Bertrand M. Roehner (with the collaboration of Leonard J. Rahilly)
Separatism and Integration: A Study in Analytical History.
Lanham, MD: Rowman and Littlefield, 2002.
xi + 351 pp. $75 (cloth),
ISBN:0-7425-1734-9.

This book is a cornucopia of useful information about separatist movements. Bertrand Roehner focuses his intriguing analysis on separatist struggles by "homeland minorities," which he defines as "minorities whose social and cultural life is associated with a specific territory." He pointedly refuses to provide a general theory of separatism. He writes that he is not interested in explaining the causes of separatist movements (the "why" question). His aim is rather to identify empirical regularities which are common to separatist struggles (the "how" question) by analyzing separatist episodes as "clusters of events." In this respect the book is a specific application of the methodological approach to history and social analysis that Roehner has vigorously argued for in his nice book ``Pattern and Repertoire in History'', written with Tony Syme, and reviewed for EH.NET by John Nye.

This monograph on separatism shares the methodological assumptions of that broader work, including a belief that social scientists should describe "forms of actions" rather than focus on "motivations." Specifically, the author argues that separatist struggles tend to follow similar patterns, and that past separatist struggles in a region are the best predictors of more recent separatist disturbances in that same region, even when explicit motivations have changed. For instance the author, building on work by J.R.G. Jenkins, stresses the elements of continuity in the separatist movement of the Bernese Jura (Switzerland) in the nineteenth and twentieth centuries. He also notices the similarity of Flemish separatist actions in the past two centuries, even though the specific issues under dispute and the economic conditions of the different groups within Belgium have been changing.

The author identifies three kinds of separatist struggles: over land, over language/culture, and over political power. He then uses this taxonomy, in conjunction with the type of society, to classify different kinds of separatist revolts. When the dispute is over land, an aristocratic revolt may occur in a feudal society (say, Hungary in the seventeenth and eighteenth centuries), while a peasant revolt may rise in a rural society where peasants are freeholders (say, Ireland around the same period). A struggle over cultural domination would involve only urban, educated people before the spread of public education (for example, Bohemia in the nineteenth century), but the mass of the population after public education has spread (twentieth century Quebec). Finally, anticolonial struggle over political power may be confined to the nation's elite (Egypt) or involve the mass of people (India), depending on the central government's level of tax pressure. The author describes and analyzes a large number of separatist struggles in these different categories, although, as he writes, space limitations prevent him from providing an exhaustive "Handbook of separatist disturbances."

Although he is reluctant to provide general theories, the author does suggest two sets of determinants of separatism: spatial and historical determinants. He constructs a "geographical index of interaction" to capture the degree of isolation of a region from the rest of the nation. The index is based on the extent to which the region and the rest of the nation share a border. For example, Scotland is more isolated from the rest of Great Britain than Andalusia is from Spain, since most of Scotland's borders are not with the rest of the UK, but with the sea. Adjustments are made for common borders with other countries that share language/culture with the "homeland minority" (e.g., Alsace's border with Germany reduces its "index of interaction" with France). The author finds that linguistic assimilation is higher and separatist disturbances lower in less isolated regions. But the lion's share in explaining separatist disturbances in a region after 1945 is given by measures of disturbances in that same region in previous periods (the "historical index").

I greatly enjoyed reading Roehner's tour de force. The book is rich in historical information, methodological discussions, and analytical observations. However, I wonder whether this book may disappoint some historians and economists, perhaps for opposite reasons. Some historians may take issue with this bold attempt to put the same hat on very different historical events. By contrast, as an economist, I have no problem with heroic generalizations. On the contrary, I wish the author had been even bolder at identifying common variables and determinants. In fact, I feel that an explicit theoretical framework could have helped rather than hindered the analysis. As often is the case with attempts to "let the facts speak" without an explicit theoretical model, assumptions about behaviors and motivations are still there, only they are kept implicit and unstated. For example, one feels that the "geographical index of interactions" could be rationalized (and possibly improved) with a more explicit and rigorous microfoundation -- perhaps linking it to the extensive literature on social interactions developed by economists and other social scientists. And the taxonomy and description of the different "types of revolts" seem pretty consistent with a politico-economic analysis of the incentives and constraints faced by different individuals and groups.

Which brings up a question of special interest to EH.NET readers: how important are economic variables and mechanisms in explaining separatism? Roehner's discussion of this question is perhaps one of the least convincing parts of his book. He observes that income per capita is not systematically correlated to separatism across and within countries: Ireland was poorer than the rest of the UK when it struggled for independence, the Basque Countries and Catalonia are richer than the rest of Spain, support for separatism in the Bernese Jura is uncorrelated to income per capita, etc. He concludes that separatist movements do not "spring from economic motives," and that economic conditions do not exert an influence on separatism. These conclusions seem unwarranted for at least two reasons.

First, the level of income per capita is not a sufficient statistic for "economic motivation." Even if separatists were motivated exclusively by material considerations (say, increasing their future income through a secession), current income per capita would hardly capture this motivation. Instead one should look at the expected effects that a secession would have on their income per capita, that is, to the expected difference between income per capita in the region before and after a secession. For example, in the classic article on the political economy of breakups by Patrick Bolton and Gerard Roland (Quarterly Journal of Economics, November 1997), a majority of people in a region may prefer a secession for economic reasons if they expect a tax and redistribution policy closer to their preferences after the breakup. In that setting, a secession may be preferred by a majority of people in a richer region, but also in a poorer region. What matters to the supporters of separatism is expected changes in tax and redistribution policies, which depend on the differences between average income and median income before and after the secession.

Second, one may believe (with plenty of good reasons) that separatists' main motives are not economic, and still find a strong influence of economic conditions on separatist movements. Again, one should look at marginal effects. People in a region may want to secede for purely cultural/linguistic reasons, even though that may reduce their income per capita. But the expected reduction in income per capita can be higher or lower depending on economic conditions. For instance, in a world of protectionism and high barriers to trade, a secession may turn a small region into a very poor, even non-viable autarkic economy, while in a world of high international openness and free trade the economic costs of secessions are much smaller. Hence, an increase in economic integration tends to strengthen separatism, although separatist preferences themselves may be motivated by purely non-economic considerations. For example, French-speaking people in Quebec may want to form their own country for purely cultural and linguistic reasons, but international economic integration with the United States (NAFTA) substantially reduces the costs of secession, and therefore makes separatism politically and economically viable within Canada. Consider an analogy with individual decisions: most artists choose their profession for non-monetary reasons (say, they love painting or composing music, even though they could make more money in other professions), but changes in the availability of scholarships and grants for artists are likely to make a difference to the supply of art (artists need to eat too).

The above examples illustrate a more general point: before ruling out "economic conditions" (or any other variables, for that matter) as explanatory variables, it is crucial to specify the mechanisms through which those variables may or may not influence separatist preferences and behaviors. Otherwise, one may rule out important variables based on implicit (and perhaps inexact) theoretical assumptions.

In spite of the above reservations and caveats, I think that economic historians and political economists can learn a great deal from Bertrand Roehner's intriguing and stimulating book, which I recommend to all EH.NET readers.

Enrico Spolaore is the author of ``The Size of Nations'' (with Alberto Alesina), MIT Press, 2003.