RULING PARTY INSTITUTIONALIZATION IN AUTOCRACIES By Fang-Yu Chen A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Political Science—Doctor of Philosophy 2020 ABSTRACT RULING PARTY INSTITUTIONALIZATION IN AUTOCRACIES By Fang-Yu Chen The majority of authoritarian regimes rely on a political party to govern, and studies show that political parties contribute to the political survival of dictators. However, there is a considerable variation on organizational developments of such ruling parties. Why do some authoritarian regimes institutionalize their ruling party and penetrate society through party organizations? Why do some dictators build up local party branches as linkage to the masses? I argue that the initial condition when a dictator comes to power determines the ruling strategies he will choose. The ruler needs to establish party organizations to control the masses when he subjectively thinks the threat to regime survival mainly comes from internal threats, especially the masses. However, if there exists a foreign enemy with territorial conflicts, the leader must direct resources to the coercive institutions, especially the military, to defend against threats. This study uses a cross-national dataset, covering all of dictators from 1950-2008, to illustrate the causes and consequences of ruling party institutionalization at the local level, accompanied by comparative historical studies of two country-pairs, Taiwan and South Korea, and Indonesia and the Philippines. Empirical evidence corroborates the hypotheses that external enemies with territorial threats impedes party institutionalization, while higher levels of religious polarization prompt the leader to develop party’s local organizations. Also, party institutionalization at the local level significantly prolongs political survival of dictators, because it helps leaders to collect essential information for the ruling and enhances legitimacy. Dedicated to my family. iii TABLE OF CONTENTS LIST OF TABLES…………………………………………………………………………..…… ii LIST OF FIGURES……………………………………………………………………………... iii CHAPTER 1 INTRODUCTION………………...…………………………………..………… 1 1-1 The Puzzle………………...…………………………………………….....…...……… 1 1-2 What Explains Causes and Effects of Party Institutionalization………..……… 8 1-3 Main Argument: A Threat-based Theory of Party Institutionalization…….….. 20 1-4 Outline of the Study………………………………………………………………… 24 CHAPTER 2 AUTHORITARIAN RULING PARTY INSTITUTIONALIZATION….. 27 2-1 The Concept of Party Institutionalization and its Multiple Dimensions……… 27 2-2 Measuring Party Institutionalization…………………………………………….. 36 2-3 Party Institutionalization at the Local Level………………….………………….. 43 2-4 Conclusion………………………………………………………………………….. 53 CHAPTER 3 CAUSES OF AUTHORITARIAN RULING PARTY INSTITUTIONAL- -IZATION…………………………………………………………………………………….. 54 3-1 Theories of Party Institutionalization………………………………………….… 55 3-2 Causes of Party Institutionalization: a cross-national study………………..… 64 3-3 Comparative Case Studies: Taiwan and South Korea…………………………… 75 3-4 Comparative Case Studies: Indonesia and The Philippines…………………… 85 3-5 Conclusion……………………………………………………………...……… 94 CHAPTER 4 AUTHORITARIAN RULING PARTY INSTITUTIONALIZATION AND ITS CONSEQUENCES ON POLITICAL SURVIVAL…………………………….……… 96 4-1 Party Institutionalization and Political Survival………………………………. 100 4-2 Explaining Authoritarian Regime Performance by Party Institutionalization.. 107 4-3 Discussion: Other Performances Of The Leader……………………………... 115 CHAPTER 5 CONCLUSION…………………………………………..……...……...…… 122 5.1 Findings……………………………………………………………………………. 122 5.2 Contributions of the Dissertation ……………………………….……………….. 124 5.3 Remaining Questions and Directions for Future Research……………...…… 129 BIBLIOGRAPHY……………………………………………………………………...… 134 iv LIST OF TABLES Table 2-1 Dimensions and Indicators of the Index of Party Institutionalization……….. 32 Table 2-2 Local Party Activities and Dimensions of Party Institutionalization…….…... 45 Table 2-3 Frequency of Local Party Presence………………………………………...…….. 48 Table 2-4 Regime Type and Party Institutionalization……………………………………. 48 Table 2-5 Party Institutionalization and Party Strength…………………………...……… 50 Table 2-6 Parties that May Have Been Mischaracterized…………………………………. 52 Table 3-1 Summary of Variables for Causes of Institutionalization………………...…… 70 Table 3-2 Logistic Regression Results of Causes of Institutionalization…………...……. 74 Table 3-3 Variations of Ruling Parties in Taiwan and South Korea………………...…… 77 Table 3-4 Values on Key Variables of Interest, Taiwan and South Korea………..……... 82 Table 3-5 Variations of Ruling Parties in Indonesia and The Philippines………….…… 87 Table 3-6 Values on Key Variables of Interest, Indonesia and The Philippines….…….. 90 Table 4-1 Summary of Variables for Effects of Institutionalization…………………….. 108 Table 4-2 Party Institutionalization and Political Survival of Leaders…………………. 109 Table 4-3 Regression Analysis: Party Institutionalization and Political Survival of Leaders………………………………………………………………………………………... 111 Table 4-4 Party Institutionalization and Party’s Outcome…………...………………….. 113 Table 4-5 Party Institutionalization and Performances of Leaders………………..……. 117 v LIST OF FIGURES Figure 2-1 Analytical Framework of Party Institutionalization…………………….…... 29 Figure 2-2 Analytical Framework of Party Institutionalization in Autocracies……….. 30 Figure 3-1 Threat Landscape and Effects to Party Institutionalization……….………… 57 Figure 3-2 Religion and Violent Collective Action………………………………….…….. 62 Figure 3-3 Predicted Margins of Territorial Claim Enemy……………………………….. 72 Figure 3-4 Predicted Margins of Territorial Claim Salience……………………………… 72 Figure 3-5 Predicted Margins of External/Internal Threats Ratio…………………..…… 73 Figure 3-6 Comparison of Number of Military Personnel……………………………….. 84 Figure 4-1 Causal Chain of Party Organization and Outcome………………………….. 102 Figure 4-2 Party’s Years in Power and Effects of Institutionalization……………..……. 114 Figure 4-3 Parties’ Years in Power After 1st Leader……………………………………... 114 vi 1-1 The Puzzle CHAPTER 1 INTRODUCTION What explains the different paths of ruling party development in autocracies? This study aims to answer the following questions: Why do some authoritarian regimes institutionalize their ruling parties and penetrate society through party organizations? Why do some dictators build up local party branches as linkage to the masses? How these local party organizations contribute to the political survival of dictatorships? The variation of ruling strategies in autocracies, especially concerning the decision whether to use the party as a primary ruling tool at the grassroot level, is the core subject of this study. The majority of authoritarian regimes rely on a political party to govern (Magaloni and Kricheli 2010), but there is a considerable variation regarding organizational developments of such ruling parties. Some authoritarian ruling parties became capable ruling machines for the dictator, as the party can penetrate to the remotest village, establishing party branches and maintaining an active role in people’s daily lives. For example, Mexico’s Institutional Revolutionary Party (PRI), Paraguay’s Colorado Party, Taiwan’s Kuomingtang (Chinese Nationalist Party, KMT), and Liberia’s True Whig Party were long-lived and extensive ruling machines. However, some parties cannot reach beyond the central city, or in some extreme cases, its own office (Geddes 2006; Hicken 2009). Some ruling parties do not have grassroots organizations, such as South Korean ruling parties, or serve merely as rubber stamps for their strongmen, such as in the case of Hung Sen and the Cambodian People’s Party. Furthermore, the ruling parties vary not only in terms of organizational developments, but also on the consequences, especially 1 length of political survival. Some parties rule for a long time, even after democratization, while some parties are short-lived. Studies show that political parties contribute to prolong the political survival of dictators. Authoritarian regimes with political parties survive longer than those without them (Geddes 2003; Magaloni 2006; Svolik 2012). In particular, parties help dictators survive under less favorable circumstances or with stronger challengers, as the dictators can use the party as a an essential ruling tool for relationships with both the elites and the masses. For the elites, authoritarian ruling party fortifies the power of the regime by facilitating elite’s co-optation (Gandhi and Przeworski 2007), providing career advancement for elites (Svolik 2012; Reuter and Turovsky 2014), assuring long-term gains and preventing elite defection (Brownlee 2007; Slater 2010). Besides distributing various rents and co-opting potential rivals by providing political positions (Gandhi 2008), one key function of political party is the credible commitment (Brownlee 2007; Gehlbach and Keefer 2008; Boix and Svolik 2013). Dictators use formal or informal institutions within the ruling party in order to make credible commitment on power-sharing deals with other elites, especially potential opponents. The party facilitates collective action of elites by providing information about dictators’ expropriation and intentions of other elites. Consequently, each side knows that deviation from the joint power arrangement will be punished, which keeps coherence of the elite alliances. For the masses, the party helps to maintain the patronage network, distributing material benefits and the privileges to the most loyal members (Diaz-Cayeros, Magaloni, and Weingast 2003; Magaloni 2006). Also, leaders utilize party organizations and staff to acquire information regarding society, especially the strength of the opposition that threatens political survival. As a result, the ruler can adopt better strategies of mobilizing supports. At the same time, the extensive party networks and its strength in electoral 2 competitions send signals to the potential oppositions that co-optation is a better option. When the party’s organizational networks are dense and when the party monopolizes valuable resources, it has the ability to “trap” the citizens into the system (Magaloni 2006). The party increases the potential cost for citizens of ousting the leader, and maintains mobilization capability against a future coup leader (Geddes 2006, 2008). Some long- lasting authoritarian parties keep winning elections at various levels even after democratization, such as the KMT and the PRI. Although political party is an essential institution adopted in authoritarian regimes, we still lack systematic understanding of the variation of party developments. In particular, there are several missing elements in the existing literature which I hope to address. First, the literature ignores an essential characteristic of the parties— institutionalization. In general, party institutionalization can be broadly defined as a process in which individual political parties experience value infusion and organizational stability (Huntington 1968; Panebianco 1988; Levitzki 1998; Basedau and Stroh 2008). This process of organizational development involves dictator’s commitments regarding specific formal rules and related power-sharing arrangements, which leads to more patterned and routinized institutional practices (Boix and Svolik 2013; Meng 2016). Some characteristics of institutionalization include establishing permanent offices and branches, hiring staffs, and holding regular party activities based on formal or informal rules. Studies on individual party institutionalization is lacking, while there is no satisfactory explanation on the causes and effects of party institutionalization in autocracies. The second missing element in the existing literature is explanations as to the reasons why some dictators establish linkage to the mass. In theory, the establishment of a ruling party serves not only a mechanism to co-opt elites, but also an incentive structure for controlling the mass (e.g., Gandhi 2008; Brownlee 2007; Cheibub, Gandhi, and 3 Vreeland 2010; Svolik 2012). However, the masses play only a secondary role, since the literature shows that the main threat for the incumbent comes from the elite. According to Svolik (2012), of all 316 dictators between 1946 to 2008 who held office for at least one day and lost power by non-constitutional means, only 11 percent were removed by popular uprising while a supermajority of them were removed by regime insiders—68% exit by coup d’état and 7% by assassination. Furthermore, less than 5 percent of all ousters in the 1960s and 1970s were related to bottom-up revolts (Kendall-Taylor and Frantz 2014a:39). This number shows that people are seldom capable of overturning dictators, as compared to regime insiders. Yet, some leaders spend considerable efforts and material costs on linking to the masses, when these strategies are seemingly unnecessary. For example, the KMT in Taiwan had party branches in every township, and the Golkar (Partai Golongan Karya, or Party of the Service Society) in Indonesia reached most of the remote islands (Ufen 2007, 2009; Tomsa 2008; Cheng and Hsu 2015). In addition to the low capability of people to overturn the regime, we must remember that for strongmen who control the military and the coercive institutions (Greitens 2016), such as Chiang Kai- shek and Suharto, these efforts in developing party organizations and activities seem irrational. To date, few studies show how the masses matter to the dictator and why an autocracy builds up connections to them. Exploring the reasons behind party institutionalization becomes even more important because one has been witnessing the changing patterns of dictators’ departure. Since the end of Cold War, more and more dictators leave power due to regime outsiders, including mass-based revolts and civil war. Popular-revolt becomes one of the major threats to autocratic leaders, as the number of it surged from less than 5 percent in the 1960s and 1970s to more than doubled after the end of Cold War, and even one quarter 4 in the 2010s (Kendall-Taylor and Frantz 2014). Therefore, it is essential to study how dictators respond to the threats from the regime outsiders, especially the masses. On the other hand, it is also a puzzle considering the positive effects of party organizations on political survival. Why do some dictators not using the party as a ruling tool, given that political parties significantly help dictators survive under less favorable circumstances? As discussed above, the party facilitates co-optation between elites and for the opposition groups, prevents elites defection, and achieves power balance between different groups of elites and subordinates (Geddes, Wright, and Frantz 2018; Boix and Svolik 2013). Despite these positive effects, some dictators do not develop their support party, while some authoritarian parties had become effective ruling machines. If the party does improve the ruling capacity and longevity of the political survival of the leader, there must be reasons why some dictators do not rely on party to rule. The third missing part in literature rests on the effects of ruling parties’ local organizations, especially leaders’ political survival. Past literatures often focus on an institutions’ effects on the elite-level interactions, such as the credible power sharing mechanisms for the elites, constraints on the leader, prevention of elites defections, and facilitating cooperation. Also, past studies often delve into the role of nominally democratic elections and the legislature. However, there seems to be no systematic examination on how local party organizations lead to leaders’ political survival. Party institutionalization in post-war East Asian authoritarian regimes is particularly a puzzle for comparative studies. Such parties were often robust in their rule, but some of them put extra efforts to develop party organizations at the local level. As American allies and part of the global anti-communist alliance, most of the dictators in East Asia not only controlled the coercive institutions but also build up relatively autonomous bureaucracies, achieving stable economic growth. On the one hand, the 5 United States established security alliances in this region as part of the Truman Doctrine of creating a global anti-communist alliance (Bush 2016) through signing bilateral treaties, such as with the Philippines (1951), Thailand (1951), South Korea (1953), and Taiwan (1954). The U.S. also formed collective security mechanisms such as the Southeast Asia Treaty Organization (SEATO). Furthermore, the U.S. provided substantial material support to countries without signing defense treaties, such as with Malaysia and Indonesia, contributing to stabilizing the regimes (IBP 2009; Head 1999). Such foreign aid gave dictators abundant resources to realize their policies and build up effective coercive institutions to rule. Therefore, additional strategies such as building up party branches at the local level seems to be irrelevant for political survival. Furthermore, American economic aid and the opening of the market helped many countries become “capitalist developmental states” (Johnson 1982; Öniş 1991; Woo- Cumings 1999). This term refers to a specific kind of capitalist economy in which all business and industry are led by a pilot agency and an efficient bureaucracy. Adopting this model of economic development, South Korea, Taiwan, Singapore, and Hong Kong, became the first-tier newly industrialized economies (NIEs) that achieved rapid economic growth through the export to and technological transfer from advanced economies such as Japan and the U.S. The emerging Southeast Asian economies of Indonesia, Malaysia and Thailand were also considered to be developmental states, although they differed in some characteristics compared with Northeast Asian counterparts (Doner, Ritchie, and Slater 2005; Vu 2007). In general, these countries have a relatively high level of state capacity, compared to other newly independent states in other regions. Taking into account East Asian dictators’ powerful foreign ally—the U.S., the strength of coercive institutions, the state capacity, and the power advantage of the dictators vis-à-vis other actors, it seems like efforts to link to society via parties are 6 unnecessary. However, one can observe that autocracies in East Asia choose different strategies on developing the ruling party, as the party plays only minor roles in some authoritarian regimes. For example, the ruling parties in the Philippines and South Korea were weakly organized and short-lived. The dictator’s party in the Philippines could not even make orders go beyond the presidential office (Hicken 2009, 2015; Lacaba 1995). On the contrary, the KMT in Taiwan, the Golkar in Indonesia, and Malaysia’s National Front (Barisan Nasional, BN) were highly institutionalized parties in the authoritarian era (Hicken and Kuhonta 2015). Besides the organizational development, the political survival of ruling parties vary significantly as well. For example, Malaysia’s BN goes through several leadership turnovers and keeps power from 1957, after Malaysia became independent, to 2018. It has long been categorized as a stable competitive authoritarianism (Levitsky and Way 2010). Taiwan’s KMT also dominates for a long time, even after democratization in the 1990s. Although it lost power to the then opposition DPP in 2000, it came back in 2008. The first time of political turnover in the legislature occurred in 2016, more than two decades after the introduce of open elections in 1992, and more than six decades since the beginning of the dictatorship. KMT still keeps dominance at the local level.1 On the contrary, some ruling parties are short-lived, frequently reorganized, or become politically irrelevant soon after losing power. For example, Filipino President Ferdinand Marcos switched from Liberal Party (Partido Liberal ng Pilipinas) to Nacionalista Party (Partido Nacionalista) 1 For example, in the 2018 Local Elections, KMT won 15 out of 22 seats of cities/counties. Also, it has never lost the majority in the number of local legislatures. See Fang-Yu Chen and Kevin Wen Kuo, “Four key takeaways from Taiwan’s recent election surprises.” Washington Post, Dec 17, 2018. https://www.washingtonpost.com/news/monkey-cage/wp/2018/12/17/four-key-takeaways-from- taiwans-recent-election-surprises. Accessed March 20, 2020; Richard Bush, “Taiwan’s local elections, explained.” Brookings Institute. Dec 5, 2018. https://www.brookings.edu/blog/order-from- chaos/2018/12/05/taiwans-local-elections-explained. Accessed March 20, 2020. 7 before joining the race for presidency in 1965. In 1978, he reorganized the ruling party and establish the New Society Movement (Kilusang Bagong Lipunan), uniting several supporting parties. After the exile of Marcos in 1986, the ruling party soon lost power and the coalition split (Hicken 2009). In sum, this study tries to solve two main puzzles, the causes of institutionalization puzzle and the consequence of institutionalization puzzle. First, I explore the key factor that leads to variations of party institutionalization by using a cross-national dataset and comparative case studies of two East Asian countries, South Korea and Taiwan. In particular, I will illustrate how the initial condition of threats when a dictator comes to power determines the ruling strategies he will choose. Second, following the causes of party the effects of party institutionalization on political survival of each leader. this dissertation examines institutionalization, 1-2 What Explains Causes and Effects of Party Institutionalization 1-2-1 Conventional Wisdom: Resource availability theory and adversity-driven theory Party institutionalization has two major characteristics: value infusion and organizational stability (Huntington 1968; Panebianco 1988; Levitzki 1998). Value infusion means that parties go beyond serving as merely an instrument of political participation, while individuals have the goal of perpetuating the organization and attach their loyalty to the party as a whole. Stability refers to a process of routinization, in which formal and informal rules and patterns of behaviors become entrenched and regularized. In short, a party becomes institutionalized when it establishes routinized and patterned 8 activities, central and local organizations, formal or informal rules, and linkage with the elites and the masses.2 To solve the first puzzle on the causes of institutionalization, some researchers have tried to explain variations of party institutionalization, but such studies are limited in number and have little explanatory power. Extant studies provide two main explanations that party institutionalization depends on either resource availability or the levels of adversity the leaders face. One explains a party’s development by the capability of the leader and another on the motivation. These two explanations do explain some cases of party institutionalization, but the causal mechanisms are contradictory to each other. Thus, it may be insufficient to generalize the theories in all cases. First, scholars argue that the leader chooses ruling strategies according to how much resource he has, or more precisely, the way money flows (Panebianco 1988; Riedl 2014). Studies show that material resource availability is necessary for the leader to develop party organizations, as maintaining party branches and regular activities do require funding and resources. “Money is indispensable to the life and functioning of every organization” (Panebianco 1988:35). The leader must invest a considerable amount of resources to build local branches and hire professional staff for the maintenance of party activities. Also, the ruler must distribute material benefits to other actors via the party, in order to keep loyalties of the local elites and the masses. Because financing is the foundation of developing the organization, the financier of the party exercises “a certain amount of power” over the organization. Consequently, the way money flows and the coalition of party financiers have significant impact on the strategic decisions of the leader. Riedl (2014) argues that regimes that adopt an “incorporation” strategy, in which leaders 2 Details on the conceptualization and measurement of party institutionalization will be in Chapter 2. 9 likely to have form alliances with local notables by maintaining clientelist networks and party organizations, are more institutionalized party systems after democratization. In her theory, the leaders need to be strong and resourceful to keep loyalties of the local elites, making the parties coherent. In contrast, leaders who “substitute” existing local elites when the regimes fail to maintain elites’ support will soon lose control when the regimes face a crisis or in progress of democratization. In short, the leader needs to be resourceful to realize institutionalization. Otherwise, even routine works of the party cannot be maintained. Many long-ruled authoritarian parties are indeed resourceful. For example, the KMT in Taiwan is allegedly one of the wealthiest parties in the world,3 as it accumulated a considerable amount of real estate and business empire, known as the “party assets” (Fell 2005). KMT utilizes state resources to build up “service stations” in every township (Gong 1998). The stations provide services such as free consulting and distribute material goods to people. Also, the KMT was able to use state resources, especially subsidies to local government, to broker deals with local faction leaders and thus control the time-share arrangement for nomination in the local elections and the positions in the party (Cheng and Hsu 2015). These actions helped the KMT become a long-lasting, hierarchical, and routinized ruling party. Another example is the Golkar (Golongan Karya, Functional Groups) in Indonesia. The dictator, Suharto, mobilized state resources generated by the oil boom economy and natural resources (Robison 1993). He nurtured the loyalties of prominent officials by distributing contracts in the petroleum industry, easy lines of credit, and concessions of the lumbering industry (Aspinall 2005). The Golkar became an 3 Michael Bristow. “Wealth probe for 'world's richest' party.” BBC News Oct 26, 2001. http://news.bbc.co.uk/2/hi/asia-pacific/1621048.stm. Accessed Feb 28 2020. 10 effective election machine and built up local branches in the remoted areas all over the country. Contrary to the resource availability theory, scholars depict an opposite causal mechanism, which I call the adversity-driven theory. This explanation mainly considers the leader’s incentives for institutionalization, arguing that a party facing resource scarcity and challenging environments tends to institutionalize. Panebianco (1988:113- 114) illustrates that the more public resources available, the less the leader’s motivation to institutionalize the party. If the quantity of public resource and fiscal power are high, the leader tend to distribute resources via the governmental bureaucracy because he can quickly “colonize” the bureaucratic elites. Also, when the resource is abundant, there will be more alternatives to party organizations, such as the clientelistic networks and the government branches. Besides resource availability, the political system’s degree of competitiveness also matters (Panebianco 1988:114-115). An unchallenged ruling party does not have incentives to develop complex organizations, while the opposition party is often highly motivated in institutionalization. Subsequent studies find that facing difficulties or environmental hostility forces the leader to institutionalize the party (Hellmann 2013; Tavits 2013). If a party enjoys advantages against potential challengers in terms of material resources, the leader tends to maintain the traditional clientelistic system by distributing benefits instead of developing party organizations or rules. Hellmann (2013) compares parties in Taiwan, Japan, and South Korea and argues that resource scarcity prompts party leaders to institutionalize their party and rely on programmatic strategies. He contends that easy access to public resources impedes party lead to the institutionalization, while difficulties of access to state resources institutionalization of formal rules, such as with the Democratic Progressive Party (DPP) 11 in Taiwan. Those without access to resources are forced to adopt a programmatic strategy for voter mobilization and thus are more likely to be formally institutionalized than parties that do have access to public resources, such as the Liberal Democratic Party of Japan, and the KMT of Taiwan. Furthermore, he contends that the difference of accessing public resource originates from the choice of monetary and fiscal policies. Taiwan focused on fiscal policies for economic development, expanding the budget of the government, while South Korea relied more on monetary policies as capital flows from the financial market. Therefore, Taiwan’s leaders have higher levels of resource accessibility to the party. Tavits (2013) also argues that environmental hostility led to pressures to post- communist party leaders, as they have to develop organizations and programmatic platforms to prolong political survival. In the cases of four central European countries— The Czech Republic, Estonia, Hungary, and Poland in 1990-2010, the former regime parties associated with communism often faced more challenges from the media and the public about their political legitimacy. Late-comers, as newly established parties, frequently need extra efforts to develop party organizations in order to attract more votes. The leaders of these parties have considerable incentives to institutionalize the party. Furthermore, scholars examine the elite-leader relationship as a micro-level dimension of environmental hostility. Meng (2016) proposes a power balance theory, arguing that, when the elites are powerful enough vis-à-vis the leader, the dictator will share powers by creating formal positions in the cabinet to institutionalize the ruling party. By doing so, the leader avoids the possibility of a coup and ensures his political survival. On the contrary, active autocrats with a low probability of being deposed from power—when opposition strength is low—are less likely to pursue strategies of ruling through institutionalized parties, because they do not have to share power with others. 12 This theory is constructed by a formal model as well as cross-national evidence. Although the power balance theory mainly concerns relationship between leader and elites at the national level, it could apply to the power-balance at the sub-national units. When local elites are influential, the leader must develop party organizations and share his power by assigning positions to local elites. That is, the leader chooses to institutionalize the party when he is weak vis-à-vis elites with high local popularity and political power. 1-2-2 Limits of the Extant Theories for the Causes Puzzle The two main explanations on party institutionalization have opposite causal mechanisms. The first theory argues that resource availability is a key factor, while the other theory illustrates that facing difficulties such as lack of material resources or environmental hostility forces the leader to institutionalize the party. Both seem to fall short on explaining party institutionalization. First, it is true that the material capability is necessary for the leader to institutionalize the party, maintaining party activities and organizational operations. However, this material capacity is not necessarily a motivation for developing parties. On the one hand, once the leader acquires resources needed, he also becomes more capable of strengthening coercive institutions and removing potential threats to political survival. In this situation, there will be fewer incentives for the leader to rely on the party or other co-optation means to rule. On the other hand, even if the ruler uses the resources in co-optation approaches, he may rely on personal networks, clientelism, targeted spending, and other approaches of distributing benefits to citizens, not necessarily via party organizations. Thus, having material resources is not sufficient to explain strategies for institutionalizing the party. 13 Take some East Asian countries, for example. A cluster of authoritarian countries has similar levels of resource availability, while it is not a guarantee of ruling party institutionalization. Due to the economic aid and the opening of the market by the U.S., many countries follow the capitalist developmental states model, in which a strong bureaucracy led economic developments through shared technologies and markets from advanced economies (Johnson 1982; Öniş 1991; Woo-Cumings 1999). South Korea and Taiwan are two prominent examples that achieved rapid economic growth in the 1960s and 1970s, while the emerging Southeast Asian economies of Indonesia, Malaysia, and Thailand were also considered to be developmental states. These governments control the state resources with effective bureaucracies, and some countries even have abundant natural resources. However, the ruling party became highly institutionalized in Taiwan, but not in South Korea. Also, in the natural-resource-endowed Philippines, the dictator did not develop the party’s organizations, while similarly resource-endowed Indonesia has extensive local party branches.4 One may conclude that resource availability is a necessary but not sufficient condition of party institutionalization. Higher levels of resource abundance can lead to higher levels of capability to develop party organizations, but capability does not guarantee the motivations of doing so. On the other hand, the adversity-driven theory of party institutionalization is also insufficient. As chapter two shows, more than half of dictators (54.6%) have extended party organizations at the local level. Although Panebianco’s theory that ruling parties facing fewer difficulties tend to be less institutionalized is supported by some comparative studies such as Hellmann (2013), it cannot explain why so many dictators 4 For comparisons of these parties and the party systems in these countries, see, for example, Hicken and Kuhonta (2015). 14 institutionalized the ruling party. In autocracies, the ruling party, by definition, has easy access to the state’s resources because it controls the executive branch and may have the power to bypass the monitors of the legislature. We might take Hellmann (2013)’s argument for example. First, in his view, the KMT of Taiwan is considered as not formally institutionalized and it lacks easy access to resources, but this theory cannot explain why South Korea had an even weaker ruling party under Park Chun-hee, who also had control over state resources. Second, there is no explanation of the causal mechanism between fiscal/monetary policies and resource accessibility. It is not clear why adopting monetary policies for economic development hinders the dictator in mobilizing public resources. Third, even if we assume that his theory on economic strategy and its effects are correct, there are contradictory predictions in the theory. The DPP of Taiwan is seen here as more institutionalized than KMT because of its difficulty in acquiring resources. KMT should be less likely to institutionalize than Korean examples because Korea, in the study, is having trouble mobilizing public resources due to the emphasis on monetary policies. However, the KMT is, in fact, more institutionalized than its Korean counterparts (Cheng and Hsu 2015). Thus, the adversity- driven theory is not sufficient to explain individual party institutionalization. Besides material resource considerations, the power-balance factor also falls short in explaining cases in East Asia. When we compare dictators with similar backgrounds, e.g., military strongmen, one finds completely different paths among authoritarian regimes. For example, Korean authoritarian regimes, all led by military strongmen, including Syngman Rhee, Park Chun-hee, and Chun Doo-hwan, did not institutionalize the ruling party. Compared to Indonesian strongman Suharto, these leaders rose from similar structures, seizing power mainly by coup d'état or elite struggle. As such, they should have faced similar pressures from other elites and tended to institutionalize the party to 15 neutralize such threats. Nevertheless, while Suharto institutionalized the party, Korean leaders did not. According to my data (see Table 2-4), 26 out of 89 (29.2%) leaders in the “military” dictatorship institutionalize the party, and 46 out of 95 (48.4%) leaders in the “personal” regimes do so. It may be challenging to find a majority of “weak leaders” vis- à-vis elites in these cases. Also, the Philippines and Taiwan are similar cases with different strategies at the party. In both countries, the local elites in rural areas are not fully controlled by the leaders (see, e.g., Looney 2020; Sidel 1999). This should be an essential motivation for the leaders to develop party organizations at the local level if the power-balance theory were correct. I did observe that the two autocracies both ranked on the top list of the ratio of authoritarian internal security personnel to population (Greitens 2016), which means that they both adopted a high level of coercive apparatus to maintain power. However, Taiwan’s dictators built up party organizations at the local level while Marcos in the Philippines did not. That is, the power-balance may be insufficient to explain the diversity of local party activities. Furthermore, in Meng’s study (2016), institutionalization is mainly measured by power-sharing—whether the dictator creates a position of vice president or a prime minister—and leadership turnover. These measures may be related to the internal structural dimension of party institutionalization—whether there exist the formal rules that constrain the leader’s power—but these cannot reflect the external structural dimension of the party in the society. Although the relative strength of the main actors may partially explain the motivation of power-sharing between leaders and elites, the theory of party role of institutionalization still lacks an explanation on the variation of the party’s developments at the local level. is absent. Therefore, 16 the mass the power-balance Taken together, the existing explanations of party institutionalizations are not satisfactory. The resource availability theory and adversity-driven theory have opposite predictions. The first school tells us that one should have observed that authoritarian regimes build up parties due to their ability to generate material benefits from such forms of rule, while the second predicts that the weakest regimes, who face difficulties in maintaining power, will try to rely on parties to gain more power. It seems like neither school of thought is sufficient to explain ruling authoritarian party institutionalization, while they do hold some cases correct. If the first school of literature, the resource capability argument, is correct, we should have observed institutionalized ruling parties in East Asian developmental states because the states fully control public resources. This theory cannot explain the absence of competent ruling parties in countries such as Korea. Were the other school of literature correct, we should have seen that East Asian dictators have little incentives to pay extra costs in connecting to the people because they have resources, with aid from the U.S., to strengthen the coercive apparatus. Therefore, both theories are insufficient to explain East Asian cases, let alone dictators all over the world. To date, there is no scholarly consensus on the origin of party institutionalization. 1-2-3 Literature on the Consequence Puzzle and its Limits According to previous studies, the establishment of some nominally democratic institutions contributes to longer political survival of autocracies. In particular, political parties should be able to enhance the ruling performances, because it provides credible commitments to elites and the masses, facilitates direct interactions between party cadres and people, maintains the patronage systems, and produces policies with better qualities (Magaloni 2006, 2008; Bizzarro, Hicken, and Self 2017). The literature on authoritarian resilience also illustrates that ruling party institutionalization at the local level is a key 17 factor for longer political survival, because it increases “inputs” to the regime and helps the dictatorship gain legitimacy (Nathan 2003). As a result, party institutionalization enhances people’s political efficacy and local governments’ responsiveness However, empirical tests about the effects of party institutionalization on political survival is still in scarce. The literature has limitations on the conceptualization, measurement, and empirical results. First, some earlier studies of authoritarian politics use regime type or the number of parties as a proxy, and they lack the proper conceptualization of party institutionalization. The assumption that ruling parties have similar effects in authoritarian regimes may be problematic in many empirical studies whose central causal mechanism is based on regime types. For example, Weeks (2012) find that party-based regimes are less likely to initiate armed conflicts against other party- based regimes because regime insiders, who also come from parties, will hold leaders accountable. This causal mechanism needs further examination, because some parties are not autonomous enough to constrain the leader. Another example is from Wright and Escribà-Folch (2012). They find that political parties increase the likelihood of democratization because it will protect the interests of the elites after democratization and thus increase their support for the transition. The critical variable in these studies is the number of parties, or more specifically, the existence of the ruling party. However, it is problematic to assume that the “one” party has the same weight in different regimes. If the party were just a rubber stamp for the dictator, lacking necessary organizational features and being short-lived, it could not become a credible mechanism for the better outcome, especially political survival. Second, for the main explanatory factor of political survival, many studies do not differentiate party strength and party institutionalization. The measurements could also be problematic. For example, the latest study by Meng (2019) may be the best example to 18 date on how to examine the outcomes of the development of individual ruling parties in authoritarian regimes. She finds that party strength—measured by whether a party goes through a leadership transition and political survival after the departure of the first leader—is associated with better regime outcomes. However, as discussed in the second chapter, party strength, as defined by longevity of political survival or popularity in the elections, is itself a consequence of party’s development. It is possible that a strong party may be made up of coercive ruling tools, such as suppressing all of the opposition. Thus, party strength (political survival over leadership turnover) may not be the best approach to explain other outcomes. In particular, the indicator cannot explain variations of political survival because it is defined by political survival. Another shortage of measurement is from the coding methods of cross-national dataset. For example, the V-Dem data of party institutionalization (Bizzarro, Hicken, and Self 2017) covers parties worldwide, but it suffers a major problem: the coding of party institutionalization is according to activities of all parties in the party system, not for the individual ruling party. Thus, it may be misleading to claim that ruling party institutionalization makes a difference on a specific outcome, as the results are not necessarily caused by a specific party. Third, empirically, cross-national data on individual party institutionalization is lacking. Data about the individual party and its organizations are scarce because parties are relatively secretive and closed organizations (Tavits 2013:13). Even in advanced democracies, there is no easy access to data on some basic characteristics of a party, such as the numbers of party members. Data availability issue is even more challenging for studies of authoritarian parties because of the lack of transparency. To date, only a few studies measure party institutionalization globally (e.g., Yardimci-Geyikçi 2013; Bolleyer and Ruth 2017), but the number of cases is still limited and they do not focus on 19 authoritarian regimes. Meng (2019) is the only one that examine authoritarian parties, but the cases only include party-based regimes. As mentioned above, her study cannot explain outcomes of political survival because the main explanatory factor is defined by political turnover. So far, there is no satisfactory test on how party institutionalization lead to variations of political survival in dictatorships. 1-3 Main Argument: A Threat-based Theory of Party Institutionalization The primary concern of this study is the dictator and his party’s structural linkage to the society. Whether a dictator establishes local branches of the ruling party and maintains regular activities have essential implications to the survival of dictatorships. First, from the party’s viewpoint, professional staffs and offices at the local level have more opportunities to reach out to citizens, and they are like radars that gather important information for the ruler. At the local level, party organization helps to collect public opinion and other information regarding society, which mitigates the Dictator’s Dilemma—rule by repression leads to lack of information on the level of support among the masses (Wintrobe 1998). When the party interacts with people’s daily lives and links with society, the ruler will be able to form better strategies regarding resource distribution and mobilization, according to information gathered directly from the society. Second, from people’s viewpoint, the presence of party organization at the local level provide a definite object for people to attach their loyalty with, especially through regular party activities and face-to-face contact. They can participate in the regular party activities as an alternatives to political participation without free and regular elections. As a result, when a party has grassroot organizations, it acquires higher levels of legitimacy and hence prolong political survival (Nathan 2003). Chapter 4 will examine 20 the relationship between party institutionalization and political survivals of the dictatorship. Empirical evidence support the claim that local party organizations help to prolong political survival. On the other hand, not all dictators establish party organizations and maintain regular activities at the local level. Under what conditions do autocracies establish local branches? For the causes of institutionalization puzzle, this study explains the variations of authoritarian ruling party institutionalization by a threat-based theory (Greitens 2016). I argue that the initial condition when a dictator comes to power determines the ruling strategies he will choose. The logic is similar to selecting strategies of developing coercive institutions in response to the main threats. The ruler needs to establish party organizations to control the masses when he subjectively thinks the threat to regime survival mainly comes from internal threats at the local level. If there exists a foreign enemy with territorial conflicts, the leader must direct resources to the coercive institutions, especially the military, to defend against threats. There are various sources of threats to autocratic political survival. Externally, a dictator faces military threats from foreign enemies or international interventions. Internally, the elites threaten the regime survival by coup d’état, assassination, and other types of non-constitutional removal of the leader, while the masses threaten the regime by demonstrations, strikes, violent riots, guerrilla attacks, civil war, or revolution (Davenport 1995; Carey 2010; Slater 2010).5 Dictators adopt various strategies to respond to these threats. When there exist foreign rivalries with territorial threat, the dictator must strengthen defense capability to defend against the possibility of attack (Rasler and Thompson 2006). 5 Figure 3-1 illustrates the landscape of threats. 21 In international politics, it is always difficult to acquire accurate information about enemies’ intention and capabilities (Fearon 1995). The best way to defend one’s own country is through strengthening the military forces for deterrence (Waltz 1979). On the other hand, the dictators have a variety of tools to handle domestic threats, including surveillance, detention, other violent approaches, or the non-violent co-optation approach to counterbalance opposition to the regime. Therefore, when external enemy threatens the territory of a country, the leader must prioritize the defense against the threat. If there are no immediate threats from external enemies, the leader has resources to deploy when facing internal threats. When the threat to regime survival mainly comes from the masses, the leader will institutionalize the party at the local level. As mentioned above, the party organization helps the dictator to acquire essential information regarding the society and thus leads to making efficient ruling strategies. Also, from people’s viewpoint, the political participation also help to gain legitimacy of the regime, as people feel that they can have inputs to the decision-making process (Nathan 2003). However, when the threat to the dictatorship comes from elites, the effect on party institutionalization is mixed. The leader might adopt multiple methods, such as adjusting the institutional design of coercive apparatus (Greitens 2016), or building up party organizations to co-opt the elites. It is not necessary for the leader to rely on grassroot organizations to relieve threats from elites. Taken together, the research hypotheses of this study in chapter 3 is that, the dictators need to rely on political institutions and co-optation strategies to establish linkage at the local level when (1) there is no spatial rivalries from outside, and (2) the masses are perceived as threats. Alternative hypotheses from conventional wisdoms are (1) resource-abundant leaders tend to institutionalize the party; (2) resource scarcity 22 results in party institutionalization; (3) leaders with higher risks of coups tend to institutionalize the party. The research hypothesis in chapter 4 is that, party institutionalization at the local level helps dictators to prolong political survival. The study of party’s local branch is connected to the broader literature in several ways. First, party institutionalization is related to the concept of the dominant party. According to Geddes (2003:72), a dominant party regime is defined as the ruling party has “some influence over policy, control[s] most access to political power and government jobs, and ha[s] functioning local-level organizations.” Scholars further differentiate the party-based regime into closed authoritarian regime and competitive authoritarianism. Dominant party exists in electoral authoritarian regimes, as it is defined as a party that has the leading role in determining access to most political offices and state resources, and the ability to maintain in power (Reuter 2017; Greene 2010). Party institutionalization may be an essential process for a party to remain dominance of power. On the one hand, the party establishes local branches and thus more capable of reaching out to the society and maintain loyalty of the citizens. On the other hand, an institutionalized ruling party tends to have more influence on the policy making process (when it becomes autonomous), and controls the appointments of more political jobs at various levels. Thus, party institutionalization may contributes to party dominance. Chapter 2 will have a more detailed discussion on the relationship between the dominance of a party—party strength—and institutionalization. Second, party institutionalization is also a coup-proof strategy, as the party organizations and the new “civilian” political positions balance the power of the military elites. Geddes, Wright, and Frantz (2018) find that dictators create support parties as a part of strategies of marginalizing military elites from policy-making. In particular, when the dictator is relatively weak, such as junior officers from the military, he is more likely 23 to create a support party. They also find that dictators tend to build up support parties when he controls the internal security services, which means that parties are essential tools to mitigate potential threats from internal elites who are in charge of coercive institutions. Taken together, party is a process that party organizations become more autonomous and thus more capable coup-proof institutions. 1-4 Outline of the Study institutionalization institutionalization first. What to conceptualize party In order to explore the causes and effects of developments in party organizations, we have is party institutionalization and how to measure the variations of party organizations? Chapter 2 discusses the concept of party institutionalization, its multi-dimensionality, existing measurements, and how it differs from party strength. I propose to use party institutionalization at the local level as the main measurement. To test my hypotheses, I create a dataset including all of the dictators who had served for more than one year after the World War II, covering from 1950-2008. The unit of analysis is leader, as my concern is each dictator’s choice of ruling strategy. I exclude island-countries with land area smaller than 1000 square feet because in these relatively small countries it is difficult to differentiate party organization in the local level. Of all 657 dictators from 123 countries, 172 dictators served for less than one year, which makes the number of cases in the dataset as 485. Then I summarize the frequencies of local party presence, its relationship with regime types, and how it differs from the concept of party strength. Chapter 3 first examines the causes of authoritarian ruling party institutionalization by using the cross-national dataset. Results illustrate that the decision maintain networks of local party organizations is impeded by the existence of foreign enemies. When there are foreign rivalries that involve threats to the territory, the dictator is less likely to have 24 local party institutions because he must strengthen the military to defend against enemies from outside. On the contrary, when the threat is from internal actors, levels of religious fractionalization is positively associated with party institutionalization. In the second part of Chapter 3, I use comparative historical analysis to compare two country-pairs, Taiwan and South Korea, and Indonesia and The Philippines, to track the causal mechanisms (Mahoney 2001, 2007), and show the conditions that lead to the institutionalization of ruling parties. In particular, I utilize the Most Similar Systems Design to eliminate potential explanations on party institutionalization. The first country- pair is Taiwan and South Korea. These two countries have similar patterns of political and economic developments, acquire alliance and assistance from the U.S., become effective ruling governments, and the mass uprising was even more fierce in South Korea. According to the conventional wisdoms, levels of party institutionalization should be equivalent in the two countries, or even higher in South Korea due to more popular uprisings. The parties developed in an opposite ways. Taiwan’s KMT built extensive party branches in the local level districts and became highly institutionalized, but South Korean ruling parties remain short-lived. I argue that the fundamental difference that leads to the divergent development of the ruling parties is the source of threats to the regime. Although both facing external rivalries that threaten regime survival, the urgency of the enemy were not the same. The Taiwan Strait made the threats from Chinese Communist Party less urgent to the KMT regime compared to threats from North Korea to the South, where the capital Seoul is just 35 miles (about 56 kilometers) to the border Korean Demilitarized Zone. South Korea has to maintain a strong army for the defense purpose, while Taiwan has been decreasing the size of army (and coercive institutions) over time (Greitens 2016). 25 The second country-pair is Indonesia and the Philippines. Specifically, I compare the ruling conditions of Suharto and Marcos. In this country-pair, there is no foreign enemy with a territorial claim, so we can trace which type of domestic threat matters. I find that religious polarization is the most likely factor that explains the opposite developments of ruling parties’ organizations. Suharto builds up and maintains an extensive network of local party organizations, while Marcos’ ruling parties remain extremely weak in every dimension. Chapter 4 conducts a cross-national examination on the effects of party institutionalization on political survival. According to previous studies, party institutionalization should be able to enhance the ruling performances, because it provides credible commitments to elites and the masses, facilitates direct interactions between party cadres and people, and produces policies with better qualities (Bizzarro, Hicken, and Self 2017). In particular, party institutionalization at the local level helps the dictatorship gain legitimacy, because it increases “inputs” to the regime. As a result, party institutionalization enhances people’s political efficacy and local governments’ responsiveness (Nathan 2003). The evidence shows that dictators with institutionalized parties indeed perform better on political survival. In addition, I differentiate party strength (defined by leadership turnover after the first leader leaves power, Meng 2019) with party institutionalization. This study proves that party institutionalization, as defined by local level developments, is a better explanatory factor for regime outcomes. In Chapter 5, the concluding chapter, I summarize the findings in each chapter and discuss my contributions to the studies of authoritarian politics. I also laid out several directions for future studies. 26 CHAPTER 2 AUTHORITARIAN RULING PARTY INSTITUTIONALIZATION In order to explore the causes and effects of developments in party organizations, we have to solve another problem first: how to measure the variations of party organizations? The literature has discussed two main concepts, party institutionalization, and party strength, but their relationships are unclear. To date, cross-national studies are limited in number, and most of them do not focus on authoritarian politics. This chapter first offers a conceptualization of party institutionalization and illustrate the analytic framework of it. I elaborate the multi-dimensionality of party institutionalization and all of potential indicators for measurement. Then I discuss several approaches of measuring party institutionalization and their limits. In particular, I differentiate the concept with party strength. This study proposes to use party institutionalization at the local level as the main measurement. Then this chapter summarizes a cross-national data with 485 dictators after the WWII, 1946-2008, describing party level. I maintain that party institutionalization and party strength are not identical. The former can be seen as a necessary condition for the latter, but not a sufficient one. Consequently, it is not proper to interchange the usage of the two concepts. 2-1 The Concept of Party Institutionalization and its Multiple Dimensions 2-1-1 The Analytic Framework institutionalization at the local The classic definition of party institutionalization comes from Huntington’s seminal study, Political Order in Changing Societies. Huntington defines institutionalization by four criteria—adaptability, complexity, coherence, and autonomy (1968: 12-24), as it can also be used to define party institutionalization in general. First, a political organization is 27 considered as institutionalized when it can adjust according to environmental challenges. A political party shows higher levels of adaptability when it maintains functioning and performing various functions according to the need from the changing environment. Second, complexity refers to the multiplication of sub-units in terms of functional, hierarchical, and differentiating types. For a political party, a greater number and variety of sub-units lead to the increased capability of maintaining loyalties of the members. With diversified purposes, an institution has a better ability to adjust itself to different roles, and it also helps to connect to members in various ways. Third, coherence is the extent of unity among party elites, which includes the recognition of the legitimacy of the party among elites. The organizational unity requires certain degrees of consensus on the decision-making process and functional boundaries of the group. Fourth, the autonomy characteristic comes from values and expectations of interactions among different actors and the institution. If a party was over-dependent on the society or representing only specific groups or individuals, e.g., the dictator, it lacks autonomy. Taken together, these criteria are often closely linked while each reflects certain important and unique characteristic of institutionalization to an institution (Huntington 1968: 22). Based on Huntington’s work, subsequent studies broadly define party institutionalization as a process in which individual political parties experience value infusion and organizational stability (Panebianco 1988; Levitzki 1998; Basedau and Stroh 2008). Value infusion means that parties go beyond as merely an instrument of political participation, while individuals have the goal of perpetuating the organization and attach their loyalty to the party as a whole. Stability refers to a process of routinization, in which formal and informal rules and patterns of behaviors become entrenched and regularized. In short, a party becomes institutionalized when it has regular and patterned party 28 organizations and activities, and is capable of seizing stable supports from different actors. Internal Structural Systemness Attitudinal Value infusion External Decisional autonomy Reification Figure 2-1 Analytical Framework of Party Institutionalization Source: Randall and Svåsand (2002: 13) The analytical framework of party institutionalization contains multiple dimensions. Randall and Svåsand (2002) categorize the characteristics of party institutionalization along two dimensions, internal/external, as well as structural/attitudinal aspects, making it a 2x2 table as in Figure 2-1. The horizontal dimension is about the main actors. Internal aspects are the development within the party—for the elites and party members, and the external aspect refers to a party’s interaction with the society and other institutions. The vertical dimension is about the behaviors of the organization and the members. In their framework, the structural component of the internal party (the upper- left cell in Figure 2-1) is labeled as “systemness” character, which comes from Panebianco (1988:13)’s definition of “increasing scope, density, and regularity of the interactions that constitute the party as a structure.” One can understand it as the formal and informal rules and organizational activities. For the attitudinal aspect of the internal dimension (lower-left cell), Randall and Svåsand use the term “value-infusion,” stated by Broom and Selznick (1968) and Levitsky (1998). It means the degree of commitment to the party 29 possessed by party figures and supporters. In the external dimension, structural component (upper-right cell) refers to “decisional autonomy,” or freedom from interference carried out by particular individuals in determining its policies and strategies. Lastly, the external attitudinal component is the party’s positive public image and a stable basis of support, which is labeled as a reification of the party. 30 in Autocracies Attitudinal (Behavioral) Coherence Value-infusion Routinization Roots in society Structural (Organizational) Internal (Elites) External (Masses) Source: adapted from Basedau and Stroh (2008: 9). Figure 2-2 Analytical Framework of Party Institutionalization This study adopts this framework and make modifications to the criteria of the two dimensions based on the characteristics of authoritarian politics. Figure 2-2 shows the two-by-two framework of party institutionalization defined in this study. On the one hand, for the dimension of the main actor, I categorize the internal and external ones by the elites and masses instead of whether one holds party membership. In an authoritarian regime, the political elites are the members of the winning coalition—whose supports empower the leader with powers over other members of the society (Bueno de Mesquita et al. 2003). Party members who are not in the winning coalition can be classified as external actors, along with the general public and those who are not in the winning coalition. In autocracies, non-elite party members have limited power to determine the activities of the party. For example, party members in democracies often have the right to choose their party chairperson by direct votes or indirectly via voting party representatives, while in autocracies party members have almost no saying on the leaderships of the party. Unlike in democracies where party members participate in party primaries to compete for the nomination, authoritarian parties often control the power of nomination for elections or for the posts as one of the key mechanisms to keep elites’ loyalty (Gong 1998). On the other hand, for the dimension of structural-attitudinal party activities, I combine the definitions of Randall and Svåsand (2002) and Basedau and Stroh (2008) in order to be better linked to Huntington’s original definition of the four criteria. That is, party institutionalization consists of its organizational development (establishments of the organization, formal and informal rules) and behavioral dimension (the attitudes different actors toward it). Structurally, the presence of permanent party organizations, party branches, programmatic linkages, and distinct party platforms are essential factors to define party institutionalization (Bizzarro, Hicken, and Self 2017). Attitudinally, a party is successfully institutionalized when people show their consistent support in elections and in regular activities such as acquiring party memberships. 2-1-3 Multi-dimensionality of Party Institutionalization Although we have the analytic frameworks for party institutionalization, the operationalization is still a challenge. We currently have little understanding of the dynamics of the ruling party’s organizational development. To date, only a handful of cross-national studies have measured the level of individual party institutionalization directly (Basedau and Stroh 2008; Yardimci-Geyikçi 2013; Hellmann 2013; Arter and Kestilä-Kekkonen 2014; Meng 2016; Bolleyer and Ruth 2017), and the cases included in the studies are often limited. One of the major problems of measuring party institutional- 31 Table 2-1 Dimensions and Indicators of the Index of Party Institutionalization Dimension Routinization (complexity, adaptability, and autonomy): the party build up the rules and organizations Coherence: The party acts as a unified organization; the party tolerates a certain level of intraparty dissidence Roots in society (complexity and adaptability): The party has stable roots in society Indicator 1. Party age relative to independence 2. Party age relative to beginning of multiparty period 3. Regular party congresses 4. Number of alternations in party leadership 5. Rules for elite selection 6. Number of professional staffs 1. Coherence of parliamentary group (no defections or floor-crossing) 2. Moderate relations between intraparty groupings (no dysfunctional factionalism) 3. Tolerance vis-à-vis intraparty dissidence 1. Membership strength 2. Nationwide organizational presence, activities beyond election campaigns 3. Number of professional staffs in local branches 4. Changes in electoral support in last and second last elections (electoral volatility) 5. Links to civil society organizations 1. Years of political survival after leadership 2. Changes in popular support after alternation in Value-infusion (autonomy): the party has legitimacy and is relatively independent from individuals within and societal change party leadership 3. Decisional autonomy from individuals and groups 4. Popular appreciation of the particular party Source: Adapted from Basedau and Stroh (2008, 12). 32 categories concerning there are framework, four main -lization may come from its multi-dimensionality. According to the aforementioned analytical the structural/attitudinal and mass/elites dimensions of party institutionalization. Basedau and Stroh (2008) try to operationalize party institutionalization according to the four criteria provided by Huntington (1968). However, some characteristics of party institutionalization may link to more than one criterion of Huntington’s original definition. I recategorize potential indicators and link them to the analytical framework in Table 2-1. In the parenthesis are the most relevant criteria from Huntington’s framework. First, for the level of structural routinization and organizational developments, Huntington suggests examining the ability of an institution to keep working and to react to the changing environment, especially by developing subunits and bureaucratic functions. Thus, the complexity and adaptability of the party refer to the procedural aspect of whether the party has formal rules such as the procedure for personnel selection, regular party congresses, and the hierarchical subunits and professional staffs within the party. The process involves bureaucratization and professionalism, in which the party develops specific types of political professionals, hiring administrators who focus on organizations’ maintenance (Pianebiacon 1988).6 The most direct sign of party routinization is the duration of a party. Party age is an essential indicator to see whether the party can implement these rules and practices, achieving higher levels of complexity and adaptability. Scholars have illustrated that a central goal of ruling parties is to serve as inter-temporal commitment devices which manage the relationship between elites and prevent conflicts (Boix and Svolik 2013; 6 For a classification of different types of party professionals, see Panebianco (1988: 232-235). 33 Magaloni 2008). Furthermore, power turnovers are critical junctures to examine whether the abovementioned organizational practices have become institutionalized (Meng 2019). Unlike in democracies, where leaders comply with the term limits and regular elections, dictators often stay in power for a longer time as long as he does not leave by unconstitutional means. Thus, the alternations in party leadership show that the party has establish rules for leadership change and the organization is not predation by the dictator. The leadership change also refers to the element of autonomy. Scholars have pointed out that besides the passage of time, party institutionalization concerns how the party survives over changes in the initial conditions of power distributions and the “momentary leaders” (Janda 1980; Levitsky and Murillo 2009). In Panebianco’s words, institutionalization requires a process of ”routinization of charisma,” transferring the authority from the strongman to the party (1988:53). A party must go beyond the influence of specific leaders or persons so that it demonstrates the organizations are able to perpetuate its own existence and become routinized. Second, the coherence dimension measures the unity of the party. An institutionalized party develops rules to incorporate elites and prevent frequent splits or defections. Indicators include the party coherence in the parliament according to voting records, the existence and tolerance of intra-party factions, and the formal or informal systemic way the elites compete and interact with each other. A high level of party unity is a result of compliance that the party elites recognize the legitimacy and commitment of the “center” of institution and they become willing to follow the rules (Panebianco 1988). Coherence is also a consequence that elites expect to see the enduring development of the party organizations, as the party realigns various interests of the elites and broadens the time horizons for different actors. Only when the party progresses on institutionalization can the elites ascertain that they can invest in time and resources to 34 climb up the hierarchical power structure at the party. Panebianco calls this process as a “vertical convergence at the center” (1988:61). When the party organizations develop to a certain degree of institutionalization, the political survival of the party becomes aligned with the interests of the leader and the elites. Without institutionalized party organization and rules, many groups of elites control important power resources, and thus the recruitments of elites become centrifugal movement, attaching to specific groups but not on the party. Third, the parties’ roots in society also refer to the complexity and adaptability characteristics, especially the local subunits and their functions for the non-members. The indicators include the strength of membership, the nationalization of the party, especially the grassroots organizations and professional staffs in rural areas, and activities in the non-electoral periods such as parties’ links to civil society groups. If a party becomes institutionalized, establishing the connections and contacts in the society, it should have higher levels of electoral support, and the support will be stable. The popular support is often measured by vote share and volatility—changes in electoral support in the last and second most recent elections. These indicators examine whether the party has multiple functions and how they work at the local level. In short, an institutionalized party should be active without the constraints of geographic and temporal conditions. Fourth, the dimension of external value-infusion mainly measures the autonomy of a party. A party becomes autonomous when it acquires political trust from the supporters and the public, not only from a small group of elites. An autonomous party is independent of individual political strongmen, interest groups, or economic elites, while the party can make decisions at different levels by the party’s rules, not by specific actor’s will. Also, it measures whether a party has a popular appreciation and legitimacy. A major indicator of autonomy examines whether the party survives over leadership 35 turnovers (Basedau and Stroh 2008), especially after the first leader. However, the transfer of political power from the leader to the organizations is not an easy task. Meng (2019) finds that between 1946 to 2008, 57 percent of ruling parties in autocracies cannot hold power after the departure of the first leader. That is, most parties cannot realize the inter-temporal promises to the elites and the masses, as they are not successfully de- personalized. Only when a party remains stable support from both internal (the coherence dimension) and external actors after leadership changes can it maintains a certain level of institutionalization. 2-2 Measuring Party Institutionalization 2-2-1 Existing Measurements There are multiple dimensions of party institutionalization and a variety of indicators for the measurement. When operationalizing party institutionalization, one must consider two main points, the number of indicators and the nature of the measurement. Existing studies often utilize one indicator or some of the indices, and there are subjective evaluation and observational indicators for cross-national studies. There is no consensus on which indicator is the best. In fact, the most serious challenge may be the data availability issue, and there is a questionable practice that empirical studies tend to see party “strength” and institutionalization as synonyms. These two concepts do have some overlapping dimensions but they cannot be alternatives to each other. The first decision for operationalization is how many indicators should be included. One way to measure organizational development is to focus on a specific dimension. For example, Gibson et al. (1983) argue that electoral success and organizational complexity (bureaucratization) are essential indicators of organizational strength. In particular, electoral performances may be the most easily accessible data researchers can acquire. 36 Besides, Roberts and Wibbels (1999) utilize the age of party as a proxy of party institutionalization. 7 The rationale is that party organizations are like biological organisms which encourage elites and members to invest resources and participate in the competitive process in democracy (Harmel and Janda 1994; Harmel, Svasand, and Mjelde 2016). Some studies measure party activities in the legislature to see the level of cohesion. For example, Thames (2007) focuses on party discipline (unity) in Duma, Russian national assembly. These studies measure party development by one or two essential dimensions. Some scholars use several indicators at the same time. For example, Basedau and Stroh (2008) utilize 15 indicators for party institutionalization of 28 African parties. Yardimci-Geyikçi (2013) measures organizational development by membership strength, financial resource, and territorial comprehensiveness. She also measures roots in society by a party’s electoral volatility, people’s party identification, and trust in political parties. Tavits (2013) employs three main indicators of organizational strength to measure parties in four post-communist East European countries, including the staff size, local branch presence, and membership size. On the other hand, regarding the nature of the indicator, some studies start to use subjective evaluation, such as expert ratings and public opinion on political parties’ organizational characteristics. For example, the V-Dem project (Bizzarro, Hicken, and Self 2017) has a party institutionalization index that covers parties in 173 countries for 116 years (1900-2016). The country experts deliver general assessments on the development of parties in each country. Another example is from Bolleyer and Ruth (2017), in which they use survey data to evaluate party activities. That is, the cross-national subjective data 7 Roberts and Wibbels (1999)’s study measures party-system institutionalization by counting the average age of the parties in a system, which implies that the age of an individual party is equivalent to its level of institutionalization. 37 is now available to supplement objective observational data such as electoral performances and the size of members. the consequence-based approach is Besides new data of subjective evaluations, among all of the observational data, one common strategy is to utilize the indicators for the consequence of organizational development. The consequences are often conceptualized as “party strength,” such as electoral vote shares, public support, and longevity of political survival. Some studies see party institutionalization and party strength as synonyms (for example, Meng 2019). Admittedly, examining the consequences of party institutionalization on each party is a practical method of operationalization, given that data on the party organization such as membership size, staff size, or internal elites cohesion are difficult to acquire. The rationale behind intuitive: were a party institutionalized, party activities should be stable and routinized, and thus it receives support from different actors and becomes “strong.” For example, many autocracies hold elections, and an institutionalized party is more likely to achieve higher levels of stable electoral support—measured by vote share and volatility. When a party acquires a high level of vote share and maintain across several elections, it performs well on the “value- infusion” dimension of party institutionalization. Another consequence of party institutionalization is the party members’ coherence of staying in the party. For example, the V-Dem Project (Coppedge et al. 2016) records the rough percentage of the members of the national legislature who change parties or abandons their party in between elections. A weakly institutionalized party should observe members’ their party. While an institutionalized party should lower incentives for party members to switch, as the party provides career advancement for elites and assure long-term gains (see, e.g., Brownlee 2007; Slater 2010; Svolik 2012). frequent switching or abandoning 38 Perhaps the most observable indicators for each party are leadership turnover and the length of tenure in power. In a recent study, Meng (2019) illustrates a simple indicator to examine institutional strength: whether it can survive over the first leadership change. As mentioned in the previous sector, a party must develop its own decision making process maintain a certain degree of autonomy away from specific leader or social group before it is categorized as institutionalized. Meng’s measurement tests how long does a party stay in power after the exit of the first leader, which best captures whether a party becomes autonomous that its function does not rely on a specific leader. Parties fail to remain in power after the first leader mean that the rules and procedures are not stable enough to achieve organizational autonomy. Her study finds that party strength is much more challenging to achieve than researchers previously thought. Of all leadership changes in authoritarian regimes from 1946 to 2008, 57 percent of ruling parties cannot remain in power for more than one year after the first leader’s exit. Even under the situation of peaceful leadership transition, a majority of ruling party, 52 percent, cannot hold power for more than one year. However, in Meng’s study, party strength and party institutionalization are two identical concepts and she uses the term interchangeably. In sum, when it comes to the measurement, scholars have to determine how many indicators should be included, and what type of data is utilized. The most common strategy is to measure the consequence of party developments, especially party strength. 2-2-2 Problems of Existing Measurement and Differentiating Party Institutionalization and Strength Despite of the efforts on measuring party institutionalization, data availability is still a challenge. Data about the individual party and its organizations are scarce because parties are relatively secretive and closed organizations (Tavits 2013:13). Even in 39 advanced democracies, there is no easy access to data on some basic characteristics of a party, such as the numbers of party members. Researchers have to collect data through archival materials, fieldwork, and many other approaches to access empirical evidence on party activities. Furthermore, data availability issue is even more challenging for studies of authoritarian parties because of the lack of transparency. The existing studies on party organizations development often measure parties in democracies, and in most of studies, the number of cases are limited. Therefore, we must come up with new approaches to measure party activities in authoritarian regimes. Another major problem of existing measurement is the mixture of party institutionalization and party strength. I argue that these two concepts do have overlapping elements, but they are not synonym. It is true that in the multiple dimensions of party institutionalization (Table 2-1), party strength and party institutionalization are analog in some indicators. For example, party age and alternations in party leadership are essential indicators for the routinization of party activities. Also, stable electoral performance is a significant outcome of party institutionalization. These indicators, which mainly measure the consequence of party’s organizational development, can be used for measuring both party strength and party institutionalization. As discussed above, measuring the party strength as the consequence of organizational development is a practical strategy for operationalization because the consequences are more observable than other indicators. However, differentiating the two concepts is necessary. I argue that party strength cannot fully replace party institutionalization, and the approach of measuring consequence will be problematic when researchers need to examine how the party influences regime performances. Measuring party institutionalization by the consequences, especially party strength or durability, has limits in several ways. First, the indicator of political survival over 40 power transition cannot capture the variation of party development strategies in the terms of the first leaders. Only until the first leader leaves his position can one identifies whether the party is strong or not. Some political strongmen remain in power for a long time, and hence this approach misses the essential information in an extended range of period in many cases. For example, the Chinese Communist Party (CCP) is a strong party, being resilient after several leadership changes. After the establisher Mao Zedong died in 1976, the party remained in power and went through several turnovers, including Hua Guofeng, Hu Yaobang, and Zhao Ziyang. However, it was until the late years of Den Xiaoping that established the term limits, age limits, and rules for cadre selection. That is, the process of institutionalization of the party occurred in the 1980s, after Den took power (See e.g., Zheng 2014; Zeng 2016). Starting from the 1990s, especially after the Fourteenth Party Congress in 1992, these rules for cadre selection becomes routine for CCP, as the senior party heavyweights could retire peacefully under this system. If we use the party strength indicator, CCP in the Mao Zedong era (1949-76) is characterized as strong or institutionalized simply because it maintains in power after Mao. However, the indicator cannot capture the timing of institutionalization. For the first leaders, even when they did not establish party organizations, the party will be coded as institutionalized or strong as long as it maintains in power. Furthermore, it is possible that during the tenure of one leader, he changes ruling strategies, adopting or abolishing certain political institutions. For example, during China’s decade-long Cultural Revolution (1966-76), the capacity of the political institutions, including the party and the governments at various levels, were overwhelmed. The ad hoc consequence approach cannot reflect such fluctuations within the tenure of a leader. 41 Second, party strength comes from multiple factors, not only from party institutionalization. A ruling party may perform well because of a strong or charismatic leader, the usage of coercive apparatus, lack of real alternatives of political parties, unfair rules for competition, among other reasons. That is, a party which acquires people’s support does not necessarily come from its institutionalization, but the strength of the leader and the state. For example, in Indonesia, the mean vote share of the ruling Golkar in the legislative elections is 67.5%, from 1971 to 1997. The mean volatility (difference of vote share between two elections) is 2.31%, which means that the high level of vote shares remain extremely stable between elections and thus we can categorize Golkar as a powerful ruling party. However, the strength of the party may come from the fact that the dictator Suharto banned most of the parties, monopolized public resources for the clientelist networks, and utilized coercive apparatus to suppress popular uprisings and mobilization of minority groups (Vatikiotis 1998; Aspinall 2005; Tomsa 2008). Therefore, measuring the electoral performance and public support may not reflect the real situation of the ruling party development. Party strength may be an outcome of the strong leader, which makes it inappropriate to explain another phenomenon such as the performance of a regime. I propose to focus on party’s organizational development at the local level as the main element of measurement. I argue that local organizations are not only essential elements of party institutionalization but also a key driving factor for party strength. That is, party institutionalization at the local level explains party strength, which avoids the endogeneity issue when using party strength as an explanatory factor for authoritarian politics. As discussed earlier, party institutionalization at the local level helps leaders to collect essential information regarding the society and thus form better decisions. Also, grassroots organizations enhances political participation of the masses and strengthen 42 the legitimacy of the regime. As a result, when a party has connections to society via party organizations, it is more likely to enjoy higher levels of supports and to keep dominance of power over time. In sum, the main obstacle of measuring party institutionalization is the data availability issue. Using party strength as an indicator, or more broadly, the consequence of organizational development, may be a solution to the data availability issue. It is conceptually overlapping with party institutionalization and is indeed more observable than many indicators of party institutionalization. However, this concept has its limits, not a perfect alternative for party institutionalization. In particular, if one plans to analyze the outcomes of party institutionalization, using party strength as the measurement will cause endogeneity issue. 2-3 Party Institutionalization at the Local Level 2-3-1 Local Level Party Activities as the Key Element of Institutionalization I propose to focus on the party’s organizational development at the local level as the primary measurement of party institutionalization. Among the indicators in the four main dimensions (Table 2-1), the nationwide organizational presence—the establishment of local party branches and maintaining regular activities—is the main subject. Not only because building up local party branches is a major puzzle not been addressed in the current literature, but also that local party organizations have essential implications for the authoritarian regime. In particular, it could help a party to advance on various dimensions of party institutionalization, and thus it is the most direct measurement. First of all, building up local branches means that the party has different levels of organizations and party cadres and staffs across regions. When the number of party workers, geographical scopes, and organizational hierarchies are increasing, the 43 organization performs a higher level of complexity. When the rules for party activities at different levels become effective, the party proceed in the “routinization” dimension. Second, the party’s activities at the local level, including hiring staff and building up branches, and maintaining regular activities beyond the elections, help the party to establish roots in society. The local level staffs have opportunities to contact different social groups and the masses and make connections with them. Also, the party can distribute material benefits to citizens. The regular party activities could help the party to attract more citizens to join the membership. In the elections, local party branches become essential mobilization hubs of personal or clientelist networks. At the same time, local staffs provide the central level elites with essential information for better decisions, while the local level cadres have direct ways to probe the public opinion, and thus, the information should be more precise. Taken together, local party branches contribute to more stable and higher levels of support in the elections. These are all essential elements for enhancing the “roots in society” dimension, including both the complexity and adaptability of a party. Third, maintaining activities at the local level is a result that the core elites of a party intentionally distribute resources and power to people who are not in the inner circle. This strategy has positive effects on the dimension of value-infusion (autonomy) in party institutionalization. When the party has roots in society and become organizationally complex, more actors within the party have decisional power at different levels. For example, the local party leaders have the mandate from the central level elites to implement policies that provide services to citizens aiming at increase size of memberships. 44 Table 2-2 Local Party Activities and Dimensions of Party Institutionalization Dimension Routinization Coherence Roots in society Value-infusion Function of Local Party Branches Increase organizational hierarchy; introduce rules for cadre selection and elites’ interaction; hire more professional staffs; provide information for making decision and have decisional autonomy for local affairs Facilitate coordination between local factions from different area; enhance tolerance vis-à-vis intraparty dissidence Increase membership strength; maintain nationwide organizational presence and activities beyond election campaigns; hire more professional staffs in local branches; link to civil society organizations Maintain stable popular supports, especially after alternation in party leadership; strengthen autonomy of the organization Fourth, party institutionalization at the local level has indirect impacts on the three indicators of the coherence dimension. Establishing local level branches means that the job positions for local affairs and management of regional party organizations become aggrandized, and the party would introduce explicit rules for party activities, such as for cadres selection and local party congress. These practices increase motivations for party members to invest in and seek promotions under this system. Local party branches also help the elites at the central level to facilitate cooperation with local factions, which in turn mitigates potential conflicts between intraparty groups or the “dysfunctional factionalism.” It is quite common that elites from different regions have a conflict of 45 interests, but the developments of party organizations lead to explicit rules for coordination. Thus, local party organizations increase the level of the indicator “the party tolerates a certain together, party institutionalization at the local level is the most direct factor that covers most of the indicators in the four dimensions. I summarize the function of local party branches and the related dimensions of party institutionalization in Table 2-2. intraparty dissidence.” Taken level of 2-3-2 Data for Party Institutionalization and its Summary To examine party institutionalization in authoritarian regimes, I create a dataset including all of the dictators who had remained in power for more than one year in the post-War era, covering from 1946-2008. The unit of analysis is leader, because I argue that the presence or absence of party local branches come from the strategic decision of the ruling elites to distribute resources and power to develop the party’s organization and maintain activities. Each leader has the choice of developing the party organizations or relying on other ruling tools. As the primary concern of this study is how parties go beyond the city of the central government, the small island-countries 8 are excluded because it is difficult to differentiate party organizations at the “local” level from the central level in these small states. According to the database from Geddes, Wright, and Frantz (2014; 2018) and Svolik (2012), I record 657 dictators from 123 countries. However, 172 dictators served for less than one year, which makes the number of cases in the dataset as 485. 8 According to the World Population Review, an island country is defined as “a country that is made up of one or more islands, or land that is surrounded completely by water.” I exclude the countries whose land area is smaller than 1000 square feet. See https://worldpopulationreview.com/countries/island- countries. Accessed: May 22, 2020. 46 In this study, the main measurement of party institutionalization for a cross-national comparison is whether the leader establishes, maintains, being without, or abandons support party’s local organizations during his term. I recode from the local party organization variable in Geddes, Wright, and Frantz’s (2018) dataset, in which they code in binary terms whether the leader’s party has local organization on a yearly bases. The data collection in their dataset adopts a minimum approach that as long as the support party “has few local organizations,” it will be coded as positive. I identify four situations of each leader: (1) Abandon: There are local party activities when the leader took power but are abandoned during his tenure (original variable changes from 1 to 0); (2) Remain absence: local party branch remains absent throughout the term (0 to 0);9 (3) Maintain presence: local party organization is maintained from the beginning until the time of leaving power (1 to 1); and (4) Build new: no local party branches at the beginning, and the dictator build up them at some moment during his rule (0 to 1).10 As Table 2-3 shows, 81 dictators (16.7%) start with no local party branches but build up new local organizations for the party during their terms. One hundred eighty-four (184) dictators (37.9%) maintain the presence of party branches at the local level. These two categories combined as 54.6 percent of all dictators. On the other hand, 42.1 percent of the rulers do not extend the party organizations to the local level (or even do not have a support party). A few dictators, 16 (3.3%), begin with local party organizations but terminate the presence of them during the term. The total number of leaders that do not 9 The category of 0 (no local branches of the support party throughout the tenure) includes the situation that the leader does not have a support party. 10 In some cases, the leaders begin with no local parties, build up new ones, and then destroy them. In this situation, it will be coded as -1, as the party organizations do not last. 47 rely on the strategy of establishing nation-wide party organizations is 220, 45.4 percent of all rulers. Table 2-3 Frequency of Local Party Presence Party local branch Build new Maintain presence Remain absence Abandon Total Frequency Percentage 81 184 204 16 485 16.7% 37.9% 42.1% 3.3% 100% Subset % Yes: 54.6% No: 45.4% 100% Table 2-4 Regime Type and Party Institutionalization Regime type Number of leaders Local party institutionalization No local party institutionalization 49 (51.6%) Personal 30 (17.8%) Party-based mainly (party-based only) (12) (10.8%) 30 (75.0%) Monarchy/Oligarchy 63 (70.8%) Military 172 (43.8%) Total Note: Row percentage in parentheses. “Party-based mainly” includes party, party- personal, party-military, and party-military-personal types in the original Geddes, Wright, and Frantz (2014) dataset. “Party-based only” is the type of party-based regime without mixed ones. 46 (48.4%) 139 (82.2%) (99) (89.2%) 10 (25.0%) 26 (29.2%) 221 (56.2%) 95 169 (111) 40 89 393 A breakdown by regime type shows that (see Table 2-4), about half of the personal dictators (48.4%) institutionalize the party, while in military regimes, only 29.2 percent do so. In the monarchies or oligarchies, 10 out of 40 leaders have institutionalized parties, 48 including Burundi’s Mwambutsa IV (1962-66), Cambodia’s Norodom Sihanouk (1953-69) and Lon Nol (1969-75), and Iran’s Mohammad Reza Pahlavi (1953-79), and 6 leaders from South Africa (1948-94). Not surprisingly, most leaders in the party-based regimes (82.2%), including party-military or party-personal regimes, institutionalize their parties at the local level. Furthermore, among all of 111 leaders of party-based regimes (the category excludes party-military and party-personal regimes), only 12 (10.8%) do not institutionalize the ruling party at the local level. The cases include Afghanistan’s Mohammed Rabbani (1996-2001), Angola’s Agostinho Neto (1975-79), Bolivia’s Barrientos Ortuna (1964-66), Colombia’s Rojas Pinillia (1953-57), Guinea-Bissau’s Amílcar Cabral (1974-80), Iran’s Ruhollah Khomeini (1979-89) and Ali Khamenei (1989-), Mali’s Dioncounda Traoré (1968-91), Mozambique’s Machel (1975-86), Niger’s Hamani Diori (1960-74) and Seyni Kountché (1974-87), and Sierra Leone’s Valentine Strasser (1992-96).11 In the party-based regimes, the percentage of not having an institutionalized ruling party is the lowest among all regime types. 2-3-3 Party Institutionalization and Party Strength leaders To clarify how party institutionalization differs from party strength, I replicate the data of all ruling parties in Meng (2019) and analyze the relationship between these two concepts. In the data, Meng records 156 ruling parties that were in power for at least three years from all of the authoritarian regimes, 1946-2008, identified by Svolik (2012). Then she counts the number of leaders (must remain in power for three or more consecutive 11 Among the 10 leaders here, Colombia’s Rojas Pinillia and Sierra Leone’s Valentine Strasser destroy the local party during the term (coded as “abandon” in my local party category), while other leaders maintain absent on local party developments. 49 years) when the party is in power, the total number of years of ruling, and number of years of political survival after the departure of the first leader. I categorize parties with successful leadership transition and a longer period of political survival (at least longer than one year) after the first leader as “party strength.” Then I add a characteristic of local party institutionalization in a binary term for each party in the data. Parties either build new local organizations or maintain the presence of local activities in tenure are coded as 1, while those without local party organizations, either remaining absence or destroying them, are coded as 0. Table 2-5 Party Institutionalization and Party Strength Party strength (Leadership transition & stay in power >1yr) Yes 54 (55.1%) (93.1%) 4 (12.5%) (6.9%) No 44 (44.9%) (61.1%) 28 (87.5%) (38.9%) 72 (100%) (100%) N (row) 98 (100%) 32 (100%) 130 Party institutionalization at local level Yes No N (column) 58 Table 2-5 is a two by two table showing the relationship between party strength and party institutionalization at the local level. If one looks at the strong parties, 93.1 percent of them are institutionalized. Only 4 parties (6.9%) that undergone leadership turnover become “strong” without institutionalization at the local level. These parties are Central African Republic’s Movement for the Social Evolution of Black Africa, El Salvador’s 50 Revolutionary Party of Democratic Unification, Haiti’s Party of National Unity, and Portugal’s National Union. On the other hand, an institutionalized party does not guarantee successful leadership turn over and political survival after the first leader. Although a majority of institutionalized parties (55.1%) achieved party strength, both have leadership transition and successfully maintain power for more than one year, 44.9 percent of them fail to become strong party defined by Meng (2019). Among parties without local branches, only 6.9 percent of them achieve political survival after leadership change, while 87.5 percent of them fail to remain in power. Therefore, party institutionalization may be seen as a necessary condition for party strength, but it is not sufficient to explain how a party becomes strong. Taken together, party strength and party institutionalization are two different concepts, and it is not proper to interchange the usage of them. While we have explored the relationships between party institutionalization and party strength, it may be possible to examine the ruling parties in each regime type and identify whether some parties are mischaracterized. Meng (2019) has done so in her study, as she points out 15 parties in the party-based regimes are weak, failing to survive at least 10 years after the first leader, and 21 parties in non-party-based regimes that may be strong enough to be considered as the same with counterparts in the party-based regimes (See Table 2-6). I examine party institutionalization for each party. Among the 15 parties considered as weak in party-based regimes, only PPN in Niger is not institutionalized. Hence, Niger (1960-1973) may be a mischaracterized party-based regime. On the other hand, among 21 strong parties in non-party-based regimes, 20 are both institutionalized and strong. Only Haiti’s National Unity (1957-58) was not institutionalized, as it is correctly put in the “personal” type of autocracy. The regime type of these 20 regimes with institutionalized and strong parties may be re-considered 51 Table 2-6 Parties that May Have Been Mischaracterized Regime Type Weak parties categorized as party-based regimes Party-based Non-institutionalized Party PPN (Niger) National Unity (Haiti) Institutionalized Party PLA (Albania) RNM (Bolivia) PDCI (Cote d’Ivoire) PPP (Gambia) PDG (Guinea) BNP (Lesotho) PSD (Madagascar) US (Mali) FSLN (Nicaragua) Parmehutu (Rwanda) APC (Sierra Leone) UNP (Sri Lanka) DP (Turkey) UNIP (Zambia) ARENA (Brazil) PCT (Congo) UPRONA (Burundi) PCN (El Salvador) RPF (Rwanda) Communist Party (Cuba) PDG (Gabon) WPK (North Korea) PCR (Romania) TPD (Turkmenistan) YAP (Azerbaijan) CPDM (Cameroon) PCT (Congo) PAIGC (Guinea-Bissau) Baath (Iraq) Liberal (Nicaragua) National Union (Paraguay) NDP (Egypt) ANR (Paraguay) Baath (Syria) Strong parties categorized as non party-based regimes Military Party-military Party-personal Personal Party-personal- military Note: adapted from Meng (2019). The list excludes authoritarian successor parties. as party-based. In general, the discussion above shows that party institutionalization and party strength are two different concepts. Parties do vary on their levels of 52 institutionalization, as one should be cautious when using regime type as an explanatory factor on other outcomes of the regime. 2-4 Conclusion This Chapter discusses the concept of party institutionalization, unveiling its multidimensionality, and the measurement issues. I illustrate why this study focuses on party institutionalization at the local level, as it is a critical ruling strategy for the dictator, and it solves some problems of the measurement. In particular, the concepts of party institutionalization and party strength have some overlapping elements, but they cannot become full alternatives of each other for measurement. Empirical evidence suggests that party institutionalization is a necessary condition but not sufficient condition of party strength. I also examine the relationship between party institutionalization and each regime type. The indicator of party institutionalization may be used to examine whether one regime is mischaracterized or not. The organizational development of individual ruling parties has essential implication to dictatorships, but so far the studies are lacking mainly due to data availability issues. Also, there is no consensus on which indicator should be used for measurement. This Chapter points out that party institutionalization at the local level is an useful measurement and it is critical to political survival (Chapter 4 will demonstrate the relationship between party institutionalization and performance of political survival). However, the empirical evidence in this Chapter is mainly descriptive, and no comparison between different indicators. Future studies are needed for comparing how indicators of different dimensions of party institutionalization measure the concept. 53 CHAPTER 3 CAUSES OF AUTHORITARIAN RULING PARTY INSTITUTIONALIZATION What explains the different paths of ruling party development in autocracies? Why do some authoritarian regimes institutionalize their ruling parties and penetrate society through party organizations? This chapter illustrates the key factors that leads to variations in party institutionalization. Extant studies provide two main explanations on party institutionalization, but the causal mechanisms are contradictory to each other. One argues that resource availability and leaders’ capability leads to party institutionalization, while another illustrates lack of capability and environmental difficulties prompts leaders to institutionalize. The two schools of thought illustrate contrary explanations and both fall short on explaining party institutionalization. To date, there is no scholarly consensus on the origin of party institutionalization. This study illustrates a threat-based theory on party institutionalization. It argues that the initial condition when a dictator comes to power determines the ruling strategies he will choose. The ruler needs to establish party organizations to control the masses when he subjectively thinks the threat to regime survival mainly comes from internal threats. If there exists a foreign enemy with territorial conflicts, the leader must direct resources to the coercive institutions, especially the military, to defend against threats. To test the hypothesis, this chapter utilizes a cross-national study covering all that party dictators after WWII, institutionalization occurs when (1) there is no foreign enemy with territorial threats; (2) the level of religious polarization is high; and (3) the relative salience of external territorial threats over internal social polarization is low. Comparative case studies Taiwan and from 1946-2008. Empirical evidence shows 54 South Korea further confirm the causal mechanism of the external threat. Furthermore, the comparison of Indonesia and The Philippines show that religious polarization is the most important factor that drives party institutionalization. The chapter proceeds as the following. I briefly discuss the conventional wisdom and the limits. Then I illustrate the threat-based theory in detail. Then I conduct a cross- national study on the causes of party institutionalization. Subsequently, I utilize the Most Similar Systems Design to compare ruling parties in South Korea and Taiwan, which confirm that foreign enemy with territorial threat is the major obstacle for party’s development. 3-1 Theories of Party Institutionalization 3-1-1 The Limits of Conventional Wisdom To date, a systematic explanation of how dictators determine the strategies of developing parties is scarce, and the existing explanations of party institutionalizations are not satisfactory. The two main school of thoughts, resource availability theory and adversity-driven theory, have opposite predictions. The first school tells us that one should have observed that authoritarian regimes build up parties due to their ability to generate material benefits from such forms of rule (Panebianco 1988; Riedl 2014), while the second predicts that the weakest regimes, who face difficulties in maintaining power, will try to rely on parties to gain more power (Panebianco 1988; Hellmann 2013; Meng 2016). It seems like neither school of thought is sufficient to explain ruling authoritarian party institutionalization, while they do hold some cases correct. If the first school of literature, the resource capability argument, is correct, we should have observed institutionalized ruling parties in East Asian developmental states because the states fully 55 control public resources. This theory cannot explain the absence of competent ruling parties in countries such as Korea. Were the other school of literature correct, we should have seen that East Asian dictators have little incentives to pay extra costs in connecting to the people because they have resources, with aid from the U.S., to strengthen the coercive apparatus. Therefore, both theories are insufficient to explain East Asian cases, let alone dictators all over the world. To date, there is no scholarly consensus on the origin of party institutionalization. 3-1-2 A Threat-based Theory of Party Institutionalization I argue that the initial condition when a dictator comes to power determines the ruling strategies he will choose. Dictators often put more efforts into dealing with the dominant perceived threat (Greitens 2016). The ruler needs to establish party organizations to control the masses when he subjectively thinks the threat to regime survival mainly comes from internal threats. On the other hand, if there exists a foreign enemy with territorial conflicts, the leader must direct resources to the coercive institutions, especially the military, to defend against threats. Under this circumstance, developing party organization becomes a secondary priority to the regime’s survival. There are various sources of threats to autocratic political survival. Externally, a dictator faces military threats from foreign enemies or international interventions. Internally, the elites threaten the regime survival by coup d’état, assassination, and other types of non-constitutional removal of the leader, while the masses threaten the regime by demonstrations, strikes, violent riots, guerrilla attacks, civil war, or revolution 56 (Davenport 1995; Carey 2010; Slater 2010). Greitens (2016: 19) illustrates an autocratic threat landscape, adapted as Figure 3-1.12 Threat to regime External Negative Internal Mass-based (popular revolt) Positive Elite-based (coup) Mixed Figure 3-1 Threat Landscape and Effects to Party Institutionalization Source: Adapted from Greitens (2016: 19) External threats When there exist foreign rivalries,13 the dictator must strengthen defense capability to defend against the possibility of attack (Rasler and Thompson 2004). In particular, among different kinds of external threats, Kim (2018) points out that territorial threats from outside lead to an environment conducive to the militarization of an autocracy. The foreign threat is the top concern for the leader for several reasons. First, it is difficult to predict an enemy’s decision to use of force even though each country has its intelligence system. In the studies of international relations, uncertainties about another state’s 12 In Greitens’ (2016: 19) original figure, there is a category of “secessionist” in the internal threat. It is not necessarily associated with political survival of the regime and is not included in the discussion. 13 Specifically, interstate rivalry is defined as a pair of states with expectation and beliefs that each side is hostile to another and is threatening national security in protracted conflicts (Colaresi, Rasler, and Thompson 2008). 57 capabilities and resolve are one of the major reasons that interstate conflicts occur (Fearon 1995). The best way to defend one’s own country is through strengthening the military forces for deterrence (Waltz 1979). By contrast, domestic information about the opposition is relatively easier to acquire via internal security systems and informants. Controlling individuals can be done by various tools such as surveillance, detention, other violent approaches, or the non-violent co-optation approach to counterbalance opposition to the regime. Compared to tools to defend against a foreign enemy, the dictator has multiple methods available to maintain control for internal actors. Furthermore, previous studies have illustrated that interstate rivalries with territorial disputes often maintained over long periods (Colaresi, Rasler, and Thompson 2008; Klein, Goertz, and Diehl 2006) and this kind of rivalries tend to “become ensconced in national memories, identities, and political discourses” (Rasler and Thompson 2011, 292). That is, the presence of territorial rivalries that are directly threatening the security of the homeland have a long-lasting and more robust sense of crisis for a regime. Colaresi, Rasler, and Thompson (2008) define rivalries involving boundary disputes and territorial claim as “spatial rivalry.” Consequently, if there is a territorial threat from outside, it is more likely that this becomes a priority and compels the leader to invest more resources in the armed forces (Gibler 2012) for defending regime survival. In this situation, party institutionalization becomes less urgent to the leader. Mass-based internal threats If there are no immediate threats from external enemies, the leader has more time and resources to evaluate ruling strategies for internal threats. The domestic threats can be classified into two main categories, from the masses and the elites. When the threat to regime survival mainly comes from the masses, the leader will institutionalize the party 58 at the local level. The party organization helps the dictator prevent potential popular uprisings because it solves the problem of lack of information among the masses, known as “dictator’s dilemma” (Wintrobe 1998). When the party interacts with people’s daily lives and links with society through professional staff and regular activities, the ruler will be able to form better strategies regarding resource distribution and mobilization, according to information gathered directly from the society. On the other hand, from people’s viewpoints, the presence of a party organization at the local level provides a specific object for people to attach their loyalty to, especially through regular party activities and face-to-face contact. Moreover, the ruler can distribute material benefits to people via party staff. As a result, when a party has connections to society through its branches and staff, this helps the dictators to counterbalance potential threats to political survival. The domestic mass-based threat is imminent to the leader particularly when the society is heterogeneous, as the identity-based mobilization is more likely to occur (Marquardt 2018). Scholars find that political parties and social groups often organize based on common social cleavages such as religion, social class, and language (Horowitz 1985; Lijphart 1984). There are two main measurements of these different dimensions of social diversity. The first one is fractionalization, which has a simple interpretation as “the probability that two randomly selected individuals from a given country will not belong to the same ethnic/religious/linguistic group” (Alesina et al. 2003). However, empirical findings indicate mixed results on the effects of fractionalization and civil wars, the most serious domestic conflict. For example, Fearon and Laitin (2003) and Collier and religious Hoeffler fractionalization has a significant effect on the onsets of civil wars. Horowitz (1985) points out that the relationship between ethnic diversity and civil wars is not monotonic. The fractionalization nor that neither ethnic (2004) illustrate 59 reason may come from higher coordination costs between different groups as a result of higher levels of diversity. Consequently, scholars further differentiate the concepts of fractionalization and polarization.14 Social polarization is the essential factor to explain how the leader faces threats from the masses. The measurement of polarization first assumes a bipolar symmetric distribution of the ethnic/religious/linguistic groups and calculates the degree of distance between the real distribution of and the bipolar one. A greater value of polarization index means that there are larger minority groups and thus higher potential of conflicts. The concept of polarization implies that ethnic or religious conflict is prone to onset when the minority is large and not divided into smaller groups. In extreme circumstances, as Collier and Hoeffler (1998) illustrate, if two ethnic (or religious) groups are similarly sized while one is loyal to the government and another with the rebel groups, the coordination cost for the collective action against the ruler will be the lowest. Empirical findings are more supportive of the statement that polarization rather than fractionalization lead to violent conflicts. For example, Esteban and Ray (1999; 2008) show that a bipolar distribution of population maximizes the intensity of conflicts. Horowitz (1985) also points out that ethnic conflicts will occur in countries where a large ethnic minority faces an ethnic majority. Among all of the dimensions of social cleavages, the religious diversity is a primary driving factor for social conflicts (Hungtington 1996; Basedau, Pfeiffer, and Vüllers 2016). It is because several elements of religion help overcome the collective action problems of organized unrest against the regime. Collective action problem can be understood as necessary conditions to initiate actions toward the target, including motivating enough 14 The concept of polarization is first applied to measuring the distribution of income by Esteban and Ray (1994) and Wolfson (1994). For more discussion, see Montalvo and Reynal-Querol (2005). 60 people to join the group, and the opportunity to launch the action (Collier and Hoeffler 2004). The relationship between religion and social unrest is illustrated in Figure 3-2. First, religious organizations provide better political opportunities for mobilization (Montalvo and Reynal-Querol 2005). When the level of polarization is high, that is, when there are two or more equally sized groups, it makes ascriptions along the religious line easier. That is, religious cleavage will become more salient. Also, religious groups offer recruitment pools of individuals with strong social ties. Compared with other social cleavages, religion has a higher level of exclusivity, as one tend to belong to only one religion. Furthermore, people with different ethnicities and languages can have the same religious belief. Thus, religiosity help overcome difficulties in organizing collective action. Second, religion leads to higher levels of motivation for collective actions. In general, the beliefs of religion determine how people view the world, social relationships, and other values. Furthermore, the worldview derived from religion is often “fixed and nonnegotiable” (Reynal-Querol 2002). When it comes to collective actions, religious ideas sometimes help to justify extreme activities against the out-groups, especially the ruling groups that cause injustices in terms of economic and political grievances. The willingness of participating in the collective action will be enhanced when the movements are theologically legitimized. For example, the Crusaders or many religious extremists such as the Islamic State in Iraq and Syria (ISIS) alike are the references of this theory (Basedau, Pfeiffer, and Vüllers 2016). Taken together, if a leader of a social group mobilizes people based on religious ideas, it has the potential to gather people with higher levels of motivation. 61 Religion Religious practice Religious structure Motive Opportunity Violent collective action Figure 3-2 Religion and Violent Collective Action Source: Adapted from Basedau, Pfeiffer, and Vüllers (2016: 231) The ethnic structure is another essential factor that drives domestic conflicts (Fearon 2003). However, previous studies find that supports for religious organization are often stronger than groups organized by ethnicity (Stewart 2012). At the same time, the level of severity between religious conflicts and ethnic conflicts are similar in terms of casualties. The most serious situation for the leader is the overlap of identities. Empirical studies show that the likelihood of armed conflict onset becomes higher when religious identities overlap with ethnic identities, or when it overlaps with economic boundaries— when horizontal economic inequality between different groups is high (Basedau, Pfeiffer, and Vüllers 2016). These social polarization-related conflicts are sources of political instability and thus threats to political survival (Annet 2000). Elite-based internal threats When the threats to the regime come from elites, that is, coup-proofing becomes the priority for the leader, there are mixed effects for the party institutionalization. On the one hand, as Geddes, Wright, and Frantz (2018) illustrate, organizing a political party is 62 a common strategy for the dictator to neutralize risks of losing support from the military. They found that dictators who come from the military with low military ranks, have no military experience, or are young before the seizure of power tend to create a support party to discipline the elites. Empirical evidence does suggest that post-seizure party creation reduces the likelihood of coups and helps autocracies secure power (Geddes, Wright, and Frantz 2018). However, party creation does not necessarily lead to party institutionalization. Threats from elites often prompt the leader to adjust the institutional design of the coercive institutions. The most common strategy of the ruler is to increase the degree of fragmentation of the security apparatus (Greitens 2016). A divided organizational structure prevents a single agency or strongman from carrying out enough political power for a coup. The ruler often creates parallel organizations, and the competition between different agencies diminishes the capabilities of a coup. These organizations can be a military branch, a paramilitary force, police, presidential guard, or a combination of security forces. That is, the priority of the leader is to deal directly with the coercive institutions, the main source of coup threats, not necessarily on the political parties. Therefore, from elites might not have significant effects on party threats institutionalization. The threat-based theory stresses on leaders’ concern on the priority of strategies. Indeed, leaders often have to face both internal and external threats, and some dictators can indeed formulate responding tools for both threats at the same time. However, I argue that the threats have different levels of urgency to the leaders. For example, as discussed above, foreign enemy is often unpredictable and it is not possible to use tools of co-optation towards external threats. The leader must maintain a strong army to defend against potential military conflicts. Regarding the scarcity of resources, the leader 63 must allocate optimal amount of resource to each ruling strategy. Therefore, the core element of the threat-based theory is to examine the relative level of threats from the external threats and various internal sources. Taken together, I hypothesize that the dictators need to rely on political institutions and co-optation strategies to establish linkage at the local level when (1) there is no spatial rivalries from outside, and (2) the masses are perceived as threats. Alternative hypotheses from the conventional wisdom are (1) resource-abundant leaders tend to institutionalize the party; (2) resource scarcity results in party institutionalization; (3) leaders with higher risks of coups tend to institutionalize the party. 3-2 Causes of Party Institutionalization: a cross-national study 3-2-1 Data and Research Design To test my hypotheses, I create a dataset including all of the dictators after World War II, covering from 1950-2008. The unit of analysis is the party leader, because the main concern of this study is each dictator’s choice of ruling strategy. I exclude island-countries whose land area is smaller than 1000 square feet because, in these relatively small countries, it is difficult to differentiate party organization at the local level. I collect data from established cross-national datasets and studies of authoritarian regimes, such as Geddes, Wright, and Frantz (2014; 2018), Svolik (2012), and Miller (2019). There are 657 dictators from 123 countries in this period. Then I exclude 172 dictators who served for less than one year, which makes the number of cases as 485. 64 Dependent variable The primary dependent variable, party institutionalization, is measured in binary terms by whether the leader has institutionalized party organizations at the local level. I recode from Geddes, Wright, and Frantz’s (2018) dataset, in which they code whether the leader’s party has local organizations on a yearly bases. The original coding is binary, 0 as no support party or no local party organization, and 1 as the support party has few local organizations. In this study, the unit of analysis is the leader so that I examine the dynamic of this variable in each leader’s terms. We can identify four mutually exclusive situations: (1) Abandon: There are local party activities when the leader took power but are abandoned during his tenure (original variable changes from 1 to 0);15 (2) Remain absence: local party branch remains absent throughout the term (0 to 0);16 (3) Maintain presence: local party organization is maintained from the beginning until the time of leaving power (1 to 1); and (4) Build new: no local party branches exist at the beginning and the dictator build them up at some moment during his rule (0 to 1). The first two situations will be coded as 0, no institutionalized parties, while the last two are 1. This approach begins with the initial condition of the local party branches when a leader takes power, then it examines how the leader develops, maintains, or destroy them. As shown in Chapter 2 (Table 2-3), the numbers of cases in each category are: 16 (3.3%) abandon; 204 (42.1%) remain absence; 184 (37.9%) maintain presence; and 81 (16.7%) build new local branches. The total number of leaders that establish or maintain 15 Some dictators both establish and abolish the local organization in his term (original variable from 0 to 1, then to 0). The situation will be coded in the first category of abandoning the local organization. 16 This category of no local branches includes the situation that the leader does not have a support party. 65 nation-wide party organizations is 265 (54.6%), while the number of leaders who do not rely on the local party organizations is 220, 45.4% of all rulers. Foreign enemy The nature of an external threat is hypothesized as the main factor that impedes party institutionalization. I utilize existing data that pointed out various types of foreign threats. First of all, I create a dummy variable for ongoing interstate war when a leader comes to power. The data mainly comes from the Militarized Interstate Disputes v4.3 dataset under the COW project (Reiter, Stam, and Horowitz 2016). This is the most straightforward variable measuring the existence of a foreign enemy. In addition, the dataset has data for the level of hostility in the interstate crisis (Gibler, Miller, and Little 2016). I also include a dummy variable for having conflicts that involve military actions, and another variable for the use of force. However, there are various types of external threats. Previous studies have found that the leaders respond more directly and forcefully to foreign actors that pose threats to a country’s territory (e.g., Gibler 2012; Kim 2018). I utilize three measurements of external threats on territory. First, I include the dummy variable for foreign territorial claims from Hensel et al. (2008)’s ICOW Territorial Claims Data. Second, Thompson and Dreyer (2011) code the presence of foreign rivalries with territorial claims, and more broadly enemies with strategic conflict. Third, from the ICOW dataset we have data for the level of the salience of territorial claim. The variable measures the importance of the claimed territory, including its resources, population, historical sovereignty, and strategic location. The variable ranged from 0 to 12, from no claims to the highest level of salience (Hensel et al., 2008). These three variables are 66 marked as territorial claim, strategic territorial rivalry, and territorial claim salience respectively. Internal threats For the potential threats from the masses, I first include the indices of social group polarization. It measures “how far the distribution of the ethnic groups is from the (1/2, 0, 0, ... 0, 1/2) distribution (bipolar), which represents the highest level of polarization” (Montalvo and Reynal-Querol 2005). The measurement of polarization ranges from 0 to 1, as a higher value means a polarized society with more potential threats. The dataset constructed by Reynal-Querol (2002) includes ethnic and religious polarization in 137 countries. Note that the data is not time series but based on resources of 1982 and 1987. It is not a perfect indicator, but in general, the distribution of religious and ethnic groups in each country does not change too much over time. An alternative measurement for ethnic polarization comes from the study of Basedau, Pfeiffer, and Vüllers (2016). Their data covers 130 countries in the period of 1990-2001. I use this data for a robustness check. I also include the indices of language fractionalization (Alesina et al. 2003). Fractionalization measures the degree to which a society is divided into different groups, as calculated by the probability of the common social group of the two randomly selected individuals.17 The variable is coded from 0 to 1, as higher values mean higher levels of fractionalization. On the other hand, for threats from the elites, the dataset includes the number of coup in the last 5 years (I count for the first year of each leader) as defined by Powell and 17 Fearon and Laitin (2003) also have alternative data for religious fractionalization, calculated by Quinn Mecham. 67 Thyne (2011). Another source of coup data is from the Center for Systemic Peace (CSP) project (Marshall and Marshall 2014). Both measurements of coups are recoded in Bodea, Elbadawi, and Houle (2017). As civil war may be one of the most severe challenges to political survival. I include a dummy variable for ongoing civil war from the COW project (Reiter, Stam, and Horowitz 2016). As my unit of analysis is the leader, I include characteristics of each dictator. Mainly, I recode from Svolik (2012) for the manner by which the leader assumed office and left office.18 In particular, those who took power by coup are more likely to lack stable base of support, while power seizure by elite consensus reflects a leader with a relatively higher level of elite’s support. I also include a dummy variable for being a military regime defined by Banks (2011). These variables can be used to test whether a relatively weak dictator tends to institutionalize the party at the local level. Threat ratio In addition to the variables measuring various sources of threat, I create a threat ratio variable measuring the relative importance of external and internal threats to the leader. It is calculated as external threats over internal threats. For the level of external threats, I use the territorial claim salience variable (Hensel et al., 2008), as it is a continuous variable ranged from 0 to 12. For the internal threats, I take the mean value of the sum of two social polarization indicators, including ethnic and religious polarization (Reynal-Querol 2002). A higher value of threat ratio reflects severe threats from outside to the country’s 18 Types of assuming power include civil war, consensus (chosen by a collective), coup, elections, foreign intervene, country’s independence, interim, revolt, succession, and others. The ways of leaving power include assassination, civil war, consensus, coup, elections, foreign, interim, natural (sickness or death), no contest (not to run for re-election), resignation, revolt, step-down (in a transition to democracy), term limit, and others. 68 territory, while society is relatively homogeneous and has less risk of a popular uprising. A lower value of threat ratio means that social polarization is high, while the potential danger from an external enemy is less salient. Alternative explanation of resource-availability theory Previous studies pointed out that resource abundance determines the development of party organization. I include the following indicators for the hypotheses. First, the data have a dummy of oil export reliance that more than one-third of export revenues from fuels (Fearon and Laitin 2003). I also include fuel export income per capita, as well as natural resource export income per capita (Haber and Menaldo 2011). These variables reflect the degree to which a country is endowed with natural resources. According to the resource availability theory, natural resource should be positively associated with party institutionalization, while adversity-driven theory predicts the opposite. Control variables The data includes a standard battery of control variables concerning the regime, including the area, population, GDP per capita, Gini index for income inequality, polity score (converted to 0-100), regional democratic diffusion (regional mean polity score and regional democratic diffusion defined by Haber and Menaldo 2011), and regional dummy (Western democracies, Eastern Europe, Latin America, Sub-Saharan Africa, Asia, North Africa, and Middle East). These data mainly come from Haber and Menaldo (2011) and the Maddison Project (Bolt and van Zanden 2014). I also include dummy variables for former British colonies, former French colonies, and communist regimes. All socio- economic factors are lagged as t-1 of each leader’s first year. Finally, I include a variable 69 for the length of tenure in years for each leader, because it often takes time for a leader to institutionalize the party. The summary of all variables is in Table 3-1. Table 3-1 Summary of Variables of Causes of Institutionalization Mean 0.55 0.72 0.46 6.14 0.55 0.57 19.96 0.05 0.10 0.28 0.43 0.01 7.71 12.65 15.85 3.81 0.31 0.22 0.24 0.23 0.16 28.31 40.08 0.16 0.35 10.83 Std. Dev. Min 0.00 0.00 0.00 0.00 0.02 0.00 0.00 0.00 0.00 0.00 0.00 -0.27 5.41 6.54 11.71 0.00 0.00 0.00 0.00 0.00 0.00 0.00 8.33 0.00 0.00 1.00 0.50 0.45 0.50 4.33 0.22 0.32 74.22 0.22 0.30 0.45 0.29 0.04 1.00 1.62 1.47 2.71 0.46 0.42 0.43 0.42 0.36 22.02 14.95 0.37 0.48 9.95 Max 1.00 1.00 1.00 12.00 0.98 0.98 800.00 1.00 1.00 1.00 0.92 0.43 11.72 16.91 20.97 10.29 1.00 1.00 1.00 1.00 1.00 100.00 93.75 1.00 1.00 49.00 Variable Party institutionalization Territorial claim Strategic territorial rivalry Territorial claim salience Ethnic polarization Religious polarization Ext./Int. threat ratio Interstate war Civil war Coups last 5yr Language fractionalization Log GDP mean year Log GDP growth Log area Log population Log resource income Coup entry Consensus entry Colony of Britain Colony of France Oil dependency Polity score Region polity score Communist regime Military regime Tenure Year Obs 484 396 415 396 382 382 319 411 411 439 442 419 427 445 442 334 485 485 484 484 483 444 445 474 431 483 3-2-2 Analysis This chapter adopts Logit models to analyze whether the leader institutionalizes the party at the local level, because party institutionalization is measured in a binary term. Table 3-2 reports the results of logistic regression models estimating how various types of threats lead to ruling party institutionalization. In general, the finding shows that external threats with territorial claims impedes party institutionalization. Interstate war 70 does not have significant effects, as the external conflicts are not necessarily accompanied by threats to a country’s territory. Among the domestic sources of threats, religious polarization is positively associated with higher odds of building up local party organizations. The finding is consistent with previous theories that religion is one of the most prominent factors that lead to domestic conflicts. When the level of religious polarization is high, the party is utilized for preventing potential identity-based mobilization. While ethnic polarization does not alter decisions on developing party organizations, language fractionalization significantly lower the probability of party institutionalization, probably because it raises the coordination cost for the collective action. Thus, the leader feels less threatened by the masses when there are multiple languages used in society. Other factors, including civil wars and coups experiences, do not significantly influence the odds. The results hold the same when we change the external threat variable from binary (model 1) to continuous (model 2). In model 1, the variable is the presence of foreign enemy. Figure 3-2 shows the predicted margins of territorial claim enemy. While in model 2, territorial claim is a continuous variable ranged from 0 to 12, as 0 represents no territorial claim. Both models have similar results. Figure 3-3 shows the predicted margins of territorial claim salience, as higher levels of salience of territorial claim leads to lower probability of party institutionalization at the local level. The empirical evidence confirms the hypothesis that foreign enemy with territorial claim could divert the priorities of ruling strategies from civilian apparatus to the military. As for other variables, countries with more population tend to witness the development of ruling parties at the local level. Also, longer tenure year lead to party institutionalization because it often takes time to build up the local branches. 71 Figure 3-3 Predicted Margins of Territorial Claim Enemy Figure 3-4 Predicted Margins of Territorial Claim Salience Model 3 examines how the relative importance of external and internal threats influence party institutionalization. Results show that the more external territorial threats weigh more than internal polarization in terms of ethnicity and religion, the less likely that the leader institutionalizes the party. Figure 3-4 presents the marginal effects of threats ratio to party institutionalization. The graph excludes three extreme values, all 72 Portuguese leaders. The society of Portugal is highly homogeneous, which leads to a high external/inter threat ratio. All three leaders, including Antonio Salazar (1932-1968), Costa Gomes (1974-1976), and Caetano (1968-1974) do not have an institutionalized party. Tunisia and Saudi Arabia are two more examples of the high level of external threats and low levels of social diversity. Leaders do not have institutionalized party. In general, for each leader, the external/internal threats ratio is significantly and negatively associated with odds of having an institutionalized parties. Figure 3-5 Predicted Margins of External/Internal Threats Ratio All factors from alternative explanations do not influence the results. For the resource availability and adversity-driven theory, factors of oil reliance, GDP per capita, and natural resource income per capita all have insignificant effects. The way leaders enter the position do not affect the strategies on party development either. That is, the cross-national evidence suggest that party institutionalization occurs when (1) there is no foreign enemy with territorial threats; (2) level of religious polarization is high; and (3) The relative salience of external threats on the territory over internal social polarization is low. 73 Party Institutionalization at Local Level (1) (2) (3) Table 3-2 Logistic Regression Results of Causes of Institutionalization Territorial claims Territorial claim salience Ext./Int. threat ratio Interstate war Civil war Coup last 5yr Religious polarization Ethnic polarization Language fractionalization Oil reliance GDP per capita Resource income per capita Coup entry Consensus entry Military regime Communist regime British colony French colony POLITY score POLITY of region Gini Mountainous terrain Area Population Tenure year Constant N pseudo R-sq Note: Exponentiated coefficients. t statistics in parentheses. Dummies for regions are not shown. Significance level: * p < 0.05, ** p < 0.01, *** p < 0.001 -0.26 *** (-3.63) -2.14 (1.85) -0.22 (0.25) -0.42 (-0.82) -4.04 ** (2.85) -0.32 (-0.18) -4.99 ** (-2.85) -0.87 (0.84) -0.18 (-0.42) -0.02 (-0.12) -0.46 (-0.64) -0.46 (0.60) -0.47 (-0.72) -1.81 (-1.57) -1.00 (-1.36) -2.18 * (1.97) -0.02 (1.50) -0.01 (-0.41) -0.54 (-0.17) -0.17 (-1.05) -0.27 (0.78) -0.74 * (2.17) -0.18 *** (4.32) -15.01 * (-2.47) 164 0.416 -0.10 * (-3.10) -1.82 (1.60) -0.08 (0.10) -0.43 (-0.88) -4.50 ** (-3.16) -0.86 (0.87) -0.46 (1.18) -0.02 (-0.12) -0.37 (-0.53) -0.53 (0.75) -0.58 (-0.93) -1.36 (-1.29) -1.14 (-1.77) -1.24 (1.29) -0.01 (1.06) -0.00 (-0.14) -1.59 (-0.53) -0.02 (-1.59) -0.06 (-0.21) -0.74 * (2.26) -0.16 *** (4.22) -6.46 (-1.26) 164 0.390 -1.88 ** (-2.92) -1.66 (1.48) -0.13 (0.16) -0.48 (-0.95) -3.00 * (2.35) -0.19 (-0.11) -4.82 ** (-2.80) -0.71 (0.71) -0.29 (-0.72) -0.06 (-0.40) -0.59 (-0.85) -0.37 (0.50) -0.51 (-0.80) -1.59 (-1.40) -1.12 (-1.59) -1.93 (1.73) -0.02 (1.42) -0.01 (-0.28) -0.48 (0.15) -0.02 (-1.32) -0.23 (0.70) -0.73 * (2.21) -0.16 *** (4.22) -13.47* (-2.33) 164 0.387 74 3-3 Comparative Case Studies: Taiwan and South Korea 3-3-1 Ruling Parties In addition to the cross-national evidence, I use comparative historical analysis to track the causal mechanisms (Mahoney 2001, 2007). In particular, I utilize the Most Similar Systems Design to compare Taiwan and South Korea. That is, the cases should be similar in many ways except the variables of interest, including the main explanatory factor (the main perceived threat to the leader) and the dependent variable (party institutionalization) (Gerring 2017). In South Korea, Syngman Rhee ruled from 1948 to 1960, while Park Chung Hee took power by coup after a short-lived democratic government in 1960, and he had ruled until being assassinated in 1979. Taiwan’s authoritarian regime is relatively stable in that Chiang Kei-shek and his son Chiang Ching-kuo both ruled as president for life. The dictators of Taiwan and South Korea share similar characteristics, especially strong state capacity, control over coercive institutions, and alliance with the U.S. However, the dictators chose completely different strategies on developing the ruling parties. Scholars illustrate that in Taiwan, individual parties and the party system are highly institutionalized by any criteria (Cheng and Hsu 2015). The parties are enduring, resilient, and well embedded in their support bases, and this high level of institutionalization occurred before the onset of democratic transition (Cheng 2006). The process of party institutionalization in Taiwan began in the early 1950s right after the KMT was defeat by the Chinese Communist Party (CCP) and retreated to Taiwan (Dickson 1993). For the reform of the party, Chiang Kai-shek established the Central Reform Committee (CRC) to reorganize the party, replacing the inept Central Standing Committee and Central Executive Committee (Meyers and Lin 2007). CRC recruited young and highly educated party members and decided to set the ideological foundation based on Sun Yat-sen, the 75 founding father of the party and the state. The attempt to reform party structure did not succeed in the first place due to the looming threat of CCP. It was the breakout of the Korean War that fundamentally changed the U.S. foreign policy (Accinelli 1996). The Truman administration sent the Seventh Fleet to guard the Taiwan Strait, and the Eisenhower administration signed a defensive treaty with Taiwan that guaranteed U.S. protection of Taiwan out of the need to defend communist expansion. The following military and economic assists stabilized the Taiwan Strait and gave Chiang a solid foundation to implement his policies; thus, a rejuvenated KMT turned Taiwan into a quasi-Leninist regime (Cheng 1989, 2006). On the other hand, parties in Korea remain weakly tied to the public and still have strong regional or personal bases, and voters’ party vote choice is still volatile until today (Tan et al. 2000; Wong 2015). Not only the authoritarian ruling parties under strongmen collapsed or reorganized following the fate of leaders but also parties in the democratic regime change names and re-aligned with every new party leader. Park Chung Hee’s Democratic Republican Party (DRP) and Chun Doo Hwan’s Democratic Justice Party (DJP) are both run by military elites. After Park took power by coup in 1961, he established the Korean Central Intelligence Agency (KCIA). He appointed his right-hand man Kim Chong P’il to lead KCIA. At the same time, while Park started to outlaw political activities, Kim Chong P’il “clandestinely and illegally” organized a new ruling party DRP as a preparation for Park to participate in the electoral campaign (Kim B.K. 2011). Kim Chong P’il originally planned to develop DRP as a “Leninist party” like KMT did, but this backfired with some military factions. In 1963, Kim resigned all of the posts and went into exile under pressures from the military. Since then, the plan of institutionalizing the DRP was never realized, as Park Chung Hee concentrated power on himself (Hoon 2015; Kim and Vogel 2011). After Park was assassinated in 1979, Major- 76 General Chun Doo Hwan took power by a coup in 1980. He dismissed all of political parties of the previous regime, and even turned the presidential election into an indirect one (Kim H.A. 2011). From Park to Chun, military factions who controlled the party all lacked vision and programs and all of them did not develop grassroots organizations for mobilization (Kim H.M. 2011, 6). The DRP was dismissed by the coup leader, while DJP merged with other two parties in 1990, soon after democratization. As mentioned in Chapter 1, several indicators of the outcome of party institutionalization show that the KMT in Taiwan is highly institutionalized under the rule of Chiang Kai-shek and Chiang Ching-kuo. In contrast, the ruling parties in South Korea under Park Chung Hee and Chun Doo Hwan are far less resilient (see Table 3-3). Table 3-3 Variations of Ruling Parties in Taiwan and South Korea Country Year Type (frequency) of Elections South Korea (1967-1988) Legislative (8) Taiwan (1954-1989) Source: counted from various sources by the author. Local (10) Mean Volatility (%) 5.15 2.32 Mean Vote Share (%) 39.39 67.18 Party Switch (%) 12.4 1.6 Survive 1 Yr After First Leader Exits N Y The KMT maintained high levels of vote share (67.2% on average), and support is stable (2.32% of volatility). Congressional elites seldom switch membership (1.6%), and the party survived over leadership turnover. However, the ruling parties in South Korea could not even win a majority of votes in the general elections (with 39.4% mean vote share and 5.15% electoral volatility), and they failed to survive the leadership turnover. In general, dictators in the two countries chose different strategies on the party’s 77 organizational development. I argue that the critical difference that led to such decisions is the nature of the main threat to political survival. 3-3-2 Similarities After WWII, Taiwan and South Korea shared similar patterns of political and economic developments.19 They were both freed from Japanese rule, faced a divided nation and an enormous security threat from the other side (China and North Korea respectively), closely tied to the U.S. as an anti-Communism alliance, sharing similar paths of economic modernization and known for high levels of state capacity, having relatively homogenous ethnic/linguistic/cultural society, and democratized after long- term dictatorships in the “third wave” of democratization. These similarities help us to eliminate possible rival hypotheses on why the ruling party institutionalized in Taiwan while not in Korea. In particular, the resource availability theory, adversity-driven theory, and the domestic threat factors all cannot explain such variation. According to the previous hypotheses, the two countries should have similar levels of party institutionalization, as the advantages of state capacity and social homogeneity either help (resource availability theory) or impede (adversity-driven theory) the leader to institutionalize the party. But the ruling parties in the two countries show significantly different developments. First, both countries have a high level of state capacity, and the dictators are political strongmen. The state capacity may be traced back to Japan’s colony before WWII. Historically, Japan colonized both Taiwan and Korea and left specific positive legacies that led to postwar economic development. Japan had colonized Taiwan since 1895, 19 The official name of the two countries are Republic of China and Republic of Korea respectively. 78 following China’s defeat by Japan in the First Sino-Japanese War (1894–95) and was made official by the Treaty of Shimonoseki. The Japanese saw Taiwan as a raw materials supplier for Japan’s industries, a market for domestic goods, the agricultural basement, and a model of economic growth (FRD 2005a). They built one of the most advanced- infrastructure and bureaucratic systems in order to obtain more resources. Later, during WWII, the wartime economy brought more construction, growth of the heavy industry, the use of modern technology, and development of a skilled industrial labor force (FRD 2005a). At the same time, Japan also turned Korea into its colony in 1910 following the Russo-Japanese War of 1904–05 (FRD 2005b). Both countries had relatively modernized infrastructures and efficient bureaucracies developed on the basis of colonial legacies (Amsden 1989; Gulati 1992; Evans 1995; Riedl 2014; Kohli 2004). To this extent, the two countries share similar relationships with the U.S. After WWII, due to the need to halt the spread of Communism, and as a result of of “containment policy,” especially after the breakout of the Korean War, Syngman Rhee and Chiang Kai- shek received American military support and massive amount of foreign aid, with formal defense treaties signed (Gaddis 1982). Out of geopolitical interests, the U.S. has maintained troops in South Korea under the U.S.-R.O.K. Mutual Defense Treaty. The U.S. also signed a defense treaty with Taiwan20 while turned the commitment to Taiwan into the “Taiwan Relations Act” as domestic law in 1979 after establishing diplomatic ties with China. For both countries, they still face divided-nation situations and enormous security threats from the other side until today, with North Korea and China respectively. However, with support from the U.S., Syngman Rhee and Chiang Kai-shek were able to consolidate their rules. 20 The official name is “Mutual Defense Treaty between the United States of America and the Republic of China”, signed in 1954 and become effective in 1955. 79 The U.S. aid, investments, and open markets fueled the economic development of two nations (Gulati 1992; Cullather 1996; Noble 1998; Latham 2011). In both countries, U.S. strongly supported a “developmental autocracy,” and two nations became classical cases of “capitalism developmental states” (Johnson 1982; Deyo 1987; Amsden 1989; Wade 1990; Öniş 1991). The state bureaucracy is autonomous and highly efficient with a centralized planning authority that intervenes economy (Evans 1995). Since the 1960s, after a military junta led by General Park Chung Hee, who took power by coup, the Korean government established the Government Economic Planning Board that promoted and implemented an export-oriented economy. In Taiwan, the leading bureau for economic development is the Council for Economic Planning and Development. In these institutions, bureaucrats obtained a high level of trust from the dictators, as well as helps from American experts. Furthermore, because of the political support from the U.S., the leaders were able to exclude major oppositional groups from sharing political power. Thus, the bureaucrats were able to make practical economic plans and efficiently implement export-led policies. This model of cohesive-capitalist states has successfully brought stable and high level of economic growth in many East Asian countries during 1960s and 1980s, especially Japan and the “four little dragons”: Taiwan, Korea, Singapore, and Hong Kong (known as “four tigers” in the English context. See Vogel 1991). They all shared some of the core characteristics of the capitalist developmental states. The World Bank termed the success as “The East Asian Miracle” (Page 1994; Gold 1986). Furthermore, the success of these economies does not depend on natural resources. Both South Korea and Taiwan do not have abundant resource-poor and are highly dependent on importing raw materials (Sachs and Warner 2001). Taken together, both Taiwanese and South Korean leaders enjoy high levels of state capacity and maintain control over the society. According to resource availability theory, 80 both countries should have observed highly institutionalized ruling parties because the leaders have the capability to develop party organizations. On the contrary, according to the adversity-driven theory, leaders in both countries should be reluctant to institutionalize the party because they can rely on other approaches to rule. Both theories cannot explain the variations of ruling party developments between two states. Besides the similar conditions of state capacity and political dominance of the leader, the domestic conditions regarding threats to the leader even predict more institutionalized ruling parties in South Korea but not in Taiwan. First, although South Korea is highly homogenous in ethnic structure, it has a higher level of religious polarization (0.78) than Taiwan (0.59) (Reynal-Querol 2002). Furthermore, some indicators for mass-based threats show that, there were 15 mass riots in Taiwan from 1968 to 1987, averaging 0.75 per year.21 On the contrary, South Korea had 81 riots from 1960 to 1987, averaging 2.9 per year. For the elite-based threats, there was no coup attempt in Taiwan,22 but both leaders, Park Chun Hee and Chun Doo Hwan seized power by coup. That is, the domestic threats from the masses and the elites should be more challenging to Korean dictators, and thus, they should have put more resources in the party organizations at the local level. However, the ruling parties in the two countries show the opposite developments. Taken together, the potential explanatory factors of party institutionalization are either similar in two countries, or predict higher levels of institutionalization in South Korea. Thus, these factors cannot explain the distinct developments of the ruling parties in the two countries. 21 Data for number of riots is from Banks Cross National Time Series (CNTS) Data Archive. In the dataset, riots are defined as “any violent demonstration or clash of more than 100 citizens involving the use of physical force.” 22 The number of coups is from Powell and Thyne (2011). 81 A summary of the similarities between the two countries on the key factors of party institutionalization is in Table 3-4. All potential factors from the conventional wisdom of resource availability theory and the adversity-driven theory are similar in the two countries. From the threat-based theory, the domestic factors are either similar in two countries or predict a higher level of party institutionalization in South Korea (religious polarization, and risk of riots and coups). Thus, the remaining factor most likely to explain the variation of party developments is the nature of the foreign threat. Table 3-4 Values on Key Variables of Interest, Taiwan and South Korea Predicted level of institutionalization Taiwan = South Korea Taiwan = South Korea Taiwan = South Korea Taiwan = South Korea South Korea > Taiwan South Korea > Taiwan South Korea > Taiwan ? Taiwan > South Korea Taiwan (1949-87) High High Low Low (0.36) Medium (0.59) Medium (0.78) Less Low risk China South Korea (1963-87) High High Low Low (0.01) More High risk North Korea Case (Years) Factors Resource / Adversity theory State Capacity Economic growth Natural resource Domestic threats Ethnic polarization Religious polarization Riots Coup Foreign threats Real level of institutionalization 3-3-3 Difference: the external threat Both Taiwan and South Korea face strong external enemies, China (the Chinese Communist Regime) and North Korea, respectively, and the two cases had experiences of armed conflicts with their foreign enemies. However, I argue that the level of perceived 82 threats from outside is the key difference that leads to the variations on party institutionalization. The main difference originates from the salience of the threat on the territorial claim. First, South Korea is geographically adjacent to North Korea, while the island of Taiwan is separated with mainland China by the Taiwan Strait. The capital of South Korea, Seoul, is only 34.8 miles (56 kilometers) away from the border. For South Korean leaders, the threat from North Korea is always the most urgent threat to political survival, as Seoul is the political and economic center. For Taiwanese leaders, after the breakout of the Korean War in 1950, the U.S. started to defend Taiwan, which broadly guarantee the security of the Taiwan island (and Penghu, which located in the Taiwan Strait). Although several seacoast small islands in Southeast China still face a considerable challenges from the Communist party,23 there is no immediate threat to the central government in Taiwan. The U.S. also helped to defend South Korea with a defense treaty and military bases. However, it cannot buffer the potential risks due to geographical proximity from North Korea. The priority of Korean leaders is to maintain a strong army against the North Korean army, while the top goal of Taiwan is to keep an elite-level of the navy and the air force. The size of the army becomes secondary in Taiwan’s defense, as the battle will occur at the sea and on the air. According to the National Material Capabilities (v5.0) dataset (Singer, Bremer, and Stuckey 1972; Singer 1987), from 1960 to 1987, the last year of Chiang Chin-kuo and Chun Doo Hwan, the growth rate of military expenditure and decrease rate of military personnel both show that Korean regimes need a strong military. The size of military personnel in Taiwan decreased from 600,000 to 406,000, a 32.3% decrease. Korean army 23 In particular, Quemoy, Matsu, and nearby islands. 83 maintained 626,000 troops in 1987, a slight decrease of 3.7% from 650,000 in 1960 (see Figure 3-6). For the military expenditure, the two countries share similar growth rate before the mid 1970s. The sharp increase of military expenditure in Taiwan might mainly reflect the switch in diplomatic recognition of the U.S., from the Republic of China to the People’s Republic of China. But after the 1980s, the Korean expenditure keeps growing. Compared with 1960, the Korean military expenditure increased from 98 million to 8.5 billion US dollar,24 an increase of 85.3 times. Taiwan’s expenditure only increased 29.1 times from 222 million to 6.7 billion, growing at a slower pace than in Korea. In short, South Korean regimes had maintained a high level of military personnel and expanded the budget faster than in Taiwan. Figure 3-6 Comparison of Number of Military Personnel 24 The currency is calculated in the year of the data. 84 3-4 Comparative Case Studies: Indonesia and The Philippines 3-4-1 Ruling Parties The comparison between Taiwan and South Korea shows that the external threat weighs more than internal threats when it comes to choices between developing party organizations and the army. In this section, I further utilize the Most Similar Systems Design to compare two leaders, Indonesia’s Suharto (1967-1998) and the Philippines’ Marcos (1965-1986), where there are no immediate foreign threats to the territory so that one can locate which type of internal threat matters for party institutionalization. President Suharto in Indonesia built up a robust ruling party Golkar, while in the Philippines, Marcos’ Nacionalista Party (1965–1978) and Kilusang Bagong Lipunan (1978–1989) remain low-level of cohesion, autonomy, and organizational development (Hellmann 2011). I argue that it is the complicated religious polarization in Indonesia that prompt Suharto to develop and maintain an extensive party organization at the local level in addition to the coercive tools. Both Indonesia and the Philippines became independent after WWII. The Philippines held democratic elections in 1965 and 1969, but President Marcos declared Martial Law in 1972, closing the elected congress, and turning the state into an authoritarian regime. He ruled until his resignation in 1986, in response to massive demonstrations against the rigged election. In this period, the Philippines is categorized as a personalist regime (Geddes 2003). In Indonesia, Suharto took power in the March 1966 coup, followed by a purge of left-leaning bureaucrats and the Communist Party (PKI) (Crouch 1978). Suharto built up a personal-military regime and ruled for 32 years until resignation after the competitive parliamentary election won by the opposition in 1998. However, his authoritarian regime still survived because he transferred power to his 85 allies, and Golkar maintained one of the major parties in the first three parliamentary elections in the post-Suharto era (Tan 2015). Under Suharto’s New Order, the ruling Golkar became an effective electoral machine, penetrating to the remotest parts of the country (Tomsa 2008). When Suharto began to introduce the New Order in 1966, political elites adopted repressive apparatus to depoliticize the society (Ufen 2008). Most of the political parties were banned or forced to merge into only two opposition parties, PDI (Indonesian Democracy Party) and PPP (Development Unity Party). At the beginning (1964), Golkar was an assemblage of several functional groups, not a real political party. After Suharto officially took power in 1968, he made “conscious choices” to structure the party system (Hicken and Kuhonta 2015), while he tried to cut linkages between parties and local strongmen, and to manipulate internal factionalization. Backed by the military, the bureaucrats, and the big business community, Golkar built up nationalized party branches, integrating local elites in every region. In 1971, the well-organized and ample-funded Golkar won 62.8% vote share and 66% seats (236 out of 360) in the first national parliament election. With the strength of economic growth as a carrot and state coercive sticks, it successfully transformed into a highly institutionalized election machine, consistently obtaining 60-75% of the vote in elections held regularly from 1971 to 1997. In the 1980s, during the chairmanship of Sudharmono (1983–88), Golkar further establishes a completely centralized command system. Suharto maintained his ultimate power by chairing the Supervisory Council (Dewan Pembina) with strict top-down decision-making processes (Tomsa 2008). On the other hand, in the Philippines, Marcos did not rely on the party to rule, while the dictator’s party could not even make orders go beyond the presidential office (Hicken 2009, 2015; Lacaba 1995). The ruling party does not have administrative structures— parties usually do not even have national headquarters, not to mention the local ones. In 86 individual local politicians, congress, the leader of the party cannot exert pressure on congressmen to ensure voting on the partisan principle. Slater (2010) illustrates that Filipino elites failed to induce collective action, while the military and the party did not develop a collective will to rule as in Indonesia during the same period. Financial and political resources were all handled through instead of informal connections between bureaucracies and parties (Hellmann 2011). Several common indicators of party performances show a significant difference between the two leaders (see Table 3-5). In Indonesia, Golkar maintains a high level of support, with 67.5% vote share on average and 2.31% of mean volatility. That is, Golkar’s popularity in the election is highly consistent. On the contrary, Marcos’ ruling party suffers fluctuations of vote share with 15.73% of volatility. Also, on average, more than one fifth (22.3%) of legislators switch their party between each election, which shows that the party lacks an autonomy to unite ruling elites. In Indonesia, only 4.6% of congressperson switch parties between elections. These number shows that Suharto had an institutionalized party while Marcos did not develop the party organizations. Table 3-5 Variations of Ruling Parties in Indonesia and The Philippines Country Year Type (frequency) of Elections Legislative (6) Mean Volatility (%) 2.31 Indonesia (1971-1997) Philippines (1965-1986) Legislative (4) 15.73 Source: counted from various sources by the author. Mean Vote Share (%) 67.50 56.75 Party Switch (%) 4.6 22.3 Survive 1 Yr After First Leader Exits Y N 87 3-4-2 Similarities The two countries share several common characteristics, especially the factors from party institutionalization theories discussed in Section 3-1. Most important of all, the two countries do not have immediate threats from a foreign enemy that endangers the territory, which makes them suitable cases for comparative studies of internal threats. Second, the two countries are both topographically and socially diverse. Both countries are island states with geographically dispersed territories. The Philippines comprises an archipelago of 7,107 islands located off Southeast Asia (FRD 2006), while Indonesia is the largest archipelagic nation in the world, encompassing more than 17,000 islands (FRD 2004). As a result, both have high levels of cultural and language diversity. According to Alesina et al. (2003)’s index of fractionalization ranged from 0 to 1, Indonesian language diversity is 0.76 (in the year of 1990) compared with The Philippines’s 0.83 (1998). Also, the two countries have similar medium-high levels of ethnic polarization, Indonesia 0.53 and The Philippines 0.53 (Reynal-Querol 2002). These two dimensions of social heterogeneity are thus ruled out as an explanatory factor of variations of party institutionalization. Furthermore, the two countries experience similar levels of domestic violence, as the frequency of riots is 1.65 cases per year in the Suharto era and 1.72 in the Marcos era (Banks 2011). That is, the threats from the masses may be similar to the leaders in terms of ethnicity, language, and frequencies of mass riots. Taking into account the resource-abundant or scarcity theory, the initial conditions of the economic development of two leaders are similar, as both countries are categorized by the World Bank as the “lower middle income” group. When the two dictators took power, the Philippines has a higher level of GDP per capita than Indonesia, 1943 to 1370 US dollars, according to the Maddison Project (Bolt and van Zanden 2014). Both countries are abundant in natural resources. Indonesia is rich in the stock of oil, natural gas, timber, 88 and valuable minerals, while The Philippines averages even more on resource-income per capita, 42.3 to 25.6 US dollars, when Marcos took power. That is, according to the resource-abundant or scarcity theory, both leaders should have similar resources on building party organizations, while Marcos has slightly more advantages over Suharto. The real party developments do not reflect the natural resource level of the two countries. As the two countries are rich in natural resources, both leaders are good at mobilizing public funds (Hutchcroft 2011), although a great portion of the resource went into personal pockets. Marcos and Suharto’s families were all in the center of corruption, which is one of the obstacles for democratization in Southeast Asia. According to the report published by Transparency International,25 Suharto and Marcos ranked the top two most corrupt leaders in the world, estimated by funds allegedly embezzled, 15 to 35 billion US Dollars, and 5 to 10 billion respectively. Although the initial conditions in terms of economic development and available resources are similar to the two leaders, their strategic choices of utilizing the resources are quite different. Suharto utilized the resource he has to build up an effective bureaucracy and a penetrating party machine, while the Filipino state machine remained weak (Hicken 2009). That is, the resource- abundant theory cannot explain the difference of party institutionalization in the two countries. In terms of ruling strategies, the two leaders also have similar strategies on the power centralization and utilization of coercive apparatus. Both dictatorships deliberately repress civil society and extend the role of military elites. Since the introduction of Martial Law in 1972, Marcos dismissed existing parties and closed the congress. He expanded the military from 65,000 to 270,000 personnel by the end of 1975, inserting military offices to 25 Global Corruption Report 2004. https://www.transparency.org/en/publications/global-corruption- report-2004-political-corruption. Accessed: March 2020. 89 various public and private corporations (Dios 1988). In Indonesia, coercion was also a key factor in the first phase of the Suharto regime, starting with violent destructions of the political left (Roosa, 2006). More than half a million people killed in the anti-communist purges of 1965 and 1966 (Cribb, 1990). Suharto’s takeover was accompanied by the military’s acquisition of the whole state. For example, military officers occupy about 80 percent of the positions of mayor, governor, and other local leaders in the early 1970s, rose from 20 percent in 1965 and 54 percent in 1969 (Bresnan 1993; Mietzner 2018). Case (Years) Philippines (1965-1986) Factors Resource / Adversity theory State Capacity GDP level (1st yr of leader) Natural Resource Domestic threats Table 3-6 Values on Key Variables of Interest, Indonesia and The Philippines Predicted level of institutionalization Indonesia = Philippines Indonesia < Philippines Indonesia < Philippines Indonesia = Philippines Indonesia > Philippines Indonesia = Philippines Indonesia = Philippines Not infected Indonesia (1967-1998) Medium->High Medium Lower middle Abundant Medium (0.53) High (0.82) Riots 1.65/per yr Coup Low risk after 71’ Higher Foreign threats no Real level of institutionalization Similar (higher) (higher) Medium (0.50) Low (0.21) 1.72/per yr Ethnic polarization Religious polarization Indonesia > Philippines no Taken the together, among all of influence party institutionalization, the initial conditions of the two leaders are similar. Table 3-6 lists the values of key variables of the comparison. Marcos and Suharto both adopt similar ruling strategies as well, attempting to centralize power and utilizing coercive institutions to that could factors 90 rule. To explain the diverse developments of the ruling parties, we can rule out factors of resource/adversity theory and foreign threats. Also, many conditions of the domestic threats are similar in these countries, including ethnic polarization, language fractionalization, and frequencies of riots. Therefore, these similarities cannot explain party institutionalization. 3-4-3 Difference Among all of the factors that could influence party institutionalization, the main difference between Suharto and Marcos is the level of religious polarization they face. The index of religious polarization (Reynal-Querol 2002) shows that religions in Indonesia is highly polarized (0.82), while it is low in The Philippines (0.21). Another calculation by Basedau, Pfeiffer, and Vüllers (2016) show that religious polarization in Indonesia in 1990 is 0.57, while in The Philippines is 0.34. 26 I argue that religious divergence is a key factor that Suharto maintains the extensive party organizations. Indonesia has a long history of various religious developments. In the seminal study by Geertz (1960), he describes four variants of Islam in Java: the animistic abangan, the orthodox Islam santri which further divided into traditionalists and modernists, and the more Hinduized priyayi. The typology is also consistent with the social class cleavage. The abangan were associated with ordinary villagers in the countryside; the santri petty were business traders and small farmers; and the priyayi were mostly state officials. The religious bases also form the foundation (aliran, Indonesian for stream or current) of Indonesian parties, as PKI for abangan, PNI for priyayi, Masyumi for modernist santri, and 26 In the study of Reynal-Querol (2002), the mean of religious polarization is 0.47 and standard deviation is 0.36. In Basedau, Pfeiffer, and Vüllers (2016), religious polarization has a mean of 0.47 and 0.29 standard deviation. 91 NU for traditionalist santri. Although Geertz’s study only covers Javanese Islam, half of the national population, the framework of religious diversity is generalized to the whole country to understand the evolution of the party system (Liddle and Mujani 2007). Suharto upheld the division of religion artificially during the New Order, while in 1973, he forced all Muslim parties into PPP and secular nationalist plus Christian parties into PDI. To minimize the significance of the established aliran structures, Suharto attempted to de-ideologize and de-politicize the society and party politics. For example, the New Order regime made into law the official ideology of all Islam parties and organizations, known as asas tunggal legislation in 1984, aiming at softening the existing Islam divisions (Tomsa 2008). However, as King (2003) noted, the attempts were not successful as the religious division is still one of the most significant factor in explaining party politics and voting behaviors before and after the New Order. Taken together, the religious division has long-term impacts on party politics and the leaders’ strategic choices (Liddle and Mujani 2007). On the other hand, in The Philippines, the formation of identity and mobilization of ethnic groups were not defined by religion, because Christianity had already widely spread nationwide in the Spanish and American colony period (Ferrer 2005). In most regions, Christianization coexists peacefully with the indigenous belief systems and practices. In the remote areas, it is indigenous cultures rather than religion that distinguish groups of people from the majority Filipino. The major internal conflicts in The Philippines often come from resentment toward the central government. In many prominent struggles, including Cordillera in the north, insurgency rises as a result of the Marcos’ government’s policies and programs toward the indigenous people. For example, the Cordillera people rally to against the building project of hydroelectric dams in the 92 mid-1970s and to protect their culture, which later developed into armed resistance movements. In the Bangsamoro region (Muslim Mindanao) in the South, it is indeed that the Moro people have a long history of resistance against the rules from the central government, even since the Spanish colony. Religious discrimination and growing economic and political grievance evolved into armed conflicts and separatism in the region (Kamlian 2004). After the independence of The Philippines, the central government maintained policies to integrate and assimilate the Moro people in the Bangsamoro area. The Marcos government encouraged Christian settlers in this region to fight against the Muslims, which further deteriorated the relationships between the central government and local groups. In 1972, right after the declaration of the Martial Law, the Moro National Liberation Front (MNLF), an armed organization, initiated attacks on the Armed Forces of The Philippines in three Southern provinces. However, starting in 1976, Marcos significantly changed the policy toward Bangsamoro, giving more political autonomy to the region. The shift was a result of the armed conflicts initiated by MNLF as well as international pressures, especially from the Organization of Islamic Conference, an organization for Islam states in the world (Kamlian 2004). Taken together, although the Bangsamoro movements mainly consist of Muslim collective action against the Christian majorities, it has long been treated as a movement for separatism or for political autonomy rather than religious or ethnic conflicts. Also, the conflict occurs in specific regions without developing into nationally religious confrontations. Therefore, this internal threat does not translate into the leader’s decisions of developing party organizations for the masses. Religious polarization is the main difference between Suharto and Marcos which leads to different levels of perceived threat, 93 while the priority of building up party organizations is higher for the New Order regime but of marginal importance for Marcos. 3-5 Conclusion Cross-national empirical evidence shows that dictators build up local organizations of the ruling party for the purpose of neutralizing threats from the masses. When the level of religious polarization is high, the leader tends to institutionalize the party. However, when there exist foreign enemies with territorial threats, the leader must prioritize strengthening the coercive institutions, especially the armed forces. Therefore, the presence of territorial threats impedes ruling party’s development at the local level. The cases of South Korea and Taiwan confirm the causal mechanism found in the cross-national study that foreign enemies with territorial claims are an obstacle for party institutionalization. The two countries have similar patterns of political and economic developments, acquired alliance and assistance from the U.S., became effective ruling governments, and the mass uprising was fierce in South Korea. Therefore, these factors cannot explain the different ruling strategies for developing the party. Ruling out all of the potential explanation from domestic politics, the perceived threat from outside of the country becomes the most viable factor that explains levels of party institutionalization in the two cases. It was the foreign enemy with an immediate territorial threat that prompts the Korean leaders to maintain a strong army to ensure political survival so that the party organizations got less attention and resources. On the contrary, Taiwan’s KMT had been on the track of institutionalization since the U.S. security commitment in the Taiwan Strait. 94 The cases of Indonesia and The Philippines show that religious polarization does matter. The two countries do not have an immediate foreign enemy with territorial claims. They are similar in many factors that could influence party institutionalization when the two leaders, Suharto and Marcos, took power, including state capacity, natural resource, ethnic polarization, and frequencies of domestic riots. However, the religious cleavage is salient in Indonesia nationally and the distribution of religious groups is close to what has been described as prone to conflicts—the level of polarization is high. On the other hand, in the Philippines, religious cleavage is not the main motivation behind the major conflicts, and religious polarization level is low. Therefore, party institutionalization is one of the ruling priorities for Suharto but not for Marcos. Taken together, the large-n studies and the comparative case studies confirm the hypotheses that the leaders institutionalize the party when (1) there is no spatial rivalries from outside, (2) the masses are perceived as threats, and (3) the relative salience of external territorial threats over internal threats is low. 95 CHAPTER 4 AUTHORITARIAN RULING PARTY INSTITUTIONALIZATION AND ITS CONSEQUENCES ON POLITICAL SURVIVAL The establishment of some nominally democratic institutions contributes to longer political survival of autocracies. Political parties, for instance, are particularly powerful ruling tools. In general, party-based regimes are more durable than others (Magaloni and Kricheli 2010), as the party manages intra-elite conflicts (Magaloni 2008; Boix and Svolik 2013), facilitates co-optation for potential oppositions (Svolik 2012; Reuter and Turovsky 2014), prevents elites defection by assuring long-term gains (Brownlee 2007; Slater 2010), and credibly distributes social welfare benefits to all of people (Knutsen and Rasmussen 2018). Moreover, regarding the diplomatic relationship between one autocracy to others, party-based regimes are less likely to initiate interstate conflicts because the party holds leaders accountable (Weeks 2012). While studies show that party-based authoritarian regimes perform better than other types of dictatorships, there are substantial variations in political performance within this regime type. Most notably, some party-based regimes rule the country for more than half of a century, such as Mexico’s PRI (Institutional Revolutionary Party, Partido Revolucionario Institucional), and China’s Communist Party (CCP). On the contrary, some party-based regimes are short-lived, such as in Afghanistan (1996-2001) and Colombia (1950-1957). The variations of leaders’ political survival within party-based regimes may be explained by the function of political parties. In particular, the organizational development of ruling parties vary significantly. Some ruling parties are just ceremonial, such as the Nigerian Progressive Party – African Democratic Rally (1960–1974, PPN- RDA), while some ruling parties are highly integrated with society by establishing local 96 branches in even remote areas, such as Taiwan’s KMT (Kuomingtang), Mexico’s PRI, and China’s CCP. What characteristics of a party’s development leads to the longer political survival of an authoritarian leader? This chapter examines whether levels of ruling party institutionalization explain differences in longevity for leaders’ political survival. Party institutionalization is a process that individual political parties experience value infusion and organizational stability (Panebianco 1988; Levitzki 1998; Basedau and Stroh 2008), in which it develops regular and patterned party organizations, sub-units, and activities. The scope of this study includes ruling parties in all of the dictatorships, because political parties exist in all kinds of regime types. literature on authoritarian ruling party institutionalization is a key factor for longer political survival (e.g., Nathan 2003). According to previous studies, party institutionalization should be able to enhance the ruling performances, because it provides credible commitments to elites and the masses, facilitates direct interactions between party cadres and people, and produces policies with better qualities (Bizzarro, Hicken, and Self 2017). In particular, party institutionalization at the local level helps the dictatorship gain legitimacy, because it increases “inputs” to the regime. As a result, party institutionalization enhances people’s political efficacy and local governments’ responsiveness (Nathan 2003). that The resilience illustrates However, existing literature on party development and the outcomes have several gaps. First, at the conceptualization level, some studies use regime type or simply the number of parties to explain variations of performances and do not measure the organizational development of the ruling party. Yet, it is problematic to assume that the party has the same weight in different regimes. Likewise, the number of parties is often determined by the ruler, as many autocracies are not open to multi-party competition. 97 it suffers a major problem: Therefore, this kind of variable may suffer from endogeneity problems. Second, at the measurement level, the discussion does not differentiate party strength and party institutionalization. For example, Meng (2019) interchangeably uses the term party strength and party institutionalization. She finds that party strength—measured by whether a party goes through a leadership transition and political survival after the departure of the first leader—is associated with better regime outcomes. This measurement may be problematic because party strength is itself a consequence of party’s institutionalization effort. Some studies measure the party system instead of the individual parties, while the coding of party institutionalization is according to activities of all parties in the party system, not for the individual ruling party. Third, at the empirical level, cross-national studies on authoritarian regimes are scarce and often limited in the number of cases. Most studies on party institutionalization focus on democracies (Tavits 2013; Yardimci-Geyikçi 2013; Bolleyer and Ruth 2017), while studies on party institutionalization in autocracies remain under-investigated. So far, there is no direct test regarding party institutionalization and political survival. Furthermore, existing studies focus less on political leaders. Instead, most studies use the regime or country-year as the unit of analysis. Therefore, the strategic choices of individual leaders’ ruling tools are seldom discussed in the literature. This study proposes another new measurement of party institutionalization--the establishment of local party branches--to examine the outcome of the organizational development of authoritarian ruling parties. This factor is the most direct approach for measuring party institutionalization, and it avoids the endogeneity issue. I argue that the local organizational presence nationwide can solve several difficulties for the quality of governance and accordingly enhance chances of political survival. First, staff in local party organizations can interact with the masses and help dictators overcome the 98 problems of the Dictator’s Dilemma, especially regarding a lack of correct information about the general public. Furthermore, with local branches, the party can distribute material benefits to citizens, and mobilize support more precisely. Local party branches increase political participation and thus strengthen legitimacy of the regime. To test the hypothesis that party institutionalization leads to longer political survival of dictators, this study creates a dataset including all of the dictators who had remained in power for more than one year in the post-War era, covering from 1946-2008. The evidence shows that dictators with institutionalized parties indeed perform better on political survival. The main contribution of this study is to provide a cross-national examination on the effects of party institutionalization on leaders’ political survival. The data include all of the dictators across different kinds of authoritarian regimes after WWII. Also, it improves the existing measurement on party institutionalization. I differentiate party strength (defined by leadership turnover after the first leader leaves power, Meng 2019) with party institutionalization. While party strength may be an easily accessible data which could fill in the gap of data availability, it is itself an outcome of dictators’ choice of ruling strategies, not an ideal factor to explain other performances. This study proves that party institutionalization, as defined by local level developments, is a better explanatory factor for regime outcomes. falls short on explaining political survival. The structure of the chapter is as follows. Chapter 4-1 discusses how the existing study I elaborate why party institutionalization at the local level is an essential factor. Chapter 4-2 empirically tests the hypotheses. Chapter 4-3 provides a discussion on the results and some preliminary findings on several other indicators of regime performances. 99 function of the primary measurement. The 4-1 Party Institutionalization and Political Survival 4-1-1 Why Focusing on Local Level Developments? To explain the effects of ruling party’s organizational development in autocracies on leaders’ political survival, this study proposes to focus on party institutionalization at the local party institutionalization is two-way. For the ruler, the nationwide organizational presence— the establishment of local party branches and maintaining regular activities—can solve the fundamental difficulties for the quality of governance and accordingly enhances chances of political survival. For the masses, the presence of grassroots organizations increases their political efficacy and loyalty to the ruler. First of all, the party’s activities at the local level, including hiring staff and building up branches, and maintaining regular activities beyond the elections, help the party to establish roots in society. With this connections, local staff are thus able to provide the central level elites with essential information for better decisions, while the local level cadres have direct ways to probe the public opinion. That is, local party organization help dictators overcome the problems of Dictator’s Dilemma, especially lack of correct information regarding the general public. With local branches, parties can distribute material benefits to citizens and mobilize supports more precisely. Therefore, party institutionalization at the local level should be positively associated with better outcomes such as longevity of political survival. Second, as Nathan (2003) points out, local level party branches are input institutions that enhances people’s political efficacy. Individuals can use these institutions and tell the officials their opinions and complaints. While there are channels to reflect their opinions to the ruling classes, more and more people believe that they are able to influence the policy process. Consequently, party’s local branches increase the legitimacy local level as 100 of the regime. Taken together, party institutionalization at the local level is an essential factor that contribute to longer political survival of dictators. 4-1-2 How Do Features of Political Parties Influence Outcomes the causal mechanisms of A more detailed discussion on local party institutionalization is as following. Existing studies have shown that the supporting party of the dictator helps the ruler to deal with the relationship with the elites and the masses, and thus it is an essential ruling tool for each leader. Empirical evidence illustrates that party-based regimes, compared with other types of autocracies, perform better in terms of durability, conflict prevention, and economic growth (Magaloni and Kricheli 2010; Gehlbach and Keefer 2011, 2012). However, the existence of political institutions does not guarantee its effects on the ruling outcomes. One of the key factors to explain the variations of regime performances, especially longevity of political survival, is party institutionalization. Figure 4-1 shows the causal chain of how party organization influences the outcome of governance. Most of authoritarian regimes have nominally democratic institutions, including political parties (and elections), but the developments of such institutions vary significantly. Some ruling parties become institutionalized (from phase I to phase II) — the party develops stable and organizational complexity, transfers power from a leader to autonomous subunits, and establishes ties with the society. These organizational developments transform the capacities and incentives of the elites (phase III), and alter ways of interactions between other actors such as the masses (phase IV). Primary mechanisms of phase III include: (1) checks and balance on the leader, (2) credible commitments that ensure unity among elites, (3) ability to collect necessary information from society (Bizzarro et al. 2018). The main mechanisms of phase IV is the “input 101 institutions” that enhances legitimacy among the masses (Nathan 2003). These mechanisms contribute to longer periods of political survival. I. Party Organization II. Party Institutionalization III. Incentives and capabilities of governing elites IV. Interactions with other actors IV. Political Survival Figure 4-1 Causal Chain of Party Organization and Outcome Source: Adapted from Bizzarro et al. (2018: 5) First, the establishments of hierarchical party organizations and formal rules create a certain degree of checks and balance to the leader. Although political strongmen may still have the ability to deviate from the rules, party organizations and explicit rules increase his cost for doing so. Scholars argue that it is the checks and balances—often lacked in autocracies—that keeps the leader accountable, and such mechanisms will attract more private investments and lead to technological innovations (Gehlbach and Keefer 2011; Gehlbach and Keefer 2012; Simmons 2016). That is, party organization constrains the leader and guarantee the safety of economic activities, and thus results in better economic performances. The mechanism of constraint also limits the incentives of leaders to initiate interstate wars, as the dictator has to respond to policy preferences of 102 to avoid being punished domestic audiences is, party institutionalization prevents unitary actions of the leader, as the subunits of the party and other elites are more autonomous in the decision-making process. (Weeks 2012). That Some might question whether setting local branch might backfire and constrain “too much” the power of dictators. It is indeed that institutionalization involves a process that the party organizations gain autonomy and the dictator share parts of his powers. However, the constraints are manageable because the dictator often controls the power of resource distribution, making local elites reliable on the central government. Also, local level activities can be managed in higher levels. For example, in competitive authoritarian regimes, the central elites control the nomination power for the elections and thus the local elites must follow the rules for intraparty competition. From the perspectives of the local elites, the hierarchical structure of the party provides incentives and entries for the development of political career. Therefore, institutionalization is also an approach to unite with local elites. On the other hand, even if party institutionalization constrains the leaders’ political power, it is more likely to have positive effects on political survival. As mentioned above, the “backfire” is actually beneficial as it brings checks and balance to the leader, which lead to higher levels of investments and economic performances. Second, party institutionalization also helps leaders to credibly commit to their “loyal friends” (Magaloni 2008, 1), as the institutionalized party organizations reassures continuity of distributing benefits and jobs to the elites. If there is no such parallel organization, or, if the organizations are not institutionalized, the elites do not know whether the benefits will continue or stop by leadership changes. Therefore, institutionalized party organizations should increase the loyalty of inner circle elites and decrease potential threats of rebellion to political survival. Furthermore, the mechanisms of credible commitment apply to other social groups, not only for core elites. Scholars 103 find that when the leader promises to distribute public or private goods to the masses and specific social groups, especially loyal civilian and military elites, it enhances the cohesion of these important actors around the regime’s survival (Magaloni 2008; Keefer 2008; Blaydes 2010). The positive result is fewer outbreaks of internal conflicts. Third, party institutionalization contributes to longer political survival because party branches and staffs help to elicit information from the society, which mitigates a significant weakness in the authoritarian regimes known as Dictators’ Dilemma (Wintrobe 1998). In autocracies, the ruler does not understand the actual preference of the masses and strength of the opposition because of the coercive apparatus and lack of free and competitive elections. An institutionalized party has branches and staff that can reach out to most places in a country geographically, and to directly contact people. The party professionals help to collect two main kinds of information. The first is regarding popular support for the regime or regime leader. Especially in diverse countries, precise regional data allows the leader to deal with distributional conflicts by targeted policies and public goods provision (Malesky and Schuler 2011). Elites in the inner circle with the capability of linking broader and varied constituencies can respond to people’s demands more precisely. The second category of information regards the potential opposition. With more knowledge about potential threats, rulers can determine whether to eliminate or co-opt the opposition (Gandhi and Przeworski 2007; Wright 2008). The leader cannot solely rely on coercive apparatus to repress potential oppositions, while doing so may be costly and increases risks of coups (Greitens 2016). That is, party institutionalization at the local level enhances the capacity of ruling elites on information collection, which in turn strengthen effectiveness of rent distribution and selections of ruling strategies. 104 Besides the capability of the elites, party institutionalization also changes the way an authoritarian regime interacts with the masses. In particular, party institutionalization at the local level strengthens the legitimacy of the regime, as the local branches and grassroots organizations provide people with “input institutions,” defined as “institutions that people can use to apprise the state of their concerns” (Nathan 2003: 14). According to Nathan, these institutions include rules or laws regulating the ways people express their opinions to the government officials and party agencies. Although there is no free elections and civil liberties are severely restricted, these input institutions at the local level endows political participation and thus become one of the main sources of resilience of China’s CCP regime. With ways that people can express their opinions, they tend to believe that they are able to influence the policy-making-process. Note that in China’s party-state political system, the CCP is the ultimate authority. These input- institutions at the local level are all part of party institutionalization because the party branches are in charge of various levels of the administration. Taken together, with higher levels of political efficacy and political trust on the regime, the dictatorship achieves longer periods of political survival. A related discussion of party institutionalization at the local level is the literature on the effect of decentralization and federalism. Federalism refers to “a constitutionally guaranteed division of competences between territorially defined governmental levels” and decentralization refers to “the autonomy of subnational levels to allocate resources within their jurisdiction” (Biela et al. 2012). Party institutionalization at the local level may have similar effects of decentralization, as the local staffs and party cadres become subnational units of the ruling elites. Therefore, the effects of party institutionalization may be equivalent to decentralization in many ways. Studies point out that the advantage of decentralization is that the policies are more responsive to the need of people (Rodden 105 and Wibbels 2002). On the economic activities, Weingast (1995) argues that decentralization is a key factor for the development of economy because it brings incentives for subnational officials to create environment for business competition. Also, decentralization limits the ability of central government to confiscate and thus protects the investment. However, decentralization also lead to the common pool problem, as every local government has the incentive to over-spending (Eyraud and Badia 2013). In democracies it also blurs the responsibility between local and central governments for the voters. Therefore, Rodden and Wibbels (2002) illustrate that the effects of decentralization on macroeconomic indicators depend on the nature of central-local government, including the level of fiscal decentralization, the nature of intergovernmental finance, and vertical partisan relations. I argue that party institutionalization has less negative effects than decentralization or federalism on the performance of the government, as the development of party organizations do not necessarily increase the power of local-level “territorial defined” administrations. Party organizations do not directly control the local executive and legislative seats, although in some places the political institutions are rubber stamps of the authoritarian ruling party. However, local party activities do help the ruling elites become more responsive to the masses. The local elites affiliated with the party can also compete with each other for higher level jobs, which is similar to the competition effects under the framework of decentralization that subnational units compete for business investments. Taken together, from the discussion of decentralization, one can expect party institutionalization has some similar effects. To sum up, party institutionalization leads to better ruling outcomes on political survival because of checks and balances on the leader, credible commitment from leader to other actors, information collection, and legitimacy enhancement. The main research 106 hypothesis of this chapter is as follows: party institutionalization at the local level is associated with longer political survival. 4-2 Explaining Authoritarian Regime Performance by Party Institutionalization 4-2-1 Hypothesis and Data To examine party institutionalization and its consequences in authoritarian regimes, this study creates a dataset, based on various sources, including all of the dictators who had remained in power for more than one year in the post-War era, covering from 1946- 2008. The unit of analysis is the leader. As mentioned before, small island-countries and city-states with a threshold of 1000 square feet land area are excluded because it is difficult to differentiate party organizations at the “local” level from the central level in these small states. The number of cases in the dataset is 485. To test the hypothesis, I examine an essential outcome of the authoritarian leaders: longevity of political survival. Data for political survival—years in power—comes from authoritarian regime datasets (Svolik 2012; Geddes, Wright, and Frantz 2014, 2018). The main explanatory factor is party institutionalization—whether the leader has local party activities or not. If a leader builds up new local branches or maintains local activities during his term, he is categorized as having an institutionalized party. On the contrary, if the local party organization remains absent or is destroyed by the leader, he does not rely on the political party to rule. Among all of the cases, 54.6% (265) of the dictators have an institutionalized party, while 45.4% (220) do not. 107 Table 4-1 Summary of Variables for Effects of Institutionalization Obs 484 468 483 442 445 427 484 484 483 470 444 431 Mean 0.55 0.37 10.83 15.85 12.65 7.71 0.24 0.23 0.16 0.50 28.31 0.35 0.50 0.48 9.95 1.47 1.62 1.00 0.43 0.42 0.36 0.26 22.02 0.48 Max 1.00 1.00 49.00 20.97 16.91 11.72 1.00 1.00 1.00 0.95 100.00 1.00 Std. Dev. Min 0.00 0.00 1.00 11.71 6.54 5.41 0.00 0.00 0.00 0.00 0.00 0.00 Variable Party institutionalization Party strength Tenure year Log pop Log area Log GDP growth Colony of Britain Colony of France Oil dependency Ethnic fractionalization Polity score Military regime For other possible explanatory factors, I first add a binary variable of the “strong party” as a competing variable. It is defined as the support party of the dictator experiences at least one leadership turnover and survives for more than three years after the departure of the first leader (Meng 2019). In theory, power turnovers are critical junctures to examine whether the party’s organizational practices have become institutionalized. In particular, successful transitions mean that the transfer of political power from the leader to the organizations, making the party autonomous from a specific leader. Meng argues that this indicator reflects the bureaucratization of the ruling parties, and empirically illustrates that strong parties are positively associated with better regime outcomes, including economic growth, coup prevention, and less war onset. Besides party strength, I include several factors that could influence GDP growth and conflicts such as the initial level of economic development, natural resource, regime type, population, and area. Data for log GDP per capita is from Haber and Menaldo (2011). Oil-dependence is defined by Fearon and Laitin (2003) that more than one-third of export revenues from fuels. The military regime is a binary term defined by Banks (2011). The 108 Polity score is from the original -10 to 10 scale converted to a score of 0 to 100. Population and area data are from the EPR project (Vogt et al. 2015). Ethnic fractionalization data is from Fearon (2003). These data show the situation of the first year of a leader. I also include former Great Britain and France colony in binary terms. The summary statistics of all variables are displayed in Table 4-1. 4-2-2 Leaders’ Performances with and without Party Institutionalization Table 4-2 summarizes the significant differences between leaders with and without institutionalized parties. The results show that party institutionalization leads to longer terms of political survival. Leaders with institutionalized parties hold power for 13.05 years on average, while those without local party presence rule for 8.07 years. Local party branches do contribute to political survival. Table 4-2 Party Institutionalization and Political Survival of Leaders Outcome Years in power Local party N 265 220 Yes No Mean 13.05 8.07 S.E. 0.35 0.57 p-value of t-test 0.000 Note: scores of coup/conflict/war are calculated by the number of the event divided by tenure. Table 4-3 shows the results of the analysis from a standard regression model for these outcomes. The regression model shows that the results of the simple t-tests hold: party institutionalization contributes to more prolonged political survival. Note that in the regression model, I added a variable of “party strength,” but it does not significantly influence the results. Leaders with a “strong party”—which is able to survive leadership 109 turnover after the first leader— do not survive longer than leaders without such a strong party. This is not because party strength or bureaucratization of the ruling party are trivial. Rather, I argue that this is simply because the measurement is not ideal, as leadership turnover and longevity of political survival are the outcome of dictatorial rules. Therefore, one cannot operationalize outcomes as explanatory factors. As illustrated in Chapter 2, party strength and party institutionalization are distinct concepts. In the analysis of political survival, the results show that party institutionalization at the local level is a better measurement than party strength. In general, the analysis points out that leaders with institutionalized parties survive longer. The results, however, might be vulnerable to problems of endogeneity. In particular, it takes times for a party to develop its organizations but for some short-lived dictatorships, the leaders do not even have time to institutionalize the party in the first place. So far, there is no instrument variable for the analysis and we can only have a partial solution here: focusing on a sub-sample of the leaders, especially those who have institutionalized party at the beginning of the tenure. As mentioned in Chapter Two, There are four kinds of leaders regarding their actions on party institutionalization: (1) abandon: there are local party activities when the leader took power but are abandoned during his tenure; (2) remain absence: local party branch remains absent throughout the term; (3) maintain presence: local party organization is maintained from the beginning until the time of leaving power; and (4) build new: no local party branches at the beginning, and the dictator build up them at some moment during his rule. The partial solution to the endogeneity issue is to compare the length of tenure between those who have institutionalized local branches in the beginning but abandon the local party organizations, and those who maintain it until exit of power. In general, 16 leaders reverse the institutionalization, while 184 leaders maintain the existing party networks. 110 Table 4-3 Regression Analysis: Party Institutionalization and Political Survival of Leaders (1) Years in power 6.190*** (5.64) 0.126 (0.11) 2.836* (1.99) -0.979 (-1.73) 0.427 (0.20) 3.304** (2.88) 1.089 (0.90) -0.101*** (-4.49) -1.154* (-2.55) -0.106 (-0.26) -3.122** (-2.97) 37.118*** (4.93) VARIABLES Institutionalization Strong party Oil dependence GDP per capita Ethnic fractionalization British colony French colony POLITY Population Area Military regime Tenure year Constant Observations Adj. R-squared 385 0.026 Note: OLS model used, and standard errors Statistical significance: *** p<0.001, ** p<0.01, * p<0.05 are reported in parentheses. With the institutionalized party, leaders hold power for 11.03 years on average with 9.44 standard deviation. Leaders who destroy local party networks survive for 6.81 years on 111 average with 4.58 standard deviation. The t-test comparison shows a significant difference at 0.05 level. Therefore, we may conclude that party institutionalization does indeed contribute to longer political survival of the leader. Leaders who do not maintain existing local party organizations suffer significantly shorter tenure. 4-2-3 Party Institutionalization and the Political Survival of the Party After examining the impacts of party institutionalization on the political survival of each leader, I change the unit of analysis from leader to party. This section will show that developments of local organizations also benefit the political survival of the ruling party. The data of parties’ outcomes are retrieved from Meng (2019), as the list for all of the authoritarian ruling parties that have been in power for more than three years is retrieved from Svolik (2012). The analysis examines three indicators initially used for testing the effects of “party strength”—years in power, leadership transition, and political survival after the first leader. Results illustrate that ruling parties with and without local level presence vary significantly on their outcomes (see Table 4-4). First, the longevity of political survival has vast gaps. The institutionalized parties averaged nearly three decades in power (28.58), while parties without local activities averaged only 9.81 years. For leadership turnover, on average, institutionalized parties witness 2.38 leaders while in power, meaning that one can expect to see successful power transitions. Of all 97 parties, 53 of them (54.6%) gone through leadership transition, while 44 (45.4%) only have one leader. 112 Table 4-4 Party Institutionalization and Party’s Outcome N 106 32 97 32 97 32 Mean 28.58 9.81 2.38 1.16 15.03 1.75 S.E. 1.92 1.52 0.22 0.08 1.94 0.93 p-value of t-test 0.000 0.002 Outcome Local party Years in power Number of leaders Yes No Yes No Yes No Years in power after first leader leaves Note: Calculation excludes ruling parties that were in power for less than three years. Original data of outcomes is from Meng (2019). 0.000 On the other hand, the non-institutionalized parties have just 1.16 leaders on average. Of the 32 parties, 28 parties (87.5%) cannot survive leadership turnover, 3 parties have one turnover (Central African Republic’s Movement for the Social Evolution of Black Africa, El Salvador’s Revolutionary Party of Democratic Unification, and Haiti’s Party of National Unity) and only one party witnessed two turnovers (Portugal’s National Union). Their total years of political survival after the departure of first leader are 13, 5, 15, and 23 respectively. On average, parties without institutionalization at the local level only remain in power for 1.75 years after the exit of first leader, while institutionalized parties keep power for 15.03 years even after the first leader departs. That is, although local party branches are not directly associated with the rules for power transition between central elites, local party institutionalization is a good indicators for the overall levels of institutionalization for the whole party. 113 Figure 4-2 Party’s Years in Power and Effects of Institutionalization Note: Original data for years in power is from Meng (2019). An outlier is excluded in the figure: the True Whig Party in Liberia, which was in power for 102 years Figure 4-3 Parties’ Years in Power After 1st Leader Note: Original data for years in power is from Meng (2019). In all three indicators of party strength utilized in Meng (2019), party institutionalization leads to significant differences. Figure 4-2 and Figure 4-3 are visualization for the variation on the years in power. The result confirms that party 114 institutionalization contributes to prolong political survival of authoritarian leaders and parties. 4-3 Discussion: Other Performances Of The Leader 4-3-1 Conclusion In this chapter, I show how the party developments at the local level explain the variations of longevity of political survival of each ruler. By using a sample including all of the dictators after World War II, I illustrate that dictators with institutionalized parties—building up local branches and maintaining regular activities—indeed perform better on holding onto the power. This study is by far the most direct test on how party institutionalization at local level contributes to political survival and covers the most cases. However, there are limits to the analyses in this chapter and I will leave them to future studies. In particular, for the measurement of party institutionalization, using only one indicator of local activities is imperfect, although this operationalization does help with the lack of cross-national data on authoritarian support parties. In particular, the data may underestimate the variations of organizational developments of parties at the local level. In the original data (Geddes, Wright, and Frantz, 2018), the variable of having local activities is coded in a minimum approach: as long as the support party has “few local organizations,” it will be coded as 1 in a given year. Several indicators that reflect the strength of such local organizations, including the geographical scope of presence, number of staffs, and membership sizes, are not recorded due to data availability issues. As discussed in Chapter two, information about individual parties is often unavailable 115 to the public even in democracies. Therefore, it takes more efforts to collect such data for further investigation. 4-3-2 Future Studies Another task for future study lies in the various outcomes of the regimes. There are several kinds of outcomes behind political survival that worth study as well, including economic growth, coups, intra-state conflicts, and inter-state wars. I briefly report the results here and leave the rest to future studies. The data for other types of regime outcomes is collected from the following sources. The GDP data is from Haber and Menaldo (2011), in which they utilize several datasets to construct the numbers each year. I take the mean value of log GDP growth for each leader, dividing the value of log GDP in the last year of a leader minus the first year by total years of tenure. I separate the oil-export countries by using the data from Fearon and Laitin (2003), in which they identify countries of oil-export-reliance that more than one-third of export revenues from fuels. For the management of conflicts and wars, I utilize three kinds of data. First, the data on coup attempts comes from Powell and Thyne (2011). Besides the number of coup attempts, the coup score adopts the methods used in Meng (2019). This calculates the number of coup attempts through dividing by the length (years) of tenure. For example, if a leader experiences one coup during his five-year rule, he has a 0.2 score. Second, I use the data of intra-state armed conflict from the UCDP Onset Dataset, Uppsala Conflict Data Program (Pettersson, Högbladh, and Öberg 2019). A conflict is defined as involving in more than 25 battle deaths. Third, I analyze the onset of inter-state wars. I use the data from the Correlates of War (COW) Project (Sarkees and Wayman 2010), in which a war is defined as “sustained combat, involving organized armed forces, resulting in a minimum of 1,000 battle-related fatalities.” I record the 116 initiator of each war for the number of onset. Similarly, the onset score is the number of war onset divided by the length of tenure. Table 4-5 Party Institutionalization and Performances of Leaders Outcome Local party N Mean S.E. p-value of t- GDP growth GDP growth w/o oil-states Number of coups Score of coup Number of intra- state conflict Score of intra- state conflict Number of inter- state war onset Score of inter- state war onset Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No 224 195 196 154 265 220 265 218 265 219 265 219 265 220 265 220 0.015 0.013 0.015 0.013 0.42 0.75 0.06 0.21 0.81 0.41 0.07 0.09 0.08 0.02 0.007 0.003 0.002 0.004 0.002 0.004 0.06 0.05 0.01 0.03 0.11 0.07 0.01 0.01 0.02 0.01 0.002 0.003 test 0.700 0.667 0.000 0.000 0.003 0.402 0.012 0.383 Note: scores of coup/conflict/war are calculated by the number of event divided by tenure year. Table 4-5 summarizes the significant differences between leaders with and without institutionalized parties. The results show that party institutionalization leads to better performances in terms of coup prevention. Leaders with local party developments 117 experience fewer coup attempts during the terms. The effect holds when considering the probability of coup attempts on a yearly bases. The coup score (the number of coups divided by years in power) is significantly lower when the leader has institutionalized parties. However, party institutionalization is not associated with higher levels of economic growth. In general, leaders with local party activities perform higher yearly GDP growth on average, but this is not significantly higher than leaders without such parties. The result holds the same when excluding leaders of oil-dependent states. The conflict management shows mixed results. Contrary to previous beliefs that party-based regimes are less conflict-prone, leaders with institutionalized parties initiate more intrastate conflicts and more interstate wars than those without such parties. It is true that, if we take into account the length of tenure for the probability of conflict onset in each year, the numbers are not significant between leaders with a different type of parties. However, 97 out of 265 (36.7%) leaders with institutionalized parties initiate internal conflicts, while 56 out of 219 (25.6%) leaders without institutionalized parties do so. To this extent, 63.4% (97 out of 153) intrastate conflicts occurred under the rule of leaders with institutionalized parties; 36.6% (56) were under leaders without institutionalized parties. Considering the proportion mentioned previously—54.6% of leaders have institutionalized parties while 45.4% do not—leaders with institutionalized parties seem to initiate more intra-state conflicts. The onset of interstate war has a similar pattern. Leaders with institutionalized parties initiate more wars than leaders without such parties. 17 out of 21 war onset (81%) occurred under dictators with institutionalized parties, while four new wars (19%) started by leaders without a locally developed party. Admittedly, the difference becomes insignificant when taking the total number of years in power into account. However, 118 descriptive numbers here seem to show a different trend compared with findings in previous studies. The regression model with variables included in the model 1 (Table 4-3) shows that the results of the simple t-tests hold: party institutionalization contributes to more prolonged political survival and the ability of coup prevention, while it does not guarantee the effects on the yearly mean of GDP growth, intra-state conflicts, and inter- state wars. As the existing literature points out how parties become credible commitments to elites or the relevant audiences, they should be able to show a better ability to prevent armed conflicts. Simple descriptive summaries of the three indicators, GDP growth, internal and external conflicts, show no difference leaders with and without institutionalized parties. And the regression analysis confirms the relationship. According to existing studies, party-based regimes have a broader base in terms of domestic audiences and credible commitments from the leader to other elites, and thus should perform better on outcomes such as economic growth and conflict prevention. The evidence shows mixed results. Party institutionalization does not lead to lower levels of intra-state conflicts onset nor inter-state war onset. Moreover, there is no significant difference leaders with and without institutionalized parties. in the economic performances between One possible explanation is that political survival is the ultimate priority of the dictators, and the quality of governance or conflict prevention might not be the top concern. The ultimate goal of the dictator is to keep the loyalty of the inner circle elites, as illustrated by Bueno de Mesquita et al. (2003)’s The Logic of Political Survival. That is, the policies might only concern a small group of people, and conflicts are not necessarily detrimental to dictators. 119 The previous finding that party organization is associated with less conflict onset may need further examination. For the domestic conflicts, an institutionalized party could help the dictator to utilize coercive apparatus more efficiently and thus less afraid of conflicts. Greitens (2016) illustrates that if coercive institutions recruit diversified personnel, they can collect information more precisely. As the level of diversity increases in the coercive institutions, so does the ruler’s capacity to combat potential uprisings. I hypothesize that the co-optative institutions such as political parties do help the dictator to utilize different ruling tools in a more sophisticated way. With the help of local party branches, he is less threatened by the mass uprising and sometimes even utilizes internal conflicts to enhance in-group identity and supports from his ingroup (Theiler 2018). Regarding interstate conflicts, previous studies have pointed out that a diversionary purpose is a common reason that dictators initiate wars (see e.g., Levy 1989). When the dictator has information about the public opinion, he could initiate external conflicts without worrying about public opposition and is thus more likely to use such diversionary tools. Again, the hypothesis needs further investigation. Another explanation may be a missing factor: the international link. It is possible that the quality of governance and conflict-prevention are associated with an external factor: Western linkage. Levitsky and Way (2010) argue that international linkage is an essential factor shaping the paths of democratization for competitive authoritarian regimes. Linkage to the US and European Union is positively associated with higher chances of democratization. In their framework, the linkage includes economic ties (exports and imports), social ties (number of citizens traveling to or living in), communication ties (international voice traffic and internet access), and intergovernmental ties (membership in Organization of American States, or potential membership in the EU). On a glimpse, competitive authoritarian countries in the 1990s with high linkage scores, for example, 120 Guyana, Croatia, Mexico, Serbia, Macedonia, Nicaragua, Albania, Dominican Republic, and Taiwan, all have highly developed parties in terms of scope (geographic scope and organizational complexity) and cohesion (Levitsky and Way 2010:374-380). However, this is merely a premature hypothesis with limited number of cases, as we need more cross-national and cross-time-span data to examine whether the linkage to major powers makes a significant difference. To sum up, this study advances our understandings of authoritarian politics and makes cross-national comparisons possible, but we need further case studies to explore how local party organizations contribute to stability and capacity of the dictators. 121 CHAPTER 5 CONCLUSION 5.1 Findings The dissertation begins with the puzzle that there are significant variations in ruling parties’ organizational development in authoritarian regimes. Given that people are not the main threat to political survival—a supermajority of regime breakdowns are caused by regime insiders—some leaders’ efforts on building party organizations and linkage to the masses are seemingly unnecessary. Yet some political strongmen spend a great deal of resources in developing extensive local party branches nonetheless. This study discusses the concept of party institutionalization and the causes and effects of such ruling strategies. Chapter 2 first discusses the multi-dimensionality of party institutionalization, its measurement, and how it differs from party strength. I propose to use party institutionalization at the local level as the main measurement, because local branches and activities could help a party to advance on various dimensions of party institutionalization. The chapter summarizes a cross-national data with 485 dictators after the WWII, 1946-2008, describing party institutionalization at the local level. In general, 54.6% of all dictators have local party organizations, while 45.4% of leaders do not rely on the ruling tools of establishing nation-wide party organizations. Furthermore, I illustrate that party institutionalization and party strength are not identical. The former is a necessary condition for the latter, but not a sufficient one. Consequently, it is not proper to mix the usage of the two concepts. Chapter 3 studies the causes of party institutionalization. I propose a threat-based theory to explain party institutionalization. It argues that the initial condition when a dictator comes to power determines the ruling strategies he will choose. The ruler needs 122 to establish party organizations to control the masses when he subjectively thinks the threat to regime survival mainly comes from internal threats. However, if there exists a foreign enemy with territorial conflicts, the leader must direct resources to the coercive institutions, especially the military, to defend against threats. Analysis of cross-national data shows that party institutionalization occurs when (1) there is no foreign enemy with territorial threats, and (2) level of religious polarization is high. Evidence also shows that the leaders institutionalize the party when the relative salience of external territorial threats over internal social polarization is low. Comparative case studies of Taiwan and South Korea further confirm the causal mechanism of the external threat. The dictators of the two countries share similar characteristics, especially regarding strong state capacity, control over coercive institutions, and an alliance with the U.S. However, the dictators chose completely different strategies on developing the ruling parties. I argue that the level of perceived threats from outside is the key difference that leads to the variations on party institutionalization. For South Korean leaders, the threat from North Korea is always the most urgent threat to political survival due to geographical proximity. The rulers must maintain a considerable military. On the contrary, Taiwan’s security threat from China is mitigated by the buffer of the Taiwan Strait. The leaders had more resources and capacities to develop party organizations, as the size of the army substantially reduces over time. The cases of Indonesia and The Philippines show how domestic mass-based threat lead to party institutionalization, as the immediate foreign enemy with territorial claims is absent in the two countries. When Suharto and Marcos took power, they are similar in many factors that could influence party institutionalization, including state capacity, natural resource, ethnic polarization, and frequencies of domestic riots. I argue that 123 religious polarization is the key factor that shape different strategies on party developments. Suharto faces highly polarized religious groups nation-wide, while religion is not a salient factor in explaining Filipino domestic conflicts. Therefore, party institutionalization becomes a ruling priority for Suharto but not for Marcos. Chapter 4 tests the hypothesis that party institutionalization leads to longer political survival of dictators. The ruling party’s local organizational presence nationwide can solve several difficulties for the quality of governance and accordingly enhance chances of political survival. First, staff in local party organizations can interact with the masses and help dictators overcome the problems of Dictator’s Dilemma, especially regarding lack of correct information about the general public. Also, with the local branches, the party can distribute material benefits to citizens, and mobilize support more precisely. Furthermore, local party branches increase political participation and thus strengthen the legitimacy of the regime. The evidence shows that dictators with institutionalized parties indeed perform better on political survival. This chapter also shows that party institutionalization is a better explanatory factor for regime outcomes than party strength. 5.2 Contributions of the Dissertation The main contribution of my study is to gain an understanding of authoritarian politics, especially about the dynamics of choosing different ruling tools for the masses. The very beginning of this project is from observations of East Asian authoritarian regimes. Studying social science in a newly democratized country, Taiwan, I try to understand the origins of authoritarian legacy and its impacts. Even after three decades of democratization, we still see a highly resilient former authoritarian party KMT maintaining dominance in local politics nationwide. Many politicians tend to promote authoritarian values, for example, praising dictatorships and criticizing democracy, as 124 the conservatives shape the political agenda, resisting reforms in various fields. I later found that Taiwan is not unique with regard to its experiences of authoritarian rule. In the authoritarian era, many authoritarian parties are highly resilient, and some of them keep dominant status long after democratization. Some authoritarian ruling parties have successfully resisted democratization until today, such as in Cambodia, Malaysia, Vietnam, and Singapore. I am always curious why people trust in the dictators (or the former authoritarian party), and how do leaders maintain a high level of support? As a famous quote by the political scientist E.E. Schattschneider (1942) points out: “The political parties created democracy and modern democracy is unthinkable save in terms of the parties.” Studies of authoritarian politics show that not only in a democracy, political parties are common ruling tools for dictatorships as well. What are the roles of the political party in an autocracy, and how does this differ from democratic counterparts? Why do some ruling parties become dominant, while some play secondary roles in the ruling? What are the long-term impacts of the ruling party’s organizational development in autocracies? These are the big questions I keep in mind. The focus of this study is on dictators’ strategies of party institutionalization at the local level. Existing studies often focus on either coercive institutions or co-optation among the elites. Research about political institutions, such as elections, parties, and formal rules, does illustrate the interactions and power-sharing between the ruler and the ruling elites. Still, the role of the masses is often neglected. Such gaps in the literature are the outcome that most of the authoritarian regime turnovers are caused by the ruling insiders but not from the masses. Even so, there is a great variation on the extent to which the dictators put efforts to invest in party organizations and connect to society. My study helps to solve one of the major puzzles in the literature—the determinant of dictators’ 125 strategic choices to develop party organizations at the local level. This study is by far the most direct examination of the origins of party institutionalization. This research question is important because ruling parties’ activities at the local level are critical factors to prolong political survival. As illustrated by Nathan (2003), authoritarian parties’ local organizations help leaders to acquire legitimacy from people. Therefore, analyzing how dictatorships establish roots in society contributes to explain the origins of authoritarian resilience. The literature has shown that the development of the ruling party in the early stages of the dictatorship substantially shapes the political development of a country.27 Some scholars refer to this as “shadows from past” (Hicken and Kuhonta 2015). However, there is no satisfactory explanation of why the ruling party institutionalized in the past. It is not clear why some rulers prioritize linking with the mass and build up an inclusive alliance while some rule mainly by coercive means. By exploring the origin of ruling party institutionalization, I show that dictators care about local party organizational developments when religious polarization is high. Also, the existence of foreign threats impedes party institutionalization. This threat-based theory explains why some leaders institutionalize the party. Empirically, the dissertation makes three improvements to the existing studies of authoritarian party politics. First, it helps to expand the scope of existing data on individual parties in autocracies. Data about the individual party and its organizations are scarce because parties are relatively secretive and closed organizations. Most of the existing studies on party institutionalization focus on democracies, while studies on the party’s organizational developments in autocracies remain under-investigated. My data 27 See, for example, Slater (2010); Hellmann (2011); Ufen (2009); Hicken (2009); Hicken and Kuhonta (2015); and Riedl (2014). 126 covers all of the dictators after WWII (but not include island-states or city-states), which provide sources for further analyses on individual parties. Second, the indicator of party institutionalization at the local level improves the existing measurement on the party’s organizational development. I differentiate party strength (defined by leadership turnover after the first leader leaves power, Meng 2019) with party institutionalization. It is true that party strength, or more broadly, measuring party development by the consequences or performances, may be an easily accessible data which could fill in the gap of data availability. For example, power turnover, electoral support, and seats in the legislature are observable data for each regime. However, these variables are outcomes of dictators’ choice of ruling strategies, not ideal factors to explain other performances. This study proves that party institutionalization, as defined by local level developments, is a better explanatory factor for regime outcomes. Third, this dissertation studies the political survival of leaders, while most extant studies of authoritarian politics focus on regime or country-year as unit of analysis. Focusing on leaders expands our understandings of strategic choices of each ruler, which could expand to more future research, adding detailed context to authoritarian politics. For example, previous study has found that leaders often introduce property rights- enhancing institutions by signing bilateral investment treaties (BITs) (Arias, Hollyer, and Rosendorff 2018). The rationale behind political institutions, especially the party, may be similar to the economic institutions because both kinds of institutions limit the power of the leader to a certain degree with the purpose of creating a better “climate” for foreign investors. With the prospect of protecting property rights and lowering the risks of confiscation, investments are more likely to inflow. That is, sometimes the dictator choose to put on some checks and balance toward himself to show credible commitment for economic development and this strategy is effective for a longer political survival. 127 The study has important implications for contemporary politics, especially the rising attention for non-violent but suppressive ruling strategies of the dictators. For example, dictators’ “learning curve” (Dobson 2012) illustrates that contemporary autocracies adopt sophisticated approaches that mix both coercive and co-optative means to people. More and more dictators rely on pseudo-democratic institutions to rule (Kendall-Taylor and Frantz 2014). My dissertation shows that dictators may be constrained by external or internal environments when choosing ruling strategies in the beginning. But once political institutions are established, especially those at the local level, they have profound effects on political survival. Besides enhancing political participation and legitimacy from people, these institutions help the dictator to collect information and to adapt ruling strategies accordingly. The learning process is thus one critical pillar for authoritarian resilience. In the current world politics, authoritarian values still enjoy high levels of supports in many countries, and the values are diffusing. One has witnessing the global trend that political rights and civil liberties around the world have been deteriorating (Repucci 2020). Some report even worse findings: autocracies became the majority in the world in 2020, for the first time since the beginning of the 21st century, as 92 countries, 54% of the global population, are living in non-democracies (Lührmann et al. 2020). The trend of concentration of power is significantly associated with rising levels of repression and worsening human rights (Frantz et al. 2020). Studying the origins of co-optation strategies will help understand the nature of authoritarian politics and the trend of democratic recession in this decade. While the dictators have their learning processes and mechanisms of adapting to new challenges, citizens in democracies, especially policymakers, also need to learn about the dynamics of democracy and autocracy. 128 5.3 Remaining Questions and Directions for Future Research Future studies include: (1) the improvements of the measurement for party institutionalization, (2) further investigations on the causes of party institutionalization, especially the role of ethnic polarization, religions, comparative cases studies for domestic threats, and various paths of party’s organizational developments, and (3) how local party organizations shape ruling outcomes and political survival. For the measurement of party institutionalization, one direction for studying the causes and effects of party institutionalization is to analyze different paths of the party developments. As discussed in chapter 2, there are four situations of party institutionalization: (1) abandoning the existing local party activities; (2) local party branch remaining absent throughout the term; (3) maintaining local party organization from the beginning until the time of leaving power; and (4) building up new local party branches during the term. The current analysis only adopts a binary variable (leaders with and without institutionalized parties), but the motivations behind two scenarios in each category may be different. For example, leaders who build up new local party branches may have stronger motives for building political institutions than leaders who maintain the existing ones. Future studies need to fill in the gap. Besides the four paths of local party branch developments, there are multiple dimensions in party institutionalization. Although the operationalization of focusing on party’s development at the local level does fill in the gap of lack of cross-national data on authoritarian support parties, using only one indicator of local activities is not perfect. In the original data (Geddes, Wright, and Frantz, 2018), the variable of having local activities is coded in a minimum approach: as long as the support party has “few local organizations,” it will be coded as 1 in a given year. I use this variable to construct party institutionalization for each leader, which may underestimate the variations of 129 organizational developments of parties. After all, the existence of local branches does not guarantee regularized and constant party activities. In the future, the measurement of party institutionalization at the local level needs to add several indicators that reflect the strength of such organizations, including the geographical scope of presence, number of staff, and membership sizes. As discussed in Chapter two, information about individual parties is often unavailable to the public, even in democracies. Therefore, data collection is still a central task for future studies. For the causes of ruling party institutionalization, the relationship between polarization and political threat to the leader is under-investigated. Scholars are still arguing on whether and how ethnic polarization leads to conflicts and the evidences seem to remain inconclusive (Wegenast and Basedau 2014). In this study, the indicator of ethnic polarization is not significantly associated with party institutionalization, which does not fit the expectation of the threat-based theory. A direct explanation is that a high level of social heterogeneity does not necessarily mean a high level of perceived threats. It is likely that between polarization and threat, there is another essential missing causal link. Also, the reasons behind the negative effects linguistic fractionalization on party institutionalization remain unanswered. One future direction is to examine determinants of internal threats due to distributional conflicts, the between-group-inequality (BGI) and within-group-inequality (WGI) (Houle 2015). I hypothesize that when BGI is high and WGI is low, that is, internal cohesion of ethnic group is high, ethnic identity-based mobilization becomes more probable. This situation will lead to higher levels of threat to the leader. First, when BGI is high, the marginalized ethnic group that experiences profound grievance is more likely to mobilize and rebel (Horowitz 1985). In modern history, ethnicity has been a fundamental element of the majority of civil conflicts (Gurr et al. 1993; Brubaker and 130 Laitin 1998), while ethnicity is one of the defining features for between-group comparisons. For the marginalized groups, the systematical discrimination they face— either by the government or the society—is based on traits that are inherited or can hardly be changed. Studies show that people who suffer such identity-based discrimination or grievance tend to engage in political activities to raise their socio-economic status, taking part in mobilization and contentious politics against the authority or other groups (Gellner 1983; Marquardt 2017). Second, the extent of within-group internal cohesion, mainly shaped by levels of WGI, influence the prospects of a successful mobilization. In particular, ethnic diversity becomes a threat to the regime when social class and economic resources are distributed along the ethnic line. In a society, various kinds of social cleavages differentiate social groups, such as religion, class, or regional origins, which can be seen as competing social identities (Stryker 2000; Snow and McAdam 2000). On the one hand, if there exist cross- cutting cleavages within social groups when WGI is high, there is less likely to be form in-group loyalty or consistent policy preference within that set of people. On the other hand, when most of the members in one group belong to similar classes or socio- economic status, that is, when the WGI is low, it will create common interests for group members and strengthen group loyalty (Houle 2015; Bodea and Houle 2017). For example, if most of the members in an ethnic group hold similar occupations, earning identical income levels, then it becomes easier for them to form collective actions vis-à-vis other ethnic groups and the ruler based on common interests toward redistribution. Taken together, when the within-group homogeneity is high, BGI makes ethnicity a salient issue, increasing the appeal and feasibility of coups and other types of conflicts. I believe that economic inequality and potential distributional conflicts explain how leaders react to the collective actions of the masses. 131 Besides ethnic polarization, the role of religion and why religious polarization leads to party institutionalization need further studies. Chapter 3 adopts a comparative case study of Indonesia and The Philippines to show that religious polarization is the most viable factor that explains the different ruling strategies of the two dictators, Suharto and Marcos. However, we may need to further investigate how the leaders perceive the threat and their responses to various religions and the activities from bottom-up. East Asian countries are particularly challenging to the conventional wisdom on party developments, as ruling parties’ organizational developments in this area show a significant degree of variations. One sees similar backgrounds in regimes, in terms of material capability and state capacity, with different outcomes of institutionalization of the ruling party, providing ideal cases for comparative study. For example, the Philippines (1965-1986) and Indonesia (1967-1998) or Malaysia (1957-) may be another ideal country-pair for comparison. In the Philippines, the ruling party cannot reach beyond the main city (Hicken 2009), while in Indonesia and Malaysia, the ruling parties both established nation-wide local branches (Tomsa 2008; Hicken and Kuhonta 2015). These countries do not have an immediate foreign military threat with territorial claims and share some similar patterns of political and economic developments. Thus, researchers can trace how domestic threats lead to various developments of party institutionalization. Last but not the least, for the effects of party institutionalization, we need further studies to clarify how the local party’s organizational developments contribute to the stability and capacity of the dictators. Chapter 4 illustrates that party institutionalization leads to more prolonged political survival, but it is not clear whether it results in better ruling qualities. Preliminary results show that party institutionalization is not significantly associated with higher GDP growth, less domestic conflicts, and external 132 armed conflicts. The relationship between party organizations and ruling outcomes needs more data and empirical examination. If party institutionalization does not contribute to better outcomes, it seems like the most plausible explanation is that party institutionalization facilitates the political participation of the people and enhances the legitimacy of the regime (Nathan 2003). Yet these causal links need further investigation. In sum, while authoritarian resilience and diffusion of autocratic values have been enduring topics in both academic and policy studies, I believe that these questions worth further investigations. 133 BIBLIOGRAPHY 134 BIBLIOGRAPHY Accinelli, Robert. 1996. Crisis and Commitment: United States Policy Toward Taiwan, 1950- 1955. Chapel Hill: University of North Carolina Press. Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and Romain Wacziarg. 2003. “Fractionalization.” Journal of Economic Growth 8(2): 155–194. Amsden, Alice H. 1989. Asia’s Next Giant: South Korea and Late Industrialization. New York: Oxford University Press. Annett, Anthony. 2000. Social Fractionalization, Political Instability, and the Size of Government. Washington, D.C.: International Monetary Fund. Arias, Eric, James R. Hollyer, and B. Peter Rosendorff. 2018. “Cooperative Autocracies: Leader Survival, Creditworthiness, and Bilateral Investment Treaties.” American Journal of Political Science 62(4): 905–921. Arter, David, and Elina Kestilä-Kekkonen. 2014. “Measuring the Extent of Party Institutionalisation: The Case of a Populist Entrepreneur Party.” West European Politics 37(5): 932–956. Aspinall, Edward. 2005. Opposing Suharto: Compromise, Resistance, and Regime Change in Banks, Arthur S. 2011. Cross-­‐‑National Time-­Series Data Archive. Jerusalem: Databanks Indonesia. Stanford: Stanford University Press. International. http://www.databanksinternational.com. Basedau, Matthias, Birte Pfeiffer, and Johannes Vüllers. 2016. “Bad Religion? Religion, Collective Action, and the Onset of Armed Conflict in Developing Countries.” Journal of Conflict Resolution 60(2): 226–255. Basedau, Matthias, and Alexander Stroh. 2008. “Measuring Party Institutionalization in Developing Countries: A New Research Instrument Applied to 28 African Political Parties.” GIGA Working Paper No. 69. Biela, Jan, Annika Hennl, and André Kaiser. 2012. “Combining Federalism and Decentralization: Comparative Case Studies on Regional Development Policies in Switzerland, Austria, Denmark, and Ireland.” Comparative Political Studies 45(4): 447–476. Bizzarro, Fernando et al. 2018. “Party Strength and Economic Growth.” World Politics 70(2): 275–320. 135 Bizzarro, Fernando, Allen Hicken, and Darin Self. 2017. The V-Dem Party Institutionalization Index: A New Global Indicator (1900-2015). Gothenburg, Sweden: The Varieties of Democracy Institute. Blaydes, Lisa. 2010. Elections and Distributive Politics in Mubarak’s Egypt. New York: Cambridge University Press. Bodea, Cristina, Ibrahim Elbadawi, and Christian Houle. 2017. “Do Civil Wars, Coups and Riots Have the Same Structural Determinants?” International Interactions 43(3): 537–561. 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. Bolleyer, Nicole, and Saskia P. Ruth. 2017. “Elite Investments in Party Institutionalization in New Democracies: A Two-Dimensional Approach.” The Journal of Politics 80(1): 288–302. Bolt, Jutta, and Jan Luiten van Zanden. 2014. “The Maddison Project: Collaborative Research on Historical National Accounts: The Maddison Project.” The Economic History Review 67(3): 627–651. Bresnan, John. 1993. Managing Indonesia: The Modern Political Economy. New York; Oxford: Columbia University Press. Broom, Leonard, and Philip Selznick. 1968. Sociology: a Text with Adapted Readings. New York: Harper & Row. Brownlee, Jason. 2007. Authoritarianism in an Age of Democratization. Cambridge, UK; New York, USA: Cambridge University Press. Brubaker, Rogers, and David D. Laitin. 1998. “Ethnic and Nationalist Violence.” Annual Review of Sociology 24(1): 423–452. Bueno de Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow. 2003. The Logic of Political Survival. Cambridge, US: MIT Press. Bush, Richard. 2016. America’s Alliances and Security Partnerships in East Asia. Washington, D.C.: The Brookings Institute. Carey, Sabine C. 2010. “The Use of Repression as a Response to Domestic Dissent.” Political Studies 58(1): 167–186. Cheibub, José Antonio, Jennifer Gandhi, and James Raymond Vreeland. 2010. “Democracy and Dictatorship Revisited.” Public Choice 143(1–2): 67–101. Cheng, Tun-jen. 1989. “Democratizing the Quasi-Leninist Regime in Taiwan.” World Politics 41(4): 471–499. 136 Cheng, Tun-jen. 2006. “Strategizing Party Adaptation: The Case of the Kuomintang.” Party Politics 12(3): 367–394. Cheng, Tun-jen, and Yung-min Hsu. 2015. “Long in the Making: Taiwan’s Institutionalized Party System.” In Party System Institutionalization in Asia: Democracies, Autocracies, and the Shadows of the Past. eds. Allen Hicken and Erik Kuhonta. Cambridge, UK; New York: Cambridge University Press, pp. 182–221. Colaresi, Michael P., Karen Rasler, and William R. Thompson. 2008. Strategic Rivalries in World Politics: Position, Space and Conflict Escalation. Cambridge, UK: Cambridge University Press. Collier, Paul, and Anke Hoeffler. 2004. “Greed and Grievance in Civil War.” Oxford Economic Papers 56(4): 563–595. Collier, Paul, and Anke Hoeffler. 1998. “On Economic Causes of Civil War.” Oxford Economic Papers 50(4): 563–573. Coppedge, Michael et al. 2016. V-Dem Codebook v6. University of Gothenburg, V-Dem Institute, and University of Notre Dame, Kellogg Institute. Varieties of Democracy (V-Dem) Project. Crouch, Harold A. 1978. The Army and Politics in Indonesia. Ithaca, N.Y.: Cornell University Press. Cullather, Nick. 1996. “‘Fuel for the Good Dragon’: The United States and Industrial Policy in Taiwan, 1950–1965.” Diplomatic History 20(1): 1–26. Davenport, Christian. 1995. “Multi-Dimensional Threat Perception and State Repression: An Inquiry into Why States Apply Negative Sanctions.” American Journal of Political Science 39(3): 683–713. Debs, Alexandre, and H. E. Goemans. 2010. “Regime Type, the Fate of Leaders, and War.” American Political Science Review 104(3): 430–445. Desmet, Klaus, Ignacio Ortuño-Ortín, and Romain Wacziarg. 2012. “The Political Economy of Linguistic Cleavages.” Journal of Development Economics 97(2): 322–338. Desmet, Klaus, Ignacio Ortuño-Ortín, and Shlomo Weber. 2009. “Linguistic Diversity and Redistribution.” Journal of the European Economic Association 7(6): 1291–1318. Deyo, Frederic C. 1987. The Political Economy of the New Asian Industrialism. Ithaca: Cornell University Press. Dickson, Bruce J. 1993. “The Lessons of Defeat: The Reorganization of the Kuomintang on Taiwan, 1950–52.” The China Quarterly 133: 56–84. Dios, Aurora J. De. 1988. Dictatorship and Revolution: Roots of People’s Power. Metro Manila: Conspectus. 137 Dobson, William J. 2012. The Dictator’s Learning Curve: Inside the Global Battle for Democracy. New York: Anchor Books. Doner, Richard F., Bryan K. Ritchie, and Dan Slater. 2005. “Systemic Vulnerability and the Origins of Developmental States: Northeast and Southeast Asia in Comparative Perspective.” International Organization 59(2): 327–361. Esteban, Joan, Laura Mayoral, and Debraj Ray. 2012. “Ethnicity and Conflict: An Empirical Study.” American Economic Review 102(4): 1310–1342. Esteban, Joan, and Debraj Ray. 2008. “On the Salience of Ethnic Conflict.” American Economic Review 98(5): 2185–2202. Esteban, Joan-María, and Debraj Ray. 1994. “On the Measurement of Polarization.” Econometrica 62(4): 819–851. Evans, Peter. 1995. Embedded Autonomy: States and Industrial Transformation. Princeton: Princeton University Press. Eyraud, Luc, and Marialuz Moreno Badia. 2013. “Too Small to Fail? Subnational Spending Pressures in Europe.” IMF Working Papers 13(46): 1. Fearon, James D. 1995. “Rationalist Explanations for War.” International Organization 49(3): 379–414. Fearon, James D., and David D. Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” American Political Science Review 97(1): 75–90. Federal Research Division (FRD). 2004. “Country Profile: Indonesia.” http://lcweb2.loc.gov/frd/cs/profiles/Indonesia.pdf. Federal Research Division (FRD). 2006. “Country Profile: Philippines.” http://lcweb2.loc.gov/frd/cs/profiles/Philippines.pdf. Federal Research Division (FRD). 2005a. “Country Profile: South Korea.” http://lcweb2.loc.gov/frd/cs/profiles/South_Korea.pdf. Federal Research Division (FRD). 2005b. “Country Profile: Taiwan.” http://lcweb2.loc.gov/frd/cs/profiles/Taiwan.pdf. Fell, Dafydd. 2005. Party Politics in Taiwan: Party Change and the Democratic Evolution of Taiwan, 1991-2004. London; New York: Routledge. Ferrer, Miriam Coronel. 2005. “The Moro and the Cordillera Conflicts in the Philippines and the Struggle for Autonomy.” In Ethnic Conflicts in Southeast Asia. eds. Kusuma Snitwongse and W. Scott Thompson. Singapore: Institute of Southeast Asian Studies, pp. 109–150. Frantz, Erica, Andrea Kendall-Taylor, Joseph Wright, and Xu Xu. 2020. “Personalization of Power and Repression in Dictatorships.” The Journal of Politics 82(1): 372–377. 138 Gaddis, John Lewis. 1982. Strategies of Containment: A Critical Appraisal of Postwar American National Security Policy. New York: Oxford University Press. Gandhi, Jennifer. 2008. Political Institutions Under Dictatorship. New York: Cambridge University Press. Gandhi, Jennifer, and Adam Przeworski. 2007. “Authoritarian Institutions and the Survival of Autocrats.” Comparative Political Studies 40(11): 1279–1301. Geddes, Babara. 2003. Paradigms and Sand Castles: Theory Building and Research Design in Comparative Politics. Ann Arbor: University of Michigan Press. Geddes, Barbara. 2008. “Party Creation as an Autocratic Survival Strategy.” Unpublished manuscript. Geddes, Barbara. 2006. “Why Parties and Elections in Authoritarian Regimes?” Unpublished manuscript. Geddes, Barbara, Joseph Wright, and Erica Frantz. 2014. “Autocratic Breakdown and Regime Transitions: A New Data Set.” Perspectives on Politics 12(2): 313–331. Geddes, Barbara, Joseph Wright, and Erica Frantz. 2018. How Dictatorships Work: Power, Personalization, and Collapse. Cambridge: Cambridge University Press. Geertz, Clifford. 1960. The Religion of Java. Chicago: University of Chicago Press. Gehlbach, Scott, and Philip Keefer. 2011. “Investment without democracy: Ruling-party institutionalization and credible commitment in autocracies.” Journal of Comparative Economics 39(2): 123–139. Gehlbach, Scott, and Philip Keefer. 2012. “Private Investment and the Institutionalization of Collective Action in Autocracies: Ruling Parties and Legislatures.” The Journal of Politics 74(2): 621–635. Gellner, Ernest. 1983. Nations and Nationalism. Ithaca: Cornell University Press. Gerring, John. 2017. Case Study Research: Principles and Practices. Second Edition. Cambridge, UK; New York, USA: Cambridge University Press. Gibler, Douglas M. 2012. The Territorial Peace: Borders, State Development, and International Conflict. Cambridge, UK  ; New York: Cambridge University Press. Gibler, Douglas M., Steven V. Miller, and Erin K. Little. 2016. “An Analysis of the Militarized Interstate Dispute (MID) Dataset, 1816–2001.” International Studies Quarterly 60(4): 719–730. Gibson, James L., Cornelius P. Cotter, John F. Bibby, and Robert J. Huckshorn. 1983. “Assessing Party Organizational Strength.” American Journal of Political Science 27(2): 193–222. 139 Goemans, Henk E., Kristian Skrede Gleditsch, and Giacomo Chiozza. 2009. “Introducing Archigos: A Dataset of Political Leaders:” Journal of Peace Research 46(2): 269–283. Gold, Thomas B. 1986. State and Society in the Taiwan Miracle. New York: M.E. Sharpe. Gong, Yi-chun. 1998. The Émigré Regime and the Indigenous Society: KMT and the Formation of Its Social Foundation After the Reform (1950-1969). Xinbei, Taiwan: Dao-hsiang Press. Greene, Kenneth F. 2010. “The Political Economy of Authoritarian Single-Party Dominance.” Comparative Political Studies 43(7): 807–834. Greitens, Sheena Chestnut. 2016. Dictators and Their Secret Police: Coercive Institutions and State Violence. Cambridge, UK; New York: Cambridge University Press. Gulati, UmushC. 1992. “The Foundations of Rapid Economic Growth.” American Journal of Economics and Sociology 51(2): 161–172. Gurr, Ted R., Barbara Harff, Monty G. Marshall, and James R. Scarritt. 1993. Minorities at Risk: A Global View of Ethnopolitical Conflicts. Washington, D.C.: United States Institute of Peace Press. Haber, Stephen, and Victor Menaldo. 2011. “Do Natural Resources Fuel Authoritarianism? A Reappraisal of the Resource Curse.” American Political Science Review 105(1): 1–26. Harmel, Robert, and Kenneth Janda. 1994. “An Integrated Theory of Party Goals and Party Change.” Journal of Theoretical Politics 6(3): 259–287. Head, Mike. 1999. US Orchestrated Suharto’s 1965-66 Slaughter in Indonesia. Part 1: New Evidence on How the October 1 Coup Was Triggered. International Committee of the Fourth International (ICFI). https://www.wsws.org/en/articles/1999/07/indo1- j19.html. Hellmann, Olli. 2013. “The Developmental State and Electoral Markets in East Asia: How Strategies of Industrialization Have Shaped Party Institutionalization.” Asian Survey 53(4): 653–678. Hensel, Paul R., Sara McLaughlin Mitchell, Thomas E. Sowers, and Clayton L. Thyne. 2008. “Bones of Contention: Comparing Territorial, Maritime, and River Issues.” Journal of Conflict Resolution 52(1): 117–143. Hicken, Allen. 2009. Building Party Systems in Developing Democracies. Cambridge, UK; New York: Cambridge University Press. Hicken, Allen. 2015. “Party and Party System Institutionalization in the Philippines.” In Party System Institutionalization in Asia: Democracies, Autocracies, and the Shadows of the Past. eds. Allen Hicken and Erik Kuhonta. Cambridge, UK; New York: Cambridge University Press, pp. 517–554. 140 Hicken, Allen, and Erik Kuhonta, eds. 2015. Party System Institutionalization in Asia: Democracies, Autocracies, and the Shadows of the Past. Cambridge, UK; New York: Cambridge University Press. Hoon, Juang. 2015. “The Historical Origins of Frail Party Politics in South Korea: The Abortive Experiment of Democratic Republican Party, 1963-1979.” Journal of Parliamentary Research 10(2): 313–346. Horowitz, Donald L. 1985. Ethnic Groups in Conflict. Berkeley: University of California Press. Houle, Christian. 2015. “Ethnic Inequality and the Dismantling of Democracy: A Global Analysis.” World Politics 67(3): 469–505. Houle, Christian, and Cristina Bodea. 2017. “Ethnic Inequality and Coups in Sub- Saharan Africa.” Journal of Peace Research 54(3): 382–396. Huntington, Samuel P. 1968. Political Order in Changing Societies. New Haven: Yale University Press. Hutchcroft, Paul. 2011. “Reflections on a reverse image: South Korea under Park Chung Hee and the Philippines under Ferdinand Marcos.” In The Park Chung Hee Era: The Transformation of South Korea. eds. Byung-Kook Kim and Ezra F. Vogel. Cambridge: Harvard University Press, pp. 542–572. International Business Publications (IBP). 1999. Malaysia Army Weapon Systems Handbook. Washington, D.C.: USA International Business Publications. Janda, Kenneth. 1980. Political Parties: A Cross-National Survey. New York; London: Free Press  ; Collier Macmillan. Johnson, Chalmers A. 1982. MITI and the Japanese Miracle: The Growth of Industrial Policy: 1925-1975. Stanford: Stanford University Press. Kamlian, Jamail A. 2004. “Ethnic and Religious Conflict in the Philippines: The Bangsamoro Experience.” In The Making of Ethnic and Religious Conflicts in Southeast Asia: Cases and Resulations. eds. Frans de Djalong and Lambang Trijono. Yogyakarta, Indonesia: CSPS Books, pp. 89–140. Kendall-Taylor, Andrea, and Erica Frantz. 2014a. “How Autocracies Fall.” The Washington Quarterly 37(1): 35–47. Kendall-Taylor, Andrea, and Erica Frantz. 2014b. “Mimicking Democracy to Prolong Autocracies.” The Washington Quarterly 37(4): 71–84. Kim, Byung-Kook. 2011. “Introduction.” In The Park Chung Hee era: The Transformation of South Korea. eds. Byung-Kook Kim and Ezra F. Vogel. Cambridge, Mass: Harvard University Press, pp. 1–31. 141 Kim, HeeMin. 2011. Korean Democracy in Transition: A Rational Blueprint for Developing Societies. Lexington, KY.: University Press of Kentucky. Kim, Hyung-A. 2011. “State Building: The Military Junta’s Path to Modernity through Administrative Reforms.” In The Park Chung Hee Era: The Transformation of South Korea. eds. Pyŏng-guk Kim and Ezra F Vogel. Cambridge, US: Harvard University Press, pp. 85–114. Kim, Nam Kyu. 2019. “External Territorial Threats and Military Regimes.” Political Research Quarterly 72(4): 863–877. King, Dwight Y. 2003. Half-Hearted Reform: Electoral Institutions and the Struggle for Democracy in Indonesia. Westport: Praeger. Klein, James P., Gary Goertz, and Paul F. Diehl. 2006. “The New Rivalry Dataset: Procedures and Patterns.” Journal of Peace Research 43(3): 331–348. Knutsen, Carl Henrik, and Magnus Rasmussen. 2018. “The Autocratic Welfare State: Old-Age Pensions, Credible Commitments, and Regime Survival.” Comparative Political Studies 51(5): 659–695. Kohli, Atul. 2004. State-Directed Development: Political Power and Industrialization in the Global Periphery. Cambridge, UK; New York: Cambridge University Press. Krain, Matthew. 1997. “State-Sponsored Mass Murder: The Onset and Severity of Genocides and Politicides.” Journal of Conflict Resolution 41(3): 331–360. Lacaba, Jose, ed. 1995. Boss: 5 Case Studies of Local Politics in the Philippines. Manila: Philippine Center for Investigative Journalism. Latham, Michael E. 2011. The Right Kind of Revolution: Modernization, Development, and U.S. Foreign Policy from the Cold War to the Present. Ithaca: Cornell University Press. Levitsky, Steven. 2001. “An ‘Organised Disorganisation’: Informal Organisation and the Persistence of Local Party Structures in Argentine Peronism.” Journal of Latin American Studies 33(1): 29–66. Levitsky, Steven. 1998. “Institutionalization and Peronism The Concept, the Case and the Case for Unpacking the Concept.” Party Politics 4(1): 77–92. Levitsky, Steven, and María Victoria Murillo. 2009. “Variation in Institutional Strength.” Annual Review of Political Science 12(1): 115–133. Levitsky, Steven, and Lucan Way. 2010. Competitive Authoritarianism  : Hybrid Regimes after the Cold War. New York: Cambridge University Press. Levy, Jack S. 1989. “The Diversionary Theory of War: A Critique.” In Handbook of War Studies. ed. Manus I Midlarsky. Boston: Unwin Hyman, pp. 259–288. 142 Liddle, R. William, and Saiful Mujani. 2007. “Leadership, Party, and Religion: Explaining Voting Behavior in Indonesia.” Comparative Political Studies 40(7): 832– 857. Looney, Kristen E. 2020. Mobilizing for Development: The Modernization of Rural East Asia. Ithaca: Cornell University Press. Luc Eyraud, and Marialuz Moreno Badia. 2013. Too Small to Fail? Subnational Spending Pressures in Europe. International Monetary Fund. Lührmann, Anna, Seraphine F. Maerz, Sandra Grahn, Nazifa Alizada, and Lisa Gastaldi. Autocratization Surges – Resistance Grows: Democracy Report 2020. Gothenburg: Varieties of Democracy Institute (V-Dem). Magaloni, Beatriz. 2008. “Credible Power-Sharing and the Longevity of Authoritarian Rule.” Comparative Political Studies 41(4–5): 715–741. Magaloni, Beatriz. 2006. Voting for Autocracy: Hegemonic Party Survival and its Demise in Mexico. Cambridge: Cambridge University Press. Magaloni, Beatriz, and Ruth Kricheli. 2010. “Political Order and One-Party Rule.” Annual Review of Political Science 13(1): 123–143. Mahoney, James. 2001. “Path-Dependent Explanations of Regime Change: Central America in Comparative Perspective.” Studies in Comparative International Development 36(1): 111–141. Mahoney, James. 2007. “Qualitative Methodology and Comparative Politics.” Comparative Political Studies 40(2): 122–144. Mainwaring, Scott, and Timothy Scully. 1995. Building Democratic Institutions: Party Systems in Latin America. Stanford: Stanford University Press. Malesky, Edmund, and Paul Schuler. 2011. “The Single-Party Dictator’s Dilemma: Information in Elections without Opposition: Single-Party Dictator’s Dilemma.” Legislative Studies Quarterly 36(4): 491–530. Marquardt, Kyle L. 2018. “Identity, Social Mobility, and Ethnic Mobilization: Language and the Disintegration of the Soviet Union.” Comparative Political Studies 51(7): 831– 867. Marshall, Monty G., and Donna Ramsey Marshall. 2019. Coup d’État Events: 1946-2018. Vienna, Virginia: Center for Systemic Peace. http://www.systemicpeace.org/inscr/CSPCoupsCodebook2018.pdf. Meng, Anne. 2019. “Ruling Parties in Authoritarian Regimes: Rethinking Institutional Strength.” British Journal of Political Science. Forthcoming. doi.org/10.1017/S0007123419000115. 143 Meng, Anne. 2016. The Organizational Weapon: Ruling Parties in Authoritarian Regimes. Berkeley: Ph.D Dissertation, Political Science, UC Berkeley. Mesquita, Bruce Bueno de, and Randolph M. Siverson. 1995. “War and the Survival of Political Leaders: A Comparative Study of Regime Types and Political Accountability.” American Political Science Review 89(4): 841–855. Mietzner, Marcus. 2018. “Authoritarian Elections, State Capacity, and Performance Legitimacy: Phases of Regime Consolidation and Decline in Suharto’s Indonesia.” International Political Science Review 39(1): 83–96. Miller, Michael K. 2020. “The Autocratic Ruling Parties Dataset: Origins, Durability, and Death.” Journal of Conflict Resolution 64(4): 756–782. Montalvo, José G., and Marta Reynal-Querol. 2005. “Ethnic Polarization, Potential Conflict, and Civil Wars.” The American Economic Review 95(3): 796–816. Myers, Ramon H., and Hsio-ting Lin. 2007. Breaking with the Past: The Kuomintang Central Reform Committee on Taiwan, 1950-52. Stanford: Hoover Institution Press. Nathan, Andrew J. (Andrew James). 2003. “Authoritarian Resilience.” Journal of Democracy 14(1): 6–17. Noble, Gregory W. 1998. Collective Action in East Asia: How Ruling Parties Shape Industrial Policy. Ithaca: Cornell University Press. Öniş, Ziya. 1991. “The Logic of the Developmental State.” Comparative Politics 24(1): 109–126. Page, John. 1994. “The East Asian Miracle: Four Lessons for Development Policy.” The World Bank NBER Macroeconomics Annual 1994(9): 219–282. Panebianco, Angelo. 1988. Political Parties: Organization and Power. Cambridge, UK; New York: Cambridge University Press. Pedersen, Mogens N. 1979. “The Dynamics of European Party Systems: Changing Patterns of Electoral Volatility.” European Journal of Political Research 7(1): 1–26. Pettersson, Therése, Stina Högbladh, and Magnus Öberg. 2019. “Organized violence, 1989–2018 and peace agreements.” Journal of Peace Research 56(4): 589–603. Powell, Jonathan M., and Clayton L. Thyne. 2011. “Global Instances of Coups from 1950 to 2010: A New Dataset.” Journal of Peace Research 48(2): 249–259. Randall, Vicky, and Lars Svåsand. 2002. “Party Institutionalization in New Democracies.” Party Politics 8(1): 5–29. Rasler, Karen A., and William R. Thompson. 2006. “Contested Territory, Strategic Rivalries, and Conflict Escalation.” International Studies Quarterly 50(1): 145–168. 144 Rasler, Karen, and William R. Thompson. 2011. “Borders, Rivalry, Democracy, and Conflict in the European Region, 1816-1994.” Conflict Management and Peace Science 28(3): 280–305. Reiter, Dan, Allan C. Stam, and Michael C. Horowitz. 2016. “A Revised Look at Interstate Wars, 1816–2007.” Journal of Conflict Resolution 60(5): 956–976. Repucci, Sarah. 2020. Freedom in the World 2020: A Leaderless Struggle for Democracy. Washington, D.C.: Freedom House. Reuter, Ora John. 2017. The Origins of Dominant Parties: Building Authoritarian Institutions in Post-Soviet Russia. Cambridge, UK; New York: Cambridge University Press. Reuter, Ora John, and Rostislav Turovsky. 2014. “Dominant Party Rule and Legislative Leadership in Authoritarian Regimes.” Party Politics 20(5): 663–674. Reynal-Querol, Marta. 2002. “Ethnicity, Political Systems, and Civil Wars.” Journal of Conflict Resolution 46(1): 29–54. Riedl, Rachel Beatty. 2014. Authoritarian Origins of Democratic Party Systems in Africa. Cambridge, UK; New York: Cambridge University Press. Robert Harmel, Lars Svasand, and Hilmar Mjelde. 2016. “Party institutionalization and de-institutionalization: concepts and indicators.” Paper presented at 44th ECPR Joint Sessions of Workshops: Institutionalization and De-institutionalization of Political Organizations. Pisa, Italy. Roberts, Kenneth M., and Erik Wibbels. 1999. “Party Systems and Electoral Volatility in Latin America: A Test of Economic, Institutional, and Structural Explanations.” American Political Science Review 93(3): 575–590. Robison, Richard. 1993. “Indonesia: Tensions in State and Regime.” In Southeast Asia in the 1990s: Authoritarianism, Democracy and Capitalism. eds. Kevin Hewison, Richard Robison, and Garry Rodan. Sydney: Allen and Unwin, pp. 39–74. Rodden, Jonathan, and Erik Wibbels. 2002. “Beyond the Fiction of Federalism: Macroeconomic Management in Multitiered Systems.” World Politics 54(4): 494–531. Roosa, John. 2006. Pretext for Mass Murder: The September 30th Movement and Suharto’s Coup D’état in Indonesia. Madison: University of Wisconsin Press. Rosenfeld, Bryn. 2020. Autocratic Middle Class: how state dependency reduces the demand for democracy. Princeton: Princeton University Press. Sachs, Jeffrey D., and Andrew M. Warner. 2001. “The Curse of Natural Resources.” European Economic Review 45(4): 827–838. Sarkees, Meredith Reid, and Frank Whelon Wayman. 2010. Resort to War: A Data Guide to Inter-State, Extra-State, Intra-State, and Non-State Wars, 1816-2007. Washington, D.C.: CQ Press. 145 Schattschneider, Elmer E. 1942. Party Government. New York: Holt, Rinehart and Winston. Sidel, John Thayer. 1999. Capital, Coercion, and Crime: Bossism in the Philippines. Stanford, : Stanford University Press. Simmons, Joel W. 2016. The Politics of Technological Progress: Parties, Time Horizons and Long-Term Economic Development. Cambridge, UK; New York: Cambridge University Press. Singer, J. David. 1988. “Reconstructing the correlates of war dataset on material capabilities of states, 1816–1985.” International Interactions 14(2): 115–132. Singer, J. David, Stuart Bremer, and John Stuckey. 1972. “Capability Distribution, Uncertainty, and Major Power War, 1820-1965.” In Peace, War, and Numbers. ed. Bruce Russett. Beverly Hills: Sage, pp. 19–48. Slater, Dan. 2010. Ordering Power: Contentious Politics and Authoritarian Leviathans in Southeast Asia. Cambridge; New York: Cambridge University Press. Snow, David, and Doug McAdam. 2000. “Identity Work in the Context of Social Movements: Clarifying the identity/Movement.” In Self, Identity, and Social Movements. eds. Sheldon Stryker, Timothy J. Owens, and Robert W. White. Minneapolis: University of Minnesota Press, pp. 41–67. Stewart, Frances. 2012. “Religion versus Ethnicity as a Source of Mobilization: Are There Differences?” In Understanding Collective Political Violence, Conflict, Inequality and Ethnicity. ed. Yvan Guichaoua. London: Palgrave Macmillan, pp. 196–221. Stryker, Sheldon. 2000. “Identity Competition: Key to Differential Social Movement Participation?” In Self, Identity, and Social Movements. eds. Sheldon Stryker, Timothy J. Owens, and Robert W. White. Minneapolis: University of Minnesota Press, pp. 21–40. Svolik, Milan W. 2012. The Politics of Authoritarian Rule. Cambridge, UK; New York: Cambridge University Press. Tan, Alexander C., Karl Ho, Kyung-Tae Kang, and Tsung-Chi Yu. 2000. “What if We Don’t Party? Political Partisanship in Taiwan and Korea in the 1990s.” Journal of Asian and African Studies 35(1): 67–84. Tan, Paige Johnson. 2015. “Explaining Party System Institutionalization in Indonesia.” In Party System Institutionalization in Asia: Democracies, Autocracies, and the Shadows of the Past. eds. Allen Hicken and Erik Martinez Kuhonta. New York: Cambridge University Press, pp. 236–259. Tavits, Margit. 2013. Post-Communist Democracies and Party Organization. New York: Cambridge University Press. 146 Thames, Frank C. 2007. “Discipline and Party Institutionalization in Post-Soviet Legislatures.” Party Politics 13(4): 456–477. Theiler, Tobias. 2018. “The Microfoundations of Diversionary Conflict.” Security Studies 27(2): 318–343. Thompson, William R., and David R. Dreyer. 2011. Handbook of International Rivalries, 1494-2010. Washington, D.C.: CQ Press. Tomsa, Dirk. 2008. Party Politics and Democratization in Indonesia: Golkar in the Post- Suharto Era. London; New York: Routledge. Ufen, Andreas. 2008. “Political Parties and Democratization in Indonesia.” In Democratization in Post-Suharto Indonesia. eds. Marco Bünte and Andreas Ufen. London; New York: Routledge, pp. 153–175. Ufen, Andreas. 2007. Political Party and Party System Institutionalisation in Southeast Asia: A Comparison of Indonesia, the Philippines, and Thailand. Hamburg: German Institute of Global and Area Studies. Vatikiotis, Michael R. J. 1998. Indonesian Politics Under Suharto: the Rise and Fall of the New Order. London; New York: Routledge. Vogel, Ezra F. 1991. The Four Little Dragons: The Spread Of Industrialization In East Asia. Cambridge, US; London, UK: Harvard University Press. Vogt, Manuel et al. 2015. “Integrating Data on Ethnicity, Geography, and Conflict: The Ethnic Power Relations Data Set Family.” Journal of Conflict Resolution 59(7): 1327– 1342. Vu, Tuong. 2007. “State Formation and the Origins of Developmental States in South Korea and Indonesia.” Studies in Comparative International Development 41(4): 27–56. Wade, Robert. 1990. Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization. Princeton: Princeton University Press. Waltz, Kenneth Neal. 1979. Theory of International Politics. Boston: McGraw-Hill. Weeks, Jessica L. 2012. “Strongmen and Straw Men: Authoritarian Regimes and the Initiation of International Conflict.” American Political Science Review 106(2): 326– 347. Wegenast, Tim C., and Matthias Basedau. 2014. “Ethnic fractionalization, natural resources and armed conflict.” Conflict Management and Peace Science 31(4): 432–457. Weingast, Barry R. 1995. “The Economic Role of Political Institutions: Market- Preserving Federalism and Economic Development.” Journal of Law, Economics, & Organization 11(1): 1–31. 147 Wintrobe, Ronald. 1998. The Political Economy of Dictatorship. Cambridge, UK; New York: Cambridge University Press. Wolfson, Michael C. 1994. “When Inequalities Diverge.” The American Economic Review 84(2): 353–358. Wong, Joseph. 2015. “South Korea’s Weakly Institutionalized Party System.” In Party System Institutionalization in Asia: Democracies, Autocracies, and the Shadows of the Past. eds. Allen Hicken and Erik Kuhonta. Cambridge, UK; New York: Cambridge University Press, pp. 436–466. Woo-Cumings, Meredith, ed. 1999. The Developmental State. Ithaca: Cornell University Press. Wright, Joseph. 2008. “Do Authoritarian Institutions Constrain? How Legislatures Affect Economic Growth and Investment.” American Journal of Political Science 52(2): 322–343. Wright, Joseph, and Abel Escribà-Folch. 2012. “Authoritarian Institutions and Regime Survival: Transitions to Democracy and Subsequent Autocracy.” British Journal of Political Science 42(2): 283–309. Yardimci-Geyikçi, Şebnem. 2015. “Party Institutionalization and Democratic Consolidation: Turkey and Southern Europe in Comparative Perspective.” Party Politics 21(4): 527–538. Zeng, Jinghan. 2016. The Chinese Communist Party’s Capacity to Rule: Ideology, Legitimacy and Party Cohesion. London: Palgrave Macmillan. Zheng, Yongnian. 2014. “The Institutionalization of the Communist Party and the Party System in China.” In Party System Institutionalization in Asia: Democracies, Autocracies, and the Shadows of the Past. eds. Allen Hicken and Erik Martinez Kuhonta. Cambridge, UK; New York: Cambridge University Press, pp. 162–188. 148