Capitalism and nationalism in the longue durée : Hegemony,crisis,and state-seeking nationalist mobilization,1492–2013

Hegemony, crisis, and nationalism

The struggles between rulers and geographically concentrated populations are not products of capitalism. They were in existence probably since the emergence of first known states in ancient Mesopotamia. Formation of the capitalist world-system, however, radically transformed the nature of these struggles in two ways. First, existing states began to operate within an interstate system whereby each state recognized and competed with each other. Second, rulers of these states started to deploy varying combinations of coercion and consent to exercise state power and to represent particularistic interests of their state as general interests of the people. Put differently, they began to exercise state power not through brute force but through what Gramsci (1971) called “hegemony” (pp. 57–58). These two developments reinforced each other. Counterbalancing coercion with varying degrees of consent not only helped rulers to mobilize masses more effectively for production, protection, and administration purposes, but also gave them comparative advantage in their economic and military competition against other states in the interstate system. Through competition, successful hegemony-building strategies were emulated by other rulers and adapted into new contexts.

These various hegemony-building strategies also helped rulers to produce more stable national states by more effectively containing internal threats and “national problems.” Existing comparative-historical research supports this view by suggesting that militarily suppression is not sufficient to produce stable polities and to avoid nationalist rebellions. To produce stable “national states,” rulers should also offer to their people—especially to geographically concentrated minority communities—public goods (Wimmer, 2013), political-economic rights (Gurr, 1993; Hechter, 2013), and certain degrees of autonomy (Hechter, 2000). We can consider these proposals efforts by rulers to establish different forms of “social compacts” (Milton, 2007; Silver, 2003) between the state and people. In states with active “national problems,” such social compacts can temporarily help the rulers to reduce minority groups’ demands for self-governance and independence and thus, strengthen their hegemony of the state over people (see Milton, 2007).

In a capitalist world-economy and a competitive interstate system, however, rulers cannot exercise an effective combination of coercion and consent at their will. This is because rulers’ capacity to establish social compacts depend in large measure on the resources available to rulers. Availability of such resources is closely linked to booms and bursts of historical capitalism and its uneven development on a world-scale. This is why periods of intense economic crisis and stagnation are not fertile for making new social compacts or maintaining existing ones. Even in regions with high levels of economic development, prolonged economic crisis, and stagnation can reduce available resources for redistribution and push rulers to unmake existing social compacts, reducing their capacity to contain social unrest (Silver, 2003) and nationalist movements (Milton, 2007).

In addition to economic crises and stagnation, geopolitical crises (such as interstate wars, inter-great power rivalries) also affect state-seeking nationalist mobilization by reducing rulers’ capacity to effectively combine coercion and consent. In the aftermath of major wars, as Gramsci (1971) once put it, “cracks opened up everywhere in the hegemonic apparatus, and the exercise of hegemony became permanently difficult and aleatory” (p. 80). During such crises and wars, rulers face increased difficulty in using coercion effectively against internal threats, including state-seeking nationalists’ movements. In these periods, imperial powers and rival states are more likely to provide material and logistic aid to SSNM to weaken their opponents (Gurr, 1993; Mayall, 1994), which serves as an opportunity for state-seeking nationalist mobilization. In addition, economic costs of war-making, war-time needs for centralization, and extraction of local resources can also push existing elites to unmake previously granted economic, social, and political privileges, causing major grievances that can be used by state-seeking nationalists to mobilize masses. Consequently, similar to economic crises, geopolitical crises, and wars also become fertile grounds for revolts, rebellions, and revolutions in general (Lenin, 1963; Skocpol, 1979) and for state-seeking nationalist mobilization in particular (Wimmer, 2013).

Escalation of social crisis and high levels of social unrest in a given territory can also increase the likelihood of state-seeking nationalist mobilization by weakening the hegemony-building capacity of rulers. The relationship between social movements and (state-seeking) nationalist movements is a complex one. Sometimes, these movements can reinforce each other by exchanging ideas, forms of public action, organizational vehicles, symbols, and slogans from other social movements (Markoff, 1996: 27–29) or by coalescing into a broader revolutionary force (see Silver and Slater, 1999) as the Bolsheviks successfully achieved during the October Revolution, and at a wider scale after the Baku Congress of 1920.

As William G Martin (2008) correctly observes, however, “networks and exchanges among [different types of anti-systemic] movements d[o] not always produce (as is often assumed) cooperative or felicitous relationships among movements” (p. 169). Thus, an analysis of the relationship between social movements and (state-seeking) nationalist movements must carefully consider various antagonisms between them. Stating this, we do not necessarily need to assume cooperative and felicitous relationships among movements to suggest that escalation of social movements can help state-seeking nationalist mobilization. As the more recent revival of secessionist movements in Donetsk and Luhansk after the 2013 Euro-Maidan protests and the 2014 Ukrainian revolution shows, state-seeking movements can also emerge as a reaction to existing social revolutions. Strong social movements, revolts, and rebellions can also provide a more fertile environment for state-seeking nationalist mobilization regardless of their intention, if they can destabilize existing state apparatuses (Goldstone, 2014; Skocpol, 1979). As Beissinger (2002) observes in the case of the nationalist movements in the USSR from 1988 to 1992, many SSNM can “ride the tide” of conflict generated by other actors and movements. More importantly, however, there is a finer line between social movements and state-seeking nationalism than is widely assumed in the literature. As Gramsci observes in Southern Italy, national problems do not necessarily manifest themselves directly and immediately as state-seeking nationalism but first as social movements that address existing inequalities and injustices (also see Smith, 1971: 82; Wallerstein, 1961: 48–52). To use Albert Hirschman’s (1970) famous categories, sometimes secession becomes an “exit” strategy for geographically concentrated social movements when they realize that their “voices” will not be heard by the rulers and “loyalty” is no longer an option.

Financialization and world hegemonic crisis

If these arguments are correct, then financialization periods of capitalist world-economy must also be very fertile for state-seeking mobilization. Although it is widely discussed as a recent and novel phenomenon by economists and economic sociologists, financialization actually is a recurrent feature of historical capitalism from 14th century to present. Building on observations and insights of French historian Fernand Braudel (1992), Giovanni Arrighi (1994) argues that all leading government-business enterprises of historical capitalism from northern Italian city-states in the 14th century to the United States in the 20th century led the world through a period of financialization when it became extremely difficult to sustain high-profit rates from production and trade due to increased competition or crisis of overaccumulation. These structural crises that push the leading business-government complexes of the capitalist world-economy to financialize the world-economy are known as the signal crises.

Financial expansion periods temporarily generate super profits mainly from financial speculation and intermediation and help produce a corresponding Golden Age (i.e. a belle époque) for the leading business-government complexes in the short run. However, they have several negative effects in the medium-run. In these periods, international political-economy turns into a crisis-ridden zero-sum game (Arrighi, 1994; Wallerstein, 1974), inter-enterprise competition and rivalry escalates (Arrighi, 1994; Arrighi and Silver, 1999; Go, 2011), social and political crises in the world intensify, and revolutions erupt (Silver and Slater, 1999). These interlinked crises become symptoms of world hegemonic crises (Arrighi and Silver, 1999: 21–36; Chase-Dunn et al., 2000: 81), signaling the beginning of the dissolution of world hegemonic orders. Instead of solving signal crises for good, financial expansions end up leading to a deeper crises, known as terminal crises, which mark the end of a world hegemony and the transition to a new world order. Financial expansion periods, and the attendant deepening systemic chaos, are also the periods during which the preconditions for a new material expansion, systemic cycle of accumulation, and hegemonic order as shown in Figure 1 and Table 1.

OPEN IN VIEWER

Figure 1.

Systemic cycles of capitalist accumulation.

OPEN IN VIEWER

Table 1.

Approximate dates for signal crisis, terminal crisis, and the end of each period.

	Beginning of the systemic cycle of accumulation	Beginning of the world hegemonic order	Signal crisis (the onset of hegemonic crisis and the beginning of the hegemonic transition)	Terminal crisis(beginning of the “Chaos” period of the hegemonic transition)	End of hegemonic transition and the systemic cycle of accumulation
|------------------Material expansion --------------->
|---------------------- Material expansion ------------------|-------------- Financial expansion --------------------------|
<----------Financial expansion ---|
Genoese-Iberian systemic cycle (no world hegemony)	1450	N/A [1492/1494] a (Peace of Tordesillas)	1560	1625	N/A [1648] a
Dutch systemic cycle and Dutch world hegemony	1560	1648 (Peace of Westphalia)	1760	1790	1815
British systemic cycle and British world hegemony	1760	1815 (Congress of Vienna)	1873/1896	1929	1945
US systemic cycle and US world hegemony	1873	1945 (Foundation of the United Nations)	1973/1980	2007/2008 b	N/A

See Arrighi (1994) for a timeline that illustrates these periods. These are approximate dates. In the statistical analysis, to assess whether or not the findings are robust to the dates chosen, I replicated the same analysis presented in this article using alternate date combinations (±5 years). Results are robust.

a

Genoese-Iberian systemic cycles did not create an Iberian-centered world hegemony in the capitalist world-economy. Thus, there are no beginning and end dates for world hegemonic orders. Yet, in our analysis, to be able to compare state-seeking movements in this period to other “hegemonic consolidation” and “hegemonic transition” periods, we will treat 1492–1560 period as era, which is analogous to an era of world hegemonic consolidation; and the period from 1560 to 1648 as an era which is analogous to world hegemonic transition.

b

Because the crisis of the US world hegemony is an ongoing process, it is not possible to assess a date for its terminal crisis or to state whether there will be a terminal crisis. Recognizing this fact, however, Silver and Arrighi (2011) suggest that the recent 2007/2008 crisis has resemblances with terminal crisis of previous world hegemonies.

Based on these observations, I argue that financial expansion and world hegemonic crisis/transition periods should also produce a more favorable macro-structural climate for state-seeking nationalist organizations to mobilize the masses than material expansion and world hegemonic consolidation periods. This is because these periods of history reduce rulers’ capacity to use coercion and consent effectively at the local level. As shown in Figure 2, some of the effects of financial expansion/world hegemonic transition on SSNM are through the intermediation of (1) economic crises, (2) geopolitical crises, and (3) social crises at the local level. Some of its effects, however, cannot be reduced to these crises alone. Financialization periods have emergent properties. That is, the combined and interactive effects of these multiple crises and conflicts are more and qualitatively different than the sum of their effects in isolation. Financialization produces contentious conjunctures at the global level due to prolonged interactions of a series of crises in the world. These contentious conjunctures are linked but not reducible to these particular crises and conflicts. These periods signal major instability at the world-system level characterized by unmaking of social compacts, declining legitimacy of governance, rapid social change, increased economic and status-based anxieties, and perceived opportunities for independence by state-seeking nationalist organizations.

OPEN IN VIEWER

Figure 2.

Financialization, crisis, and state-seeking nationalist mobilization.

However, these structural opportunities disappear with the emergence of a new world hegemonic order. The reconstitution of the interstate system (i.e. after the Peace of Westphalia in 1648, the Vienna Congress in 1815, or the establishment of the United Nations in 1945), the establishment of a new world hegemony (i.e. Dutch world hegemony, British world hegemony, and US world hegemony), and the simultaneous demise of financialization and the beginning of a new period of material expansion of trade and production create favorable conditions for rulers to defeat secessionist movements through brute force and/or to co-opt them through the distribution of new rights or privileges. Consequently, the frequency of SSNM declines in these periods in regions that benefit from geopolitical stability and economic growth.

In short, the regular alternation of periods of material and financial expansion (or hegemony and hegemonic crisis) at the world level produces successive global waves of SSNM over the longue durée of historical capitalism. If this argument is correct, it would bring not only a new explanation to the global clustering of nationalist movements—observed by many macro-historical sociologists such as Frank and Fuentes (1990), Calhoun (1997), Silver and Slater (1999), Boswell and Chase-Dunn (2000), and Martin (2008) to name a few—but also a new interpretation of Lenin’s (1963) theory linking the rise of national liberation movements to domination of finance capital during the age of imperialism (also see Smith, 1971: 79–71). Similar to Lenin’s theory, the perspective presented in this article also explains why periods of capitalism dominated by finance capital are prone to social revolutions and national liberation movements. Different from Lenin’s approach, however, in this conceptualization, domination of finance capital is not the highest stage but a recurrent phase of historical capitalism. Thus, instead of an age of national liberation, we see successive global waves of state-seeking nationalist mobilization.

Agency, uneven development, and colonization

It is important to note that these structural opportunities provided by multiple forms of crises will not automatically turn into SSNM unless there are existing nationalist organizations on the ground, waiting to exploit such opportunities for mobilization. Existence of nationalist organizations is necessary to articulate these problems through the lenses of a nationalist ideology (emphasizing the necessity to form a separate state as a solution), and to mobilize people for an exit strategy. However, existence of such organizations does not necessarily guarantee that people will follow these organizations. While multiplicity of crises explained above serve as important structural opportunities, another equally important aspect of such mobilization is people’s perception of independence—or formation strong movements—as an actual possibility. Historically speaking, people are more easily mobilized for independence when other nations start to gain their independence or start strong movements.

Finally, regions under imperial domination are more likely to produce SSNM because imperial domination relies mostly on coercion but not on consent. We cannot properly explain the temporal-spatial patterning of anti-colonial forms of SSNM, however, without dealing with uneven development of capitalism in space and time. The literature focusing on decolonization finds compelling evidence for an increase in decolonization movements during periods of world hegemonies (Bergesen and Schoenberg, 1980; Boswell, 1989; Strang, 1991). However, quantitative research that applies this theory to all forms of nationalism in the world-system does not find robust evidence to support these claims (Wimmer, 2013: 98). This discrepancy emerges because macro-level dynamics of peripheral decolonization cannot be directly applied to macro-level dynamics of state-seeking nationalism movements in core (or semi-peripheral) regions, because of their different structural positions in the capitalist world-economy and the interstate system. Since the peripheral regions of the world-economy do not share fully in the economic and political benefits of the material expansion (Arrighi, 1994; Wallerstein, 1974), elites in peripheral colonies are unable to increase their consent-making capacities during material expansion periods. During world hegemonies, they do not benefit from a stable geopolitical environment either. After all, when peace is established among great powers during world hegemonies, the periphery often turns into a battlefield for great power rivalries and proxy wars. Moreover, when a new world hegemony is established, new hegemonic powers attempt to liquidate the peripheral colonies of former or contemporary rival hegemonic powers and thus to weaken their opponents and to gain the leadership in a new interstate system. Hence, due to the uneven geographical development of capitalism, SSNM in peripheral regions of the world-economy—especially anti-colonial nationalist movements—do not decrease during material expansion and world hegemonic consolidation periods but further increase.

Hypotheses

All of these arguments can formally be expressed in terms of the following hypotheses:

SSNM dataset 1

The first dataset includes articles reporting on a wide spectrum of state-seeking nationalist activities throughout the world—including democratic rallies, peaceful protests, referendum proposals, and violent actions as well as nationalist-secessionist warfare—from 1804 to 2013 using The Guardian/Observer (1804–2013) and The New York Times (1851–2013) newspapers. Among all newspapers available in the ProQuest Historical Newspapers digital archives, ¹ I chose The Guardian/Observer and The New York Times as the sources of this dataset due to their (1) longer temporal scope, (2) wider geographical scope of coverage, and (3) ability to produce a higher frequency of articles published on nationalist mobilization on a world-scale. An additional consideration was that they were major newspapers of the world hegemonic powers (i.e. the United Kingdom and the United States) in the last two centuries covered by the dataset. World hegemonic powers, by definition, take the entire world as their sphere of interest or influence; hence, their major newspapers report global affairs more than newspapers of other countries (see Silver, 2003: 191).

Reliability and validity of SSNM dataset 1

Data collection from newspaper archives is a widely used strategy in the social sciences (Beissinger, 2002; Franzosi, 1987; McAdam, 1982; Tilly, 1978), and some scholars find newspaper reports more reliable than official statistics and alternative methods like household surveys (Silver, 2003; Varshney et al., 2008; Wilkinson, 2006). Despite its advantages, however, use of news reports for analyzing forms of social unrest contains potential biases (Earl et al., 2004; Franzosi, 1987). While it is impossible to completely eliminate all potential biases in any data collection strategy, researchers can detect and use strategies to minimize them. Given the limited scope of this article, I will briefly mention four of these strategies I applied to assure the reliability and validity of the dataset below.

First, to avoid potential data collection bias (Earl et al., 2004: 68) due to faulty data collection schemes, I did not rely on any sampling scheme but coded all news reports produced by the keyword string. Likewise, I relied neither on indexes and subject categories produced by newspapers, nor on fully-automated data collection techniques, that can produce a high-rate of false positive results. Instead, all data are coded manually by individually reading each of these articles. To avoid potential researcher bias, I did not rely on my own knowledge on these cases but only on the information available in these news reports. To avoid other forms of researcher biases due to manual data-entry and coding errors in the data collection process (Franzosi, 1987), all articles were double-coded by the same researcher (99.9% overlap in decisions); and a randomly selected 10% of the articles were double-coded by assistant researchers who received 20-hour long training about the data collection process through coding instructions (97.3% overlap in decisions).

Second, to avoid possible description bias due to incorrect (or missing) information in news reports (Earl et al., 2004: 2–73), I did not collect information such as “who did what to whom?” “why did the conflict/protest start?” or “how many people participated (or died)?” Considering that geopolitical interests would seriously affect the way these newspapers would frame and report about the actions by SSNM in different parts of the world, I limited the data collection only to essential information—which would be reported more or less in a similar fashion by the supporters, opponents, or neutral observers of these movements—such as date, name of the state-seeking nation (including alternative names, spellings, etc.) and name of the state against which state-seeking nationalists mobilized.

Third, I constructed weights for the dependent variable to account for the potential temporal selection bias due to different frequencies of news reports produced by newspapers at different points in history. In historical research using news reports as data sources, increase in reporting capabilities of newspapers—and the corresponding increase in the number of pages and words newspapers dedicate for reporting international news—over time can be a matter of serious concern. To illustrate, on average, total number of pages of The Guardian/Observer in a given week was 12 in the 1830s, 60 in the 1880s, 130 in the 1930s, 64 in the 1950s, 160 in the 1970s, 420 in the 1990s and approximately 1000 in the 2000s. Such changes in the physical format of the newspapers can produce an arbitrary increase of the frequency of reported events. Alternatively, there are periods of history—such as during the Second World War—when newspapers’ reporting capacities decline due to various misfortunes including paper shortages.

The weighted form of the dependent variable aims to avoid any arbitrary inflation (or deflation) in frequency of SSNM due to the increasing (or decreasing) pages of newspapers over time. To establish these weights, I sampled and examined papers published from 1st June to 30th June each year (from 1804 to 2013 for The Guardian/Observer, and from 1851 to 2013 for The New York Times) and calculated their total number of pages and number of pages which report about international news. Based on my analysis, I constructed weights that are inversely proportional to the logarithm ⁷ of the average number of pages these newspapers had per week each year as shown in Figure 3. The statistical analysis presented in the article is conducted using both the weighted and unweighted data. Findings are robust.

OPEN IN VIEWER

Figure 3.

Average number of pages in The Guardian/Observer (in logarithmic scale) and weights used.

Fourth, I created a binary version of the dataset to minimize the potential geographical selection bias that can emerge due to different frequencies of news reports produced by newspapers about different countries in the world. ⁸ As mentioned earlier, when choosing the potential news sources for this project, I already identified the sources with the widest geographical coverage capacity. The problem is that even newspapers with the widest geographical coverage capacity might produce more news reports about some regions of the world than others. For instance, most American newspapers have the tendency to produce more news about Latin America than British newspapers. Whereas most British newspapers produce more news about their former colonies than most American newspapers. By the same logic, many French newspapers have a tendency to produce more news about North Africa, French Indochina, and French speaking parts of Europe than the American and British papers. This problem is important because if newspapers tend to report more news about some regions than others, it may arbitrarily increase (or decrease) the frequency of movements we observe in the database in such regions.

The binary version of the SSNM dataset 1 aims to account for this potential geographical selection bias. In this binary version of the dataset, SSNM in a particular territory gets a value of “1” if there is at least one mention of a SSNM in a given year, and a value of “0” if there is no mention at all. By deflating all frequencies into a binary, this version of the dataset makes sure that there are no arbitrary fluctuations in the frequency of state-seeking nationalist mobilization due to potential geographical selection bias. As I will document in the analysis section, replication of the statistical analysis using the binary version of the SSNM dataset 1 produces the same results.

Further reliability tests also show that the SSNM dataset 1 includes all incidents of state formation when states are founded by SSNM (as opposed to non-nationalist social revolutions or international agreements) and overwhelming majority of secessionist/ethno-nationalist warfare as reported by other existing datasets. ⁹ However, other existing datasets of state formation and ethno-national warfare do not capture majority of SSNM as reported by the SSNM dataset 1.

SSNM dataset 2

It is impossible to use the same data collection method for the period that spans from the late-15th to early-19th centuries. Hence, the second dataset of the SSNM database utilizes a different methodology. Building upon Tilly’s (1993, 1994) data on revolutionary situations in Europe, I calculated the number of revolutionary situations and high-intensity conflicts involving state-seeking movements from 1492 to 1829, using the same territorial units I used in the first dataset.

Going over each revolutionary situation in Tilly’s (1993) data from 1492 to 1829, I coded revolutionary situations involving state-seeking movements but I extended (and revised) these data in two significant ways. First, I extended Tilly’s list of “national revolutionary situations” by including instances of state formation and high-level of conflicts where subordinate groups demanded independence. For this task, I relied on four volumes of Minahan’s (2002) Encyclopedia of Stateless Nations, Brecke’s (2012) Conflict Catalog, Hewitt and Cheetnam’s (2000) Encyclopedia of Modern Separatist Movements, historical nationalist movements listed by Smith (1971) and a wide spectrum of secondary historical resources on these conflicts. Because the SSNM database focuses on state-seeking nationalism, I excluded conflicts and reactions against state-led nationalist policies of homogenization if there is no evidence for demands of independence or autonomy by subject nations. Second, Tilly’s data include Western Europe and Eastern Europe, but does not include regions in North America or Latin America that have already been incorporated into the capitalist world-economy in this era. Therefore, I expanded Tilly’s data geographically in order to include state-seeking movements in the Americas as well.

The complete list of state-seeking nationalist revolutionary situations and high-intensity conflicts from 1492 to 1829, covered by the SSNM dataset 2 can be found in Appendix A. For the purposes of this article, I aggregated the data at the territory-year level using the same territorial units used in SSNM dataset 1. The final outcome is 446 territory-years of state-seeking nationalist revolutionary situations and high-intensity conflicts produced by 86 distinct SSNM.

It is important to note that from the 16th century to the 19th century, the SSNM database includes revolutionary situations and high-intensity conflicts involving state-seeking movements. After the 19th century, however, we start to observe a wider repertoire of state-seeking activities, including protests, rallies, demonstrations, and independence referenda, as well as those secessionist armed conflicts, revolutionary situations, and secessionist wars that managed to make it into the international news reports. On the one hand, this distinction is imposed by technical and methodological difficulties, as we do not have a reliable singular news source with world-level reporting capacity for the earlier period. On the other hand, this distinction captures a major historical transformation that took place in the 19th century: the dual rise of democracy and democratic social movements from below (Markoff, 1996: 24–26). As nationalist ideologies started to spread, nationalist organizations emerged, democratization movements became successful, national self-determination was embraced as a democratic principle, and as nationalist movements became social movements, a wider repertoire of actions (including non-violent and normative ones) were used by nationalist groups and organizations. Our two datasets capture this major transformation, and the overlapping years in both datasets, the 1804–1829 period, is intentional. As Figure 4 illustrates, there is no major discrepancy in the patterning/frequency of state-seeking movements in both datasets in this overlapping period, suggesting a smooth transition from one form of action to the other.

OPEN IN VIEWER

Figure 4.

Comparison of SSNM datasets 1 and 2, overlapping period (1804–1829).

Analytical strategy

Assessing the explanatory power of competing theories of nationalism (or any other topic in social sciences) using quantitative methods is a major challenge. Although some scholars believe that large-scale datasets are becoming very helpful in falsifying theories that are not supported by empirical/historical evidence, the problem is that many serious theories of nationalism (as in many other subject matters in comparative-historical sociology) do not easily lend themselves to quantification. From Ernest Gellner to Anthony Smith, to Eric Hobsbawm, most theorists of nationalism provide us with assemblages of extremely complex arguments, insights, and observations that cannot be turned into a few testable hypotheses without oversimplifying the subject matter at hand. It is also very difficult to find high-quality historical data that can be used as variables and valid proxies in the analysis. That is why existing skepticism of some comparative-historical sociologists toward testing competing theories of nationalism using quantitative methods must not be disregarded. Rather, it must be seriously taken into consideration by making clear that the quantitative analysis used in comparative-historical studies does not aim to establish causality or to falsify existing theories in a Popperite fashion, but merely to unearth some broad patterns and relationships hidden in large-scale data, which should later be the object of a serious historical sociological inquiry. Thus, the methods and analyses used in this article must be seen as a first macro-quantitative cut for the beginning of a broader and rigorous historical and empirical inquiry, rather than the last word, on a discussion on some of the existing theories of nationalism.

With this caution in mind, I will proceed in two steps. First, I will analyze formation and dissolution of global waves of state-seeking nationalism from 1492 to present, using the two datasets of the SSNM database. In this section, both the unit of analysis and the unit of observation are the world-system observed at individual years. This analysis will provide us with a bird’s-eye view to the historical trajectory of state-seeking nationalism in the longue durée. Second, I will use SSNM dataset 1 to analyze the explanatory power of the theoretical perspective presented in this article against some of the existing approaches in the literature using multivariate regression analysis. In this section, the unit of analysis will still be the world-system but the unit of observation will be territory-year. Thus, we will explore the dynamics of state-seeking nationalism in the world by analyzing how it is affected by different variables at the local level. Statistical analysis in this section focuses on the effects of a series of covariates derived from some of the competing theories of nationalism on frequency (and hence, likelihood) of SSNM. The dependent variable is frequency of SSNM (i.e. count data), which shows properties of overdispersion (i.e. its variance is higher than its mean). Therefore, I will use a negative binomial regression analysis to estimate how independent variables affect the frequency of state-seeking movements.

Table 2 shows explanations and descriptive statistics of the variables used in the models. Most independent variables are available only for the 1816–2001 period and for 137 territories. All regression models include a 1-year lag for all independent variables. Pearson correlations (Table 3) and variance inflation factor scores (see Table 6) suggest no multicollinearity problem in the models used.

OPEN IN VIEWER

Table 2.

Variables and descriptive statistics.

Variable	Explanation	Observations	Mean	Standard deviation
SSNM	Frequency of state-seeking movements in the territory in a given year as measured from international news reports (SSNM dataset 1)	25,482	0.42	3.21
Ethnic diversity	Ethnic fractalization index score (Fearon and Laitin, 2003; Wimmer and Min, 2006)	25,482	0.41	0.28
Religious diversity	Religious fractalization index score (Fearon and Laitin, 2003; Wimmer and Min, 2006)	25,482	0.37	0.22
Iron-steel production	Iron and steel production (thousands of tons) as calculated by Singer (1987) and updated by Correlates of War project.	25,345	1260.49	8560.98
GDP per capita	GDP per capita in Geary-Khamis (PPPs) as calculated by Maddison (2003). Missing years are imputed using linear interpolation and extrapolation based on the growth rate of the closest neighbor with a similar geo-economic standing (Karataşlı 2017).	25,345	1874.82	2847.88
Globalization	Global trade openness index score as calculated by Chase-Dunn et al. (2000) as world average of the country-level ratio of the level of external trade (e.g. imports) to the GDP.	25,345	0.11	0.04
State power	Composite index of national capabilities score as measured by Singer (1987) using military expenditure, number of soldiers, iron-steel production, energy consumption, urbanization and population size in a given country relative to the world, updated by Wimmer and Feinstein (2010). Centered around the mean.	24,972	0.00	0.07
Recent nation-state formation in the empire	Number of nation-states formed in the empire in the last 5 years (Wimmer and Feinstein, 2010).	25,482	0.15	0.84
Recent nation-state formation in neighborhood	Number of nation-states formed in neighbor territories in the last 5 years (Wimmer and Feinstein, 2010).	25,482	0.15	0.55
Nationalist organizations	Dichotomous variable (1 = there is at least one nationalist organization; 0 = no nationalist organization) calculated by Wimmer and Feinstein (2010).	25,345	0.58	0.49
Duration of nationalist propaganda	Years passed since first nationalist organization is founded, calculated by Wimmer and Feinstein (2010).	25,482	41.00	53.00
Imperial dependency	Dichotomous variable (1 = the territory is an imperial dependency; 0 = the territory is not an imperial dependency) as calculated by Wimmer and Feinstein (2010).	25,345	0.42	0.49
Number of wars in the empire	Number of wars interstate wars fought in the empire as calculated by Wimmer and Feinstein (2010).	25,345	0.36	0.98
Number of wars in the territory	Number of wars interstate wars fought in the territory as calculated by Wimmer and Feinstein (2010).	25,345	0.09	0.32
Economic crisis	Dichotomous variable (1 = crisis, 0 = no crisis) calculated using Maddison’s GDP estimates as years where a country’s GDP per capita growth rates fall below 0 or below the 25th percentile in the world.	25,345	0.25	0.43
Social unrest	Frequency of non-nationalist social unrest measured from international newspapers, calculated by Global Social Protest Research Group (Arrighi Center, 2019).	25,345	0.18	1.17
Financialization	Dichotomous variable (1 = financial expansion, 0 = material expansion) calculated by the author based on Arrighi (1994) and explained in Table 1.	25,345	0.42	0.49
State-seeking nationalist unrest in neighborhood	Frequency of SSNM in neighbor territories (SSNM dataset 1)	25,345	2.45	8.58

GDP: gross domestic product; PPPs: purchasing power parities; SSNM: state-seeking nationalist movements.

Control variables are year of the analysis and regional controls (Middle East, Eastern Europe, Africa, Asia, Oceania, Latin America), where territory gets a value of 1 if it belongs to the region, and a value of 0 if it does not. Western Europe/North America is the reference group.

OPEN IN VIEWER

Table 3.

Correlation matrix.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)	(19)
(1) SSNM	1.00
(2) Ethnic diversity	0.03	1.00
(3) Religious diversity	−0.02	0.37	1.00
(4) Iron-steel production	0.03	−0.05	0.05	1.00
(5) GDP per capita	0.04	−0.20	0.02	0.40	1.00
(6) Globalization	0.04	0.00	0.00	0.15	0.38	1.00
(7) State power	0.01	0.09	0.10	0.10	−0.07	−0.17	1.00
(8) Recent nation-state formation in the empire	0.11	0.06	0.03	0.00	−0.03	−0.05	0.03	1.00
(9) Recent nation-state formation in neighborhood	0.06	0.01	0.00	0.03	0.02	0.04	−0.08	0.12	1.00
(10) Nationalist organizations	0.09	−0.21	−0.12	0.13	0.39	0.28	−0.23	0.08	0.17	1.00
(11) Duration of nationalist propaganda	0.06	−0.25	−0.13	0.28	0.65	0.37	−0.21	−0.05	0.05	0.66	1.00
(12) Imperial dependency	0.01	0.14	0.10	−0.12	−0.23	−0.23	0.53	0.15	−0.08	−0.34	−0.42	1.00
(13) Wars in the empire	0.03	0.08	0.03	−0.04	−0.10	−0.16	0.30	0.19	−0.01	−0.09	−0.17	0.36	1.00
(14) Wars in the territory	0.16	0.01	−0.02	−0.02	−0.06	0.04	−0.01	0.02	0.08	0.10	0.04	−0.07	0.07	1.00
(15) Economic crisis	0.04	−0.09	−0.00	−0.03	−0.04	0.00	0.05	−0.04	0.05	0.15	0.06	−0.01	−0.01	0.03	1.00
(16) Social unrest	0.21	−0.03	−0.01	0.22	0.14	0.02	0.05	0.02	0.04	0.12	0.20	−0.09	−0.01	0.09	0.02	1.00
(17) Financialization	0.03	−0.00	−0.00	0.04	0.12	0.54	−0.00	−0.05	−0.01	0.08	0.12	0.01	0.02	0.00	−0.00	0.03	1.00
(18) SSNM in neighborhood	0.15	−0.04	−0.05	0.09	0.09	0.10	−0.04	0.06	0.12	0.12	0.09	−0.05	0.03	0.09	0.03	0.08	0.05	1.00
(19) Year	0.07	0.00	0.00	0.17	0.47	0.58	−0.24	0.07	0.13	0.63	0.58	−0.19	−0.07	0.07	−0.00	0.10	0.28	0.17	1.00

GDP: gross domestic product; SSNM: state-seeking nationalist movements.

Regional control variables (Middle East, Eastern Europe, Africa, Asia, Oceania, Latin America), where territory gets a value of 1 if it belongs to the region, and a value of 0 if it does not, are not presented in the matrix.

Resurgence of state-seeking nationalism in the 21st century

The theoretical frame and empirical evidence presented in this article brings some fresh insights to the unexpected resurgence of nationalism in recent years and decades. Our findings suggest that the unexpected escalation of SSNM since the 1970s is linked neither to “enduring ethnic identities” nor to “reactions against globalization” but is instead linked to the escalation of economic, social, and geopolitical crises that has gone hand in hand with the fourth major wave of financialization of the capitalist world-economy and the crisis of the US world hegemony.

While the unexpected resurgence of secessionism around the world in the last quarter of the 20th century was linked to the signal crisis of the US systemic cycle, the multiplicity of secessionist movements that we see in front of our eyes today appear to be linked to the terminal crisis (i.e. chaos phase) of this US-led financial expansion period. Since the turn of the 21st century, we seem to have entered into the chaos phase of the US world hegemonic decline (Silver and Arrighi, 2011), whereby existing nationalist organizations use economic crisis conditions surrounding their states to propagate the benefits of independence (as Catalan, Scottish, Padanian or Flemish nationalist are doing), struggle to turn geopolitical crisis and wars into an opportunity for independence (as Kurds in Iraq and nationalists in Donetsk and Luhansk regions of Ukraine are doing), and ride the tide of social movements (i.e. social crisis) to achieve further autonomy and independence (as nationalists in the USSR did in the 1988–1992 period, Kurds in Rojava did during the recent Syrian Arab Spring, and Donetsk and Luhansk did in the aftermath of the Euro-Maidan protests in 2014). In regions where developmentalist states collapsed in this current era of financialization, the declining legitimacy of existing regimes in the eyes of citizens also provide nationalist organizations with various structural opportunities for mobilization. If some of these movements succeed in establishing new states, it has the potential to spread especially in states and in regions with high concentrations of SSNM. A longue durée perspective on nationalism suggests that we are living in an era that is very fertile for the rapid escalation of nationalist unrest, and the reconfiguration of the interstate system that came into being during the long 20th century.

Resurgence of chauvinist, exclusionary, and authoritarian forms of state-led nationalism

While the limited scope of this article only focused on the effects of crisis and hegemonic decline on state-seeking nationalism, the theoretical approach presented in this article can also help us make sense of the sudden resurgence of aggressive, exclusionary, and expansionist forms of state-led nationalism in recent years. The current resurgence and rapid proliferation of authoritarian, right-wing populist and chauvinist movements and leaders around the world, also have its roots in the declining hegemony of the rulers. This is not the first time we have experienced such rapid revival either. On the contrary, during every world hegemonic crisis and transition period, hegemonic forms of state-led nationalism were dismantled and paved the way for more authoritarian, exclusionary, and chauvinist forms of state-led nationalism. The rise of Bonapartism during the terminal crisis of the Dutch world hegemony and the rise of Fascism during the terminal crisis of the British world hegemony were different historical examples of the same process. Today, as we enter into the terminal crisis of the US world hegemony right-wing nationalists are gaining power in many parts of the world.

This relationship is not accidental because intensification of economic, political, and social crisis and the corresponding decline in the consent-making capacity of rulers also provide opportunities for the rise of a section of state-led nationalists who rely more on coercion than consent. Because contentious conjunctures make it very difficult for rulers to maintain their hegemony on all people, it becomes attractive for factions of the ruling elite who aim to mobilize a particular section of the people (e.g. racial, ethnic, religious, or class-based groups) against others. Likewise, the co-escalation of anti-systemic social movements and state-seeking nationalist unrest in these periods pressure these elites to adopt more aggressive and authoritarian strategies to maintain their power. When we consider all these dynamics altogether, the transformative character of these intense crisis periods—such as the one we are living in right now—would become more explicit. These are periods of rapid social change whose direction is contingent on the actual struggles on the ground.

Implications for global governance patterns

The theoretical frame and the findings of this article also bring new insights into the evolution of global governance patterns. They show that unlike suggested by political modernization theories, the rise of nationalism did not cause a gradual transition from a world of empires and city-states to a world of nation-states. Rather, the rise of nationalism transformed our world through successive global waves of nationalist uprisings that occurred during contentious conjunctures of historical capitalism from the 16th century to the present. These global waves not only produced a series of quantum-leap like changes in the political-territorial landscape of the interstate system but also transformed existing conceptions of nationhood and cultural and ideological content of nationalism. Examination of these transformations is beyond the scope of this article. ¹⁰ Let me suffice to mention that the clustering of nationalist unrest during contentions conjunctures of historical capitalism produced a series of creative destructions in the modes of governance used by state elites because each global wave pressured state elites to invent new forms of hegemony-building strategies that would keep subjects loyal to their states. Consequently, nations and nationalism transformed over space and time in synchrony with systemic crises. Unlike it has generally been claimed, however, none of these processes ended up producing homogeneous nation-states on a global scale. Most SSNM ended up producing mini-empires which were almost as heterogeneous as the ones they seceded from.

Global waves of nationalism also played a key role in the formation of world hegemonic orders and global governance regimes. Waves of state-seeking nationalism led to the creation of new states in the interstate system and cycles of state-led nationalism provided existing states with greater command over production, protection, and administration, that is, greater political and economic power. Accordingly, the world hegemonic powers that aimed to lead the states in the interstate system had to expand their sphere of influence into a wider geography over time and to command over greater resources from their predecessors. This is why the increase in the size, scope, and complexity of the interstate system from one global wave to another (and from one world hegemony to another) went hand in hand with the growth in size, scope, and complexity of world hegemons over time (also see Arrighi, 1994: 217). The United Provinces was slightly more than a collection of city-states but still less than a nation-state. The United Kingdom—built upon the subjugation of Ireland, Scotland, Wales, and so on—was more than a typical nation-state that controlled a world-scale commercial and territorial empire. At the height of its hegemonic power, the United States was a huge military-industrial complex with unprecedented continental size, informal surplus appropriation networks of an empire, and direct control over a global network of military bases. None of the world hegemonic powers were nation-states. They all combined elements of national and imperial forms of rule.

Moreover, different ways in which these hegemonic powers aimed to contain national problems around the world changed the qualitative character of each world hegemonic regime. This is why it would be wrong to assume that the macro-structural dynamics of world hegemonies produce global waves of nationalism in a top-down fashion. Indeed, the direction of the relationship has always been both ways. From the 16th century to the present, nationalist movements from below (together with social movements) actively shaped the functioning of the capitalist world-economy, the interstate system, and the world hegemonic regimes. Nationalist movements were not only affected by the contentious conjunctures but they became an active agent of these vicious cycles. The recent resurgence of multiple manifestations of nationalism in our era—from SSNM demanding self-determination and independence to authoritarian, expansionary and exclusionary forms of state-led nationalism—must be interpreted from these lenses. Ongoing struggles on the ground will not only redefine what kind of relationships people will have vis-à-vis their states and vis-à-vis other groups residing in these polities, but also what kind of world we will be living in once this chaos is over.