Chapter One presented secessionist vote shares in 11 regions of Western democracies for the period 1970-2001 (Table 1.1). The first chapter also hinted at a distinction drawn in this dissertation between two types of secessionist parties: “radical-autonomist” and “independentist” parties, or “moderate secessionist” and “radical secessionist,” respectively.[1] The empirical work of this dissertation does not make direct use of this distinction in the regressions, but the classifications may be important for interpretation of the results: for example, in analyzing whether a moderate strategy has benefited a particular party. Therefore, I will undertake a brief description of the characteristics of both types of parties.
“Moderate secessionist” or “radical-autonomist” parties are those that endorse a region’s “right to self-determination” and advocate direct regional participation in international organizations, but remain silent on the issue of full political independence. Plaid Cymru/The Party of Wales is one such political party, advocating “self-government for Wales” and Welsh representation in the United Nations but stopping short of endorsing the breakup of the United Kingdom. The new, much smaller Cymru Annibynnol/Independent Wales Party does support Welsh independence. The Catalan CiU (Convergence and Unity) and the Basque PNV (Basque Nationalist Party) are also considered moderate secessionist parties in this dissertation; they both advocate direct representation for their regions in the Council of the European Union but hold that political independence, while a right, should not yet be considered.
“Radical secessionist” or “independentist” parties are those parties that do endorse full political independence as both a right and the appropriate step to be taken. At the same time, independentist parties may well advocate retaining a shared military or currency with the former country or with another country. The Scottish National Party has always been such a party, as has the Bloc Québecois. The Parti Québecois has been independentist during all election contests, though it went through a brief period in the 1980s de-emphasizing sovereignty.[2] The SNP supports the European Union (but opposes European federalism), while Quebec endorses a concept of “sovereignty-association” with Canada (and possibly the U.S.) that involves a customs union and common defense. In the United States, the Alaskan Independence Party is an example of an independentist party, while the Libertarian Party of Alaska is a moderate secessionist party (the only state Libertarian affiliate to endorse self-determination for its state). Some parties change from moderate to radical or radical to moderate over time. The Galician National Bloc (BNG) even changed from moderate secessionist to non-secessionist (or “regionalist”) after 1989.[3]
There are also parties that support the withdrawal of existing countries from shared currency, military, or customs union arrangements. The United Kingdom Independence Party favors the withdrawal of the United Kingdom from the European Union.[4] These parties are not treated in this dissertation, nor are regionalist parties that favor some degree of autonomy but explicitly oppose any consideration of independence.
Table 3.1 presents the full classification of secessionist parties included in the dataset used in the regressions of Chapters 4 and five, along with the regions in which they operate in parentheses.[5] The “moderate” versus “radical” classification is not useful for the purpose of the empirical work of this dissertation, but the division is presented here for reference purposes.
Table 3.1 Secessionist Parties Used
in Analyses
|
“MODERATE” |
“RADICAL” |
|
Western Canada Concept (British
Columbia, Alberta, Saskatchewan, Manitoba) |
Western Independence Party
(British Columbia, Alberta, Saskatchewan) |
|
Alliance Démocratique du Québec
(Quebec) |
Parti Québecois (Quebec) |
|
Folkeflokken (Faroe Islands;
through 1994) |
Bloc Québecois (Quebec) |
|
Selvstyrepartiet (Faroe Islands;
through 1994) |
Diverse independentists (Quebec) |
|
Diverse autonomists (Alsace) |
Republikanske Parti (Faroe
Islands) |
|
Union Democratique Bretonne (Brittany) |
Folkeflokken (Faroe Islands;
after 1994) |
|
MPA (Corsica) |
Selvstyrepartiet (Faroe Islands;
after 1994) |
|
Diverse autonomists (Corsica) |
Parti pour l’Organization d’une
Bretagne Libre (Brittany) |
|
Mouvement Social Occitan, Partit
Occitan, & diverse Occitan autonomists (Midi-Pyrenees, Limousin,
Languedoc-Roussillon, Auvergne, Aquitaine, Provence-Alpes-Cote d’Azur) |
UPC (Corsica) |
|
Bayernpartei (Bavaria) |
Corsica Viva (Corsica) |
|
Liga Veneta (Veneto) |
Corsica Naziune (Corsica) |
|
Lega Lombarda (Lombardia) |
A Manca Naziunale (Corsica) |
|
Lega Nord (Veneto, Lombardia,
Friuli-Venezia Giulia, Liguria, Piemonte) |
Rinnovu Naziunale (Corsica) |
|
Partito Sardo d’Azione
(Sardinia) |
Partidu Indipendentistu Sardu
& variations (Sardinia) |
|
Lega Sard (Sardinia) |
Noi Siciliani – FNS (Sicily) |
|
MAS (Sardinia) |
Nacion Andaluz (Andalucia) |
|
SiE (Sardinia) |
PSM (Balearic Isles) |
|
Sardigna Natzione (Sardinia) |
Esquerra Republicana de
Catalunya (Balearic Isles) |
|
Mesa Sardos Liberos (Sardinia) |
Estat Catala (Balearic Isles) |
|
Partido Democratico do Atlantico
(Azores) |
P.N. Canario (Canary Islands) |
|
Partido Democratico do Atlantico
(Madeira) |
CNC-FREPIC-ANAWAC/CCI (Canary
Islands) |
|
PPC (Canary Islands) |
Esquerra Republicana de
Catalunya (Catalonia) |
|
UPC (Canary Islands) |
Herri Batasuna/Euskal
Herritarrok (Euskadi) |
|
CCNC (Canary Islands) |
Euskadiko Ezkerra (Euskadi;
through 1984) |
|
AC-INC (Canary Islands) |
Eusko Alkartasuna (Euskadi) |
|
UNI (Canary Islands) |
Frente Popular Galego (Galicia) |
|
Convergencia i Unio (Catalonia) |
Herri Batasuna/Euskal
Herritarrok (Navarre) |
|
Euskadiko Ezkerra (Euskadi;
after 1984, through 1990) |
Euskadiko Ezkerra (Navarre;
through 1984) |
|
Partido Nacionalista Vasco
(Euskadi) |
Eusko Alkartasuna (Navarre) |
|
BNPG/BNG/Bloque PSG (Galicia;
through 1989) |
Esquerra Nacional Valenciana
(Valencia) |
|
Partido Nacionalista Vasco
(Navarre) |
Esquerra Republicana de
Catalunya (Valencia) |
|
Euskadiko Ezkerra (Navarre;
after 1984) |
Estat Catala (Valencia) |
|
BLOC & div. Catalan
nationalists (Valencia) |
Scottish National Party
(Scotland) |
|
Plaid Cymru/Party of Wales
(Wales) |
Alaska Independence Party
(Alaska) |
|
Libertarian Party of Alaska (Alaska) |
Partido Independentista
Puertoriqueño (Puerto Rico) |
|
Conseyo/Conceju Astur (Asturias) |
Vlaams Blok (Flanders) |
|
Conceju Nacionaliegu Cantabru
(Cantabria) |
|
The “panels” for the data are geographically first-tier sub-state territorial units, as defined in existing legislation. The definitions were carried back in time: thus, Flanders and Wallonia did not exist as unified regions before 1980, but they are the units of analysis for the entire time period of the regression analyses. Countries included in the analysis are those that score at least an “8” on the Polity IV dataset’s Democracy measure and no higher than “0” on the Autocracy measure.[6] In other words, they are highly advanced democracies. This criterion is an extremely strict one and for the most part restricts the dataset to Western industrial countries (data limitations also serve to eliminate the non-Western cases). The reason for this strict criterion is that when governments do suppress political parties, secessionist parties are usually the first to be suppressed. Observed electoral secessionism is thus extremely sensitive to the democratic nature of domestic political institutions. (In fact, some countries that openly suppress secessionist movements, such as Turkey and India, actually pass the Polity criterion in some years and had to be removed by the author.) In addition, there are good reasons to expect that the causes of secessionism will differ between democracies and autocracies, as the first two chapters have argued. Not all the findings of this dissertation are to be extrapolated to non-democratic countries.
The
dataset also does not include capital regions (Washington, D.C., Île-de-France,
and England, for example), and countries without significant sectional
variation (Iceland and Austria, for example) are eliminated. In the theory of
center and periphery, the capital region represents the core of the state: it
cannot therefore secede from itself, by definition. If a “secessionist” party were to arise in one of these regions,
it would really be an “expulsionist” party, seeking to expel peripheral regions
from the country. Regardless, to my
knowledge there is not a single historical example of a secessionist/expulsionist
party of any size arising in a capital region.
This strong correlation would seem to indicate that by and large capital
regions are quite content with their positions in their respective countries.
Table
3.2 lists the countries and regions used in the raw dataset. Not all of these regions appear in the
regression analysis, due to missing data, usually on some of the independent
variables.
Table 3.2 Provinces Used in Analyses
|
Country |
Provinces |
|
Australia |
New South Wales, Northern
Territory, Queensland, South Australia, Tasmania, Victoria, Western Australia |
|
Belgium |
Flanders, Wallonia |
|
Canada |
Alberta, British
Columbia, Manitoba, New Brunswick, Newfoundland, Northwest Territories, Nunavut,
Nova Scotia, Prince Edward Island, Quebec, Saskatchewan, Yukon Territory |
|
Denmark |
Arhus, Bornholm,
Fredericksborg, Fynen, Copenhagen (province, not city), North Jutland, Ribe,
Ringkoping, Roskilde, South Jutland, Storstrom, Vejle, West Zealand, Viborg,
Faroe Islands |
|
Finland |
Ahvenanmaa/Aaland,
Lappi (other regions inconsistently defined, data unavail.) |
|
France |
Alsace, Aquitaine,
Auvergne, Basse-Normandie, Bourgogne, Bretagne, Centre, Champagne-Ardenne, Corse,
Franche-Comte, Haute-Normandie, Languedoc-Roussillon, Limousin, Lorraine,
Midi-Pyrenees, Nord-Pas-de-Calais, Pays de la Loire, Picardie,
Poitou-Charentes, Provence-Alpes-Cote d’Azur, Rhone-Alpes |
|
Germany |
Baden-Wuerttemberg, Bayern,
Bremen, Hamburg, Hessen, Niedersachsen, Nordrhein-Westfalen, Rheinland-Pfalz,
Saarland, Schleswig-Holstein, Brandenburg (after 1990),
Mecklenburg-Vorpommern (after 1990), Sachsen (after 1990), Sachsen-Anhalt
(after 1990), Thueringen (after 1990) |
|
Greece |
East
Macedonia-Thrace, West Greece, West Macedonia, Ionian Isles, Epirus, Central
Macedonia, Crete, South Aegean, Peleponnesus, Central Greece, Thessaly, North
Aegean |
|
Italy |
Valle d’Aoste,
Liguria, Lombardy, Piedmont, Emilia-Romagna, Friuli-Venezia Giulia,
Trentino-Alto Adige, Veneto, Marche, Toscana, Umbria, Abruzzi, Basilicata,
Calabria, Campania, Molise, Puglia, Sicily, Sardinia |
|
Japan |
Hokkaido, Okinawa
(data on other prefectures unavailable) |
|
Netherlands |
Drenthe, Friesland, Gelderland,
Groningen, Limburg, Noord-Brabant, Overijssel, Utrecht, Zeeland, Flevoland |
|
Norway |
Akershus, Aust-Agder,
Buskerud, Finnmark, Hedmark, Hordaland, Moere og Romsdal, Nord-Troendelag,
Nordland, Oestfold, Oppland, Rogaland, Sogn og Fjordane, Soer-Troendelag,
Telemark, Troms, Vest-Agder, Vestfold |
|
Portugal |
Azores, Madeira,
Norte, Centro, Alentejo, Algarve |
|
Spain |
Andalucia, Aragon,
Asturias, Balearic Isles, Canaries, Catalonia, Ceuta y Melilla, Pais Vasco/Euskadi,
Galicia, Cantabria, Castilla y Leon, Castilla-La Mancha, Extremadura, Murcia,
Navarre, La Rioja, Valencia |
|
Sweden |
Blekinge, Dalarna,
Gaevleborg, Gotland, Halland, Jaemtland, Joenkoeping, Kalmar, Kronoberg,
Norrbotten, Oerebro, Oestergoetland, Skaone, Soedermansland, Uppsala,
Vaermland, Vaesterbotten, Vaesternorrland, Vaestmansland, Vaestra Goetaland |
|
Switzerland |
Aargau, Appenzell
Ausser-Rhoden, Appenzell Inner-Rhoden, Basel-Landschaft, Basel-Stadt, Bern,
Fribourg, Geneve, Glarus, Graubuenden, Jura, Luzern, Neuchatel, Nidwalden,
Obwalden, Sankt Gallen, Schaffhausen, Schwyz, Solothurn, Thurgau, Ticino,
Uri, Valais, Vaud, Zug |
|
United Kingdom |
Northern Ireland,
Scotland, Wales, Isle of Man, Jersey, Guernsey |
|
United States |
Alabama, Alaska,
Arizona, Arkansas, California, Colorado, Connecticut, Delaware, Florida,
Georgia, Hawaii, Idaho, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana,
Maine, Maryland, Massachusetts, Michigan, Minnesota, Mississippi, Missouri,
Montana, Nebraska, Nevada, New Hampshire, New Jersey, New Mexico, New York,
North Carolina, North Dakota, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode
Island, South Carolina, South Dakota, Tennessee, Texas, Utah, Vermont,
Virginia, Washington, West Virginia, Wisconsin, Wyoming, Puerto Rico |
Dependencies of a “colonial” nature are not included in the dataset, as independence movements in those territories would be of an anti-colonial rather than strictly secessionist nature. Thus, French overseas departments and territories, British dependencies in the Pacific and Caribbean, and U.S. commonwealths other than Puerto Rico are not included. Also, Greenland is not included, even though the other crown dependency of Denmark, the Faroe Islands, is. What marks out a territory as “colonial” for the purposes of this determination is whether it has a majority-aboriginal population or a majority-settler population. If the territory’s population is mostly composed of the descendants of settlers from the country currently possessing sovereignty over that territory, then the territory is included; if the territory’s population is mostly aboriginal, then it is not included. Puerto Rico is the only exception: it is included because of its long, close ties to the United States and because full statehood is a live issue there.
In
some cases the frames of reference for secessionist parties are historical
regions much smaller than existing first-tier regions as defined above. For example, the Savoy League wants Savoy to become an independent country:
currently Savoy is made up of two departments within the much larger
administrative region of Rhône-Alpes in France. Rhône-Alpes is the unit of analysis in the dataset, and results
for the Savoyan secessionists are not reported, as there is no reason to expect
the political, economic and cultural conditions of the much larger Rhône-Alpes
to correlate with those of Savoy.
Likewise, results for Catalan secessionists in the Roussillon region of
Languedoc and results for Basque secessionists in the Pyrenées-Atlantiques département
of Aquitaine are not reported in the dataset.
Accordingly, the independent variables do not take into account, for
example, Basque speakers in Aquitaine or Catalan speakers in Languedoc, since
those languages are not associated with the whole, or even greater part, of the
region represented in the analysis. It
is unfortunate that these micro-secessionists could not be included in the
analysis, but as their vote totals are uniformly small, the omission should not
affect the results and interpretation.
It is now time to describe more
exactly how secessionist vote shares are measured for the purposes of the next
two chapters. The percentages of the
vote obtained by all secessionist parties in elections to the lower houses of
central and regional legislatures constitute the raw data for the dependent
variable.[7] In cases in which both proportional
representation and single-member-district constituencies exist (Italy and
Germany), results for the proportional constituencies are used. For French data, results are from the first
round of elections. In the case of
Puerto Rico, an odd electoral system prevents results for legislative elections
on an island-wide scale from being calculated, and instead results for the
Resident Commissioner and Governor races are used for countrywide election and
regional election observations, respectively.
All data for both independent and dependent variables are at the provincial
level, unless otherwise noted.[8] Thus, in the case of the Basque country
(presented below), for example, vote percentages are percentages of the Basque
vote in elections to the Spanish parliament and percentages of the Basque vote
in elections to the Basque autonomous legislature.
Table 3.3 below presents a view of
the dataset’s treatment of Basque election years. VOTE represents the vote share of secessionist
parties. From 1979 to 1984, as well as
the 1986 Spanish election, VOTE includes the vote for Herri Batasuna
(HB), Euskadiko Ezkerra (EE), and the Basque National Party (PNV). Eusko Alkartasuna (EA) split from the PNV in
1986 and contested the Basque election that year and every election thereafter. It is included in VOTE from 1986
onwards. Before the 1993 election EE
merged with the Spanish Socialists and is not counted at all in that election
and thenceforth. In 2000, HB organized
a boycott of the election in both Euskadi and Navarre, so that for this
election in these two provinces, VOTE is calculated by counting the
abstentions and blank and null ballots in the 2000 election minus the
abstentions and blank and null ballots in 1996 as votes for HB in 2000. This was done to ensure that the dataset contained
a roughly accurate and consistent measure of secessionist support in 2000 in
Euskadi and Navarre. This was the only
election in the entire dataset for which this sort of procedure needed to be
done. The patterns of nationalist party
politics in the Basque country are complex, but secessionist vote share remains
fairly constant over the entire period.
Table 3.3 Secessionist Vote in
Euskadi, 1979-2000
|
Region |
Country |
Year |
NATELEC |
PROVELEC |
VOTE |
|
Euskadi |
Spain |
1979 |
1 |
0 |
49.7 |
|
Euskadi |
Spain |
1980 |
0 |
1 |
64.4 |
|
Euskadi |
Spain |
1982 |
1 |
0 |
53.0 |
|
Euskadi |
Spain |
1984 |
0 |
1 |
64.7 |
|
Euskadi |
Spain |
1986 |
0 |
1 |
67.6 |
|
Euskadi |
Spain |
1986 |
1 |
0 |
53.7 |
|
Euskadi |
Spain |
1989 |
1 |
0 |
59.1 |
|
Euskadi |
Spain |
1990 |
0 |
1 |
66.0 |
|
Euskadi |
Spain |
1993 |
1 |
0 |
48.1 |
|
Euskadi |
Spain |
1994 |
0 |
1 |
55.4 |
|
Euskadi |
Spain |
1996 |
1 |
0 |
45.2 |
|
Euskadi |
Spain |
1998 |
0 |
1 |
53.6 |
|
Euskadi |
Spain |
2000 |
1 |
0 |
47.1 |
Notes: NATELEC is coded “0” for data points
in which there is not a Spanish election, “1” for data points in which there is
a Spanish election. PROVELEC is
coded “0” for data points in which there is not a Basque election, “1” for data
points in which there is a Basque election.
Since Chapter 4
deals with the cross-sectional variations in secessionist vote, the dependent
variable is averaged for each type of election, regional (or “provincial”) and
countrywide, for those election years in which data for all independent
variables are available.[9] In the case of Euskadi, independent
variables are available for 1982-2000, so the data for these years are averaged
(the mean is taken), yielding a single figure for each region and election
type. The entries for Euskadi in the
new dataset are presented in Table 3.4.
Note that the year figures have been averaged as well.
Table 3.4 Average Secessionist Vote
in Euskadi
|
Region |
Country |
Year |
NATELEC |
PROVELEC |
VOTE |
|
Euskadi |
Spain |
1991 |
1 |
0 |
51.0 |
|
Euskadi |
Spain |
1990.4 |
0 |
1 |
61.5 |
Of course, for
the purposes of Chapter Five VOTE remains unaveraged, since Chapter 5
deals with secessionist vote over time.
The averaged vote is transformed as follows:
LNVOTE=ln(3+VOTE)
The natural logarithm is intended to smooth out skewness in the
dependent variable, and adding 3 was done to ensure an appropriately even
distribution.[10] VOTE remains untransformed for the
analysis in Chapter 5, since a logarithmic transformation was judged to
actually worsen skewness in the variable’s distribution.
The reason skewness is a bigger
problem for the analysis in Chapter 4 is that it includes all regions, whether
or not they have a secessionist party.
In fact, the vast majority of regions did not have a secessionist party
at any point during the period under investigation.
Figure 3.1
Figure
3.2
This skewness is apparent in the
above figures. Figure 3.1 shows the
skewness of VOTE in all regions and after the mean is taken,
and Figure 3.2 shows the skewness of the VOTE variable in those regions
with strong secessionist parties. (The
lines in these figures indicate the hypothetical normal distribution, taking
the mean and standard deviation from the existing data.) The size of the first histogram in Figure
3.1 indicates that over 94% of the observations have an average secessionist
vote below about 13%.
After logistic transformation, the skewness in the averaged variable for all regions remains high, but somewhat improved. Only 69.6% of observations now fall into the first histogram. Figure 3.3 shows the distribution of LNVOTE for all regions, averaged.
Figure 3.3
This chapter has described patterns of secessionist vote in advanced democracies and introduced the general methods of Chapters 4 and 5. The next chapter describes the independent variables used to test the hypotheses advanced in Chapter 2 and performs a cross-sectional regression analysis to determine which regions have more successful secessionist parties.