the age limit for officials but do not explicitly indicate whether such rules apply to the Politburo.8 Since the ninth party congress, the general norm has been to limit top leaders to not running past 65 years of age. Discussions around the age limit have the potential to become hot topics of debate around party congress or national assembly elections. Age limits can be used as a tool for political competition (Abuza 1998; Giang 2020). One example is the revelation of Tran Dai Quang in 2020, who was racing to the position for Minister of Public Security in 2010, with two profiles that specifying two different birth years spread 6 years apart (1950 vs. 1956). The older one would disqualify him from running for the position. Quang managed to suppress the profile with older age and won the Minister of Public Security seat under the pretense of a younger age. He then used the same age specification to become President in 2016.9 Another example is the contested battle between Party General Secretary Nguyen Phu Trong and then Prime Minister Nguyen Tan Dung in 2016, both exceeded the age limit to continue running for another term. Yet the discussion on age limit was soon silenced. The Party General Secretary called for a vote in the Party Congress which he extended to a larger number of delegates to gain favor on his side.10 In other case, the age limit for members of the National Assembly was set in 2021 to is 63 for men and 53 for women.11 This rule was exploited to weed out unfavorable candidates for each party at the local level.12 However, the rule did not prevent 43 people in the National Assembly winning seats whose ages exceeded the limit.13 All of these examples point to the fact that the 8 https://thuvienphapluat.vn/tintuc/vn/thoi-su-phap-luat/chinh-sach-moi/25779/ket-luan-cua-bo-chinh-tri-ve-do-tuoi-tai-cu-nhiem-ky-toi 9 https://www.voatiengviet.com/a/tran-dai-quang-qua-doi-o-tuoi-62/4581427.html 10 Trong then called on a new card by expanding the size of the Party Congress. 11 https://www.quanlynhanuoc.vn/2021/03/15/do-tuoi-ung-cu-dai-bieu-quoc-hoi-va-hoi-dong-nhan-dan-cac-cap-nhiem-ky-2021-2026/ 12 https://usvietnam.uoregon.edu/viet-nam-quoc-hoi-khong-the-mai-mai-la-mat-tran-to-quoc-thu-hai/ 13 https://baucuquochoi.vn/dai-bieu/quoc-hoi-khoa-XV-37.vnp 53 age limit, a publicly observable signal of leaders’ commitment on power-sharing, is violated. In the section below, I examine the extent to which age limit norms apply to the Politburo and the government cabinet. Violation of Age Limit Rule The plot below shows the distribution of “time budget”, i.e., number of years that individuals have before reaching the age limit, by the beginning of each working term. The negative values indicate a violation of the retirement age. Figure 5: Number of working years left The distribution shows that most members of the Politburo and the government had less than 10 years left to work before reaching their retirement age. The politburo has more members exceeding the retirement age limit when they started their term than did those in government. In this sense, it is worth noting that the government accepts more often younger members than does the Politburo. All members who exceeded 7 or more years of age were in the Politburo, while all members who still had 17 or more working years were in the government. The plot shows that the average age of Politburo members is consistently higher than those of government. Year 2011 marks the rise in power of Nguyen Phu Trong as the party Secretary 54 General. While four members of the Politburo (Nguyen Phu Trong, Nguyen Sinh Hung, To Huy Rua, Ngo Van Du) exceeded the age limit of 62 years of age, none in the government did. In contrast, the 2011 – 2016 government term had 22 positions filled by 20 members whose working time before the age limit was 10 years or more.14 Noticeably, the average age in the most recent 2021 – 2026 term was the highest in 20 years, reflecting a gridlock in the 2021 Party Congress over transitioning power to a younger generation. Figure 6: Average age In short, contrary to the view that institutionalization processes have taken place in Vietnamese politics, I found that the age limit – one of the few rules established to smooth out power transitions – were violated frequently at the high level. The CPV’s Thirteenth Party Congress defied the institutionalist expectations with respect to retirement ages limit. Also, there has been no meaningful power transfer to younger generations in recent years. At some levels, the government branches have had a chance to introduce younger individuals into their networks, but this has not constituted a trend in recent terms. The 2011 – 2016 term marked the lowest average age in the government but also included a rise of average age in the Politburo. The 2021 - 2026 term currently includes individuals whose average age has been the highest in 25 years. Not surprisingly, the Politburo continues to out-age the government cabinet. Factions: Analysis of Hometown Networks 14 Pham Binh Minh held two positions Vice minister and Minister of Foreign Affairs. Vu Duc Dam was Vice Minister and Chairman of Governmental Office. 55 Transforming the Networks I use social network analysis to examine the relationships between individuals across working term, branches, and hometowns. This analytical approach is beneficial for understanding the power relations among elites, and among the political bodies that govern the country. First, it allows us to display the importance of nodes, i.e., hometown power hubs in this study, based on their centrality score. This enables us to understand the relative position of the core and the peripheral nodes. Second, the network demonstration allows us to trace the flow of power holders along with their positions overtime (Robins 2015; Carrington and Scott 2011). An important part of this method is visualizing the connections between actors and events. The visualization itself is an analytically meaningful way of deciphering the data as it reduces otherwise very complicated relationships and helps to highlight meaningful data. To focus on the core of power, I use a subset of 90 individuals who held two or more seats in either one or both branches, the Politburo or the government cabinet, leaving out 69 less important seats. Using the information of membership, working term, age, and hometown, I constructed networks displaying the connections between individuals and the working terms across time, between working terms to each other, and between hometowns and the working terms. Person-to-working Term Network First, I create an incidence matrix with the rows being names of individuals who are either members of the six Politburo boards or the six government cabinets. The columns are the working terms in each branch from 1996 – 2022 (see appendix). If a person is a member in a working term, their corresponding value is 1, otherwise 0. A person can have multiple values of 1 if they work for multiple terms. This matrix is then converted to a bipartite network where both persons and working terms are the nodes of the network graph. A linkage (edge) is only displayed from a person to a working term if said person worked during that term. The bipartite network of person-to-working term is plotted below. There are two kinds of nodes in this graph. Each person is represented by a circle 56 whereas each working term is a square. There are 90 individuals (90 circles) and 12 working terms (12 squares). The gray lines represent the linkages between people and working terms. Figure 6: Person-to-working term network. Note: The squares are working terms (Names of politburo terms start with “pol”. Names of government terms start with “gov”). The circles are individuals. The edges between person and working terms denote that said person works in that term. Person-to-person Network Next, I transform the data to a person-to-person matrix, where each node represents a person. The nodes are connected if two people working in the same term in the politburo or the government. This is done by taking the incidence matrix above and multiplying it by its transpose matrix (see appendix). Nodes of individuals are observably linked in the same working term, either with a Politburo board, a government cabinet, or both. The network then becomes a “hairball” of interwoven connections between individuals in which each node represents a 57 person, and each line represents the connection between the individuals through their affiliations with each working term. The person-to-person network above based on working term affiliation gives a clear impression of the highly dense and tightly bound core of power in the country. The circle nodes denote individuals. There are 90 circles in the graph denoting 90 people who held more than 1 seat in either or both branches. Put another way, after twenty-five years of market economic without political reform, the power to govern a country of ninety million people has largely been kept to less than a hundred individuals. Term-to-term Network The person-to-person network above paints an overall picture of power circulation. To better understand the extent to which power was preserved and traveled over time, I created another network showing working term-to-working term. First, to get a working term-to-working term matrix, I multiply the transpose of the working term matrix to a working term matrix. This resulted in the 12 by 12 term-to-term matrix below, where both rows and column are working terms (see full matrix in appendix). The value where the rows and columns intersect are the number of shared seats. For example, during the 2011 – 2016 government term there were 52 seats while the 2016 – 2021 government term had 34 seats. These two working terms had 32 seats occupied by the same people, hence the value of 32 at their intersection (row 4 and column 3).15 Figure 7: Term-to-term Network 15 Note that the matrix counts number of share seats, not unique persons. For example, government term 2011 – 2016 had 52 seats, held by 48 persons; government term 2016 – 2021 had 34 seats, held by 33 persons. The 32 seats share between these 2 terms were held by 26 persons. 58 The term-to-term matrix was then transferred into an adjacency network, where each edge denotes a shared seat that two working terms share with each other. Finally, this network is plotted in the Figure “Flow of power across time and branches” shown in the next section. Identification of Hometown Networks So far, I have constructed networks using 2 kinds of nodes: seats/individuals and working term. In such networks, people are connected through their working term affiliation. In this section, I incorporate hometown as another kind of node, making our network tripartite. That is, individuals are connected, or affiliated through their hometowns. Again, to not crowd the plot, I filtered out individuals who had only one seat during 6 working terms between 1996 – 2022. First, I created a person-to-hometown matrix, where the rows represent persons and columns represent hometowns. The matrix is rather simple, where 1 denotes a person born in a hometown. No person (as represented by rows) has more than one value of 1, but each hometown (as represented by columns) can have multiple values of 1. Next, I created a term-to-hometown matrix by multiplying the transpose of person-to-hometown matrix by the working term matrix. The terms-to-hometown matrix has rows as hometowns and columns as working terms. The values where rows and columns cross are the number of seats that each hometown has in each working term. Next, the matrix was converted into a network, where every hometown that has a seat in any term would form a linkage. The network is plotted as in the Figure “The core and the peripheral provinces”. Regionalism and Hometown Factions In this section, I examine the graph of the core and peripheral regions in the Figure “The core and peripheral provinces”. The graph captures hometowns connected to the Politburo and the government over the last 25 years. The pink colored nodes represent hometowns whereas those in purple are working terms. The size of the nodes corresponds to the number of total seats each hometown has, and each linkage represents one seat connection. 59 Figure 8: The core and peripheral provinces. The most important hometowns which originates the most seats are in the middle of the graph. The satellite nodes surrounding the graph are hometowns that had only a few seats during the time frame examined (1996 – 2022). At the center of the graph are the most important nodes that originated the most power holders in the regime: Hanoi, Nam Dinh, Thanh Hoa, Nghe An, Quang Nam, Ben Tre, Ha Tinh, Ca Mau, Long An. These power hubs are further examined in the section below. There’s great variation in the characteristics of the satellite nodes. Among those is Ho Chi Minh City, the largest locale in terms of population size and gross domestic product. Binh Duong, Da Nang, Bac Ninh are provinces with the highest economic growth rates in the country, home of the largest industrial complexes and also situated in the periphery. These locales are represented by the same number of seats as the poorest and least populous provinces, namely Lao Cai, Tuyen Quang, Son La. The historian K.W. Taylor has identified six major military conflicts over a span of five centuries in Vietnam, from the early fifteenth century to the late nineteenth century, of which two are between Dong Kinh (Red River area centered upon Hanoi, including Nam Dinh) and Thanh 60 Nghe (Thanh Hoa, Nghe An, Ha Tinh). These two regions were at war with each other for almost two centuries. Dong Kinh was the home of kings that “bears the countenance of a coherent and dominant place from the eleventh to the fourteenth centuries”, while Thanh Nghe seems to have “claimed moral precedence in pan-Vietnamese discourse for centuries” (Taylor 1998). I found compelling evidence to support the continuation of such historical legacies in the current faction network. The graph below shows that hometowns account for the origination of at least 5 seats in the Politburo and the government between 1996 and 2022. These 23 provinces accounted for 274 seats. This means that 36% of the total provinces held 85% of the total seats. Table 6: Number of seats held by the 5 provinces across 6 terms Provinces 1996 – 2001 - 2006 - 2011 – 2016 - 2021 - 2001 2006 2011 2016 2021 2026 Ha Noi 6 6 9 9 5 3 Nghe An 8 5 4 3 2 5 Ha Tinh 5 4 1 5 4 3 Thanh Hoa 5 4 2 4 2 3 Nam Dinh 3 3 2 7 5 7 Seats held 27 22 18 28 18 21 by 5 provinces Total seats 59 54 44 68 51 46 Seats share 46% 41% 41% 41% 35% 46% When breaking down the composition of seats by the Politburo and the government, it is clear that the top 6 provinces are the biggest driver of seats in government se provinces - Ha Noi, Nghe An, Nam Dinh, Ha Tinh, Thanh Hoa, and Quang Nam - maintained their dominance in both branches. The next two provinces, Ben Tre and Thai Binh, had only 2 seats each in the Politburo. The following three, Ha Nam, Hai Phong, Da Nang had only 1 seat each, while Quang Ngai never had a seat in the Politburo. Among the top 6 provinces, five are above the 17th parallel: Hanoi, Nam Dinh, Nghe An, Ha Tinh, Thanh Hoa and the last, Quang Nam, is an adjacent province to the 17th parallel. These provincial linkages collectively underly 134 out of 61 322 seats, or 42% of the positions in the Politburo and the government between 1996 and 2022. Seat occupations by Northerners are examined more closely in the plot below. Figure 9: Change in number of seats In the 1996 – 2001 term, Nghe An had the biggest share of seats, followed by Hanoi, Thanh Hoa, Ha Tinh and Nam Dinh. The five provinces held 27/59 (46%) of the total seats. In the next term 2001 - 2006, only Hanoi maintained its 6 seats while the other four lost their seats to other provinces, making up 22/ 54 seats. Between 2006 and 2016, Hanoi occupied the largest share of seats before losing almost half and then 2/3 of its seats in 2016 and 2021. The rise of Nam Dinh and Ha Tinh in the 2006 – 2011 term is a remarkable feat on its own. Following this term, Nam Dinh continued to rise, gaining more seats, while Ha Tinh reversed course to having the same number of seats as Thanh Hoa and Ha Noi in 2021. The table below details the numbers of seats each province held each term and their collective share in total seats for the whole country. They maintained their share of more than 40% in four out of five terms. Term 2016 – 2021 demonstrates a decrease with only 35% seats shared. Degree of Centralization Over Time I then calculated the magnitude of centralization in each working term. Degree of centralization measures “how variable or heterogenous the actor centralities are (where) the larger it is, the 62 more likely it is that a single actor is quite central, with the remaining actors considerably less central” (Wasserman 1994, 176). Figure 10: Degree of centralization The term 2006 – 2011 is strikingly centralized compared to the others which may explain the consolidation of power for the Hanoi group. This group gave rise to the party secretary general, Nguyen Phu Trong, in 2011 who has remained in the top position for three party congresses. The 2011-2016 term, on the other hand, reflects a severe fight between the party secretary general, Nguyen Phu Trong, and the then prime minister, Nguyen Tan Dung, before the party congress in 2016. This resulted in the lowest degree of centralization where the seats are somewhat more dispersedly allocated. 63 Figure 11: Seats at the hometown tables Note: Each plot represents one working term, combining the Politburo and the government cabinet. The number on top of each plot denotes the working term year. For example, “96_01” is term 1996-2001. In each plot, the green circles represent hometown, the orange squares denote the seats that originated from that hometown. The size of the green circles is proportional to their number of seats. For example, the plot “06_11” denotes working term 2006-2011. In this term, Hanoi is the biggest table, occupying 9 government seats and 3 politburo seats. The figure above shows the seats organized around each hometown table in each working term. Each plot represents one working term, combining the Politburo and the government cabinet. The number on top of each plot denotes the working term year. For example, “96_01” is term 1996-2001. In each plot, the green circles represent hometown, the orange squares denote the seats that originated from that hometown. The size of the green circles is proportional to their number of seats. The plots assist us to see which table stands out as the biggest player at the time, whether it dominates the whole political landscapte or the tables balance against each other. For example, the plot “06_11” denotes working term 2006-2011, in which Hanoi is the biggest 64 table, occupying 9 government seats and 3 politburo seats. In contrast, the working term 2016- 2021, illustrated in plot “16_21”, has a relatively more balancing power compared to others. Flow of Power Across Time and Branches In the network below, a non-directional link is drawn between a seat in the Politburo and the government to a hometown. Viewing these networks across political terms reveals several interesting insights about changes of power in Vietnamese politics in the last two and a half decades. This supports the view that regardless of the economic development in the South, regionalism persists in Vietnamese politics. The network below captures the flow of power from one working term to another where the nodes are working terms and their size represents the number of seats in each term. The labels of each node indicate either a Politburo or a government term and the edges (gray lines) represent the connections between one term to another by how many seat holders they share. The 2011 – 2016 government term stands out as the biggest node. Out of its total 52 seats, 6 seat were held from term 2002 – 2007 and 13 seats from term 2007 – 2011. Furthermore, this term carried 32 seats to the following term – 2016-2021 – and 10 seats to the term after that from 2021 – 2026. The oldest party term in the dataset, term 1996 – 2001 is connected the current Politburo term 2001 – 2016 through current party secretary general Nguyen Phu Trong. Meanwhile, the oldest government term in the dataset 1997 – 2002 has not carried anyone over after 2016. Only 1 person, Nguyen Tan Dung he lasted only up until the 2011 – 2016 term. 65 Figure 11: Flow of power across time and branches Note: the circles are working term. The size of circle represents number of seats in each working term. Both branches are shown. (Names of politburo terms start with “pol”. Names of government terms start with “gov”). Two working terms are connected by an edge if they have common individuals. In summary, this analysis reveals the significance of hometown networks for understanding power hubs among Vietnamese political elites. I identify 90 individuals who held the most important positions in the Politburo and the government over the last 25 years, governing a population of 90 million people. I show that factional politics prevailed through regionalism, defying the expectations of earlier theories purporting the importance of regional balancing. Regardless of the fact that the South is the strongest economic force, the major power holders remain Northerners. The share of power for balance representation did not go deeper than the top three positions (party secretary general, prime minister, president). I found six provinces originating half of all power seats: Ha Noi, Nam Dinh, Thanh Hoa, Nghe An, Ha Tinh, and Quang Nam. The first five are above the 17th parallel whose seats were occupied by about 40 people most of the time. VII. Conclusion 66 This paper examines two tenets of institutionalization that institutionalists often rely on to explain the durability of authoritarian regimes, the institutionalized power access and rule-based succession. I found that these processes are rather superficial in the case of Vietnam. First, by tracing individuals through their working terms, I show that, in contrast to the expectation of the institutionalists, top ranking Vietnamese elites were not bound by the rule of retirement ages. Second, contrary to the expectation of institutionalized power access, i.e., balancing power between important regions in the country, I found that regional balance does not penetrate deeper than the top three positions. Forty-five years after northern communists took over the South in 1975, the North still has the upper hand and continues to hold dominating positions. Despite Southern provinces being the major engine of the economy, the southern factional group has not been able to claim a any significant share of political power. Perhaps, a less overestimating the progress in institutionalization view and more understanding the durability of factional politics might help us understand Vietnamese politics from a different angle from what institutionalists offer. I also show how the historical context of the country underlies current factional politics. That is, persistent regionalism in Vietnam is an extension of the historical warfare that took place for centuries. I identify six provinces that originated half of the power seats for 25 years: Ha Noi, Nam Dinh, Thanh Hoa, Nghe An, Ha Tinh, Quang Nam. The first five are above the 17th parallel which used to divide the country during the Vietnam War and occupied 40 percent or more seats most of the time. They also represent the current version of Dong Kinh – Thanh Nghe warfare that historically shaping the power clusters of the country. The sixth powerful hub, Quang Nam, is just adjacent to the 17th parallel, meaning that it is not hugely representative of the southern region. This finding suggests that rather than universally applying the institutionalist argument to all cases, we need a meso-level theorical examination to better understand authoritarian regimes. The empirical finding reminds us once again of the market reform dilemma in Vietnam since the collapse of communism – the fear that an economic opening would challenge the role of the Communist Party. The model of economic development without political reform has allowed the regime to bear the fruits of the free market without sharing political power with the opposition. The dilemma today seems less than tangible as at some point, the “Vietnamese model” of 67 development was even advertised to North Korea to follow.16 Yet, there were numerous times the party leadership showed great hesitation to economic reforms. The dilemma, in fact, stays real at all times. The feedback loop goes something like this: had the party, originally consisting of the Northern faction, allowed the economy, mainly in the South, to develop to its full potential, the Southerners may have demanded more power, which in turn could have threatened the Northern faction and the stability of the whole regime. By this logic, the durability of the communist regime up to this point can possibly be attributed to the rigidity of the factions which remain in place. This paper echoes Gainsborough’s view that, in Vietnamese politics, the state “remains little changed in terms of its underlying political philosophy and many of its practices […] Nepotism rather than meritocracy is the norm in appointments to the government and civil service, which are themselves governed by informal, discretionary rules rather than the “’rule of law’” (2010), 157, 169). Those who care about Vietnamese politics, with respect to the prospects of democratization, may worry about the most recent formed Politburo and government cabinet. Both have the highest average age for members in the last 20 years. The degree of centralization in both branches is high. The power succession procedure has not been specified. The retirement age rule remains weak. The party general secretary is holding on to power and not having signaled a transition. These observations could signal the verge of a political crisis. Reference 2011. The SAGE handbook of social network analysis, Handbook of social network analysis. London ; Thousand Oaks, Calif.: London ; Thousand Oaks, Calif. : SAGE. Abrami, Regina, Edmund Malesky, and Yu Zheng, eds. 2013. Vietnam through Chinese Eyes: Divergent Accountability in Single-Party Regimes. Edited by Martin K Dimitrov, Why communism did not collapse : understanding authoritarian regime resilience in Asia and Europe. New York: New York : Cambridge University Press. Abuza, Zachary. 1998. "Leadership Transition in Vietnam since the Eighth Party Congress: The Unfinished Congress." Asian survey 38 (12):1105-1121. doi: 10.2307/2645824. 16 https://www.cnbc.com/2019/02/13/north-korea-may-choose-to-follow-vietnams-economic-model.html 68 Amstrong, Charles K. 2013. Ideological Introversion and Regime Survival: North Korea's "Our- Style Socialism". Edited by Martin K Dimitrov, Why communism did not collapse : understanding authoritarian regime resilience in Asia and Europe. New York: New York : Cambridge University Press. Anja, Osei. 2018. "Elite Theory and Political Transitions: Networks of Power in Ghana and Togo." Comparative politics 51 (1):21-40. doi: 10.5129/001041518824414610. Anthony, W. Pereira. 2007. "Political (In)Justice: Authoritarianism and the Rule of Law in Brazil, Chile and Argentina (Pittsburgh, PA: University of Pittsburgh Press, 2005), pp. xvi+262, $27.95, pb." Journal of Latin American studies 39 (4):883-885. doi: 10.1017/S0022216X07003276. Balme, Stéphanie, and Mark Sidel. 2007. Vietnam's new order : international perspectives on the State and reform in Vietnam. 1st ed. New York: New York : Palgrave Macmillan. Ben, Kerkvliet, Chan Anita, and Unger Jonathan. 1998. "Comparing the Chinese and Vietnamese Reforms: An Introduction." The China journal (Canberra, A.C.T.) 40 (40):1-7. doi: 10.2307/2667451. Beresford, Melanie, and Bruce McFarlane. 1995. "Regional inequality and regionalism in Vietnam and China." Journal of contemporary Asia 25 (1):50-72. doi: 10.1080/00472339580000031. Bernstein, Thomas P., ed. 2013. Resilience and Collapse in China and the Soviet Union. Edited by Martin K Dimitrov, Why communism did not collapse : understanding authoritarian regime resilience in Asia and Europe. New York: New York : Cambridge University Press. Blaydes, Lisa. 2011. Elections and distributive politics in Mubarak's Egypt. New York: New York : Cambridge University Press. Bo, Zhiyue. 2007. China's elite politics political transition and power balancing, Political transition and power balancing. Hackensack, N.J.: Hackensack, N.J. : World Scientific. Bo, Zhiyue, ed. 2017. Factional Politics in the Party-State Apparatus. Edited by W.W.-L Lam, Routledge Handbook of the Chinese Communist Party: Routledge. Boix, Carles, and Milan W. Svolik. 2013. "The Foundations of Limited Authoritarian Government: Institutions, Commitment, and Power-Sharing in Dictatorships." The Journal of politics 75 (2):300-316. doi: 10.1017/S0022381613000029. 69 Brownlee, Jason. 2007. "Hereditary Succession in Modern Autocracies." World politics 59 (4):595-628. doi: 10.1353/wp.2008.0002. Bueno De Mesquita, Bruce. 2003. The logic of political survival. Cambridge, Mass.: Cambridge, Mass. : MIT Press. Bunce, Valerie, and Sharon L. Wolchik, eds. 2013. Bringing Down Dictators: Waves of Democratic Change in Communist and Postcommunist Europe and Eurasia. Edited by Martin K. Dimitrov, Why communism did not collapse : understanding authoritarian regime resilience in Asia and Europe. New York: New York : Cambridge University Press. Chen, Cheng. 2016. The Return of Ideology - The Search for Regime Identities in Postcommunist Russia and China: Ann Arbor: University of Michigan Press. Choi, Eun Kyong. 2012. "Patronage and Performance: Factors in the Political Mobility of Provincial Leaders in Post-Deng China." The China quarterly (London) 212 (212):965- 981. doi: 10.1017/S030574101200118X. Dell, Melissa, Nathan Lane, and Pablo Querubin. 2018. "THE HISTORICAL STATE, LOCAL COLLECTIVE ACTION, AND ECONOMIC DEVELOPMENT IN VIETNAM." Econometrica 86 (6):2083-2121. doi: 10.3982/ECTA15122. Dimitrov, M. 2013. Why communism did not collapse : understanding authoritarian regime resilience in Asia and Europe. New York: New York : Cambridge University Press. Dittmer, Lowell, and Yu-Shan Wu. 1995. "The Modernization of Factionalism in Chinese Politics." World politics 47 (4):467-494. doi: 10.1017/S0043887100015185. Domes, Jurgen, ed. 1984. Intra-Elite Group Formation and Conflict in the PRC. Edited by David S. G Goodman, Groups and politics in the People's Republic of China. Cardiff, United Kingdom: Cardiff, United Kingdom : University College Cardiff Press. Gainsborough, Martin. 2010. Vietnam rethinking the state. London ; New York: London ; New York : Zed Books. Gallagher, Mary, and Jonathan K. Hanson, eds. 2013. Authoritarian Survival, Resilience, and the Selectorate Theory. Edited by Martin K Dimitrov, Why communism did not collapse : understanding authoritarian regime resilience in Asia and Europe. New York: New York : Cambridge University Press. 70 Gandhi, Jennifer. 2008. Political institutions under dictatorship. New York: New York : Cambridge University Press. Gandhi, Jennifer, and Adam Przeworski. 2006. "COOPERATION, COOPTATION, AND REBELLION UNDER DICTATORSHIPS." Economics and politics 18 (1):1-26. doi: 10.1111/j.1468-0343.2006.00160.x. Geddes, Barbara. 1999. "WHAT DO WE KNOW ABOUT DEMOCRATIZATION AFTER TWENTY YEARS?" Annual review of political science 2 (1):115-144. doi: 10.1146/annurev.polisci.2.1.115. Giang, Nguyen Khac. 2020. "SUCCESSION POLITICS AND AUTHORITARIAN RESILIENCE IN VIETNAM." Southeast Asian affairs SEAA20 (1):411-426. doi: 10.1355/aa20-1v. Handlin, Samuel. 2016. "Mass Organization and the Durability of Competitive Authoritarian Regimes: Evidence From Venezuela." Comparative Political Studies 49 (9):1238-1269. doi: 10.1177/0010414016628186. Huang, Jing. 2000. Factionalism in Chinese Communist politics. Cambridge, UK ; New York, NY: Cambridge, UK ; New York, NY : Cambridge University Press. Keller, Franziska Barbara. 2016. "MOVING BEYOND FACTIONS: USING SOCIAL NETWORK ANALYSIS TO UNCOVER PATRONAGE NETWORKS AMONG CHINESE ELITES." Journal of East Asian studies 16 (1):17-41. doi: 10.1017/jea.2015.3. Koesel, Karrie J., and Valerie J. Bunce. 2013. "Diffusion-Proofing: Russian and Chinese Responses to Waves of Popular Mobilizations against Authoritarian Rulers." Perspectives on politics 11 (3):753-768. doi: 10.1017/S1537592713002107. Li, Cheng. 2001. China's leaders : the new generation. Lanham [Md.]: Lanham Md. : Rowman & Littlefield Publishers. Magaloni, Beatriz. 2008. "Credible Power-Sharing and the Longevity of Authoritarian Rule." Comparative political studies 41 (4-5):715-741. doi: 10.1177/0010414007313124. Malesky, Edmund. 2009. "Gerrymandering—Vietnamese Style: Escaping the Partial Reform Equilibrium in a Nondemocratic Regime." The Journal of politics 71 (1):132-159. doi: 10.1017/S0022381608090099. 71 Malesky, Edmund J. 2008. "Straight Ahead on Red: How Foreign Direct Investment Empowers Subnational Leaders." The Journal of politics 70 (1):97-119. doi: 10.1017/S0022381607080085. Malesky, Edmund, and Paul Schuler. 2010. "Nodding or Needling: Analyzing Delegate Responsiveness in an Authoritarian Parliament." The American political science review 104 (3):482-502. doi: 10.1017/S0003055410000250. Meyer, David, Victor C. Shih, and Jonghyuk Lee. 2016. "FACTIONS OF DIFFERENT STRIPES: GAUGING THE RECRUITMENT LOGICS OF FACTIONS IN THE REFORM PERIOD." Journal of East Asian studies 16 (1):43-60. doi: 10.1017/jea.2015.5. Nathan, Aj. 2003. "Authoritarian resilience." Journal Of Democracy 14 (1):6-17. Nathan, Andrew J. 1973. "A Factionalism Model for CCP Politics." The China quarterly (London) 53 (53):34-66. doi: 10.1017/S0305741000500022. Pye, Lucian W. 1988. Asian power and politics : the cultural dimensions of authority. Cambridge, Mass.: Cambridge, Mass. : Belknap Press. Robins, Garry. 2015. Doing social network research : network-based research design for social scientists, Network-based research design for social scientists. Los Angeles: Los Angeles : Sage Publications Ltd. Schuler, Paul. 2021. United front : projecting solidarity through deliberation in Vietnam's single-party legislature. Stanford, California: Stanford, California : Stanford University Press. Shi, Tianjian. 1997. Political participation in Beijing. Cambridge, Mass.: Cambridge, Mass. : Harvard University Press. Shih, Victor, Christopher Adolph, and Mingxing Liu. 2012. "Getting Ahead in the Communist Party: Explaining the Advancement of Central Committee Members in China." The American political science review 106 (1):166-187. doi: 10.1017/S0003055411000566. Solomon, Peter H., Jr. 2007. "Courts and Judges in Authoritarian Regimes.". World Politics. Svolik, Milan W. 2012. The Politics of Authoritarian Rule. New York: New York: Cambridge University Press. Taylor, K. W. 1998. "Surface Orientations in Vietnam: Beyond Histories of Nation and Region." The Journal of Asian studies 57 (4):949-978. doi: 10.2307/2659300. 72 Tismaneanu, Vladimir, ed. 2013. Ideological Erosion and the Breakdown of Communist Regimes. Edited by Martin K Dimitrov, Why communism did not collapse : understanding authoritarian regime resilience in Asia and Europe. New York: New York : Cambridge University Press. Trinh, Duy. 2020. "Explaining Factional Sorting in China and Vietnam." Problems of post- communism:1-19. doi: 10.1080/10758216.2020.1739540. Tsai, Kelle S., ed. 2013. Cause or Consequence? Private-Sector Development and Communist Resilience in China. Edited by Martin K Dimitrov, Why communism did not collapse : understanding authoritarian regime resilience in Asia and Europe. New York: New York : Cambridge University Press. Vuving, Alexander. 2010. "VIETNAM: A Tale of Four Players." Southeast Asian affairs 2010 (1):367-391. doi: 10.1355/SEAA10U. Vuving, Alexander L. 2017. "THE 2016 LEADERSHIP CHANGE IN VIETNAM AND ITS LONG-TERM IMPLICATIONS." Southeast Asian affairs 2017:421-435. Wasserman, Stanley. 1994. Social network analysis : methods and applications. Cambridge ; New York: Cambridge ; New York : Cambridge University Press. Zhao, Suisheng. 1998. "A state-led nationalism: The patriotic education campaign in post- Tiananmen China." Communist and Post-Communist Studies 31 (3):287-302. doi: 10.1016/S0967-067X(98)00009-9. 73 CHAPTER 3: RISING MEDIA ENCLAVE EXTREMISM THE CASE OF THE EPOCH TIMES AND ETHNIC MOBILIZATION ON THE FAR RIGHT I. The Rise of Media Enclave Extremism Does the abundance of voices lead to the diversification of alternative viewpoints and an expansion of the democratic public sphere? Not necessarily. This paper demonstrates a negative effect when a media channel enclaves their audience into an extreme belief. Using 19,000 articles published between 2013 and 2022 in the “Politics” section of a New York-based media outlet affiliated with the Falun Gong religious group, The Epoch Times (hereafter ET), I examine the production of media enclave extremism on the far-right. Enclave extremism is defined as “a special case of the broader phenomenon of group polarization, which extends well beyond politics and occurs as groups adopt a more extreme version of whatever view is antecedently favored by their members” (Sunstein, 2007).1 I expand on this concept by providing a deeper understanding of enclave extremism that captures a wider phenomenon, showing how enclave extremism not only occurs when group members externally conform to a belief that was antecedently favored by other members, but also when the idea and belief are externally imposed on groups by the media producers. That is, beyond the original conceptualization of enclave extremism that emphasizes the consumption side, the paper considers the supply side – how the enclave was made. Scholars of media enclave extremism point out that repressive radical media outlets exist as a sort of “imitated counterpublic” with ideological closure. While it is normal for people to self- select membership in groups with like-minded individuals that share similar beliefs, it is alarming that more and more audiences get caught in media niches that limit their abilities for civil dialogue. Although enclave extremism could theoretically take place at either end of the political spectrum, new studies suggest that this phenomenon has been more common to the far- right in recent years (Faris, Roberts, and Benkler 2018, 14, 384). The rise of media enclave extremism is arguably driven mainly by a majority population who fear losing out to emerging social segments (Sik 2015; Back 2002). But a recently burgeoning phenomenon worth investigating is how a newly activated population of ethnic minorities joins a far-right circle. To 1 https://www.chronicle.com/article/the-polarization-of-extremes/ 74 what extent is the rise of media enclave extremism fueled by these new occurrences? I contribute to the literature by providing analysis of the supply side of media enclave extremism through a case of The Epoch Times. The central question in the paper is how the media outlet successfully mobilized a historically inactive ethnic minority population into enclave extremism, and how they engaged them politically using far-right discourse. I will argue throughout the paper that the ET’s discourse is an example of enclave extremism whereby the media company makes up an ideological bubble and bakes readers into an extreme belief. As further analyzed below, ET utilizes two major discursive strategies. On the one hand, the outlet strategically bundles their traditional enemy with the other targets of the far-right to fuel hatred and discrimination. On the other hand, ET tailors their own category of identity and bundles it with the social conservative issues that are supported by far-right circles. Both strategies grow their audience and allow them to promote a one-sided belief towards an “alternative truth.” Overall, ET functions less as a representative voice of the minority Asian community but more as a radical ethnic media news outlet selling its political agenda by opportunistically jumping into the far-right niche. An ethno-religious media platform based in New York since 2000, ET is mainly circulated among a small group of Falun Gong sympathizers who criticize the Chinese government for its brutal suppression of Falun Gong practitioners. ET could have been a channel for the “unheard” or the “silent minority,” contributing to an expansion of the democratic public sphere. But it has instead headed in an entirely different direction. As this paper points out, ET has formed a far- right political niche that motivates Asian minorities to stand up against an imagined socialism, communism, or “collective minds”. These made-up enemies, which parallel the trend of far-right resurgence in the US, worked to get the outlet a sharp increase in readership. The rise of ET was, thus, not accompanied by an expansion of the democratic sphere, where a media platform gives rise to the voices of a historically inactive population, represents their community interests, or expands their political views. 75 II. Radical Media and Democratic Values Radical Media “The liberal order is rigged” (Colgan and Keohane 2017). In the US, “liberal democracy itself appears fragile and polarized, vulnerable to far right populism and backlash politics” (Ikenberry 2018). In today’s America, there exist both “the unheard” Americans of Martin Luther King and a “silent majority” of Americans in the Nixonian sense that await to be heard. Thus is born “radical media” that “poses challenges to existing power structures, empowers diverse communities and classes, and enabling communities of interest to speak to each other” (Downing 1984, 2). But not all radical media was created equal. Media that raises the voice of this population could go in two opposing directions, by either expanding the democratic sphere or damaging democracy in a serious way. Downing (2001) distinguishes two kinds of radical media, namely democratic and repressive radical media. The distinction relies on the notion of “developmental power” as a basis to evaluate media, where “developmental power” is defined by Macpherson as the opportunity for members of the public “to use and develop [their] capacities” (in Downing 2001, 43). Democratic radical media - a democratic alternative to media monopolies - exists towards enhancing democratic processes. To serve this function, Downing argues, radical media must “sit lightly to government or corporate or religious oversight”, and “responsive to and critical of constructive social forces”(Downing 2001, 94). Democratic radical alternative media can play a role of “developmental power” agents in four ways: expanding the range of information, being responsive to the voices of the excluded, being free from state or religious power’s censorship, democratically arranging their own organization (Downing 2001, 44). The second kind, which is the focus of this paper, is repressive radical media and exists in service of repression. It “not only fail(s) to enhance but actually maims the public’s ability to develop its powers. Neither critical reflection nor any genuine increase in personal or collective freedom are on the radar screen of such media”(Downing 2001, 89). Throughout modern history, we have seen various kinds of far-right movements: religious, racist, elitist, reactionary, modernizing, populist, etc. The resurgence of the far-right in Europe and the United States 76 together with an unprecedented large number of followers around the world in recent years has been accompanied and empowered by far-right media. The explosion of the internet has allowed far-right followers to massively develop their media channels. They appear in various forms: political party forums (Atton 2006), cloaked websites to disguise ownership (Daniels 2009), pseudo-social science and academics’ personal websites (Mihailovic 2015b), civil organizations (Padovani 2016; Tischauser and Musgrave 2020; Cammaerts 2009), social media platforms (Sik 2015), music band and personal sites (Simpson 2015). These various media platforms vary in operation, but they all share the common themes of white supremacy, anti-immigration, anti- refugee, “neo-Aryan”, and anti-Semitic, among others.2 How Media Enclave Extremism Rise For many decades, conservatives have accused mainstream media of being “biased” in favor of liberals (Dagnes 2019; Jamieson 2008). As of 2018, there were 19 major players in the right- wing media circle, whose purpose is “to support conservative ideals but even more to oppose those outside of their bubble”, as identified by Will Sommer and Dagnes (Dagnes 2019). Drudge sits on the top on internet, Fox News on television, and Rush Limbaugh on talk radio. What all of these outlets have in common is to feed “argumentative and unrepentant” discourse, leading their audience to be “adamant of their virtues, faithful adherents to their messaging” regardless of accuracy, keeping these consumers within their bubble, and “dividing the public between the faithful and the non-believers” (Dagnes 2019, 169). Moreover, news and opinion websites developed by the far-right outsized any other ideological strands. A group of extreme sites including Breitbart, Infowars, Truthfeed, Zero Hedge, and the Gateway Pundit radicalized the right-wing media ecosystem (Faris, Roberts, and Benkler 2018). Noticeably, where Right-wing media most sharply diverges from the Left is in “their easy acceptance of manipulative tactics, of demagogic slogans, as opposed to challenge and debate. They particularly like to present themselves as crying out for the right to free speech, a free speech they would cheerfully refuse to any opposition, were they in power” (Downing 2001, 93). 2 I use repressive radical media, far-right media, ultra-right media interchangeably in this project. 77 Several characteristics of the repressive radical media that contribute to enclave extremism are depicted in the literature. First, the nature of the public sphere upon which the far-right media exercises is actually a type of “imitated counterpublic” due to its “ideological closure of discursive space, the disavowal of dialogue and deliberative engagement with broader publics”(Tischauser and Musgrave 2020). These are “discursive and material spaces that look inward and build oppositional identities, while also strategically entering into deliberations in the dominant domain” (Tischauser and Musgrave 2020). A study of the North Belgian post-fascist movement and activists shows the rise of anti-public space, voicing hatred and essentialist discourse (Cammaerts 2009). This discursive space “does not allow debates, argumentation, criticism, and justification of validity claims, only alignment with the dominant position” (Sik 2015, 149). In other words, one of the central tenets of the far-right media is to “construct or sustain a community with closure” (Couldry in Atton 2006). Second, there is a dogmatic discursive logic of friend and enemy - us versus them - where “critical race rhetoric is coopted to mobilize white supremacists’ sentiments and organize white tribal politics” (Tischauser and Musgrave 2020). The basis for the far-right worldview is “the feeling of being threatened: as an individual and as a nation, from the inside and the outside, below and above” (Sik 2015). The whites present their cultures and identity as being diluted by other subgroups of population (Atton 2006). The ultra-conservatives portray emergent ideas of feminism or LGBTQ rights as destructive forces bent on destroying families and tradition. Common themes of illegal immigration and Black crime is repeatedly pronounced. Above all, the far-right media spreads a pervasive fear that “whiteness” is going to be superseded on a global scale and that White people need to do something to prevent that from happening (Back 2002). Third, there is always attempts in far-right discourse to disguise the public, to cover the surface with naïve narratives and to load repressive content underneath. A study of the British National Party’s forum shows that their media is “anti-multicultural, anti-equality and anti-freedom, yet its discourse uses the tropes of multiculturalism, equality and freedom to maintain an ideological space where racism and repression may appear natural and commonsensical” (Atton 2006). An investigation of “cloaked websites” - those published by ambiguously declared individuals or 78 groups - shows that they “conceal political agendas intended to subvert civil rights and affirm white supremacy through an audacious deployment of civil rights movement rhetoric (Daniels 2009). The techniques of the far-right activist meme wars involved “releasing embarrassing documents, hateful drowning-out of opponents, and other substantial personal offenses” (Faris, Roberts, and Benkler 2018, 12). Fourth, speaking of news media in particular, far-right news media inherited all biases of partisan journalism where news stories that are “more likely to be framed, spun, and slanted so that certain political agendas are advanced”, as opposed to detached journalism that is “rooted in the values of objectivity, fairness, and balance” (Project 2007). Numerous studies have found links between the consumption of partisan media and the lack of confidence in democratic institutions and political processes (Moy, Pfau, and Kahlor 1999; Hollander 1995, 1996; Pfau et al. 1998). Partisan media can cause polarization among boths its regular consumers and inadvertent audiences (De Benedictis-Kessner et al. 2019; Baum and Groeling 2008; Baum 2010). In particular, news analyses of the 2016 election showed how right-wing media outlets were “the primary culprit in sowing confusion and distrust in the broader American media ecosystem” (Faris, Roberts, and Benkler 2018, 13). In short, although the media terrain is broader in all dimensions, its expansion has not been followed with the expansion of the democratic sphere. In fact, as Jamieson et al. show in their analysis of Limbaugh’s political talk radio program, the cable network Fox, and the Wall Street Journal, “conservative media create(s) a self-protective enclave hospitable to conservative beliefs”(Jamieson 2008, 134). A large-scale observation study by Faris, Roberts and Benkler shows that the right-wing media ecosystems “conforms to the very worst fears of those critical of the effects of the internet on democracy. It exhibits all the characteristics of an echo chamber that radicalizes its inhabitants, destabilizes their ability to tell truth from fiction, and undermines their confidence in institutions” (Faris, Roberts, and Benkler 2018, 383). In other words, the spread of repressive radical media did not bring about wider representation. The abundance of voices, inadvertently created closed spaces in which each subgroup had their own “facts” and “truth”, all together creating a “nation of distrust” (Dagnes 2019). 79 Enclave Extremism in a Far-right Minority Ethnic News Media The overview of the literature above shows how far-right media gives rise to enclave extremism. Ethnically oriented far-right media platforms have also increased but have not received much attention from scholars. Some ethnic minority media outlets do promote inter-cultural dialogue and spaces of community do exist. For example, Native American videos depicting cultural survival (Seubert 1987, in Downing 2001) or Spanish TV in the US disseminating South American accents to foster pan-Latino ethnic identity (Rodriquez 1999 in Downing 1984). Nonetheless, many other ethnic media oriented toward different interests. For example, most Spanish-language media in New York City in the 1980s were certainly neither tasked to contribute to ethnic cohesion and cultural maintenance nor to voice the needs of Latino communities, but rather to serve business purposes of the owners (Riggins Ed., 1992). In the last few years, an unprecedented number of Asian-Americans have become subsumed by far-right media.3 The “us versus them” rhetoric has activated deep divisions among immigrant and refugee populations. These followers have adopted the idea of a "good versus bad immigrant" which was initiated originally by the Trump administration.4 Even among Vietnamese Americans who comprise a mainly immigrant and refugee population, there were deep divisions between those who supported Trump and those who did not.5 Many well- educated members of the Chinese community were attracted to the rhetoric of the populist right during the debates on global political events such as the European refugee crisis and the American presidential election. They appropriated the rhetoric of “Western-style” right-wing populism to criticize Western hegemony, combining those vocabulary and argument with the pre-existing expression of cyber-nationalism, cyber-racism and Han supremacism”. In so doing, Chinese internet users “discursively construct China’s ethno-racial and political identities” (Zhang 2020, 89, 107). The Epoch Times has become one of the most popular news outlets among Asian-Americans in the last few years. ET proclaims its motto as “Truth & Tradition. Fact Bases. Unbiased. 3 https://www.bbc.com/news/world-us-canada-57587364 4 https://www.nbcnews.com/news/asian-america/asian-american-voters-support-biden-54-trump-30-numbers-don-n1240263 5 https://www.nbcnews.com/news/asian-america/who-are-asian-americans-still-voting-trump-spite-his-china-n1244849 80 Reporting important news other media ignore”. Within just a few years, The Epoch Times, with dozens of editions in multiple languages, remarkably gained its popularity and influence among right-wing audiences, becoming “one of the most mysterious fixtures of the pro-Trump media universe”.6 According to a New York Time’s investigation, until 2016 ET was just a small newspaper whose Chief Editor was the leader of the Falun Gong spiritual movement banned in China and its circulation was mostly among its religious practitioners. But ET pivoted its position by making a big bet on Trump, treating him as an ally to fight the Chinese communists (The New York Times, 2020).7 In other words, “Falun Gong came to see Trump as a kind of killer angel, summoned from heaven to smite the Chinese government” (The Atlantic, 2021).8 In 2020, The Epoch Times was the top advocate of Donald Trump on Facebook by number. Between February and August 2020, they spent about $1.5 million on 11,000 pro-Trump advertisements, an amount that was more than any organization outside of the Trump campaign itself. This figure also outsized the amount of campaign spending on most Democratic presidential candidates (NBC News, 2019).9 ET soon gained attention from Trump and his family and attended the White House press briefing.10 From a position of “near irrelevance” for more than a decade after it was founded in 2000, by early 2021 the ET “claim(ed) the most popular Apple newspaper app” in the US (The Atlantic, 2021).11 The Epoch Times Group has been identified by media scholars such as Dr. Donovan of Harvard Kennedy School, political observers, and journalists at Nieman Foundation and many mainstream media outlets as a far-right media outlet spreading disinformation.12, 13 Many of their video talk shows, YouTube accounts, or other affiliated channels have blatantly spread disinformation about the “deep state”, the QAnon conspiracy theory, and so on. However, in this project I only consider ET’s political news reports in which the language and the semantics are supposedly more “subtle” than their opinion sections. Compared to detecting the falsehood in other ET sections, the task of identifying reporting biases and political agendas behind the news reports are somewhat more challenging, requiring deeper analysis of the discourse and subtle 6 https://www.nytimes.com/2020/02/05/technology/epoch-times-youtube.html 7 https://www.nytimes.com/2020/10/24/technology/epoch-times-influence-falun-gong.html 8 https://www.theatlantic.com/politics/archive/2021/01/inside-the-epoch-times-a-mysterious-pro-trump-newspaper/617645/ 9 https://www.nbcnews.com/tech/tech-news/trump-qanon-impending-judgment-day-behind-facebook-fueled-rise-epoch-n1044121 10 https://www.nytimes.com/2020/10/24/technology/epoch-times-influence-falun-gong.html 11 https://www.theatlantic.com/politics/archive/2021/01/inside-the-epoch-times-a-mysterious-pro-trump-newspaper/617645/ 12 https://www.theguardian.com/us-news/2021/apr/30/falun-gong-media-epoch-times-democrats-chinese-communists 13 https://www.niemanlab.org/2021/02/the-dark-side-of-translation-the-epoch-times-is-now-spreading-disinformation-through-new-brands/ 81 semantics. But I believe it is more meaningful for the public to realize how seemingly objective reporting could contain systematic hidden agendas. In the same vein, it is important to show how the biases hidden in the news reports can misled enthusiastic supporters for the rule of law and democracy. Figure 13: Number of articles published 2013 – 2021 Note: Two peaks are years of US presidential elections 2016, 2020. ET makes an ideal case to study far-right ethnic media in the US because of its minority ownership and religious affiliation, its commitment to conservative issues, and its inclusion in far-right circles. ET worked aggressively to politically activate minorities who have historically not engaged in the political realm and quickly converted them to loyal consumers of ET reporting, especially during presidential election cycles. ET is particularly popular among Vietnamese readers in America and in Vietnam. Indeed, the Vietnamese edition was reportedly among the most successful branches.14 I randomly checked hundreds of articles in both editions to verify that most political articles in Vietnamese are translated directly from the English edition. Therefore, although this analysis was only conducted on the English edition, we can safely assume that Vietnamese readers received similar ET content. 14 https://www.nytimes.com/2020/10/24/technology/epoch-times-influence-falun-gong.html 82 III. Data and Method Data Collection The dataset consists of 19278 articles published in the “Politics” section of The Epoch Times (http://theepochtimes.com) between 2013-06-20 and 2021-03-07. I used R and the following packages to scrape all text articles available in the section in March 2013: rvest, tidyverse, xlm2, httr, tidytext, stringr.15 At the time of content scraping, this section only contains articles that were published from June 20, 2013, or later. The text data were then processed and converted into a data frame with 19278 observations and 7 fields. Data Partition I use a combination of quantitative and qualitative methods to partition the data into subgroups. First, in familiarizing myself with ET’s coverage and semantics, I regularly read the political section for several months. Next, I manually constructed lists of themes and lists of keywords for each theme. I then used these lists of keywords to search article titles. The search is conducted using regular expression, a kind of sequence of characters that specifies patterns of interest.16 Because the corpus of this study is news articles, we can assume that titles are quite illustrative for signaling the main subject of each article. This means that the title search results would also be inclusive of articles covering the same theme. For example, to determine a subset of articles that directly mention China in the headlines, I use these key words: CCP, China, Chinese, Communist and regular pattern: '.*(?i) ccp.*|china|Communist|Chinese*', which result in 933 articles. Any news coverage on China would certainly include at least one of these key words in its title. Thus, my sample would ensure that no article on China gets missed. The procedure above is repeated for each subset of data to further investigate the discourse in detail.17 For example, within the subset of 933 article on China, I read through the list of titles to 15 The whole dataset that has been collected, processed, and cleaned for this project will be made available freely to the public on a Github repository. This will be particularly meaningful for the academic community, where data collection is often one of the most challenging and costly tasks. All the codes that were used to collect and to clean the data will also be made available. This will help the researchers to go past data manipulation phase quickly and proceed to applying appropriate algorithms and statistical models to answer their research questions. 16 https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_Expressions 17 I attempted to implement topic modeling to partition the data, a method that relies a lot more on algorithms functioning and a lot less on human knowledge from reading the materials. The result was not inclusive. It’s probably because such approaches require a much more homogenous and larger dataset. With a great amount of time invested in reading the site, I am convinced that a combination of human reading experience and computer tools are the best option for correctly partition the data. 83 tentatively identify common themes, construct a list of keywords, build a regular expression pattern, then search against the whole list of titles. R packages used in these tasks were: lubridate, skimr, janitor, ggplot2, ggrepel, wordcloud, tidyverse, dplyr. Once I identified key themes and collected sufficient subsets of articles for each theme, I qualitatively read through the descriptions and the main contents of the articles. During this reading step, I relied on a discourse-historical analysis approach to analyze the text. Discourse-historical Approach I apply a discourse-historical approach (Reisigl 2018) to analyze the text of ET’s political news articles. This method was adopted from critical discursive analytical approaches which focus on “the discursive reproduction of power abuse and social inequality, with a detailed examination of the concept of power” (Dijk 2008, 1). Reisigl (2018, 52) points to five analytical questions, or five discursive strategies, through which discourse-historical analyses could be done systematically: 1) nomination – how persons and things are named and referred to, 2) predication – what characteristics are attributed to actors, 3) argumentation – what argument are employed, 4) perspectivization – from what perspective are the nominations, attributions, arguments expressed, and 5) mitigation and intensification – to what direction are the utterances articulated. IV. Analysis This section analyzes two main strategies that ET employs in their enclave construction: bundling enemies and bundling identities, illustrated in the graph below. Bundling enemies are tactics that ET uses in undermining its own traditional enemy, the Chinese Communist Party (CCP), in making them appear as the US’s ultimate foe while tying them together with subjects that are considered adversaries of the far-right. Bundling identities is a way for ET to align with the political Right, labeling themselves conservatives, “good” immigrants, naturalized citizens, and Trump supporters. 84 Figure 14: Analytical framework illustrates two main discursive strategies deployed in ET Bundling Enemies: The CCP, Democrats, “Bad” Immigrants, Antifa, BLMs In this section I investigate ET’s strategy of bundling enemies. First, ET’s traditional enemy, the Chinese Communist Party (CCP), is painted with two main strokes. ET contends that CCP is the US’s number one security threat aggressively intervening in presidential elections to sway outcomes. ET also claims that the CCP intentionally spread the global pandemic to other countries, the US in particular. Next, ET actively associates the common targets of the far-right media such as Democrats, Liberals, “illegal” immigrants and refugees, Antifa, and Black Lives Matter proponents with the CCP. Selling the Ultimate Enemy: The Chinese Communist Party ET exploits the anti-communism sentiment in the USA to sell the idea that the Chinese Communist Party is the ultimate enemy of the Americans. That is, “the CCP is deceiving and poisoning the minds of young students on U.S. campuses with its socialist and communist ideas and luring them to resort to any means to subvert the current American government” (Historian: Why Today’s America Is Akin to 1949 China, Publish 2020-12-10). 85 The outlet develops a narrative around Chinese interference in the US election begins in June 2020. Before that, there is only 1 article of the same topic published in 2017 and 3 others in 2018. China is defined as the “US’s top election security threat”, playing the “most active role in US election interference”. China’s actions are defined as “an act of war” that must be recognized as such in holding China “responsible for election fraud”. The nature of China’s intention is to “sow discord” and to “subvert America”. On September 3, 2020, several Democratic senators urged the Trump administration to impose sanctions on individuals linked to Russia. ET published an article reporting the event with the headline: “Senate Democrats Urge Trump Admin to Sanction Russia Not China Over Election Interference”. In the same day, ET also ran an article titled “Intelligence Shows China, Not Russia, Is Top US Election Security Threat: Barr”, in which it asserted that CCP, not Russia, poses the biggest threat to U.S. election security. The juxtaposition of the two articles does little in the way of presenting an alternative view as a regular practice of objective journalism. In contrast, it acts as a guide for readers to see their “truth” – that Russia was brought up by the Democrats, while China remains the real threat to America. Next, ET claims the CCP is solely responsible for the pandemic. Following Trump’s tweet on March 16, 2020, where he used the term “Chinese virus” instead of coronavirus, the ET began using the term “CCP virus” only two days later on its coverage of the Covid-19 pandemic. Only three articles published in early March 2020 used the term “coronavirus”. After March 18, 2020, the term “Covid-19” was only used on three occasions when referring to an official report. In the rest of the 197 news articles that report on the pandemic, the term “CCP virus” is consistently used. ET explains its choice of the term: “The Epoch Times refers to the novel coronavirus, which causes the disease COVID-19, as the CCP virus because the Chinese Communist Party’s coverup and mismanagement of the outbreak in Wuhan allowed the virus to spread across China and create a global pandemic.” (Trump: Spread of CCP Virus Not the Fault of Asian Americans, March 24, 2020) 86 Using the words of Rudy Giuliani, Trump’s personal attorney, it insisted that CCP “intentionally kept its borders open to maximize damage to the rest of the world from the CCP virus.” (Giuliani: Chinese Regime Let CCP Virus Escape to Damage the World in ‘Act of War’). Furthermore, the blaming game was waged at all levels. At a macro level, CCP was accused of “ruining economy in just 60 days due to its handling of the CCP virus outbreak”, “cover(ing) up of the virus outbreak and allow(ing) it to spread by continuing to let people from China travel to other parts of the world” (White House Trade Adviser Says Chinese Communist Party Took Down US Economy in 60 Days). At a micro level, ET ran stories of Chinese individuals who had “malicious joy in the misfortune of others”, by highlighting some business owners in China who “celebrated” Trump’s infection by offering discounts to customers (Chinese State Media Editor Mocks Trump Over Virus Diagnosis). Overall, ET created a discursive battle where, on one hand, the Republicans were doing their best for the public interest showing how “Trump expressed calm, (and) didn’t downplay CCP virus in February”, or how “Rep. Mark Walker introduced act to hold China accountable for CCP virus”. On the other hand, in referring to the Democrats, it suggests that if there was any reason that the terms “China virus” or “Wuhan virus” didn’t stick in the US, it’s because a lot of influential people consider commercial relationships with China more important than America. (Giuliani: Chinese Regime Let CCP Virus Escape to Damage the World in ‘Act of War’). Associating the Democrats/Liberals, the Mainstream Media with the CCP With the CCP comprising the ultimate enemy, ET exploited all opportunities for equating the common targets of the far-right to the CCP. The CCP’s system became the benchmark by which the Democrats/Liberals, mainstream media, and “other part” of America would be measured against. ET’s most pronounced narrative was that the mainstream media and the Left are hospitable to the emergence of leftist ideas, which will lead America to experience what happened in China. For instance, Chinese historian Xin Haonian, whose work was published by ET as a book, likens the present situation in America to China in 1949 when “the communists controlled the media, published lies in the media, cheated the people, and criticized the government under the pretext of free expression. […] Literature, history, philosophy, social 87 science, anthropology, the study of law, media, and other concentrations have become inundated with various derivatives of Marxist theory” (Historian: Why Today’s America Is Akin to 1949 China, Publish 2020-12-10). The forceful comparison between today’s America and 1949 China was essentially a total attack of the part of America that is under the influence of the Liberals and Democrats. ET echoed loud and full the voices in right-wing media by relating Twitter’s ban on Trump and the criticism of Fox News to Chinese dictatorship, “cancel culture” to CCP’s Social Credit System, and Trump’s impeachment to public shaming in China during the revolution. ET did not mind taking both sides of an argument. On the one hand, they lined up with the conservatives to claim that private companies such as Twitter and Facebook who banned Trump’s and some others’ accounts were comparable to CCP censorship (Pompeo Says Twitter’s Ban on Trump ‘Un-American,’ Compares It to CCP Censorship). On the other hand, ET took opposing sides by comparing Democratic House members who requested private companies such as television service carriers to combat misinformation with China’s oppression of private entities. In the first case, they accused private companies of political motivations. In the second case, they accused the Democrats of intervening in the operations of the private sector (Top Republican Compares House Democrats’ Attempt to Deplatform Fox News With CCP Censorship). Citing Judicial Watch President Tom Fitton, they concluded that “The political left in the United States is leveraging the pandemic caused by the Chinese Communist Party (CCP) to promote fraud-prone mail-in balloting in order to steal elections”, spreading the belief that “mail-in voting is particularly susceptible to fraud” (Conservatives Say Left Using CCP Virus as Pretext to Impose Nationwide Mail-in Voting). They further proposed that “American media outlets have stepped up content censorship to keep people from the truth; and people are experiencing communist brainwashing […] Those media even willingly serve up fake news, serve the CCP and the U.S. Democratic Party, and push America further into ruin” (Historian: Why Today’s America Is Akin to 1949 China, Publish 2020-12-10). The associating tactic described above shares some similarity with Jamieson et. al ’s findings in their examination of Limbaugh’s political talk radio show. The authors found that Limbaugh evoked moral outrage that “take advantage of what psychologists call a core theme associated 88 with the production of anger (sic) - namely, a transgression, often moral, against the self or someone close to the self”. For example, the show host claimed that Democrat senator Robert Byrd was a grand wizard of the Ku Klux Klan then severely attacked the Democrats for failing to criticize Byrd. The show also criticized Al Gore because his father did not support the 1964 Civil Rights Act (Jamieson 2008, 133). Such associations evoking moral outrage were apparent in ET’s discourse, but on a whole new level. ET mobilized the Chinese, Vietnamese and other Asian-American communities against the force associated with one of the worst experiences in modern history – Maoist communism. In this way, the intended audience did not gain representation for their interests, cultures, or political beliefs, but were rather misled in arousing fierce anger. Demonizing “Illegal” Immigrants and Refugees, Antifa, Black Lives Matters The far-right public sphere “seems to be based on negative sentiments. It is grounded by the feeling of being threated on (an) individual and collective level. This generalized fear is channeled into the public sphere creating a dogmatic semantics of friend/enemy” (Sik 2015). The author identifies five thematic clusters in far-right Hungary media: ‘anti-Hungarianism’, ‘Gypsy crime’, ‘gene cemetery’, ‘politician crime’, and ‘Jew crime’. Hasell and Weeks (2016) posit that “partisan media may drive online information sharing by generating anger in its audience. Tischauser and Musgrave (2020) identify three inter-related themes that represent racial meaning, namely an-white whites, race realism, “cultural Marxism”. The discourse on the extermination of the historic American nation, consisted of four sub-themes: leftist violence, black crime, media bias, and deep state operatives (289). I found these themes repeatedly in ET political news articles. ET relied on white supremacist narratives asserting that BLM and Antifa are threatening to innocent lives. Far-left groups were accused of instigating violence in the general public and towards the police by organizing riots nationwide. Terms such as “anarcho-socialist”, “violent anarchist”, “communist groups” were frequently used as a default attribution to the left. The fear of losing economic interest to “others” is visible in the way they report the second round of Covid-19 government aid as “mixed-status immigrant families eligible for stimulus checks” (Mixed-Status Immigrant Families Eligible for Stimulus Checks Under New Relief Bill). The emphasis was that the 89 “socialist politician” Bernie Sander planned to “nationalize the health care industry and cost U.S. taxpayers an estimated $32 trillion during the first 10 years” to pay for all immigrants. (Bernie Sanders: Illegal Immigrants ‘Absolutely’ Covered in His Medicare for All Plan). The burden of illegal immigration at the border is made bold in terms of cost, i.e., taxpayers’ money to stop illegal crossings. Again, partisan actors were put in dual opposing roles, where Trump “enhanced border security and sent active-duty military to help stop illegal crossings”, while “when Obama was in office, these illegal immigrants were very cocky with us. They went on about ‘the Dream Act, Obama, the Dream Act”. The argument explicitly blames Democratic politicians “who often side with illegal immigrants (when) looking for future votes (In a Small Texas County, a Sheriff Battles With Illegal Immigrant Crime). In short, the bundling enemies strategy worked to heighten fear and anger in ET discourse. The common enemies were built up through accusatory language and arguments. They heightened the fear of violence in the country by, for example, mispresenting the Antifa movement as an “extremist organization”, and Black Lives Matter as a pro-communist organization. More generally, they often aligned the Left with communist groups, socialist groups, and the broader collective political mindset. There were virtually no alternative views reported. The ideas and “facts” were presented in a very limited spectrum revolving around an extreme narrative: that the Chinese government was an evil tyrant, the Democrats and the Liberals were compromised, and that the mainstream media was misleading the public for the worst. Bundling Identities: “Our” Knowledge, “Our” Values, “Our” People “Our” Shared Knowledge, “Our” Shared Values ET particularly relied on the use of “experts” in delivering political comments. The label “expert” was tacked onto individuals from various backgrounds (see table 1), with a strict common denominator of their identities as Trump-supporters, China-condemners, and far-right proponents. There were no independent experts considering balanced views with representation from both sides. They had their own election experts claiming that “voter fraud is real”, data experts that talked about “vote(r) irregularities in Pennsylvania and Georgia”, cybersecurity experts revealing “35,000 Fraud Votes Embedded for Democrat Candidates in Arizona”, and 90 billionaires warning “stock market dip(s) if Democrats beat Trump in 2020”. When someone from the Left was included in ET commentary, it was only to the extent that such person criticized the Democrats or praised Trump. For example, Antony Blinken, a Biden appointed diplomat, agreed that Trump was right in “taking a tougher approach to China”, and Gina Raimondo, Biden’s Commerce Chief, acknowledged that Trump’s China tariffs on certain goods were “effective” (‘Trump Was Right’ With Tougher Approach to China: Biden’s Top Diplomat Nominee, Biden’s Commerce Chief Gina Raimondo Praises Trump’s China Tariffs as ‘Effective’). Apart from such exceptions, the possibilities of communism in America were constantly reinforced by experts who contended that, by the push for open borders”, was really just a socialist experiment, family structures could be destroyed by the progressives, or today’s US could be comparable to China in 1949. In some case, both sides of experts could be quoted, but not to expand viewpoints or diversify opinions. On the contrary, it was more to lock readers to a one-sided political belief. On the talk of voter fraud in 2016, ET quoted accredited academics that “voter fraud is so incredibly rare that it has no impact on the integrity of our elections”, reassuring their readers that “you are more likely to be struck by lightning, more likely to see a UFO, than to be a victim of voter fraud.” They brought in a study by a Loyola Law School professor that found “31 instances involving allegations of voter impersonation out of 1 billion votes cast in U.S. elections between 2000 and 2014.”, a study by the Brennan Center for Justice at New York University Law School that found “many reports of people voting twice or ballots being cast on behalf of dead people were largely the result of clerical errors that suggested wrongdoing when none had occurred’ (“Amid Talk of ‘Rigged’ Election, Experts Say Fraud Is Rare”). However, on a report in the exact same topic in 2020, none of these researchers were consulted again. Instead, ET published an article titled “Voter Fraud Is Real, Elections Expert Says”, quoting Hans von Spakovsky, a manager of The Heritage Foundation’s Election Law Reform Initiative. This “expert” pointed to a “database managed by the foundation that has 1,298 proven instances of voter fraud”, in which “nearly 350,000 dead registrants remain on voter rolls across 41 states” and “the total potential cases of fraud that they uncovered is over 140,000 cases, just from the last two federal elections”. Obviously, the same topic was framed in two completely different ways. There was a double standard in evaluating the quality of the US election system. The former was to refute any 91 suspicion in the 2016 election result when Trump won. The latter was to question the integrity of Trump’s election loss in 2020. ET included in their coverage a lot of far-right activists who often appeared or had their works cited in other far-right media. One example is Edward Feser, a philosophy professor at Pasadena City College in California, who instilled fear that socialists would destroy family traditions. Accordingly, “socialists push for plans such as a universal basic income, single-payer health care, state-backed education for all, and heavy taxes on inheritance”. This means “all basic material needs will thereby be provided by the state rather than by the father or even the father and mother together […] The state becomes the de facto breadwinner for all” (Socialism Ruins the Traditional Family Structure, Warns Expert). Indeed, ET hunted a fair number of academics, researchers, and activists on the far-right spectrum to speak to their end. The table below provides an illustration of types of interviewees, speakers, actors who made headlines on ET. Table 7: Examples of “Experts” interviewed or quoted in ET political news Authors as Examples of interviews, quotes on Article introduced on ET ET 1 China analyst Gordon “The Chinese Communist Party is Cooperation With Chang, working to overthrow the government Communist China Author of “The of the United States […] inciting Impossible—It Coming Collapse of violence on American streets […] Seeks to Overthrow China” connection with Denver’s riots.” America 2 Christian Adams, “Mail ballots are uniquely subject to Push for Mail-In president of the fraud and undue influence” […] Balloting During Indianapolis-based Pandemic Is Public Interest Legal Partisan and Foundation Political, Experts Sa 3 Hans von Spakovsky, “Voter fraud isn’t a myth, with nearly Voter Fraud Is The Heritage 1,300 proven instances… And that’s Real, Elections Foundation’s Election not a comprehensive list. It’s just a Expert Says Law Reform Initiative sampling of cases” 4 Chinese historian Xin American media outlets have stepped Historian: Why Haonian, author of up content censorship to keep people Today’s America Is “One Country Two from the truth; and people are Akin to 1949 Systems and the Civil experiencing communist China, Publish War of America” brainwashing. 2020-12-10 92 Authors as Examples of interviews, quotes on Article introduced on ET ET 5 Michael Rectenwald, “The formation of a totalitarian state Ideological retired liberal arts is just about complete in America” Alignment Pushing professor at New York America Toward University18 Totalitarianism, Experts Warn 6 Edward Feser, “The rhetoric and politics of Socialism Ruins the professor of socialism have seen a revival in the Traditional Family philosophy at United States. […] Liberalism has Structure, Warns Pasadena City College come to regard the basic premises of Expert in California feminism and of the sexual revolution as matters of justice”. 7 Phil Waldron, retired “an anonymous individual claiming Cybersecurity army colonel19 to have witnessed a detailed plan to Expert: 35,000 embed 35,000 fraud votes to each Fraud Votes Democrat candidate’s total votes” Embedded for Democrat Candidates in Arizona 8 Trevor Loudon, Tens of thousands of illegal aliens Communism Expert communism expert cross the southwest border every on What’s Behind week […] is an orchestrated, the Push for Open communist assault on America […] Borders 15, 16, 20, 25 million new Democratic voters is going to do to this country. You will lose Texas, you’ll lose Florida, you will lose Georgia, Arizona, North Carolina. There will never be, ever, another Republican or conservative president in our lifetimes. You will have a one- party state in America […] This is Marxist. 9 Jeff Nyquist, author, China is using fentanyl as a form of China Is Using researcher of Chinese chemical warfare. […] It opens up a Fentanyl as and Russian strategy number of opportunities for the ‘Chemical penetration of the country […] uses Warfare,’ Experts the money generated […] to Say effectively “influence political parties”. 10 Brian Kennedy, “One immediate way is to make sure Time to Take a chairman of the that the Chinese government makes ‘Tougher 18 Retired from NYU, invited alt-right provocateur Milo Yiannopoulos to discuss Halloween culture and politics in one of his writing classes. https://nyunews.com/2019/01/28/rectenwald-antipcprof-retires-from-nyu/ 19 https://www.nytimes.com/2021/12/21/us/politics/phil-waldron-jan-6.html 93 Authors as Examples of interviews, quotes on Article introduced on ET ET Committee on the good on their defaulted sovereign Approach’ Toward Present Danger: China debt held by 20,000 American CCP, Experts Say families”20 11 Anders Corr, the […] Xi Jinping is hoping for a Biden Time to Take a publisher of the presidency […] Chinese state media ‘Tougher Journal of Political have openly voiced support for a Approach’ Toward Risk Biden presidency, saying he would be CCP, Experts Say “smoother” for the regime to deal with than Trump. The above described phenomenon is related to what Mihailovic (2015a) called “hijack authority.” Through an analysis websites of two academics, Keven MacDonald, a retired psychology professor from California State University - Long Beach and Aleksandr Dugin, a professor of Sociology at Moscow State University in Russia, the author finds that the two used their academic positions to “construct academic identities with the goal of taking on the mantle of “conventional” authority”, making themselves a source of authority to spreading anti-Semitic propaganda while establishing a “neo-Aryan” nationalist authority. In short, these selected right-wing experts created an “authority” around the issue they purported to know best. They were put in a position to interpret the hidden agendas of the Left for readers, while leading readers to more extreme positions by strategically keeping out important information. In so doing, ET created information cocoons to isolate readers with only one perspective. If any alternative view was presented, it was only to be ridiculed. Readers, therefore, voluntarily select themselves into the political enclaves and niches that they become a part of. As this enclaving occurs, people move further into their niches and further away from alternative viewpoints, producing rampant disinformation and falsehoods. “Our” People - The “Good” Immigrants, Naturalized Citizens, and Conservative Americans If, as in ET’s script, China was the number one threat to the US and the cause of the pandemic while immigrants were to blame for social chaos in the US, then how could Chinese immigrants 20 Refer to the resolution introduced by Rep. Mark Green, demanding that China's government immediately pay $1.6 trillion: https://markgreen.house.gov/2020/8/rep-mark-green-introduces-resolution-demanding-china-repay-1-6-trillion-in-debt 94 justify living in America for good? Being an ethnic-minority media while choosing to brand itself as a conservative outlet associated with the far-right, ET faces an identity dilemma. Given the fact that the core proponents of far-right circles are white, anti-immigrant, anti-refugees, and anti-Asia, how did ET make sense of the existence of their audience which mainly consists of Chinese, Vietnamese, and other Asian immigrants and refugees in this country? First, ET positions itself and many interviewees as first-hand witnesses knowing from personal experience what communism and socialism look like. They believe that they’re telling Americans to wake up and that they are fighting against this devil with Americans, for Americans. In the words of a Chinese American who fled China years ago, ET justified: “I can’t just watch this country go down the way of China, a communist country, because that is where I came from, and I left searching for freedom.” Against this background, it is “no longer about whether you’re picking Trump or Biden as the president, this is a contest between good and evil. If Americans let this fraud happen, then it’s basically already a totalitarian state” (Chinese American Drives Three Hours to Protest Election Fraud, Tells Americans Not to Stay Home). Another woman expressed a similar sentiment when she joined a demonstration for the first time in her life. Going to the “Stop the Steal” rally in Harrisburg, Pennsylvania, she declared she “does not want to see America become a socialist country”, and yet felt “America has come to a critical point and freedoms are at stake” (Chinese Immigrant Says She Doesn’t Want to See America Lose Her Freedoms). Second, to pin Asian Americans, who are immigrants themselves, against other immigrants, as well as to get the non-whites, who are victims themselves, to support white supremacy, ET had their readers locked into a mentality that they were the “good” immigrants – i.e. those who came to the US “legally” and followed the naturalization procedures to become US citizens - as opposed to the “illegal” immigrants coming in from the southern border. ET covered a rally of “naturalized citizens” at Sunrise Highway on Nov. 6, 2020, the day before the result of the 2020 presidential election. Attendees stood side-by-side with Trump-supporters in the midst of “growing allegations of ballot-counting mishaps”. One revealed she “moved to the United States just to enjoy its freedom and democracy, its universal values. But right now, they are all gone” (Long Island Chinese-American Voter Concerned about Election Fraud). Immigrants are 95 welcome, but only if they support Trump: “Immigrants, voters or not, should think about joining such demonstrations because no matter what candidate they like, they are on American soil, and their homes and their livelihoods are on American soil. […] It’s not about the candidate it’s about whether this is a fair election [..] We have to use our voices today, support Trump, support a fair election,” (Chinese Immigrant Says She Doesn’t Want to See America Lose Her Freedoms). In addition, while trying to link the onset of the pandemic with the Chinese government, ET also made efforts to distinguish Asian Americans, a major part of their readership, from the dictatorial authority. Borrowing Trump’s words, “Asian Americans aren’t responsible for spreading the CCP virus and that they need protection […] It’s very important that we totally protect our Asian-American community […] They are amazing people.” (Trump: Spread of CCP Virus Not the Fault of Asian Americans, March 24, 2020) Another image that constitutes Asian immigrant identity in ET discourse was the troops who defend conservative ideals, side-by-side with other conservatives. They adopted languages such as “abortion industry”, “late-term abortion”, “cancel culture”, etc. At the same time, they proactively attacked the opposition, standing shoulder-to-shoulder with the right-wing, to denounce abortion right, LGBTQ rights, feminism, as long as the idea of viewing individuals as units of society rather than nuclear family. Among these is the biggest straw-man argument around the idea that socialism, communism, and the collective mindset are growing in American. ET even joined the ultra-right in denouncing the reality of racism in the US. In a news article, a young Asian-American pointed to the relative success of Asian-American minorities in the US to “debunk” any claim that the United States is systemically racist. This person also held that targeted police brutality on Black people in the United States is simply not true, suggesting that there are so many more White people who die by police than Black people. The article ends with a call to action, proposing that: “as an Asian […] you don’t need to conform to the Left. And conservatism is the way to go. And we must go back to the values, the American values that made this country so great to begin with, if you want to see progress and a better future for this country” (‘Wokeism Is the Next or the New Secular Religion’: Young Asian-American Conservative). 96 All in all, identity bundling in ET discourse was a process of band-wagoning. Far from genuinely helping an ethnic community to recognize its true identity and bringing new voices to the public sphere, what ET actually did was bundle loyal readers to a prevailing faction. Nowhere in ET discourse did we see the interest of the readership presented independently from the outlet’s political agenda. Instead, the whole discourse was a mobilization of a historically inactive group into a political extreme faction. V. Conclusion Scholars have long warned about the danger of political enclaving. Once people get wrapped in a bubble of an “alternative truth”, they tend to slide further to the extreme, making dialogue with people having opposing views unlikely. It becomes even more dangerous when “echo chambers ringing with false news make democracies ungovernable” (Faris, Roberts, and Benkler 2018). History saw great consequences of those who flock together without sufficient consideration of counterarguments and civil dialogues; The rise of fascism and terrorism are two such examples. Most recently, the image of Chinese, Vietnamese and other Asian-Americans “wholeheartedly” participating in the January 6th Capitol riot thinking that they were doing some good for America, suggests a new level of peril. It signals that a historically inactive population can be mobilized to political events that subvert democratic processes. Expanding beyond Sunstein’s original conceptualization of enclave extremism, which emphasizes the consumption side – how people tend to conform to their group members’ view, I have examined the supply side – how the enclave was made by media producers. The study provides new insights into how an ethno-minority news outlet mobilized its target ethnic population by opportunistically jumping into the far-right circle, baking the audience into clave extremism. From a religio-ethnic minority outlet historically circulated among a small group of religious practitioners, The Epoch Times entered the far-right media circle during the time such ecosystems were quickly expanding. But ET’s ambitions, and its subsequent growth, were not tied to diversifying the democratic public sphere or enhancing democratic values. In fact, the news articles were presented to further isolate their readers in a far-right bubble of extreme “truth”, prejudice, and hatred. On top of its strategic leverage of far-right languages and 97 narratives, ET heightened the blame game, further demonizing common targets of the right-wing media as well as its own ultimate target – the Chinese government. The discourse analysis of the political news articles in this paper demonstrates two major strategies that The Epoch Times deployed in strategically advancing its goals: bundling enemies and bundling identities. The first strategy, bundling enemies, involved setting up an ultimate enemy, the Chinese Communist Party, with all the linkages of the major problems in the US today, including election fraud, the global pandemic, and its internal and external security threats. This enemy was associated with other common targets in the far-right media such as the Democrats/ Liberals, mainstream media, and Black Lives Matters. On top of that, other progressive phenomena or events in the US were compared to dictatorial China. The second strategy, bundling identities, comprised of building a notion of shared-knowledge, shared-values, and distinguishing “us” – the “good” immigrants, “naturalized citizens”, “conservatives” - from “them” – the “bad” immigrants looking for free social services while being exploited by corrupt politician for votes. Reference 1992. Ethnic minority media : an international perspective. Edited by Riggins. Newbury Park, Calif.: Newbury Park, Calif. : Sage Publications. Atton, Chris. 2006. "Far-right media on the internet: culture, discourse and power." New media & society 8 (4):573-587. doi: 10.1177/1461444806065653. Back, Les. 2002. "Aryans Reading Adorno: Cyber-Culture and Twenty-First-Century Racism." Ethnic and racial studies 25 (4):628-651. Baum, Matthew A. 2010. War Stories : The Causes and Consequences of Public Views of War. Princeton: Princeton : Princeton University Press. Baum, Matthew A., and Tim Groeling. 2008. "New Media and the Polarization of American Political Discourse." Political communication 25 (4):345-365. doi: 10.1080/10584600802426965. Cammaerts, Bart. 2009. "Radical pluralism and free speech in online public spaces: The case of North Belgian extreme right discourses." International journal of cultural studies 12 (6):555-575. doi: 10.1177/1367877909342479. Colgan, Jeff D., and Robert O. Keohane. 2017. "The Liberal Order Is Rigged: Fix It Now or Watch It Wither." Foreign affairs (New York, N.Y.) 96 (3):36-44. Dagnes, Alison. 2019. Super mad at everything all the time : political media and our national anger, Political media and our national anger. Cham, Switzerland: Cham, Switzerland : Palgrave Macmillan. 98 Daniels, Jessie. 2009. "Cloaked websites: propaganda, cyber-racism and epistemology in the digital era." New media & society 11 (5):659-683. doi: 10.1177/1461444809105345. De Benedictis-Kessner, Justin, Matthew A. Baum, Adam J. Berinsky, and Teppei Yamamoto. 2019. "Persuading the Enemy: Estimating the Persuasive Effects of Partisan Media with the Preference-Incorporating Choice and Assignment Design." The American political science review 113 (4):902-916. doi: 10.1017/S0003055419000418. Dijk, Teun A. van. 2008. Discourse and power, Discourse & power. Houndmills, Basingstoke, Hampshire [United Kingdom] ; New York, N.Y.: Houndmills, Basingstoke, Hampshire United Kingdom ; New York, N.Y. : Palgrave Macmillan. Downing, John. 1984. Radical media: The political organization of alternative communication. Boston: South End Press. Downing, John. 2001. Radical media rebellious communication and social movements. Thousand Oaks, Calif. ; London: Thousand Oaks, Calif. ; London : SAGE. Faris, Robert, Hal Roberts, and Yochai Benkler. 2018. Network Propaganda: Manipulation, Disinformation, and Radicalization in American Politics. New York: New York: Oxford University Press. Hasell, A., and Brian E. Weeks. 2016. "Partisan Provocation: The Role of Partisan News Use and Emotional Responses in Political Information Sharing in Social Media." Human communication research 42 (4):641-661. doi: 10.1111/hcre.12092. Hollander, Barry A. 1995. "The influence of talk radio on political efficacy and participation." Journal of radio studies 3 (1):23-31. doi: 10.1080/19376529509361971. Hollander, Barry A. 1996. "Talk Radio: Predictors of Use and Effects on Attitudes about Government." Journalism & mass communication quarterly 73 (1):102-113. doi: 10.1177/107769909607300110. Ikenberry, G. John. 2018. "The end of liberal international order?" International affairs (London) 94 (1):7-23. doi: 10.1093/ia/iix241. Jamieson, Kathleen Hall. 2008. Echo Chamber : Rush Limbaugh and the Conservative Media Establishment. Cary New York ; Oxford: Cary : Oxford University Press, Incorporated. Mihailovic, Alexandar, ed. 2015a. Hijacking Authority : Academic Neo-Aryanism and Internet Expertise., Digital media strategies of the far-right in Europe and the United States. Lanham, Maryland: Lanham, Maryland : Lexington Books. Mihailovic, Alexandar, ed. 2015b. Hijacking Authority: Academic Neo-Aryanism and Internet Expertise. Edited by P. Simpson & H. Druxes, Digital media strategies of the far-right in Europe and the United States. Lanham, Maryland: Lanham, Maryland : Lexington Books. Moy, Patricia, Michael Pfau, and LeeAnn Kahlor. 1999. "Media use and public confidence in democratic institutions." Journal of broadcasting & electronic media 43 (2):137-158. doi: 10.1080/08838159909364481. Padovani, Cinzia. 2016. "The media of the ultra-right Discourse and audience activism online." Journal of language and politics 15 (4):399-421. doi: 10.1075/jlp.15.4.02pad. Pfau, Michael, Patricia Moy, R. Lance Holbert, Erin A. Szabo, Wei-Kuo Lin, and Weiwu Zhang. 1998. "The Influence of Political Talk Radio on Confidence in Democratic Institutions." Journalism & Mass Communication Quarterly 75 (4):730-745. doi: 10.1177/107769909807500408. 99 Project, Annenberg Democracy. 2007. A republic divided. Edited by Annenberg Democracy Project. Oxford ; New York : [Radnor, PA]: Oxford ; New York : Oxford University Press ; Radnor, PA : Annenberg Foundation Trust at Sunnylands. Reisigl, Martin. 2018. "The Discourse-Historical Approach." In The Routledge handbook of critical discourse studies, edited by John Flowerdew and John E. Richardson. Milton Park, Abingdon, Oxon ; New York, NY: Milton Park, Abingdon, Oxon ; New York, NY : Routledge. Sik, D., ed. 2015. The imitated public sphere: The case of Hungary's far right. Edited by P. Simpson & H. Druxes, Digital media strategies of the far-right in Europe and the United States. Lanham, Maryland: Lanham, Maryland : Lexington Books. Simpson, Patricia Anne, ed. 2015. Digital media strategies of the far-right in Europe and the United States. Edited by P. Simpson & H. Druxes, Pure Hate: The Political Aesthetic of Prussian Blue. Lanham, Maryland: Lanham, Maryland : Lexington Books. Tischauser, Jeff, and Kevin Musgrave. 2020. "Far-Right Media as Imitated Counterpublicity: A Discourse Analysis on Racial Meaning and Identity on Vdare.com." The Howard journal of communications 31 (3):282-296. doi: 10.1080/10646175.2019.1702124. Zhang, Chenchen. 2020. "Right-wing populism with Chinese characteristics? Identity, otherness and global imaginaries in debating world politics online." European journal of international relations 26 (1):88-115. doi: 10.1177/1354066119850253. 100 Appendices A. List of Countries Coded as Federal States List of 26 countries coded as federal states using method 1: "Argentina", "Australia", "Austria", "Belgium", "Bosnia and Herzegovina", "Brazil", "Canada", "Comoros", "Ethiopia", "Micronesia", "Germany", "India", "Malaysia", "Mexico", "Nigeria", "Pakistan", "Palau", "Papua New Guinea", "Russia", "Saint Kitts and Nevis", "South Africa", "Switzerland", "United Arab Emirates", "United States", "Venezuela" List of 11 additional countries coded as federal states using method 2: "Spain", "United Kingdom", "Chile", "Italy", "New Zealand", "Philippines", "Serbia", "Denmark", "Finland", "Netherlands", "France" B. Special Cases in Coding Hometown Vu Khoan, Do Quang Trung, Pham Khoi Nguyen were born in Ha Tay, which officially became a part of Hanoi in 2008. The discussion of integrating Ha Tay into Hanoi was made public several years before. For the purpose of understanding hometown as a powerbase, I chose to code hometown of Pham Khoi Nguyen, who was government member during 2007 - 2011 term as Ha Noi, as it reflects the true hometown – power connections. Meanwhile, I kept the hometown of Vu Khoan and Do Quang Trung, who were in the government during term 2011 – 2006, as Ha Tay, because their network then would be distinguishable from those from Hanoi at the time. Nguyen Phu Trong was born in Dong Anh in 1941. Although Dong Anh district belonged to Bac Ninh province then, it was integrated to Hanoi in 1961. I therefore code Trong’s hometown as Hanoi. Although these decisions do add weight to Hanoi, they do not shift the ranking of hometown power base in the analysis as a whole. Matrix 1: Person – to – working term matrix. Example of the first 10 observations. 101 Matrix 2: Person – to – person matrix. Example of the first 10 observations. Matrix 3: Working term – to – working term matrix 102 103