The Political Economy of Secession

Chapter One: Introduction

A paradox in the world system of states has become apparent in the post-World War II era: as markets have become more integrated and political borders have become less important, the number of borders and nation-states has increased. With the alleged decline of the nation-state has come a proliferation of nation-states. To be sure, much of this proliferation is attributable to the process of decolonization. However, since the 1960s secessionist movements – movements that seek to detach a territory from an existing state in order to form a new, independent state – have been gathering steam even in the developed West, often accompanied by political decentralization short of secession. In the United Kingdom, the Scottish National Party and Plaid Cymru/Party of Wales achieved almost unprecedented success[1] in legislative elections to new Scottish and Welsh Parliaments. In Canada Quebeckers almost voted in a referendum to secede from the Confederation in 1995, and the Bloc Québecois currently holds about 13% of the seats in the Canadian Parliament. The Canadian government has offered Quebeckers new decentralization several times over the last twenty years, in order to forestall secession. Belgium has moved from a unitary to federal state during the 1980s and 1990s, while the radical Flemish secessionist party, Vlaams Blok (Flemish Bloc), increased its vote share in Flanders from 2.0% in 1978, when it was founded, to 15.9% in 1999. In Italy, the Lega Nord (Northern League) burst onto the scene with 19.5% of the northern vote in 1992 and participated in government in 1994 and again in 2001. The Faroe Islands is the European territory most likely to secede in the near future, as secessionist parties form a majority coalition in territorial government and are preparing a referendum on independence from Denmark. Secessionist politics has also been important recently in democratic countries such as Spain and the Republic of China (Taiwan), and to a lesser extent France (Corsica, Brittany, Occitania, the Basque Country, Roussillon/Northern Catalonia, and Savoy), Germany (Bavaria), St. Kitts and Nevis, Trinidad and Tobago, and the United States (Puerto Rico and Alaska). In less democratic countries there has also been a trend toward political fragmentation (Indonesia, Yugoslavia, the Soviet Union, and Ethiopia are a few countries that have at least partly broken up recently).

Existing comparative research on secessionist politics is meager, and most of what does exist is methodologically flawed. In the past, political scientists have considered secessionist and radical-autonomist movements to be just another expression of “nationalism.” This dissertation makes the case that secessionism is a unique category of political action, in that the factors affecting the rise and success of secessionist movements are different in important respects from the factors affecting the rise of non-secessionist regionalism or minority nationalism. In particular, voters take into account the cost and benefits of a policy of independence when they decide whether to vote for a secessionist party. This unique policy of independence (or quasi-independent “self-government”) unites secessionist parties and separates them from other kinds of territorial movements. The politics of secession in advanced democracies has little to do with ethnic conflict or hatred; it has more to do with the prospective benefits of a government “closer to home,” as well as the risks attendant with a policy change of this magnitude. Many secessionist parties are not ethnically based at all, and this dissertation finds that, by and large, the same political, economic, and demographic factors influence non-ethnic and ethnic secessionists.

There are significant differences in the electoral success of secessionist parties, both across regions and over time. Table 1.1 below presents the electoral results for secessionist parties in 11 regions from 1970-2002, for those election-years for which data are available. Even among these regions that already possess established secessionist movements, significant differences in secessionist support exist. Secessionist parties are weak in Sardinia and Puerto Rico, while they are strong, even dominant, in Quebec, the Faroe Islands, Catalonia, and Euskadi.

Table 1.1 Secessionist Parties in 11 Regions, 1970 to present

Year

Flanders Quebec Faroes Corsica Lombardy Sardinia Catalonia Euskadi Scotland Wales Prto.Rico

1970

1971

1972

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

23.1 11.4 11.5

0 43.7

0 0 0 5.0/5.4

30.2 49.7

0 0 3.1 21.9/30.4 10.7/10.8

47.1

41.4 0 0 5.1/5.7

0 39.1

2.0 45.4 0

0 41.1 0 1.9/3.3 20.2 49.7 17.3 8.1

0 46.3 36.6 64.4 5.3/5.4

1.8 49.3 44.4 0

12.6 26.1 53.0

0 1.7 11.8 7.8

2.5 50.6/49.7 14.4 51.0 64.7 3.8/3.6

2.2 39.2 0.5

7.0/8.9 34.4 53.7/67.6

3.1 49.3 3.8 12.6 14.0 7.3

0 49.1/49.5 49.9 4.5/5.5

40.3 12.9 35.2 59.1

45.9/45.4 18.9 66.0

10.3

21.1 23.0 1.5 53.9 21.5 8.8 3.4/4.2

49.3 21.0 36.8 48.1

44.9 33.5/35.4 22.4 2.1/7.8 55.4

19.6/21.3 17.8 50.3

25.5 6.2 33.7 45.2 3.5/3.8

37.9 22.0 10.0

42.9 55.5/52.8 21.8 53.6

25.0/24.8 23.5 6.3 46.2 28.8 28.4

39.9 15.5 34.2 37.6 4.8/5.2

12.1 3.4 20.1 14.3

48.9

Notes: both radical-autonomist and independentist parties included in vote totals. Regional assembly election results in italics. Results for Corsica from first-round balloting. Results from Lombardy, Sardinia, Wales 1999, and Scotland 1999 for proportional representation constituencies. Results for Puerto Rico from resident commissioner race (countrywide) and governor race (regional). Euskadi includes only the three provinces of Païs Vasco, not Navarre. For sources see Appendix.

What explains the cross-regional differences in secessionist support? A regional culture or language does not seem to be the main cause of stronger secessionism, for by any measure Puerto Rico is culturally and linguistically more distinctive from the rest of the U.S. than Catalonia is from the rest of Spain, yet secessionism (of a moderate variety) is much stronger in Catalonia. Likewise, Wales has a stronger regional language than does Scotland, yet Scotland is more nationalistic and has a stronger secessionist party. Why does Flanders have two secessionist parties, but Wallonia has none? This dissertation finds that non-cultural economic, political, and demographic factors have a significant impact on whether a region develops a (stronger) secessionist party. Relative regional affluence and population are just two of the most important factors explaining why secessionism is weaker in Scotland than in Wales or in Wallonia than in Flanders. Both factors track popular conceptions of post-independence viability.

A cursory examination also reveals two apparent time trends in Table 1.1: for the most part, vote shares are increasing over time, and vote shares in regional assembly elections (in italics) are usually higher.

Trends are clearly upward over time in Flanders, the Faroes, Corsica, Lombardy (with the exception of the last two elections), and Wales, and seem to be more upward than downward in recent years in Catalonia and Scotland. Only in Euskadi and Puerto Rico is the time trend decidedly downward. Of the 19 region-years in the table in which both a regional assembly and countrywide election were held, in only five did the secessionist party do better in the countrywide election.

Could the upward time trend in secessionist electoral success reflect the influence of globalization, which has likewise grown significantly by any measure in the 1980-2000 period? The old view on globalization and secessionism was that nationalist minorities were threatened by globalization, and that secession was seen as a way of thwarting its advance.[2] A new conventional wisdom has emerged that globalization presents an opportunity for nationalist movements, both in that secession is no longer as disruptive economically under globalization and in that globalization allows minority nationalists to forge economic links with other territories and thus become “functionally independent” of the central state.[3]

It is thus by now well known that globalization has increased the potency of regions’ exit threats toward central governments, and this trend has important implications for political institutions. Voters in a region may use the threat of exit to obtain greater benefits from the center for the region. Those benefits may take the form of greater self-government, more representation in national institutions, cultural and linguistic programs, and a more favorable distribution of central expenditures. In particular, national governing parties will make concessions to regions in order to keep them within the nation when such concessions benefit the parties electorally – by decreasing the support for secessionist parties.

The gap between the real-world significance of secessionism and the lack of adequate scholarly treatment of the phenomenon deserves to be remedied. This dissertation takes a fundamentally comparative-politics approach to the topic, while taking into consideration prior economic and area-studies accounts. Economists and area-studies specialists have difficulty explaining both cross-national differences in secessionist support and short-term fluctuations in secessionist political fortunes. Many of the economic explanations of the “equilibrium size and number of countries” we would not expect to hold true except over many generations: they certainly cannot explain well variations in secessionist support between elections, nor do they attempt to account for cross-sectional differences in secessionist support. Area studies, on the other hand, often overlook both economic and institutional explanations, seeking to explain demands for secession as flowing out of primordial cultural sentiments or “social constructions,” and lack the explanatory benefits of large quantitative studies. While identity is perhaps malleable over long periods, that variable cannot alone explain the decisions that precipitate or alleviate demands for secession.

Therefore, this dissertation studies the demand for secession from diverse perspectives, including economic, political-institutional, and cultural-demographic accounts. The result is an empirically robust model of secessionist demands that can be used to predict when and where secessionist movements will arise and achieve electoral success – and what the implications of secessionist electoral success are for the balance of power between central governments and regions (decentralization and federalism).[4] It can also serve as advice to secessionist movements on strategies for cultivating success among voters or to national governments on strategies for preventing secessionist successes, although the assumption behind much of the theory in this dissertation is that secessionist parties and central governments are already doing everything they can to pursue their respective interests effectively.

As already intimated, the dissertation employs secessionist party vote share as the indicator of secessionist demands and studies these parties only in democratic countries, where vote shares can be expected to correlate reasonably with popular support. There are several reasons why this variable has been chosen over the alternatives. First, alternative measures are lacking. One measure is the actual incidence of secession.[5] This measure fails to capture the successes of secessionist movements short of actual secession. Moreover, it conflates regional demands with central government response. Whether secession occurs depends a great deal on how the central government reacts to secessionist demands. A related reason for using secessionist party vote shares over actual secession is that using the latter would mean forgoing the empirical benefits of using quantitative tests in which the dependent variable is continuous. Another measure is the level of secessionist violence. This measure could be a worthy one for a study focused exclusively on non-democratic countries, but it conflates popular demands for secession with secessionist movement orientation. Some secessionist movements use violence, but the vast majority in advanced democracies explicitly eschews violence in favor of a strategy of independence through the ballot box. Second, while secessionist party vote share does not correlate simply and perfectly with demands for secession, the deviations from this correlation are readily predictable. Secessionist party vote share might be affected by electoral rules, by current economic conditions, and by party strategy. Fortunately, we can control for most such variables fairly well in quantitative tests. Third, using secessionist party vote share can allow us to see the give-and-take of central government response to regional demands, which in almost all democratic countries with secessionist movements has resulted in fiscal and constitutional appeasement rather than secession. Quantitative tests that include measures of appeasement among the independent variables should show us what works to reduce secessionist demands.

The disadvantage of using secessionist party vote share is that non-democratic countries are excluded. Such division between democratic and non-democratic countries is inevitable, in my view. Non-democratic countries frequently suffer from major secessionist violence; democratic countries rarely do.[6] Actual secession occurs much more frequently in non-democratic countries than in democratic countries. Moreover, there is reason to expect that the explanations for secessionism differ between democratic and non-democratic countries. For example, in democratic countries, wealthier regions are more likely to be disaffected by centralism than are poorer regions, because in democratic countries wealthier regions tend to subsidize poorer regions because of progressive taxation. On the other hand, there exists a well-known correlation in non-democratic countries between poor regions and secessionist regions.[7]

An aside on terminology is appropriate. I have avoided the terms “separation,” “separatism,” and “separatist” in favor of “secession,” “secessionism,” and “secessionist.” “Separatist” is viewed as a pejorative term, as it implies a desire for some kind of ethnic segregation or a backlash against international cooperation. As will become clear, most secessionist parties are keen to portray themselves as friendly to the global economy and opposed to ethnic discrimination of any kind. A secessionist party is defined as one that seeks full rights of self-determination for the geographical unit in which it operates, perhaps but not necessarily including full political independence. There are many nationalist or autonomist parties that fudge the issue of independence in various ways, either by saying that it should be decided at an indefinite date in the future, that it will become irrelevant in a “Europe of the regions”,[8] or that sovereignty should be “shared” according to some nebulous formula. These parties are treated in this dissertation, though sometimes in slightly different ways from clearly independentist parties. Calling these parties “secessionist” is admittedly to impose more clarity on their positions than they would desire. Regionalist parties that merely serve as territorial lobbies or explicitly oppose independence are not treated. Irredentist parties – parties that seek to attach their territory to an already existing state[9] – are likewise not treated, as the causes of irredentism are not the same as the causes of secessionism.[10]

To refer to the sub-state territory for which independence is or might be sought, I have used the term “region” or sometimes “province.” Secessionist leaders would no doubt argue against such a usage on the grounds that their territories are “nations without states,” not mere regions. But the intent is that the term “region” be taken in its most neutral sense; to use the term “nation” is to prejudice the issue, and moreover to invite confusion with independent states, who are sometimes called “nations” in popular usage. I have thus tried to avoid the term “nation” altogether. Already independent states are called “states,” “countries,” or “the central government,” depending on the context.

The second chapter of the dissertation deals with the political-economic theory of secession. Assuming some cultural difference that makes a region fertile ground for secessionist demands, what economic and political factors should affect those demands and whether they will translate themselves into political outcomes?[11] In short, what should make support for secessionist parties higher across regions? Once a region does have a secessionist movement, how does the central government respond to it, and what factors cause declines and increases in its popular support?

The third, fourth, and fifth chapters operationalize these factors with quantifiable variables and test the theory across the population of democratic countries for the last two decades, using secessionist party vote shares as the dependent variable. The third chapter is occupied with describing the dependent variable (secessionist vote shares, logarithmically transformed), the units of analysis (geographically top-tier provinces), and the methodological strategies. The use of sub-state provinces as data units is an innovation in political science research. It is the obvious choice for research on secessionism, given that secessionist parties’ relevant environments are their respective sub-state territories, but the method may well extend to studies of other regional political phenomena, especially since it provides more observations than using countries as the units of analysis. The fourth chapter uses regression analysis to test the hypotheses on cross-sectional determinants of secessionist vote. The fifth chapter uses regression analysis to test the hypotheses on over-time determinants of secessionist vote.

The fifth chapter also examines the implications of the theory for decentralization across the sample of democratic countries, using a diversity of empirical methods. If the theory is correct, world economic trends should be inducing not only increases in votes for secessionist parties, but also increases in concessions such as autonomy intended to reduce the appeal of secessionism. The fifth chapter also seeks to answer the question of whether offers of autonomy or other policy changes really do reduce the appeal of secessionism.

The sixth and final chapter draws lessons for future political developments from the results of the empirical work and suggests avenues for further research.

The general theory of secessionism developed here consists of three stages. In the first stage, cultural, demographic, economic, and political factors explain whether regions develop secessionist movements and how strong these movements are, on average. Once a region fulfills the conditions for having a secessionist party or parties, changes in economic conditions and changes in public policy affect the parties’ support over time. The third stage of the theory has to do with the effects, rather than the causes, of secessionism. Once a region develops a secessionist party, it is likely to receive more autonomy and at earlier times than regions without such parties.