Abstract
Why does the presence of coethnics across the border sometimes lead to categorizing that land as part of the homeland, but sometimes not? We argue this variation is shaped by whether regimes use ethnic logics to elicit domestic legitimacy. Relying on shared ethnicity for legitimacy elevates ethnicity’s political salience, making ethnic groups living across borders socially meaningful and enabling continued claims to their land as part of the homeland. Using survival analysis, we demonstrate that coethnics’ presence on lost lands significantly influences whether those lands maintain their homeland status largely in contexts where ethnic legitimacy is prominent, such as autocracies, states that marginalize populations along ethnic lines, and countries where the government’s legitimacy cannot be based on economic performance. We also illustrate this phenomenon with a case study of Croatia. Our findings have important implications for understanding how ethnicity interacts with domestic politics to shape territorial conflict.
Introduction
Russia’s attempt to conquer Ukraine and its propaganda claiming Ukrainian territory as part of the Russian homeland because it is populated by its coethnics (e.g., Putin, 2021) has resurfaced questions about the relationship between ethnicity and the area nations understand as part of their homeland. Indeed, the notion that coethnics’ presence on territory consecrates that land as homeland territory is widespread and often accurate. Yet, there is significant variation in this regard, and land populated by coethnics is not always considered homeland territory. For example, the presence of more than four million Hindus in Pakistan’s Sindh province does not lead Indians to consider that adjacent territory as part of the Indian homeland. Likewise, Germans no longer consider Austria part of their homeland, and Croats no longer claim Herceg-Bosna despite the continued presence of coethnics on these (once) homeland territories.
Why does the presence of coethnics on territory sometimes lead to categorizing that land as part of the homeland, but sometimes not? We argue that this variation is shaped by whether regimes use ethnic logics to elicit domestic legitimacy. Prior work shows that how regimes generate legitimacy affects a wide range of outcomes, including the definition of the national interest, what counts as a threat, which domestic publics are available for mobilization, and the range of issues over which such mobilization is deemed appropriate (e.g., Goddard & Krebs, 2015; Scott, 1985). We synthesize this insight with findings that strategies of domestic legitimization also impact territorial claims (Goddard, 2010; Saideman & Ayers, 2008; Shelef, 2020a) to argue that logics of legitimization also influence whether ethnicity plays a role in categorizing territory as part of the national homeland.
We theorize that where appeals to shared ethnicity are used to generate domestic legitimacy, ethnicity itself becomes politically salient. This increases the likelihood that having coethnics across the border is understood as politically meaningful, that mobilization along ethnic lines and for ethnic causes is seen as permissible (even in otherwise repressive regimes), and that asserting continued national ownership of lost lands because of their demography is socially and politically acceptable. Where ethnic logics of legitimization are less prominent, the political salience of coethnics across the border may be reduced, allowing societal categorizations of those territories as part of the homeland to wane.
We provide systematic cross-national evidence consistent with this theoretical claim through a survival analysis of the continued homeland status of lost lands populated by coethnics. Our analysis focuses on three settings where states are especially likely to appeal to shared ethnicity for legitimacy: autocracies, states excluding populations from power along ethnic lines, and regimes with less credible access to performance legitimacy. Additionally, we present qualitative evidence from an (out-of-sample) illustrative case study of Croatia. Although the observational nature of our study constrains our capacity to assert direct causality, the consistency of our findings across conceptually distinct and independently assessed contexts, and through diverse analytical methods, suggests that the identified connection between how legitimacy is established and the importance of coethnics in determining the homeland’s boundaries is not spurious.
Our argument makes a number of substantive contributions. First, an important literature details how legitimacy affects domestic politics, including in authoritarian contexts (e.g., Geddes & Zaller, 1989; Wedeen, 1999; Slater, 2010; Schmidt, 2015; Brusis, 2016; Dukalskis & Gerschewski, 2017; von Soest & Grauvogel, 2017), but has not applied this lesson to territorial issues. We extend this scholarship by showing that relying on ethnic legitimation also influences whether having coethnics outside the state shapes the homeland’s scope.
Second, our findings nuance the conventional wisdom about the relationship between regime type and territorial conflict. A significant scholarship expects that democracies, because they allow for more competition and hence more ethnic outbidding, would experience more irredentism (e.g., Horowitz, 1985; Jenne, 2006; Saideman, 1998; Siroky & Hale, 2017). Such outbidding and electoral dynamics certainly occur. However, autocracies’ relatively greater reliance on ethnic strategies of legitimization and the attendant persistence of social understandings of lost land populated by coethnics as part of the homeland help explain why these contexts also maintain irredentist potential.
Third, our argument extends prior analyses about the factors shaping variation in the homeland’s extent over time (Shelef, 2020a; 2020b). Specifically, while the case studies explored by Shelef (2020a) suggest that the impact of ethnicity on a territory’s homeland status could vary with the strategies used to legitimate territorial claims, the statistical analysis in that work does not go beyond the demonstration that “ethnicity matters” and the use of ethnicity as a control variable. Addressing that lacuna, this article investigates the conditions under which the presence of coethnics across the border shapes the homeland status of lost lands.
More broadly, our analysis redirects attention to the underlying logics of domestic legitimacy as a critical factor in determining ethnicity’s political salience and, as a consequence, whether ethnicity shapes the homeland’s contours. The endurance of homeland claims, in other words, is not an automatic function of ethnic ties but is intricately linked to how states seek domestic legitimacy.
Theory
All nations designate a certain territory as their homeland because the desire for sovereignty that animates nationalism requires a geographical location in which this sovereignty can be exercised. When nations control the land in question, its designation as homeland territory is applied through both concrete and discursive acts like settling co-nationals on it, erecting border controls, and deploying maps, place names, and history lessons that categorize the land as “ours”. Where part of the national homeland lies under someone else’s sovereignty, homeland territoriality cannot be applied concretely and the discursive labeling of territory as part of the homeland is the primary means of maintaining its homeland status (Hensel, 2001; Newman, 2001; Shelef, 2020a). While the application of homeland territoriality through these concrete and rhetorical acts imbues land with tremendous emotive power, land once included in the homeland can nonetheless lose its homeland status (Shelef, 2020b).
Ethnicity often plays an important role in determining which territories are designated as part of the national homeland (Gellner, 1992; Miller, 2007; Toft, 2003). Indeed, the conventional understanding is that irredenta, those “unredeemed” parts of the homeland, “typically extend only as far as there are ‘brethren’” (Fearon, 1998, p. 110).
Yet, there are good theoretical reasons to expect that “ethnic brethren’s” impact on a territory’s homeland status can vary. First, the domestic relevance of ethnic groups varies (Cederman et al., 2010; Fearon, 2008; Mylonas, 2012; Posner, 2004; Vogt et al., 2015), as does ethnicity’s salience as a domestic political cleavage (Chandra & Wilkinson, 2008; Akturk, 2012; Mor, 2022). This raises the possibility that, where ethnicity is less relevant domestically, it could play a smaller role in shaping the scope of the national territory.
Second, the way coethnics abroad are viewed by the “home” state also varies over time (e.g., Han, 2022) and across states (Waterbury, 2021). This ranges from varieties of legal inclusion (e.g., Bauböck, 2005, 2007; Joppke, 2003), to viewing coethnics abroad as political or financial resources (e.g., Varadarajan, 2010), to their denigration and exclusion as “traitors” or security risks (see, respectively, Smith, 2003; Brand, 2006). This diverse spectrum is institutionalized in a wide array of “diaspora governance” institutions (Gamlen, 2008, 2014) and “diaspora management policies” (Adamson, 2020) that, in some cases, maintain the link between the diaspora and the state, but that, in others, effectively define coethnics abroad as less than full group members. This variation reinforces the theoretical expectation that whether coethnics outside the state are seen as integral parts of the nation is neither automatic nor constant.
Since both ethnicity’s role in domestic politics and the relationship between states and their diasporas vary, ethnicity’s role in determining the homeland status of particular parcels of land likely also varies. While this variation could be shaped by international pressure (Mylonas, 2012), domestic politics often shapes perceptions of coethnics abroad, and even whether a group outside the state counts as part of the nation’s diaspora at all (Adamson, 2012; King, 1998; Levitt & de la Dehesa, 2003; Mügge, 2013; Waterbury, 2006, 2010). We extend this perspective to argue that how regimes generate domestic legitimacy is one source of variation in whether the presence of coethnics on land outside the state transforms “terra” into “terra irredenta”.
This argument begins from the premise that all states seek legitimacy, if only because relying solely on material incentives to elicit compliance results in an “unstable situation” (Weber, 1978, p. 213) and is quite costly to maintain (Hurd, 1999; Schmidt, 2015). Among other benefits, legitimacy reduces the costs of eliciting compliance, mobilizes popular support, and enhances elite unity (Pakulski, 1993; Barker, 2004; Cummings, 2006; Grauvogel & von Soest, 2014; Dukalskis, 2017; Kailitz & Stockemer, 2017). Regimes can appeal to several logics to generate legitimacy, including shared identity, procedural legitimacy, performance legitimacy, a particular belief system, and/or a leader’s charisma. The choice of which to include in a regime’s portfolio of legitimating logics fundamentally shapes the range of issues over which political contestation is considered appropriate, the topics deemed socially and politically meaningful, and the publics available for mobilization (Beetham, 1991; Bukovansky, 2002; Dukalskis & Gerschewski, 2017; Goddard & Krebs, 2015; Krebs, 2015; Scott, 1985, p. 338).
Which logics are used to generate legitimacy also shapes ethnicity’s political relevance. Appealing to a shared identity increases its salience and renders the state-nation incongruity (Miller, 2007) caused by having group members outside the state politically meaningful. Where shared ethnicity is used to generate domestic legitimacy, the presence of coethnics outside the state is more likely to be understood domestically as a legitimate issue around which to mobilize and to limit a regime’s ability to repress mobilization on behalf of coethnics abroad (Alagappa, 1995; Brubaker, 1996; Jenne, 2009). Both of these factors, in turn, increase the likelihood that political actors in those societies would apply homeland territoriality to areas populated by coethnics. Where other logics of legitimization dominate, ethnicity may be comparatively less salient, leading a territory’s ethnic composition to become relatively less important in determining its homeland status.
There are strong theoretical reasons to expect that ethnic appeals would constitute an especially significant component of a regime’s portfolio of legitimating strategies in a number of contexts. These contexts, including authoritarian regimes, states that use ethnicity to define the polity’s membership boundaries, and regimes with less credible access to performance legitimacy, can thus serve as reasonable proxies for the use of ethnic logics of legitimization in the absence of direct global measures of the use of ethnic logics to generate legitimacy across time and space.
Regime type may be especially consequential in this regard both because regime type structures important aspects of how states engage with diaspora populations (Adamson, 2020; Mirilovic, 2016; Shain & Barth, 2003; Smith, 2003) and because there are good reasons to expect different regime types to emphasize different logics of domestic legitimization. Critically, democracies are, on average, likely to emphasize procedural legitimacy based on standardized, transparent, and rule-based decision-making, because it is the cheapest and most durable form of legitimacy. Although some democracies do appeal to ethnicity, the availability of procedural legitimacy and the democratic norm of equality regardless of subgroup membership mean that such ethnic appeals are, on average, likely to be less prominent in democracies.
Autocracies, on the other hand, are “structurally disadvantaged” in their ability to appeal to procedural legitimacy because they “lack the procedures which link political decisions with citizens’ preferences” (Croissant & Wurster, 2013, p. 7; see also Kailitz, 2013; von Soest & Grauvogel, 2017). 1 Indeed, prior work finds that authoritarian regimes are significantly less likely to refer to legal norms and regulations to generate legitimacy than democracies and more likely to deploy alternative strategies (Mayer, 2001; von Haldenwang, 2017; von Soest & Grauvogel, 2017; Tannenberg et al., 2021).
Among these alternatives, autocratic regimes are likely to emphasize appeals to shared identity for a few reasons. First, the other available logics of legitimation can be both costly and risky. Performance legitimacy, for example, is expensive to implement, increases regimes’ vulnerability to economic shocks, and, when successful, creates a middle class that can make demands that are increasingly hard to satisfy (Schmidt, 2015). Charismatic legitimacy, because it is bounded by the lifespan of a particular leader, amplifies the risk to the regime in times of leadership transition. As a result, even states that stake much of their legitimacy on economic performance and the personal charisma of their leaders, like contemporary China and Putin’s Russia, often bolster it by appealing to a shared ethnic identity (e.g., Gerschewski, 2013; Mazepus et al., 2016). Indeed, China’s active promotion of the notion that the Chinese people are “one family” that belong to a single ethnic group regardless of any other divisions (Brady, 2012; Shue, 2010; Stroup, 2020, p. 661) shows that even authoritarian states with credible claims to performance legitimacy will deploy identity-based strategies of legitimization.
Second, appeals to a shared identity facilitate the elision of domestic status and power differentials (Hobsbawm, 1992). Such appeals, like those based on ethnicity, can undermine the emergence of domestic challengers mobilizing around that shared identity (Rothschild, 1977; Schlumberger, 2010; Grauvogel & von Soest, 2014; Dukalskis, 2017). As illustrated by Middle Eastern monarchies’ appeal to religious legitimacy to undermine Islamic opposition movements (Schlumberger, 2010), autocracies may be especially concerned with undercutting identity-based avenues of counter-regime mobilization.
Finally, scholars have argued that diversity creates pressure for states committed to “liberal-democratic precepts” to base membership on residence rather than filiation. By implication, where the commitments to the “democratic principles and human rights considerations” that promote this “de-ethnicization” of citizenship (Joppke, 2003, pp. 430, 455) are weaker, as they are in autocracies, ethnic notions of membership are more likely to define the bounds of the political community and to be used to legitimate existing arrangements.
Importantly, autocratic regimes are not the only contexts where ethnicity is likely to feature prominently in a state’s portfolio of legitimating strategies. Ethnic appeals for legitimacy are also likely where ethnicity is used to define the polity’s membership boundaries. In both ethnocracies (Yiftachel, 2006) and in what Aktürk (2012) labels as “monoethnic regimes of ethnicity”, for example, minorities are systematically excluded to maintain a particular ethnic group in power. Since manifest inequality requires some justification, these states are likely to deploy ethnic logics of legitimization regardless of regime type. For example, Israel’s self-conception as “both Jewish and democratic” appeals for domestic legitimacy on both ethnic and procedural grounds (Smooha, 2002). More broadly, the exclusion of a domestic ethnic minority can create an “other” against which to mobilize, thus justifying and sustaining the link between rulers and ruled (Marx, 2002; Wimmer, 2002).
States with less access to other types of legitimacy are also more likely, ceteris paribus, to rely on ethnic appeals. Since their portfolio of legitimating strategies is comparatively constrained, these states are relatively more likely to rely on any of the remaining logics of legitimation, including shared ethnic identity. Indeed, it is quite common to witness regimes turning to ethnicity to generate support in difficult economic times. This was the case, for example, in 1980s Hungary where falling living standards led the regime to focus on the issue of Hungarians abroad to bolster its legitimacy (Waterbury, 2010).
In summary, there are good reasons to expect that a regime’s reliance on ethnic logics of legitimization for domestic legitimacy creates political space for actors in a society (individual leaders, the regime, and the regime’s opponents) to continue categorizing lost land as part of the homeland when that land has coethnics on it. As a result, our main hypothesis is that:
H: Societies will be more likely to apply homeland territoriality to lost homeland territory populated by coethnics when their regimes derive domestic legitimacy from appeals to shared ethnic identity.
Empirical Strategy
Building on Shelef (2020a), we evaluate this hypothesis in all cases where pre-existing ethnic homelands were truncated by new international borders drawn between 1945 and 1996. 2 We identify ethnic homelands according to whether coethnics were present on both sides of the border according to ethnographic maps drawn prior to the border’s demarcation. 3
The focus on previously defined ethnic homelands has two advantages. First, it sets a particularly high bar for this study because actors in such contexts continue to have credible appeals to an ethnic past in their repertoire of potential homeland claims. Second, as we elaborate in the alternative explanations section, it also helps address concerns that having coethnics across the border impacted the use of ethnic logics of legitimization in the first place. Since all cases had coethnics across the border at one point, this constant cannot account for subsequent variation.
We focus on cases where states lost homeland territory because a land’s homeland status is unlikely to change when it remains under a group’s control regardless of its ethnic composition. Moreover, as noted above, where part of the national homeland lies under someone else’s sovereignty, its discursive labeling as part of the homeland is the primary means of maintaining its homeland status. This enables the assessment of whether particular parcels of land maintain their homeland status over time. Mirroring the markers of norm internalization in other contexts (Finnemore & Sikkink, 1998; Lustick, 1993), we interpret the persistent absence of domestic categorization of lost lands as part of the homeland as marking the loss of their homeland status.
Data
Our dependent variable is the time between the creation of a border through ethnic homelands and the last social categorization of the lost land as part of the homeland in a particular country. While states often use a variety of approaches (e.g., dual citizenship, cultural promotion, etc.) to reach out to kin abroad, our focus is on the literal claim of the territory itself (independently of claims to the population). Building on Shelef (2020a), homeland territoriality was coded as applied to lost lands in a year if states, state executives, or other sociopolitical organizations in the state that lost access to the land, claimed it as part of their homeland at least once in that year. 4 We include the categorization of land as part of the homeland by relevant political organizations even if they are not part of the government because, although legitimating strategies are deployed by the regime, they open avenues along which other domestic actors can pursue their agendas. 5
Our key independent variable is the interaction between coethnics’ presence on lost homeland territory and contexts in which ethnic logics of legitimation are theoretically likely to be especially prevalent. We identify the presence of coethnics on lost lands using a binary variable largely based on whether the EPR-TEK dataset (Vogt et al., 2015) identified an ethnic group in one state as having coethnics across the border. 6
As noted, we use contexts where there are strong theoretical reasons to expect ethnic legitimation logics to play a prominent role as proxies for the use of these logics. We rely on these proxies because, although scholarship has made significant advances in measuring the sources of domestic legitimacy cross nationally (e.g., von Soest & Grauvogel, 2016; 2017; Tannenberg et al., 2021), these path-breaking efforts have not tracked the role of appeals to shared ethnicity. For example, the important work by Grauvogel and von Soest (2014, 2016, 2017), recognizes that identity can be used in this way. However, it operationalizes identity-based claims to legitimacy in terms of appeals to a foundational myth, ideology, or personalism, rather than ethnicity. The Varieties of Democracy (V-DEM) project’s impressive tracking of legitimating strategies globally similarly overlooks the use of shared identity to generate legitimacy. 7 These efforts, moreover, tend to rely on retrospective expert surveys. While reasonable, concerns remain that hindsight bias might inadvertently color assessments of the use of a particular legitimation strategy in a particular year (Levick & Olavarria-Gambi, 2020).
In the absence of direct, global, time-varying measures of ethnic legitimation, we use as proxies the three conceptually distinct contexts discussed above where there are strong theoretical expectations that ethnic legitimation strategies play prominent roles. Regime type is coded using a binary version of V-DEM’s ordinal electoral democracy index (e_v2x_polyarchy_3C) (Coppedge et al., 2021). 8 States excluding ethnic groups from power are coded by the proportion of the population excluded from political power along ethnic lines (logged) (using EPR’s exclpop variable (Vogt et al., 2015)). 9 Finally, lack of credible access to performance legitimacy is coded using (low) GDP growth (as coded by Coppedge et al., 2021). 10
All models also account for other factors that might influence the persistence of societal claims to lost homeland territory, including past and current conflict across the border, the relative material capabilities of each state, regime age, the economic value of the land across the border (including the presence of ports, rivers, diamonds, and petroleum reserves), the distance of the lost land from a state’s capital, and the extent of a state’s integration into the global economy. 11 Accounting for the extent of integration into the global economy helps address both secular changes in territories’ value and the possibility that greater integration into the global economy shapes how states view their coethnics abroad (e.g., Mirilovic, 2016; Varadarajan, 2010). Table A.1 in the on-line Supplemental Material provides descriptive statistics.
Modeling Approach
We evaluate our hypothesis using Cox conditional shared frailty models because our data is both right- and left-censored, and because cases are included in the data for varying time periods. Cox models avoid the error of interpreting cases observed for short periods as equivalent to those that are in the data for much longer periods (Box-Steffensmeier et al., 2007).
Conditional shared frailty models, in particular, also address two potentially problematic sources of correlation between event times (Box- Steffensmeier et al., 2007; Box-Steffensmeier & Zorn, 2001; Garson, 2013). The first is the possibility that claims may be more likely to be made across some borders than across others (e.g., there may be an underlying difference between the Israel-Syria border and the Mexico-Belize one). The second source of potential correlation stems from the logic of sequencing (i.e., for a third claim to exist, there had to be a second, etc.). We also include dummy variables for the various strata (groups of observations that share the same baseline hazard rate) to correct for the different baseline hazard rates for each stratum. 12
We use a conditional risk set up as we are interested in understanding the likelihood that the societal application of homeland territoriality to lost lands in a particular year (“failure”) is the final one. In this set up, an actor must fail a first time before they can fail a second time, and so on. 13 The unit of analysis is the state-border-year (rather than state-year) because homeland territoriality might be applied to land across multiple borders that vary independently. Since each “failure” constitutes its own observation, the 153 instances of truncated ethnic homeland territory between 1945 and 1996 yield 1180 observations. 14
As noted, the data and modeling approach build on (Shelef, 2020a), but extend the analysis in important ways. Most significantly, our main independent variable is the interaction of proxies for the use of ethnic logics of legitimization and the presence of coethnics across the border. While Shelef (2020a) focuses on the interaction of regime type and regime age, that work does not investigate the interaction of the presence of coethnics and any of our proxies for the use of ethnic logics of legitimization.
Results
Figure 1 compares the risk of homeland territoriality being applied to lost land populated by coethnics in contexts where ethnic logics of legitimization are likely to be especially prominent versus contexts where they are less likely to be prominent, with all other variables held at their means.
15
A higher hazard function denotes a higher likelihood that a society would apply homeland territoriality to lost lands populated by coethnics. The solid lines in Figure 1 indicate contexts more likely to be characterized by ethnic logics of legitimization, while the dashed lines denote contexts where ethnic logics of legitimization likely play a less prominent role. Risk of societal applications of homeland territoriality to lost lands with coethnics (Cox conditional hazard). (a) Regime Type (b) Proportion of (log) population belonging to excluded ethnic groups (c) GDP growth rate.
Consistent with our argument, lost land populated by coethnics is significantly more likely to be considered part of the homeland where ethnic logics of legitimization are likely to be prominent. Figure 1(a) shows that autocracies are much more likely than democracies to continue applying homeland territoriality to lost lands populated by coethnics. Figure 1(b) shows that the same is true of states that exclude significant portions of the population from power along ethnic lines. Finally, an examination of contexts where states are less able to credibly appeal to performance legitimacy—and are therefore more likely to rely on shared identity for legitimacy—yields similar results. Figure 1(c) shows that being less able to appeal to performance legitimacy because of low economic growth rates is associated with a greater likelihood of continuing societal applications of homeland territoriality to land populated by coethnics. 16
We argue that this relationship exists because these contexts are ones in which ethnic logics of legitimation are likely to feature prominently in a regime’s portfolio of legitimating strategies. This argument is bolstered by Figure 2, which compares the risk that homeland territoriality would be applied to lost lands in societies whose regimes are likely to rely on ethnic logics of legitimization when those lands are (solid lines) or are not (dashed lines) populated by coethnics, with all other variables held at their means. If the observed relationship in Figure 1 was the product of something unrelated to ethnicity, this comparison should show no meaningful difference. Figure 2, however, shows that this is not the case. Rather, societies where ethnic logics of legitimization are likely to be prominent categorize lost lands as part of the homeland largely when those territories are populated with coethnics but not otherwise.
17
Risk of applying homeland territoriality to lost lands in contexts likely to rely on ethnic legitimization, by whether that land is populated by coethnics (Cox conditional hazard). (a) Authoritarian regimes, (b) 90th percentile of (log) population belonging to excluded ethnic groups, (c) Low GDP growth.
This conclusion is reinforced when further examining the impact of different types of authoritarian regimes. Table A12 shows that the impact of autocracies is driven by party-based autocracies rather than personalist dictatorships, monarchies, or military regimes (based on Geddes et al.’s (2014) categorization). The latter types of autocracies, notably, have other relatively easily accessible sources of legitimacy (charismatic and security-based performance legitimacy, respectively) and so have comparatively less need to appeal to ethnically-based logics of legitimization. Overall, consistent with arguments that nationalism can constrain territorial ambitions (Schultz & Goemans, 2019), our findings suggest that the use of ethnic logics of legitimation domestically may also limit territorial ambitions where territories are not populated by coethnics.
Figure 3, which combines Figures 1(a) and 2(a), focuses on the regime type proxy to illustrate the magnitude of the differential impact of being “treated” with the use of ethnic legitimization strategies and coethnics across the border.
18
The black lines depict the risk that autocratic societies would continue to claim lost land as part of the homeland with (solid) and without (dashed) coethnics on the other side of the border. The gray lines depict the risk that democratic societies would do so with (solid) and without (dashed) coethnics on the other side of the border. Risk of applying homeland territoriality to lost lands (Cox conditional hazard), combined.
Consistent with ethnicity’s role in shaping the homeland, both democratic and autocratic societies are more likely to continue claiming lost land as part of the homeland when it is populated by coethnics. At the same time, autocratic societies are much more sensitive to the presence of coethnics across the border. The finding that the distance between the two black lines is much bigger than the distance between the two gray lines is consistent with our argument that being treated with coethnics on the other side of the border would have a much bigger impact on the homeland categorization of the territory they live on where ethnic logics of legitimation are likely to dominate (authoritarian contexts) than in contexts where other logics of legitimation are likely to matter (democracies).
These results are also consistent with the importance of outbidding dynamics in driving the continued homeland status of lost lands, at least in democracies. As expected by the “Macedonian syndrome” (Weiner, 1971), democracies with coethnics on lost lands are more likely to continue categorizing that land as part of the homeland than when coethnics are absent. 19 Democracies, where outbidding dynamics have more purchase, are also more likely to continue claiming lost lands without coethnics than autocracies. At the same time, the finding that autocracies—where outbidding dynamics are less likely due to reduced political competition—are more likely to continue categorizing lost land populated by coethnics as part of the homeland suggests that factors beyond outbidding dynamics, such as domestic logics of legitimization, may also be at play.
Changing Logics of Legitimacy in Croatia and the Declining Salience of “Greater Croatia”
We now turn to illustrating the impact of changing logics of legitimation on whether coethnics abroad are used to define the homeland’s extent in Croatia. This is a hard case in which to show the declining social salience of ethnicity for categorizing lost land as homeland territory for a number of reasons. First, both the history of irredentist claims embedded in visions of ‘Greater Croatia’ and the ethnically-based violence in the 1990s politicized ethnicity and focused attention on the treatment of Croatian coethnics abroad. Second, the massive out-migration of Serbs from Croatia after the war meant that Croatia became even more ethnically homogeneous. 20 As we discuss in the Alternative Explanations section, other things being equal, such demographic homogeneity likely reduces the domestic cost of relying on ethnic logics of legitimization thereby increasing its political appeal. Nonetheless, as Croatia increasingly relied on procedural rather than ethnic legitimacy after its democratization in 2000, the political salience of coethnics abroad declined (relatively speaking), and conceptualizations of the land on which they resided as part of the Croatian homeland faded from political discourse.
The scope of the Croatian homeland conventionally extended well beyond Croatia’s international borders to encompass all Croatian population centers. Often referring to the borders of the Croatian Banovina of 1939, visions of a “Greater Croatia” encompassed the Croatian populated areas of Bosnia-Herzegovina (Herceg-Bosna), parts of Vojvodina and the Sandzak of Novi Pazar in contemporary Serbia, and parts of Montenegro. Even more expansive versions of “Greater Croatia” stretched from the German border to Macedonia, incorporating Croatian-populated territory from Bosnia-Herzegovina, Serbia, and Slovenia within the scope of the Croatian homeland (Binder, 2009, p. 47; Irvine, 1996; Vasilije, 2014). 21
At independence in 1991, the domestic legitimacy of Croatia’s competitive authoritarian regime was based on a combination of ethnic appeals and the charisma of its leader, Franjo Tudjman. The reliance on ethnicity to generate legitimacy meant that it was commonplace to highlight the injustice posed by the Croatian state-nation incongruity and to categorize the Croatian-populated territories outside the state as part of the homeland. Indeed, Tudjman made no secret of his desire for a “Greater Croatia” that would bring Croatian populated land under Croatian sovereignty (see, e.g., Macdonald, 2003, p. 228; Bellamy, 2003, p. 93) and implemented policies, including supporting Croatian separatists in Bosnia-Herzegovina, aimed at achieving this goal (Robinson & Pobrić, 2006, p. 244; Ramet, 2008, pp. 38–40). Consistent with these policies, the governing Croatian Democratic Union (HDZ), demanded “the territorial integrity of the Croatian people within its historical and natural borders”; a demand widely understood to mean the inclusion of the Croat populated parts of Bosnia-Herzegovina in Croatia (Fisher, 2006, pp. 31, 41; see also Irvine, 1996). The International Criminal Tribunal for the former Yugoslavia thus concluded that Tudjman’s government’s goal was “to incorporate the ‘Croatian provinces’ [in Bosnia-Herzegovina] into a united Croatian state” (Ramet, 2008, p. 39). As Macdonald (2003, p. 220) summarized, expanding Croatia’s borders to encompass these territories and their coethnics “was central to the legitimacy” of Croatia’s regime.
The homeland status of these territories, however, waned after Croatia’s democratization. As expected by the theoretical primacy of procedural legitimacy in democracies, this regime change came hand-in-hand with the marginalization of ethnic logics of legitimization in favor of appeals to procedural legitimacy and “European” principles (Finn, 2021; Fisher, 2006; Maldini, 2012; Malešević, 2002; Massey et al., 2003; Ramet, 2008; Zakošek, 2007; Śedo, 2010). Figure 4 illustrates the change as captured by V-DEM’s coding of the legitimization strategy used in Croatia. As it shows, with democratization in 2000, the relative salience of nationalist legitimization (which maps directly onto ethnic legitimization in the Croatian case) and procedural legitimacy flipped. As a result, observers of Croatian society noted that manifestations of ethnocentrism ebbed “to the level. . . usual in Western democratic societies” (Maldini, 2012, p. 16). Changes in Croatian legitimization strategies.
As ethnicity’s role in legitimating the regime declined, so did the role of ethnicity in defining the homeland’s extent and with it the centrality of demands for including those lost lands in the Croatian state (Robinson & Pobrić, 2006, p. 244). Indeed, observers noted that “the most radical nationalist aspirations which had called for annexing [the Croat populated] parts of Bosnia-Herzegovina” have been “completely marginalized” (Pavlaković, 2014, p. 19; see also, Binder, 2009, p. 47; Stojarová, 2010, p. 45; Śedo, 2010). Even the HDZ shed its advocacy of a “Greater Croatia” and reframed what had been its overt territorial expansionism as a concern with the wellbeing of Croatians abroad (Matić, 2008, p. 176; Stojarová, 2010, 2012; Śedo, 2010). A similar transformation characterized even the more radical Croatian Party of Rights (HSP). This party, which historically demanded ‘Hrvatska do Drine’ (Croatia to the Drina), dropped its demands for annexing those parts of the homeland beyond Croatia’s borders (Pavlaković, 2014, p. 45; Stojarová, 2010, 2012). While some smaller radical right parties maintain their expansive territorial claims, they “remain irrelevant and neither succeed in gaining seats in parliament nor in achieving stronger membership. Currently, we may conclude that Croatian nationalism is fading away from the party system” (Stojarová, 2010, p. 45).
A content analysis of the inaugural addresses by Croatian presidents since 1997 reinforces the declining salience of appeals to ethnicity and the concomitant “demotion” of the homeland status of the land on which Croatians abroad reside. As elsewhere, these addresses are highly symbolic occasions used by newly elected leaders to articulate their ideological vision and to underscore the legitimacy of the new government and its leadership. Inaugural addresses are thus primary sites in which the use of different legitimating strategies can be traced over time.
The baseline is Franjo Tudjman’s (1997) inaugural address. Tudjman repeatedly emphasized the importance of Croatian-ness and the link with Croatians abroad. Indeed, he openly directed his address first to “Croatians in the Homeland and in the world” and only secondarily to “citizens of the Republic of Croatia.” The importance of Croatians (distinct from citizens) was reinforced in his promise “to faithfully serve the Croatian people and all Croatian citizens.” While paying lip service to the Dayton Accords and to the territorial completeness of the Croat state, he simultaneously highlighted the constitutive nature of the Croats in Bosnia-Herzegovina for the Croatian nation and declared that “it is up to us to complete the full integration of the Croatian national being, both territorially and spiritually.”
The emphasis placed on Croatians abroad, their role in the nation, and the imperative for territorial integration declined after the transition to democracy in 2000. Stjepan Mesić (2000, 2005), for example, directed his inaugural addresses primarily to “Croatian citizens” rather than to Croatians wherever they may be. His 2000 address still affirmed the constitutive importance of Croatians in Bosnia-Herzegovina for the Croatian nation, but rather than calling for territorial integration, he invited them to support the Croatian state by moving back to it or supporting it from their “new homeland.” The 2005 address dropped even this nod to Croats abroad and focused instead on themes of democracy and European values.
Like his predecessor, Ivo Josipović (2010) reiterated his interest in continuing towards European integration and underscored the importance of justice and human rights. He also continued the trend of addressing Croatian citizens (rather than all Croatians). The Croatian diaspora was now reframed primarily as a bridge to neighboring states rather than a constitutive part of the Croatian nation or requiring territorial integration.
The HDZ’s return to power in 2015 brought with it a renewed appeal to ethnicity, but these themes were greatly muted compared to Tudjman’s appeals. Like Tudjman, Kolinda Grabar-Kitarović (2015) opened by addressing “Croats in the homeland and in the world, and all Croatian citizens” and by highlighting her responsibility to serve both “the Croatian people and all Croatian citizens.” However, while acknowledging Croats in the near abroad (explicitly in Bosnia-Herzegovina), rather than including them and their territory in Croatia, she thanked them for their (external) support of the state and promised to protect their interests wherever they lived. Reflecting the reduced salience of the territory on which they reside, the support of Croatians in Bosnia-Herzegovina was now explicitly framed in terms of consolidating Bosnian statehood rather than claiming that those areas should be part of Croatia.
The 2020 inauguration speech by Zoran Milanović (2020), from the center-left, not only emphasized democratic values and patriotism, but redefined Croatian patriotism in terms of procedural democratic norms rather than the ethnic ones that animated the Tudjman regime. Milanović explicitly argued that patriotism, “unlike raw nationalism, is a critical attachment and negation of exclusivity” that centers around the value of constitutionally-bound procedural justice. By 2020, then, patriotism is no longer equated with uniting all Croatians in their historic homeland, but with maintaining Croatia as a democracy. Not accidentally, the Croatians abroad are entirely absent from his address.
In sum, some partisan differences notwithstanding, over time there is a clear shift in emphasis from appeals to ethnic identity to appeals to democratic procedural legitimacy and to Croatia’s membership in the European club across the political spectrum. 22 Croatians abroad remain salient, especially for the Croatian right, but even they now frame the issue in terms of protecting the interests of Croatians abroad where they reside rather than including them in the nation-state and categorizing the territory on which they live as part of Croatia’s homeland. Their presence on territory, in other words, no longer sanctifies it as part of the homeland.
The Croats in Bosnia-Herzegovina also felt the change. Whereas in the 1990s their inclusion in Croatia was indisputable, starting in 2000, they increasingly perceived rhetoric in Croatia as excluding them (and the land on which they resided) from the body-politic (Subašić, 2023). Perhaps most concretely, the route of the Pelješac Bridge embodied the exclusion of the land on which they resided from the Croatian homeland. This bridge was constructed to facilitate transit from the Croatian exclave of Dubrovnik-Neretva to the rest of the country (by removing the need to go through border checkpoints into and out of Bosnia). Yet, in so doing, the bridge literally passes over, thereby symbolically and practically excluding, the Croat populated areas in Bosnia-Herzegovina that were once seen as part of the Croatian homeland.
Alternative Explanations
This section addresses potential concerns about endogeneity arising from reverse causality and omitted variable bias as well as potential alternative explanations for the observed relationship between contexts in which ethnic logics of legitimization are likely to be used and the role of coethnics across the border in shaping whether land retains its homeland status.
Controlling for Sources of Endogeneity.
†p < .10, *p < .05, **p < .01, ***p < .001.
Exponentiated coefficients. Cox conditional shared frailty proportional hazards model. Strata dummies and their nonproportional hazard corrections are also included (not shown). Baseline controls include: prior conflict, current conflict, economic value, capability ratio, distance from capital, regime age, and integration into global economy.
An important potential source of endogeneity is the presence of coethnics outside the state in the first place. It could be that this constellation influences both the use of ethnic logics of legitimization and the perception of the lost land as part of the homeland. As noted above, this concern cannot account for our results because all of the cases included in the analysis are ones where ethnic homelands were truncated. In other words, all cases had coethnics on the other side of the border before the border was drawn. As a result, this constant cannot explain the variation in the subsequent application of homeland territoriality to lost lands.
Conceptually, however, the presence of coethnics on a territory may be the product of prior settlement projects conducted precisely because that land was seen as part of the homeland (Haklai & Loizides, 2015). It is also plausible that demographic changes (such as the ethnic cleansing of a territory in extreme cases) could affect both the claims to that territory and the domestic prominence of ethnic logics of legitimization. As a result, column 2 of Table 1 adds a control for the (time-varying) size of coethnics of the ruling group(s) across the border to capture the possibility that demographic changes shape both the homeland status of territory and the domestic reliance on ethnic logics of legitimization. 24 As it shows, controlling for the size of coethnics across the border does not substantively change the main results.
Domestic ethnic demography poses a second potential source of endogeneity. This concern builds on the intuition that a society’s demographic composition shapes both whether ethnic logics of legitimization are deployed and whether homeland territoriality would be applied to lost lands populated by coethnics. The impact of ethnic fractionalization, however, is likely nonlinear. Adapting Siroky and Hale’s (2017) argument, both ethnic logics of legitimization and homeland claims are least likely in both highly homogeneous and highly heterogeneous states. In homogeneous states, the political cost of using ethnic logics is low, but there is also less need to use them. In highly heterogeneous states, ethnic logics of legitimization and homeland claims are also less likely because deploying them could disproportionately benefit one group, disrupting the ethnic balance of power and increasing domestic political costs, thus making such logics less viable.
At moderate levels of ethnic fractionalization, however, ethnic logics and homeland claims may be more likely because groups have an incentive to deploy these strategies to mobilize ethnic support but ethnic divisions are not so extreme as to make such logics politically unfeasible. This calculus may also interact with regime type, as the demographic composition’s influence on ethnic logics may be more pronounced in democracies, where political competition and the mobilization of diverse constituencies are more transparent.
Table 1 addresses this possibility by controlling for a country’s domestic ethnic diversity in two ways. First, column 3 controls a country’s ethnic fractionalization (squared). 25 As Table 1 shows, the impact of contexts likely to feature ethnic logics of legitimization on whether the presence of coethnics on lost land is associated with the continued homeland status of those territories is consistent even after controlling for a country’s diversity (the impact of the interaction of regime type and coethnics on the other side is significant). Second, column 4 controls for the number of ethnic groups since domestic ethnic polarization - reflecting the existence of a few large groups - might be more relevant in this context than fractionalization per se. As it shows, accounting for the number of ethnic groups in a society does not substantively affect the results. 26
State weakness poses a third potential source of endogeneity. Shain and Barth (2003, p. 464), for example, suggest that weak states invite the influence of diasporic communities because they need that support to survive. Weak states may thus be more inclined to use ethnic logics to maintain the links with their diaspora and with the territory they inhabit. The main model already addresses this possibility by controlling for the capability ratio of the states across the border. However, it is possible that absolute state capacity matters more than relative capacity in this context. Column 5 of Table 1 shows, however, that additionally controlling for state capacity does not significantly affect the results. 27
Fourth, regimes in the process of consolidation might be more likely to use ethnic logics of legitimization and to apply homeland territoriality to lost lands as they seek to cement the connection between leaders and masses (Mansfield & Snyder, 2005). More specifically, Waterbury (2010, p. 8) argues that “[r]eaching out to those abroad and establishing cross-border ties based on ethnic kinship is a way for governments and political parties to redefine the bases of their legitimacy, particularly at times of economic or political crisis, or when the sovereignty of the state is weakened.” Koinova (2008) similarly argues that ethnic nationalism is especially like to rear its head during times of regime transition because the otherwise restraining institutions of the kin-state are weakened in these contexts. This weakness, in turn, facilitates the creation of alliances with coethnics across the border by political entrepreneurs seeking to advance their own domestic agendas.
The main analysis already incorporates the potential impact of such times of transition by controlling for regime age. Nonetheless, following Mansfield and Snyder (2005), column 6 of Table 1 controls more directly for periods of regime transition by additionally controlling for whether a state is within five years of entering the international system, the end of an occupation, or a transition to democracy. As the table shows, the results are substantively unchanged.
Fifth, Fearon (2008) argues that countries where politics revolves primarily around patronage distribution should have more ethnic politics because the relatively static character of ethnicity makes identifying losers from the winning coalition and excluding them relatively easier. Column 7 of Table 1 shows, however, that the results are robust to controlling for the degree to which politics are based on clientelistic relationships. 28
Finally, superpower influence could also plausibly impact both the domestic use of ethnic logics of legitimization and whether homeland territoriality is applied to lost lands. This could be the case because international patrons can shape the domestic importance of ethnicity (Mylonas, 2012). The fact that our data largely covers the Cold War period also raises the possibility that being either a Soviet or US client could introduce a source of endogeneity. Client states may have had other readily available sources of legitimacy—ideologically based (e.g., Communism) or procedural legitimacy—and may have been under pressure from their patrons to avoid applying homeland territoriality to lost lands that could have triggered conflict between the superpowers. The final two columns in Table 1 show, however, that controlling for whether the states facing each other across the border are proxies of the same superpower or of opposing ones does not substantively change the results. 29
While these factors are among the most theoretically likely sources of endogeneity, it is still conceivable that some other unobserved confounder could explain away our results. 30 To assess this possibility, Table A15 in the Supplemental Material reports the results of a sensitivity analysis exploring how strongly such a variable would have to be related to both our independent variable (the interaction of the proxy for the prominence of ethnic logics of legitimization and the presence of coethnics) and to our outcome (applications of homeland territory to lost lands) to explain away the results (VanderWeele & Ding, 2017). It shows that, to eliminate the impact of our weakest finding (coethnics across the border interacted with the proportion of the population belonging to excluded ethnic groups), an omitted confounder that was associated with both the independent and dependent variable would need to have a risk ratio of 1.419-fold each, above and beyond all the measured confounders in the main model. To eliminate the impact of all of findings, a confounder would need to have a risk ratio of 11.86-fold each over and above measured confounders. While impossible to discount completely, such a large but still theoretically unanticipated confounder is relatively unlikely.
Alternative Explanations.
†p < .10, *p < .05, **p < .01, ***p < .001.
Exponentiated coefficients. Cox conditional shared frailty proportional hazards model. Strata dummies and their nonproportional hazard corrections are also included (not shown). Baseline controls include: prior conflict, current conflict, economic value, capability ratio, distance from capital, regime age, and integration into global economy.
One possibility is that FBIS, as a US-based open source intelligence gathering source, paid more attention to homeland claims in authoritarian contexts than in democracies. If this were the case, the observed impact of the interaction of regime type and having coethnics across the border in shaping the homeland may reflect the greater likelihood that FBIS would detect such claims in authoritarian contexts in the first place (though, importantly, it would not affect the other proxies). As Shelef (2020b) notes, however, once the outlier of the Soviet Union is excluded, there is no significant difference between the average number of annual FBIS articles from democracies and autocracies (t = −0.0760, df = 6, 223, p = .9394). Column 2 of Table 2 shows that the impact of the interaction of the proxies for using ethnic logics of legitimation and the existence of coethnics across the border and whether lost homeland territory retains its homeland status does not meaningfully change when excluding Soviet cases (it is significant at p = .052). This increases our confidence that the greater role of coethnics in shaping the homeland in societies under authoritarian rule is not a product of a systematic bias against democracies in FBIS.
It is also possible that the size or treatment of coethnics abroad could trigger the application of homeland territoriality to the land on which they reside. Larger groups of coethnics across the border, for example, may simply be harder to ignore. Likewise, the treatment of coethnics living on lost homeland territory could shape mobilization on their behalf and therefore whether the land they populate continues to be categorized as part of the homeland. Where coethnics abroad face repression or discrimination, we might expect more pressure to protect them, and social actors to use homeland territorial claims to spur and justify such interventions (e.g., Brubaker, 1996; Jenne, 2006; Saideman & Ayers, 2008; Weiner, 1971).
Column 2 of Table 1 already showed that controlling for the size of the group across the border did not substantively affect the results. Here, we account for the potential impact of how coethnics are treated in two additional ways. First, we control for whether coethnics of the ruling coalition are discriminated against, powerless, or “self-excluded”, in the neighboring state (based on Girardin et al., 2015). 33 Consistent with scholarship highlighting the triadic relationship between neighboring states and the ethnic groups they share, political discrimination against coethnics of the ruling coalition in neighboring states is associated with an increased risk of continued applications of homeland territoriality to those lost lands (See Tables A6-A8). However, controlling for this factor does not substantively change the impact of the interaction of contexts where ethnic logics of legitimization are likely to be prominent and the presence of coethnics (column 4 of Table 2). 34
Second, we control for whether a group’s coethnics across the border seek self-determination. 35 As numerous scholars have persuasively argued, the political mobilization of ethnic kin abroad can trigger the “Macedonian syndrome” (Weiner, 1971, Brubaker, 1996; Jenne, 2006; Saideman & Ayers, 2008). Despite this expectation, the existence of a self-determination movement by coethnics of ruling groups is not significantly associated with the continued application of homeland territoriality to the land across the border. In any case, as column 5 of Table 2 shows, controlling for the presence of a self-determination movement by coethnics of ruling groups does not substantively affect the main results.
It is also possible that factors associated with the democratic peace could account for the relationship observed above. This could be the case if democracies, rapidly growing states, and states that do not exclude groups from power based on ethnicity are less likely to apply homeland territoriality to their neighbors. Column 6 of Table 2 explores this possibility by controlling for democratic dyads. As it shows, the main results remain substantively unchanged.
Finally, it is possible that regional dynamics are at play. Indeed, in the Croatian example, the desire to join the European Union played an undeniable role in the decline of applications of homeland territoriality to lost lands despite the continued presence of coethnics on them. Other regional associations could conceivably have similar effects. Existing scholarship, importantly, disagrees about the direction of the impact such associations could have. For example, it is possible that shared membership in security organizations, which are frequently regional, could both vary systematically by our proxies (especially regime type) (Mansfield & Pevehouse, 2008) and shape the existence of societal claims to lost territory (making them perhaps less likely). On the other hand, the opposite may also be the case. Waterbury (2020, p. 802), for example, argues that “[m]embership in supranational organizations can create incentives for more intra-ethnic coordination across borders, and home state actors can strategically frame pressure from regional bodies to justify more interventionist or exclusive policies.”
We address these possibilities in three ways. First, while shared membership in a security organization does reduce the likelihood of continued homeland claims by social actors (see Supplemental Material Tables A6-A8), column 7 of Table 2 shows that it nonetheless does not substantively affect the impact on the continued application of homeland territoriality to lost lands by the interaction of the proxies for ethnic legitimization and the presence of coethnics on territory. 36 Second, Tables A9-A11 show that our main results are robust to controlling for European borders, suggesting that our findings are not driven by the arguably distinctive European context. Finally, Tables A9-A11 also show that controlling for all regions individually also does not substantively affect the results. As a whole, these findings suggest that, even though the European norm against irredentism surely played a role in Croatia and other cases, our findings are unlikely to be driven by the particularities of the European (or any other specific regional) context.
Conclusion
When does the presence of coethnics across the border signify the continued homeland status of the land they reside on? We argued that the impact of ethnicity on whether lost land remains part of what a society views as their homeland is shaped by previously underexplored variation in the logics of legitimation states use to generate support. States more likely to rely on ethnically based legitimation strategies are also more likely to continue categorizing lost land populated by coethnics as homeland territory. Where ethnically-based logics of legitimation are less prominent, territory once seen as part of the homeland can lose this status even when coethnics continue to be present. While the observational nature of this study limits our ability to make strong causal claims, the consistent results across conceptually distinct and independently coded contexts provides greater confidence that the relationship between logics of legitimization and the relevance of coethnics abroad for defining the homeland’s shape is not spurious.
Highlighting the role of ethnic legitimization strategies in conditioning when ethnicity shapes the homeland serves as a corrective to the common assumption that the mere presence of coethnics across the border is sufficient to trigger irredentist notions (if not necessarily acting on them). This assumption overlooks important variation in what is understood as politically meaningful in a society; variation shaped by regimes’ strategies of legitimization. While ethnicity is often important, it is not inherently meaningful. Rather, the role of coethnics across the border in shaping notions of the homeland depends on the prior activation and maintenance of ethnicity as a meaningful political category; an activation that can vary across time and space.
Our findings also suggest that arguments that the normalization of the existence of coethnics across borders reduces borders’ importance may not necessarily signal the death knell of irredentism. Such “virtual nationalisms” (Csergo & Goldgeier, 2004) are sometimes framed as alternatives to nationalist policies intended to change borders by offering less costly, often largely symbolic, ways of maintaining links with coethnic abroad (see also King, 1998; Ragazzi, 2009; Waterbury, 2010). Our findings suggest that such symbolism may have a darker side as well. When these links are maintained, perceptions of the territory that coethnics inhabit as part of the homeland are more likely to persist, thereby maintaining the irredentist potential of such constellations. While virtual nationalisms may lower the flames, they simultaneously keep the irredentist ember alive.
While our findings hold across multiple contexts where ethnic logics of legitimization are especially likely, the finding in authoritarian contexts has some additional implications. First, it nuances the existing conventional expectation that democracies, because they allow for more competition and hence more ethnic outbidding, would experience more irredentism. Our finding that the prominence of ethnic legitimization leads the presence of coethnics of lost lands to stay politically and socially relevant helps explain why irredentist notions persist in authoritarian contexts as well.
Our finding also adds another pathway linking regime type and territorial conflict to the roles played by external threats and centralizing authority (Gibler, 2012). The centralization of authority, where it slides into authoritarianism, makes the reliance on ethnic logics of legitimation more likely, and therefore the continued categorization of lost lands populated by coethnics as part of the homeland. Likewise, our results also suggest another mechanism for the operation of democratic peace, whereby autocracies may be more likely to fight because their societies are more likely to continue considering lost territory as part of the homeland, while this reason for conflict wanes in democratic contexts (Shelef, 2020a).
Third, the regime-based aspect of our findings implies an important scope condition on the operation of “diversionary peace” by authoritarian regimes (Fravel, 2005). Autocrats may indeed attempt to neutralize domestic opposition by resolving territorial disputes, but they are likely to face countervailing pressures when coethnics are present on those territories. In those cases, territories are likely to remain seen as part of the homeland. Combined with the reliance on ethnic legitimization, concessions over such territories could open the door to even more domestic contestation by feeding into opposition narratives that the state is not nationalist enough. This may be why even China, which has tended to peacefully resolve its border conflicts by ceding territory, has not done so when it comes to land perceived as part of the Chinese homeland (e.g., Taiwan) or over land populated by Han Chinese (Fravel, 2008).
Finally, our study underscores the need for future scholarship to develop direct measures of legitimization strategies, including of ethnically-based appeals. Such measures are perhaps especially important since the data used in this analysis is disproportionately drawn from the Cold War period. Future work can more concretely evaluate the possibility that the relationship between logics of legitimization and territorial claims varies across historical eras. 37 Such measures would also help nuance studies of legitimacy (including ours) by enabling more sophisticated operationalizations of the portfolio of legitimating strategies states use, changes in these portfolios over time, and the consequences of those changes. They may also facilitate a more direct investigation of the causal impact of various legitimization strategies. While this study established the existence of a relationship between domestic logics of legitimization and the role ethnicity plays in determining the homeland status of territory, the causal process requires additional elaboration. We hope that observing this relationship opens the door for future work in this vein.
Supplemental Material
Supplemental Material - Domestic Legitimacy, Coethnics Abroad, and the Shape of the Homeland
Supplemental Material for Domestic Legitimacy, Coethnics Abroad, and the Shape of the Homeland by Nadav G. Shelef, and Anne S. Jamison in Domestic legitimacy, coethnics abroad, and the shape of the homeland
Footnotes
Acknowledgments
The authors are grateful for the comments of the anonymous reviewers, participants in the International Relations Colloquium at the University of Wisconsin, Madison, the Princeton Niehaus fellows 2020-21, Adeline Lo, Marko Kljajic, and Mincong Pan. We are also grateful for research assistance by Marko Kljajic.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
Notes
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
