Theory of Metropolitan Response

Administrative lines and categories shape secessionist outcomes in three mutually reinforcing ways. First, laying claim to a specific territory with an identified boundary helps the relevant parties—typically the sovereign and the breakaway region—to overcome the basic problem of where to place borders. Administrative boundaries provide clear bargaining solutions and reduce the possibility of conflict.Footnote ³² They fit Schelling's definition of focal points in that they coordinate the participants' expectations of where borders should be drawn.Footnote ³³ This is important given that nations are typically fuzzy, overlapping groups rather than discrete, easily delineated units.

Second, administrative organization creates categories. Just as the Soviets had the so-called nationalities ladder of ordered types, the United States has its incorporated territories (core fifty states), Native American reservations, and unincorporated territories (for example, Puerto Rico). Consider the cases of Hawaii and Puerto Rico, both island units that were acquired by the United States in 1898. Although both became administrative regions and eventually generated a secessionist movement, their difference in type has profound consequences for their chances of secession. Puerto Rico was admitted as an unincorporated territory, a category that constitutes the outer belt of the various American units. As with other units in that category, Puerto Rico has been given the right to secede. Hawaii, on the other hand, was inducted as an incorporated territory. As a type of unit that is subject to the full weight of the US Constitution, Hawaii's chances for eventual secession were dramatically reduced more than 100 years ago when the relevant actors decided to incorporate it. Whereas its secession would set a precedent for the other forty-nine states, the secession of Puerto Rico would not.

Finally, both of these points are further reinforced by international law that, since at least the dissolution of the Spanish Empire in the Americas, has emphasized internal administrative territories as a guide for recognizing new states. Based on the principle of uti possidetis, international law holds that emerging states should inherit their administrative boundaries at the time of independence.Footnote ³⁴ Shaw writes that “the primary justification of uti possidetis … has been to seek to minimize threats to peace and security, whether they be internal, regional, or international. This is achieved by entrenching territorial stability at the critical moment of transition.”Footnote ³⁵ Although the application of the principle has received criticism, legal theorists typically defend it as a least-bad solution. As Ratner argues, uti possidetis is an example of an idiot rule—a simple, clear norm that offers an acceptable outcome in most situations, “especially in the absence of any other solution.”Footnote ³⁶

My theory synthesizes several directions in the existing literature. First, I emphasize administrative units, much like Bunce and Roeder.Footnote ³⁷ But their work focuses on the generative capacity of administrative units that, they argue, are states in the making. In contrast, I shift the focus from the region to the metropole to argue that states use administrative lines and categories when selecting which regions can secede. Whereas the first account helps explain where many secessionist movements come from, my theory explains why secessionist movements, even ones with a well-developed sense of nation, almost never secede unless they have the appropriate administrative status. A good example is Moscow's response to secessionism in the early 1990s, when it chose to release its first-order units (union republics) and keep its second-, third-, and fourth-order units (autonomous republics, autonomous oblasts, and autonomous okrugs), even though these latter categories were organized on an ethnonational basis and many of them had secessionist movements. Although a unit's rank often changed over the decades as Moscow periodically upgraded and downgraded units—sometimes as a punishment and sometimes for efficiency—what mattered most was the rank when the state did finally downsize. Unlike the lower-ranked units, the first-order republics had been guaranteed the right to secession under the Soviet constitution, and this was the legal and historical distinction upon which Moscow and the emerging central governments of the former republics could erect a barrier to further secession.Footnote ³⁸ Moscow could not release any lower-order units because doing so would set a precedent, and thus it denied independence to Chechnya even as it permitted the secession of the union republics.

This theory brings the state into the center of the analysis. The states-in-the-making literature rightly notes the nation-generating effects of administrative design.Footnote ³⁹ But central governments are not passive players in this process, merely permitting the fittest nations to secede. Rather, they often deny independence to well-developed secessionist movements that lacked the right administrative status, and they have often permitted the independence of administrative units that possessed weak national identities. States are strategic when dealing with secessionists and they incorporate administrative lines and categories into their strategies.

To model the strategic behavior of states I utilize Walter's logic of entry deterrence, but I develop it further by highlighting the importance of administrative organization. Walter argues that metropoles will fight secessionists as a way to deter others from attempting to secede in the future, just as a chain store will engage in a costly price war with an entrant to establish a tough reputation.Footnote ⁴⁰ This is a useful analogy and several key dynamics can be identified. First, the number of potential secessionist movements matters. If there is only one, then there is no risk of setting a precedent by letting the region secede and therefore no reason to demonstrate resolve. Second, secessionist commitment is variable, and one of the reasons the metropole needs to demonstrate its resolve is to deter potential secessionists who adjust their estimates of success with signals coming from the metropole. Third, some territories are less valuable than others and the cost of administering a unit can outweigh the benefits.Footnote ⁴¹ However, metropoles cannot always pick and choose which units they want to keep. Releasing a low-value territory that is actually not worth keeping can start a cascade effect that would eventually reach a high-value territory.

But the analogy falls short in one critical way: the assumption that the regions are like types and that the metropole cannot discriminate between them. Indeed, the relevant economics literature recognizes that entry deterrence works only if entrants perceive that each market is sufficiently similar; the more variegated the market types, the less deterrence will have an effect.Footnote ⁴² In the context of the chain store analogy, metropoles are usually playing on multiple boards (see Figure 2).

Figure 2.

Schemata of proto-state types

The varied typology of proto-states effectively compartmentalizes secessionism in such a way that cascades often do not travel across type. There are two general reasons. First, the conceptual distinction between types can become quite salient. For example, the United States could permit the secession of Puerto Rico without fear of inspiring a movement in one of its fifty states. That it could not so easily permit Hawaii to secede without setting a precedent for Alaska or Texas demonstrates the cognitive salience on the issue. This explanation maps onto Lustick's theory regarding the importance of ideological thresholds in state contraction.Footnote ⁴³ Despite its distance from the mainland, Hawaii's attachment to the core is such that no serious contender for political power would currently suggest otherwise. Meanwhile, Puerto Rican secession is openly discussed and considered nonthreatening to the core. Lustick is right that these thresholds of attachment to territory govern how states downsize, but I go further to argue that there is a tight correspondence between these thresholds and administrative organization. Therefore, my theory lays out predictions about where these thresholds will be located.

Second, it is upon these distinctions that metropoles typically draw a line between those they will permit to secede and those they will deny, as the UK did when it discriminated between colonial units such as Burma and Uganda, and other colonized nations such as the Karens and the Baganda. Such distinctions are focal in nature, and they clearly sort out the respective groups. Moreover, these distinctions are backed by the principle of uti possidetis, which shapes the legal and strategic environment for aspiring nations. It is when these distinctions are particularly salient that the metropole will have an easier time and potential secessionists will be dissuaded; conversely, the weaker the salience, the more likely the metropole will actually have to demonstrate its resolve.

My theory yields three core hypotheses. The first holds that secessionist movements lacking a proto-state stand a poorer chance of being recognized by their metropole.

In my data set, some 38 percent of the secessionist movements lack a proto-state and these are the least blessed. Most of these movements claim territories that are incongruent with any existing administrative region.Footnote ⁴⁴ Their secession typically presents a problem for the metropoles about where boundaries should be drawn and how the nation should be delineated, and their lack of status makes the precedent-setting problem more acute because their secession would send a signal to all potential takers.

The remaining hypotheses address comparisons between different types of proto-states. Drawing on the chain store analogy, metropoles are often playing on multiple boards. They may choose to permit the secession of one type of unit—see, for example, the Type 3 unit in Figure 2—but fight to ward off any other type of challenger. This is particularly so when the unit in question is the sole member of its type, as Norway was for Sweden, Syria was during its brief federation with Egypt, and as Bougainville currently is for Papua New Guinea.

Importantly, these cases are not exclusive to so-called binary states. What matters is that there is an administrative unit that is considered unique in type. There are plenty of administrative regions in Papua New Guinea, but Bougainville's special status after 2001 made it unique.

The third hypothesis focuses on levels of administrative organization. States organize their space in a variety of ways. For example, India used to have three types of units: its core states, its union territories, and its protectorates (until 1975). It was much easier for India to permit the independence of its protectorates, as it did with Bhutan, because these were regarded as outer-belt territories, different in status from neighboring states such as Assam. In contrast, Assamese independence was denied because it was a core state, because its secession would set a dangerous precedent, and because opening the door to this sort of fragmentation risked the dissolution of the country. My data set identifies these political architectures for all countries between 1816 and 2011 and sorts the constituent proto-states for each country by type.

The unifying logic behind these hypotheses is that the administrative architecture of states actually structures how they respond to secessionist demands. These architectures may have been created for other reasons. They may be arbitrary, poorly drawn, or the product of geopolitical calculations, but they nevertheless present a blueprint for how states can downsize. They provide states with clear lines and status categories, and the strategic use of these distinctions is backed by the international legal emphasis on uti possidetis. This administrative focus is a common mechanism in political life. The mechanism was evident during decolonization, when the colonial powers relied on an administrative snapshot to recognize about fifty administrative units in Africa and simultaneously deny or dissuade perhaps 500 nations.Footnote ⁴⁵ It was used by a beleaguered Yeltsin to erect a bulwark to state dissolution.Footnote ⁴⁶ Administrative distinctions continue to feature in both the legal arguments of aspiring nations and in the responses of states and international organizations who decide their fate.

Critics may contend that the theory suffers from endogeneity because administrative distinctions are given to only robust nations, and such distinctions only foment further nationalism. Thus, proto-states are permitted to secede not for strategic reasons on the part of the metropole, but because they are well organized and deserving. I concede that these dynamics do operate, and I model them as an alternative hypothesis in the analysis,Footnote ⁴⁷ but I contend that the picture is incomplete unless one includes the strategic realities of the metropole. Catalonia is a hard case for my theory because it was once a sovereign state, and because its current autonomous status is both the consequence and cause of nationalist ambition. Yet even here one can see the importance of political organization. Madrid cannot recognize Catalan independence without setting a precedent for other autonomias such as the Basque Country. The Spanish administrative architecture shapes Catalonia's possibilities.

But the Catalan case is hardly the standard where administrative design is concerned. Many higher-level administrative units were the product of top-down decisions that had little regard for the contours of local nations. This is a common critique of decolonization and the application of uti possidetis.Footnote ⁴⁸ As a result, many territories with weak national identities find independence foisted on them, just as many historic nations bereft of status find that independence is denied them. In all, this administrative focus is a least-bad solution that often yields odd, unfair, and near-miss outcomes. For example, Eastern Karelia was a Union Republic until 1956, when it was downgraded to the level of autonomous republic. Had it kept its status for the remainder of the Soviet era, there is every reason to expect that it would have seceded peacefully along with the other union republics and would now be a sovereign state.