EXTERNAL THREATS, INTERNAL CHALLENGES, AND STATE BUILDING IN EAST ASIA

Abstract

This paper empirically tests bellicist theories of state building in the East Asian context, paying attention to the interplay between external threats and internal challenges and their implications for these states’ extractive power. How much variation in state building in the region can be attributed to war and war preparation as a result of both external threats and internal challenges? In particular, it provides more fine-grained analysis on the different types of internal challenges and their impact on state capacity building. The article argues that in the East Asia region, both external threats and internal challenges are crucial to explaining the variation in state capacity across the region. However, we also find that different types of internal challenges have different effects. Particularly, communist insurgencies seem to have both an immediate and long-term positive effect in compelling the state to respond with more extraction to engage in state-building efforts.

INTRODUCTION

The bellicist theory of state building that emphasizes the crucial role war plays in state capacity has been well established ever since Tilly's work on the European experience (Tilly 1975, 1992). Since then, the theory has been tested in different regional contexts around the world (Centeno 2003; Downing 1992; Ertman 1997; Herbst 1990; Hui 2005; Lu and Thies 2013; Spruyt 1994). Additionally, modifications of Tilly's original theory have sought to differentiate the processes of war making and war preparation as they affect variations in state capacity, including new emphases on military threats in lieu of actual war fighting. Particularly, scholarship that focuses on rivalry processes has made significant contributions to our understanding of the mechanisms of state capacity building (Diehl and Goertz 2001; Thompson 2001). Distinctions have also been made between interstate and civil warfare in order to capture their drastically different implications for state building in many developing countries. However, existing literature tends to emphasize the external threat dimension. In our view, not enough attention has been placed on internal challenges, as different types of internal challenges may produce different effects on state capacity building.

Bellicist theoretical approaches have been tested in a number of contexts and geographical regions; however, so far there has been comparatively little quantitative tests done for East Asia.¹ More importantly, East Asia is an ideal region in which to test the effects of external threats and internal challenges, because it is a region that has experienced a tremendous amount of interstate and civil warfare during the modern period. The region not only suffered from Western imperial expansion and colonialism, it was also subject to Japanese imperial expansion and aggression during World War II. In the post-World War II period the region also became one of the key battlegrounds between global communism and anti-communist coalitions. The two major wars that the United States fought in the post-World War II period to prevent the spread of Communism—the Korean War and the Vietnam War (the Second Indochina War)—occurred in this region. The end of World War II and the onset of the Cold War also ushered in territorial divisions of the region's states, as we can see in the separation of North and South Korea as well as the division between mainland China and Taiwan, which continue to characterize much of the region's strategic rivalry until the present day.

Additionally, the decolonization process in the region in the post-World War II period introduced a strong dose of nationalism among the ethnically diverse population within those supposed “nation-states.” As a result, many of the post-independence states suffered from internal ethno-nationalist insurgencies by groups of people demanding self-determination. Many of these conflicts are still ongoing, for example in the southern Philippines, Myanmar, and southern Thailand. In addition to ethno-nationalist insurgencies that demand secession and autonomy, the region also experienced long periods of communist insurgencies during the Cold War period.

Indeed, the Cold War has been credited as particularly conducive for internal insurgencies as a result of external military support, revolutionary ideologies, and specific military doctrines that emphasized guerilla warfare (Kalyvas and Balcells 2010, 420). In East Asia, the Soviet Union, and especially the People's Republic of China (PRC) with Mao's emphasis on “People's War,” stimulated much of the communist insurgencies in the region (Han 2019a, 2019b). Thus, for states that faced such communist insurgencies, they not only faced these internal challenges to their rule, but they were also vulnerable to the linkages these internal insurgencies had with the external great powers (Cunningham 2016; Salehyan, Gleditsch, and Cunningham 2011; Schultz 2010; Shelton, Stojek, and Sullivan 2013). Indeed, for many East Asian countries within the capitalist camp during the Cold War, their domestic communist mobilizations were both perceived to be and actually were supported by external communist threats to the country. For example, this was the case with South Korea after its independence, when rural rebellions that were supported by North Korean government broke out in the southern part of the country (Cumings 1997). Similarly, the communist insurgencies carried out by Viet Cong in South Vietnam were supported by the North Vietnamese government. It is this type of linkage that we believe makes communist insurgencies pose a higher level of threat to the ruling elites than the ethno-nationalist variety that, comparatively speaking, only had limited political aims. Did such variations in threat perception from these different types of internal challenges, in addition to variation in their intensity, also produce different effects on state capacity building?

Thus despite the fact that the East Asian region has experienced a significant amount of stability and economic dynamism in recent decades, defined by the lack of actual war fighting, especially compared to areas such as Africa or the Middle East, this is still a region with prolonged external rivalries and internal challenges (Solingen 2007). The nuclear crisis on the Korean Peninsula, the Taiwan Strait, territorial disputes in the East and South China Seas, just to name a few, continue to define many of the region's security tensions. This article thus attempts to empirically test bellicist theories of state building in the East Asian context, paying particular attention to the effects of external threats and internal challenges and their implications for these states’ extractive power. How much variation in state building in the region can be attributed to war and war preparation as a result of both external threats and internal challenges? In particular, we aim to provide more fine-grained analysis of the different types of internal challenges, as well as the variation in their intensity, as they affect state capacity building. We argue that in the East Asia region, both external threats and internal challenges are crucial to explaining the variation in state capacity across the region. However, we also find that different types and intensities of internal challenges have different effects over time. Particularly, communist insurgencies seem to have both an immediate and long-term positive effect in compelling the state to respond with more extraction to engage in state-building efforts. This effect is also found when examining the intensity of a communist insurgency, rather than just its type. On the other hand, ethno-nationalist insurgencies that do not aim to capture the whole state have no significant, contemporaneous effect on extraction, though they exhibit longer-term effects. Higher intensity insurgencies of this type exhibit similar positive effects on state capacity, but not of the same magnitude as communist insurgencies.

Perhaps most surprising, we find that a conventional statistical test of bellicist theory that examines the effects of interstate and civil war with a 1,000-casualty threshold produces strong, positive contemporaneous effects on state building as measured by extraction. This finding is unique among the regions of the world, but perhaps theoretically expected given the extreme threat pressures faced by the region during the Cold War era. This effect dissipates over time and becomes negative, which is more consistent with previous findings. Our empirical tests of interstate or strategic rivals finds that they also enhance the state's ability to extract revenues from society in the short and long term. Thus, all in all, most forms of external threat and internal challenges seem to augment East Asian states’ ability to extract with some variation over time.

The article is structured as follows. We first discuss the bellicist theories of state building, laying out the theoretical foundation for conceptualizing interstate rivalries as well as internal challenges and their different implications for state building. In particular, we draw attention to the differences between two broad types of internal challenges, communist insurgencies versus ethno-nationalist secessionist ones, and their potentially different implications on threat perception on the existing states. The next section introduces existing literature on state building in the East Asian region and presents the plethora of regional interstate rivalries and internal challenges. We then introduce our data and research design for the empirical testing. The article concludes with reflections on the empirical and theoretical findings and a call for further research outside the East Asian context.

BELLICIST THEORIES OF STATE BUILDING

The literature on state building, as pioneered by Charles Tilly and others, has consistently emphasized the crucial role that war plays in developing the modern state. As Tilly argues, “to the extent that they are successful in subduing their rivals outside or inside the territory they claim, the wielders of coercion find themselves obliged to administer the lands, goods, and people they acquire; they become involved in extraction of resources, distribution of goods, services, and income, and adjudication of disputes” (Tilly 1992, 20). In the European context, in order to effectively fund the war-making enterprise, European rulers became more efficient in their revenue collection, improved civil administration in exchange for civilian cooperation, and established nationalist symbols to unify the population they governed. In such processes, foundations of modern bureaucratic states were built. As the famous phrase goes, “states make war, and war makes the state” (Tilly 1975, 42).

Empirical studies have generally found support for this relationship among early modern European states and great powers (Cohen, Brown, and Organski 1981; Glete 2001; Rasler and Thompson 1985a, 1985b, 1989). Bellicist theories of state building have also been theorized and tested in non-European contexts. Miguel Centeno, in his study on the relationship between war making and state building in Latin America, confirms how the lack of “total wars” in Latin America historically stunted the growth of the bureaucratic state in the region. Due to the lack of need for mass mobilization for total wars, “limited wars rarely leave positive institutional legacies and often have long-term costs,” such as fiscal or debt crisis, professional military rather than popular participation, alienation from patriotic symbols, and economic downturn (Centeno 2003, 23). However, he also cautions against a straightforward causal relationship between the two, since depending on specific configurations of domestic political and social structures, war provides no guarantee that states would encroach on society. In the Latin American context, “the easy availability of external funding allowed the state the luxury of not coming into conflict with those social sectors that possessed the required resources” (Centeno 2003, 28).

Many other scholars have examined the interplay between external threats and internal challenges and their different effects on state building (Taylor and Botea 2008). Particularly in the post-World War II period, international normative changes meant that states are far less concerned with external conquest than before, and state death as a phenomenon has diminished significantly (Fazal 2011). Some scholars have therefore used warfare as a kind of foil for the failure of states to develop significant levels of capacity in the contemporary developing world, while touting their own particular theoretical explanation (Clapham 2008). Others continued to argue that war should produce stronger states even in the rare instances it is now observed (Desch 1996). Finally, some scholars suggest that instead of looking at actual interstate war fighting, we should look for the effects of other forms of external threat, such as interstate rivalry, on state building. Thies (2004) has argued that interstate rivalry can have effects on state building similar to those of actual wars. In his analyses, he finds interstate rivalry overall has a positive effect on a state's extractive capacity. There are also studies that look at the level of external threat posed by transnational rebels instead of interstate wars and their effect on state building (Kisangani and Pickering 2014). Generally speaking, there seems to be support for a more generalized proposition that external threats in the contemporary world can spur state-building efforts.

Studies on the effect on state building of internal challenges have not reached much consensus. In the Latin American context, civil warfare such as guerilla insurgencies, often lead to the delay of the consolidation of the state, and “state makers had to make important tax concessions to the landed elite and traders, especially in regions of high rural mobilization” (Lopez-Alves 2001, 169). Thies's work (2005, 2006, 2007) that measures internal rivalry has produced a variety of results, in part due to inconsistency in measurement. For example, in a cross-regional, cross-national comparison, internal rivals as measured by the Political Instability Task Force (PITF) measures (ethnic war, genocides/politicides, adverse regime changes, and revolutionary wars) do not have a significant effect on the tax ratio (Thies 2004). In Central America, the combined PITF measure has a significant, negative effect on the tax ratio. When decomposed into ethnic rivals (ethnic war and genocides/politicides) and political rivals (adverse regime changes and revolutionary wars), there is a significant, positive effect of ethnic rivalry on the tax ratio in a cross-regional, cross-national study, though political rivals did not have a significant effect. However, using the ethnic war component alone to measure internal rivals produces a significant, positive effect on the tax ratio. In the Middle East, the measure of enduring internal rivalry (Derouen and Bercovitch 2008) produces a consistently significant, positive effect on extraction. We consider these results on the whole as somewhat inconclusive. We are not convinced that the PITF measures, which reveal only instances of actual fighting with the state, truly captures the full range of internal rivals. Nor is it clear that the political versus ethnic rivalry categorization is necessarily the most nuanced approach to understanding internal rivalries. Furthermore, the effects of civil conflicts on state capacity in East Asia seem mixed where there were clearly spectacular failures in the face of domestic insurgencies, most notably in Vietnam, as well as those where they were defeated, such as in Malaysia, and in those continual ongoing low-intensity ones such as in Myanmar. We thus attempt to forge a regionally sensitive approach to understanding the role of internal challenges in the state-building processes in East Asia by coding the two main forms of internal challenges faced in the region: communist insurgencies and ethno-nationalist ones.

EXTERNAL THREATS, INTERNAL CHALLENGES AND STATE BUILDING IN EAST ASIA

East Asia is now often described as a region of economic dynamism. For example, this region hosts the second and third largest economies of the world (China and Japan), as well as many vibrant economies that define the East Asian developmental state model (Beeson 2009; Naughton and Tsai 2015; Vogel 1991). Despite the overall economic dynamism and the ongoing regionalization efforts linking Northeast Asia with Southeast Asia, the region is nonetheless extremely diverse internally (Ba 2009; Kim 2004). The economic gap between the rich and poor is glaring: some countries, such as Myanmar, Cambodia, and Laos, remain deeply mired in poverty and are consistently defined as least developed nations (United Nations, Economic and Social Commission for Asia and the Pacific 2007), while some countries such as Japan, South Korea, Taiwan, and Singapore have already entered the developed countries club. It is also a region with diverse regime types. There are established democracies such as Japan, South Korea, Taiwan, and Indonesia, together with long-lived communist states such as China, North Korea, and Vietnam. There are also many countries in the region with a long history of military dictatorships, such as Indonesia, Myanmar, the Philippines, and Thailand. With such different historical patterns, these states also penetrated their respective societies to varying degrees. State capacity measured by its extractive power shows that the average tax ratio (taxes as a share of GDP) for the region from 1960 to 2011 is 14 percent, though this historical average ranges from a low of two percent for Myanmar in the early 2000s to 48 percent for Vietnam in the years immediately following the war with the US. These forty-two-year averages also mask tremendous variation across time within each country, as China's tax ratio has steadily increased over time, while Vietnam's has dropped from the high point of the end of the Vietnam War. This variation is illustrated graphically in Figure 1. What role might various forms of external threats and internal challenges play in determining the amount of extraction in which states have engaged?

Figure 1 Variation in the Tax Ratio across East Asia by Country

Quite different from the Latin American and African contexts, East Asia as a region suffers from perennial devastation by both interstate and civil warfare. Indeed, there are many scholars who have compared the East Asian experience to that of Western Europe in terms of warfare and historical development of state systems (Hui 2005; Kiser and Cai 2003; Zhao 2015). Despite the current relative peace in the region and the overall economic dynamism, East Asia is in fact a region that was devastated by Japanese aggression during World War II, the Chinese Civil War, the Korean War, wars in Indochina, and communist and ethno-nationalist insurgencies throughout Southeast Asia, both during the Cold War and post-Cold War years. In addition, East Asia was also one of the core theaters of operation during the Cold War in the battles between communism and anti-communism. The communist victory in the Chinese Civil War ushered in a sustained effort to spread communism in the region as well as the subsequent US-led effort to counter the possible “fall of dominos.” The Cold War also caused the separation of North and South Korea, China and Taiwan, and North and South Vietnam. The unification of Vietnam was achieved through a prolonged and bloody process, while the other two bifurcations of the body politic have continued to define some of the major security tensions in the region (Cha 2002; |Christensen 2002, 2006). Table 1 provides a list of major interstate conflicts in the region since the end of World War II.

Table 1 Major Interstate Conflicts in East Asia since end of World War II

In addition to such actual war fighting, there was also continuing war preparation among states as a result of interstate rivalries, such as between North and South Korea, China and Taiwan, China and Japan, China and Vietnam, Malaysia and Singapore, Malaysia and Indonesia, Thailand and Vietnam, Cambodia and Vietnam, as well as with states outside of the region, such as China and the United States or India (Thompson 2001).

With such wide prevalence of war and war preparation in the region, it is thus worthwhile to examine exactly how much impact these interstate rivalries have on state building in East Asia. With the exception of a few studies that look at the relationship between Cold War dynamics and the East Asian economic miracle (Haggard 2018; Woo-Cumings 1999), there has been little systematic and region-wide quantitative analysis looking at the dynamic relationship between war and war preparation and the variation in state building in the region. In a study on the variation of political regimes in Southeast Asia, Dan Slater argues that the endurance of some authoritarian regimes depends on whether a protection pact can be formed among elite coalitions. Only “when a mass movement with revolutionary aims penetrate the urban sphere, threatening to explode to a communal powder keg in the process,” can elite coalitions be frightened into supporting increased concentration of public authority (Slater 2010, 42). Indeed, Slater's ratchet logic is similar to the bellicist theory of state building in demonstrating that the willingness and capacity of Southeast Asian state elites to extract taxes corresponded with different types and levels of internal threat. Yet, much of his emphasis is on domestic power challenges rather than external security threat in producing the ratchet effect on elites.

Richard Stubbs, in his study of the relationship between war preparation and economic development in East Asia, points out that “nowhere has this lack of attention to the economic consequences of war and preparation for war detracted more from our understanding of events than in East and Southeast Asia” (Stubbs 1999, 337). Although Stubbs’ main focus is to explain the effect of war on economic development, he does argue that in the seven economically successful states of East Asia—Japan, South Korea, Taiwan, Hong Kong, Singapore, Malaysia, and Thailand—institutional states became strong and autonomous as a result of a series of wars in the region (Stubbs 1999, 341). Similar to Stubbs’ study, Doner, Ritchie, and Slater compare South Korea, Taiwan, and Singapore with Malaysia, the Philippines, and Thailand, and point out the developmental state with impressive capabilities in the former three countries “emerged from the challenges of delivering side payments to restive popular sectors under conditions of extreme geopolitical insecurity” (Doner, Ritchie, and Slater 2005, 327). However, both Stubbs and Doner et al. only look at countries that were in the anti-communist camp and did not include in the analyses communist ones, which arguably faced similar survival pressure in making and preparing for war. Certainly there is a large literature that looks at state-building efforts in communist states, particularly in the case of the PRC (Brown and Pickowicz 2007; Pieke 2009); however, almost all such studies are single-country analyses. So far there has not been any study that includes both communist and anti-communist states in East Asia and studies the mutual security threat on their respective domestic state-building processes.

In addition to such external threats, there are also a tremendous number of internal challenges in the region. See Table 2 for a list of communist and ethno-nationalist insurgencies in the region.² As we can see, many countries in the region have suffered from long-term internal challenges. We do notice there is substantial difference between Northeast and Southeast Asia, in that Japan did not face an insurgency in the post-World War II period, while South Korea defeated its own communist insurgency in the late 1940s. However in Southeast Asia, ideologically based communist insurgencies posed serious challenges to these states during the Cold War period (Ettinger 2007; Lintner 1990; Ong 2015). What is peculiar about the communist insurgencies in Southeast Asia was that they were all heavily supported by external great powers, particularly the People's Republic of China. For example, the PRC supported communist insurgencies in Malaysia, Myanmar, Thailand, and Vietnam through both ideological and material means (Baker 2003; Han 2018; Jian Chen 2001; Jie Chen 1994; Lintner 1990). Therefore, such internal challenges were deeply tied to the external threat to those states in Southeast Asia. At the same time, several states in the region also experienced ethno-nationalist challenges to their rule, as a result of the ethnic compositions of such societies and their unequal access to power (Bertrand 2004; McCargo 2008; Smith 1999). Taken all together, East Asia is a region where states have a plethora of external and internal rivalries. Yet, do these forms of rivalry affect state-building efforts in a systematic, measurable way? If so, they may provide additional support to bellicist theory as applied to the contemporary developing world.

Table 2 Internal Insurgency Years in East Asia by Type

Data and Methods

We examine these potential empirical manifestations of bellicist theory with a dataset consisting of 12 East Asian countries from 1960–2011 (unless otherwise noted). The countries include Cambodia (1964–73, 1998–2011), China, Indonesia, Japan, Laos, Malaysia, Myanmar, Philippines, Singapore, South Korea, Thailand, and Vietnam (1979–2011). We were unable to incorporate Brunei, North Korea or Taiwan into the analysis due to missing tax data and other indicators of the economy. Sporadic missing data across the other countries reduces the number of country-years in the final analyses to a maximum of 470 observations.³ Descriptive statistics for all of our variables can be found in Table 3. The choice of statistical method was driven by considerations about the nature of the data. Pooled cross-sectional time-series models often involve violations of the ordinary least squares (OLS) assumptions of homoskedasticity and uncorrelated error terms. While OLS estimates are unbiased in the presence of autocorrelation, these estimates are not efficient, and the variability of OLS coefficients affects the tests of statistical significance. We estimate the following pooled cross-sectional time-series models using the Beck and Katz solution for these problems in OLS: panel corrected standard errors (PCSEs) (Beck and Katz 1995). This approach has frequently been used in the quantitative literature on state building (Thies 2007).

Table 3 Descriptive Statistics

We gauge the extent of state building in each country using a measure of extraction. The tax ratio is the state's tax revenue as a percentage of GDP, which is the conventional gauge of the state's extractive capacity (Lieberman 2002; Thies 2010). The tax ratio reflects the ability of the state to extract resources from both individuals and corporate actors in society. It is by no means a perfect measure of state building or state capacity, as others have suggested.⁴ Yet, without revenue or the “sinews of the state” in Cardinal Richelieu's terms, nothing else can be accomplished by a ruler. By using the tax ratio, we can also directly compare our results to those found in previous studies of the effect of external and internal threats on state building across the globe.

In terms of our independent variables of interest, we use several measures of external threats and internal challenges to examine how these bellicist pressures have affected state building in the region. We start at an extreme level of threat by examining the effects that interstate war and civil war have on state building in the region. We measure interstate and civil war using the Correlates of War data (Sarkees and Wayman 2010). An interstate war involves sustained combat between organized armed forces that results in a minimum of 1,000 battle-related combatant fatalities within a 12-month period. Interstate war is coded dichotomously, such that a state in our dataset that is engaged in an interstate war in a given year receives a score of one and zero otherwise. Similarly, a civil war must meet the same thresholds for interstate war mentioned above and involve combat between the recognized government of a state and some internal rebel group(s). It is also coded dichotomously with a score of one assigned for years in which a civil war occurs and zero otherwise. We examine the contemporaneous, as well as five- and ten-year lagged effects of war and civil war on the tax ratio.

We also consider the claim of Lu and Thies that we should modify traditional bellicist approaches to state building given the declining frequency of interstate war to examine interstate rivalry as an indicator of external threat (Lu and Thies 2013). We use Thompson's concept and operationalization of strategic rivalry in this study (Thompson 2001). Strategic rivalries occur when states view each other as competitors/enemies and the source of actual or potential militarized threats. The perceptions of the decision makers in both states are used to determine the beginning and ending of a strategic rivalry. States experiencing at least one strategic rivalry in a given year receive a score of one and otherwise zero. We examine the contemporaneous, as well as five- and ten-year lagged effects of strategic rivalry on the tax ratio.

Finally, we attempt to bring greater nuance to the understanding of internal challenges in this paper, building off of previous work by Thies, who coded internal challenges as engaged in ethnic conflict (ethnic wars/genocides) with the state or political conflict (abrupt regime changes/revolutionary wars) (Thies 2004). While we think this was a useful first step, we aim to focus on two forms of internal challenges most salient to East Asia. We therefore engaged in our own coding of internal challenges in this region by both type and intensity. The first is the threat posed by the type of communist insurgency.⁵ States experiencing an internal threat from a communist insurgency in a given year receive a score of one and zero otherwise. The second type of internal challenges is that posed by ethno-nationalist insurgencies. States experiencing an internal threat from an ethno-nationalist insurgency in a given year receive a score of one and zero otherwise. The data for internal insurgencies comes from two sources. The first is the UCDP/PRIO Armed Conflict Dataset (Pettersson and Wallensteen 2015). We use the UCDP/PRIO data to identify whether countries experienced insurgency and the name of the insurgent groups in a particular year. We then use the IISS Armed Conflict Database to specifically identify the nature of these insurgencies, whether they are communist or ethno-nationalist.⁶ We examine the effects of communist and ethno-nationalist types of insurgency contemporaneously, as well as their five and ten year lagged effects on extraction.

In addition, we consider the possibility that it is not just the type of insurgency that matters, but their intensity as well. The UCDP/PRIO dataset provides two different measures of intensity. The first we will call intensity, which scores an insurgency as a 1 if the total battle-related deaths are between 25 and 999 and a 2 if the battle-related deaths are 1,000 or more. The second we will call cumulative intensity, which scores an insurgency a 1 when in the history of the conflict it achieves 1,000 battle-related deaths and for every year thereafter. We examine the contemporaneous and lagged effects of ethnic and communist insurgency by intensity on extraction.

Taken as a whole, we have two measures of external threat (war and strategic rivalry) and two sets of measures of internal challenges (civil war and the two forms of insurgency and their intensity levels). They give us the contemporaneous effect of external and internal threat on state building. In addition, we examine their effects over a five- and ten-year period of time to see if temporal dynamics are an important feature of external and internal threats on state building in the region.

We control for a number of factors that are known to influence the opportunity and willingness of states to extract revenue from their subject populations. We include the well-known polity score as a measure of regime type that varies from −10 to + 10. Rulers in different regime types likely face different discount rates and transaction costs in their extraction efforts, and one can easily theorize how democracies might extract more or less than their autocratic counterparts and vice versa. Previous empirical research finds mixed effects for the relationship between democracy and state capacity (Cheibub 1998; Fauvelle-Aymar 1999; |Thies 2004, 2005). US aid per capita captures the effect of an alternate source of revenues to the states of the region. It is therefore typically seen as a substitute for tax revenues, thus it may have a negative effect on our dependent variables. Foreign aid is one mechanism through which the United States attempted to address security threats in the region, whether they were from communist insurgents during the Cold War or potential terrorist groups in the post-September 11 Era. Such projects often had the intention of strengthening the capacity of the state, whether or not that was their actual effect. The data are from the AidData project, which contains detailed time-series data on aid projects funded by the US in constant US dollars in the region for the entire period under study (Tierney et al. 2011). We aggregate these project-level data into country-year amounts and standardize them by the state's population to capture the effect of those aid project dollars per capita. We also create an aid change variable based on the yearly change in aid per country in order to account for variability in such funds with the expectation that increases might be more closely related to attempt to deal with security threats, thereby strengthening the state in the short term.

Ethnic fractionalization and religious fractionalization each attempt to capture the transaction costs associated with extracting in more or less homogeneous societies (Collier and Hoeffler 1998, 556; Thies 2010, 326). Ethnic fractionalization is the probability that two randomly chosen individuals belong to different ethno-linguistic groups. Religious fractionalization is the probability that two randomly chosen individuals belong to different religious groups. Both measures are from Fearon and Laitin (2003). Higher fractionalization should generally produce lower levels of extraction, ceteris paribus. Similarly, GDP per capita (in 2005 constant US dollars), exports as a percentage of GDP, agriculture as a percentage of GDP, and mining as a percentage of GDP also reflect transaction costs that accrue to the ruler attempting to extract from society. GDP per capita, exports as a percentage of GDP, and mining as a share of GDP should increase state revenues, since they reduce transaction costs. Agriculture as a share of GDP is thought to generally increase transaction costs, since it is difficult to tax in rural areas and in predominantly agriculturally based economies. All of these measures are based on typical sources, such as those provided by the International Monetary Fund, World Bank, and United Nations, and contained in the dataset generated by Kugler and Tammen and maintained on their website (Kugler and Tammen 2012).⁷

Analysis

We begin our analysis by looking at the effect of interstate and civil war on extraction in East Asia. This represents a more direct test of bellicist theory, since war has always been thought to be the prime example of external threat that allows a state to extract from society. As show in the first column in Table 4, interstate war has a significant, positive effect on the tax ratio when measured contemporaneously. In years when a state is at war, it achieves on average a bump of almost two percent in the tax ratio. This may not seem large, but the average tax ratio in this region is only 14 percent. East Asia is thus different from Central and South America and the Middle East where the contemporaneous effect of interstate war on the tax ratio is negative (Lu and Thies 2013; Thies 2005, 2006). Yet, by the time five years has passed the positive effect of war on the tax ratio is no longer significant, and by ten years it is actually negative. Our interpretation is that the positive effect in the short run is likely a result of the Cold War pressure. Our expectations about the strong effect of the Cold War turning the region into a major site of contestation is therefore born out in the data. Then, over time, the effect of war begins to resemble the results found in longer term analyses of the rest of the world that show a negative effect of war on state building (Thies and Sobek 2010).

Table 4 The Effect of War and Civil War on State Building in East Asia

Note: All significance tests are two-tailed: *p < .05, ** p < .01, ***p < .001.

Somewhat surprisingly, civil war also has a significant, positive effect on extraction in East Asia. The magnitude of the effect is actually larger than that of interstate war. East Asia is thus different once again from Central and South America and the Middle East where the effect of civil war on the tax ratio is negative. Large-scale internal conflicts that resulted in at least 1,000 battle-deaths thus had an even greater effect on extraction than interstate war. These effects also grow larger over time, as the contemporaneous increase in the tax ratio is 1.4 percent, growing to 4.20 percent over a ten-year period.

The control variables also produce some interesting results for discussion. Overall, GDP per capita is positively related with states’ extractive power, which indicates that wealthier states in this sample extract more from society. This reflects these states’ ability to capture some of the revenues in the economic miracles created by the developmental state. Interestingly, polity scores are negatively correlated with extractive power, which means that more democratic states actually extract fewer resources from society in the short run, though this effect disappears over time. This is similar to Central America (Thies 2006), but different from South America and the Middle East where a positive effect has been previously identified (Lu and Thies 2013; Thies 2005), and Africa where there is no significant finding (Thies 2007). Higher levels of US aid per capita reduce the level of extraction, which has been consistently reported in previous analyses. However, increases in the amount of aid actually contribute to extraction. Thus, while increases in US aid may build capacity, its cumulation into higher levels of aid per capita serves as a substitute for states penetrating their societies further for the purpose of extraction. Ethnic fractionalization reduces extraction consistently across time, while religious fractionalization produces a positive short run effect. Agriculture as a share of GDP reduces the tax ratio, though that effect disappears over time. Mining and exports as a share of GDP both increase extraction in the tax ratio across time.

If we move away from the traditional interpretation of the bellicist approach to state building and focus on the role of internal and external rivalries, we obtain a more nuanced understanding of the dynamics at play in the region. Table 5 contains the contemporaneous effects of external threats and the types/intensity of internal challenges on extraction in East Asia. External threats, as measured by the strategic rivalry variable, produce a significant, positive effect on the tax ratio. The substantive effect is not quite as strong as that of interstate war on the tax ratio coming in close to 1 percent. The general finding is much the same as has been found in Latin America, the Middle East, and Africa. It suggests that although interstate wars may be on the decline, the interstate rivalry mechanism may still serve as a commensurate level of external threat to generate state extractive capacity.

Table 5 The Contemporaneous Effect of External and Internal Threats on State Building in East Asia

Note: All significance tests are two-tailed: *p < .05, ** p < .01, ***p < .001.

However, internal challenges tell us two different stories in terms of contemporaneous effects on the tax ratio. Communist insurgencies, whether classified by type or intensity, seem to drive significant, positive increases in extraction. Measured by type, the increase is an average of 2.5 percent of the tax ratio in a year when a communist insurgency is ongoing, which is an even larger effect than external threats. We expected this, since communist insurgencies posed a greater level of threat since they were also directly connected to the Cold War dynamics at play in the region with heavy involvement by the global superpowers. It is this linkage that we believe makes communist insurgencies pose a higher level of threat to the ruling elites than the ethno-nationalist variety in the short run. We argue that the communist insurgency can be interpreted as a proxy for both internal and external conflicts, which is not true to the same extent with the ethno-nationalist ones. Such heightened existential threats thus propelled these states to extract more from the societies. The effect is present when measured using the intensity or cumulative intensity versions of the communist insurgency variable. Ethno-nationalist insurgencies have no consistent effect in the short run. When measured by type, there is no significant effect. The intensity version of the variable indicates that higher intensity ethno-nationalist insurgencies have a positive effect, but using the cumulative intensity measure they show no significant effect. As expected, since these internal challenges are more limited in their aims, they do not pose a similar level of threat to the existing state in the short run. The control variables tell largely the same story as the models in Table 4.

We also consider the temporal dynamics potentially associated with our measures of insurgency type and intensity. We first examine the five-year lag in Table 6. Consistent with the contemporaneous models, we find that strategic rivalry still exerts a positive effect five years later. The substantive effect is slightly larger in these models than in the contemporaneous versions. Five years on, we also find that both communist and ethno-nationalist insurgencies as measured by type or the two forms of intensity all exert positive effects on the tax ratio. The substantive effects are larger across the board for both as well. What is important to note is that, as we would expect, the substantive effect of communist insurgencies is still much larger than those of ethno-nationalist insurgencies. The five-year effect of communist insurgency (by type) on the tax ratio is 3.3 percent, while that of ethno-nationalist insurgency is 1.5 percent.

Table 6 The Five-Year Effect of External and Internal Threats on State Building in East Asia

Note: All significance tests are two-tailed: *p < .05, ** p < .01, ***p < .001.

If we move ten years out, as show in Table 7, we find that strategic rivalry is no longer an important determinant of the tax ratio. Its effects seem to be largely short-term in East Asia. Communist and ethno-nationalist insurgencies still exert profound effects at the ten-year mark, whether by type or intensity. Communist insurgencies are still more important substantively, but it is clearly the internal threats that are driving long-term state-building efforts in the region. Thus, without considering both the effects of external and internal rivals, previous attempts at testing bellicist theory were missing an important dimension. Internal rivals, especially communist insurgencies, are extremely important for East Asian state building.

Table 7 The Ten-Year Effect of External and Internal Threats on State Building in East Asia

Note: All significance tests are two-tailed: *p < .05, ** p < .01, ***p < .001.

Conclusion

This article has produced the first statistical test of bellicist theory in the East Asian region. While a number of case studies and a few comparative case studies have previously suggested that war has had a transformational effect on the states in the region, we can now add systematic cross-country, over-time comparisons using statistical control to bolster that evidence. Although we admit there might be complex endogeneity problems in our statistical model, particularly with our inclusion of both interstate and intrastate rivalries that might be correlated with each other, we contend our research findings provide a solid and comparable evidence for the bellicist theory of state building. We find strong evidence that measures of external threat associated with traditional or modified approaches to bellicist theory, such as interstate war and interstate rivalry are positively correlated with a state’ extractive capabilities. On average the substantive effect is quite large—a one to two percent increase in the tax ratio in years when states experience an interstate war or an interstate rivalry. Yet, our extension of tests of bellicist theory beyond the typical contemporaneous findings also showed that the positive effect of interstate war disappears by year five and becomes negative by year ten. The effect of strategic rivalry persists through at least year five, but disappears by year ten. Thus, the effects of external threats are relatively short-term in the East Asian region, pointing to the need to explore these dynamics across the rest of the world.

Similarly, we find that large-scale civil war associated with 1,000 battle deaths or more also produces an almost two percent contemporaneous increase in the tax ratio that only grows larger over time. Civil wars in East Asia clearly produce strong responses from states attempting to neutralize the internal challenges they represent. Our attempt to be more precise at coding internal challenges produced the expected finding that communist insurgencies prompted increased extraction, as they pose a similar existential threat to external rivals. Further, they were most likely tied into the Cold War dynamics associated with the superpowers and regional contention for supremacy, in that they produced an immediate contemporaneous effect not found with the ethno-nationalist insurgencies. Ethno-nationalist insurgencies, on the other hand, had no consistent contemporaneous effects. Yet, over time, both ethno-nationalist and communist insurgencies produced increased extraction. Even so, the substantive effects of the communist insurgencies were nearly twice as high as that of the ethno-nationalist ones. Considering the type of insurgency, the intensity of those types of insurgency as well as temporal dynamics are all innovations in the bellicist approach. The results found for East Asian internal threats demand comparisons with other regions. Whether communist insurgencies elsewhere also have such positive effects on state capacity building would need further empirical testing.

East Asia has been compared to early modern Europe in terms of the heavy role that warfare has played in the development of its states. East Asia turns out to be quite different from other contemporary regions in this regard. We have demonstrated that the role of external threats appears to be similar to other regions, and our nuance in measuring internal challenges makes it an interesting comparison with other regions. Indeed, a recent study comparing Latin America with Southeast Asia in terms of the role of internal warfare on taxation illustrates a very similar logic as the one discussed here (Rodríguez-Franco 2016). Overall, we believe this article has produced an accurate picture of the role of external threats and internal challenges on the extractive capacity of the East Asian state since 1960. Having said that, we also admit there are a tremendous number of intra-regional variations between Northeast and Southeast Asia in terms of state capacity. Previous scholarship has noted, for example, the different colonial legacies on economic development in the region, by pointing out the differences between Japanese colonial policies in Korea and Taiwan versus the ones implemented by European powers in Southeast Asia (Cumings 2002; Kohli 1994). Thus, different patterns of investment in infrastructure, education, resource extraction, and agricultural development during the colonial period set some of the foundations for later divergence of state capacities between Northeast and Southeast Asian states.

Given current developments in the region, we would expect these processes to continue in some fashion. The Cold War overlay is gone, and indeed almost all communist insurgencies in the region have gone, but many other conflict processes continue unabated. China's aspirations as a rising power globally and in the region have led to increased confrontation in the South China Sea with overlapping territorial and exclusive economic zone claims. North Korea continues to intensify its nuclear activities and provocations with missile tests over South Korea and even Japan. China and Japan continue to dispute the Senkaku/Diaoyu Islands, as well as Japan's responsibilities for its World War II activities. And all this is happening as Japan removes its constitutional restraints on the use of force. The United States remains an active participant in these security dilemmas with its ongoing rebalance to Asia, but it is not exerting anything like the same impact it had on the region while fighting wars on the Korean Peninsula or in Vietnam. Regardless, our expectation is that the region will continue to be awash in external threats as well as ongoing internal challenges that will continue to shape the extractive capacity of East Asian states.