Geography has been a blind spot for political scientists.
Poor countries are not uniformly poor.
Introduction
Socioeconomic inequality is very high in African countries, but very little systematic scholarly attention has been given to the political effects of this. Many attributes of African economies predict high inequality, including natural resource dependence, trade openness, and low levels of economic development. Economic theory long held that, under these conditions, development itself would increase socioeconomic inequality (Kuznets Reference Kuznets1963; Williamson Reference Williamson1965). Because so much social science research takes inequality as an indicator of latent social conflict, simple deduction might lead one to expect rancor and division over economic policy and redistributive issues to structure national political life in African countries. Yet much of political science writing on African countries argues that structural socioeconomic disparities are not politicized in any systematic way, except in some extreme cases of civil war, and has pointed a finger at cultural or communal differences to explain political division.Footnote 1 This presents a puzzle that has bedeviled several generations of research on African political economy. Do socioeconomic cleavages and economic interests structure national political life in African countries, as theory would predict? If so, how, and why?
Existing work in Africa-centered political science has attempted to solve the inequality puzzle by arguing that in African countries, political elites manipulate deep-rooted ethnic identities to channel politics along the lines of patron–client relationships, undercutting possibilities for the mobilization of class-based or issue-based politics. Ethnic clientelism focuses the political energies of individuals and communities on the micropolitics of retail-level provision of clinics and boreholes, sidelining policy issues of national consequence. Most scholarly work thus suggests that socioeconomic cleavages do not structure national political debate, or may do so only at the margins, because of the intense salience of ethnic politics. What scholars have read as the weakness of electoral pressures for development-oriented policies in African countries is taken as evidence in support of this view.
This book advances a different theory. I argue that while scholars have not found strongly politicized class cleavages in most African countries, they have missed the dominant form of inequality politics in African countries. At the national level, the most politically salient form of economic inequality is spatial inequality. I theorize that spatial inequalities between regions go far in structuring political competition in national elections, and that these same regional cleavages underpin the enduring salience of competition around territorially targeted economic policy and issues of constitutional design.
This theory is rooted in comparative political economy (CPE) literatures on why, how, and which geographic inequalities become salient in national politics. Most such work has been developed through research on European and Latin American countries. In this book, I show that this regional perspective on political economy resonates strongly in most African countries.
In most African countries, regional economic differentiation and inequality are strongly accentuated by territorial institutions that channel both political representation and state action. Strongly regionalized economies have developed within the framework of strongly territorial political institutions. Much theory in CPE predicts that, where spatial inequalities are high and territorial institutions are strong, regionalist interests and political strategies tend to predominate over programmatic, state-wide strategies across many policy and governance arenas (Beramendi Reference Beramendi2012; Rogers Reference Rogers2016). I find that this is indeed the case in much of Africa. What classic works in social cleavage theory call the “spatial–territorial dimension of politics” is a critical but largely unobserved and undertheorized driver of political competition in African countries (Lipset & Rokkan Reference Lipset, Rokkan, Martin Lipset and Rokkan1967; Rokkan Reference Rokkan1971; Caramani Reference Caramani2004). These regional dynamics transcend the spatial and temporal scale of individual patron–client ties and local ethnic identities that are observed in individual acts of voting or one-off clientelistic exchanges within electoral constituencies. By placing microlevel behavior in a wider context, this book brings the larger stakes and purposes of national politics into focus.
Significant research programs in comparative politics associate stark regional inequalities with distinctive sets of political and economic challenges. In countries as diverse as Italy, Spain, Germany, Indonesia, and China, regionalized competition exerts a pull on the overall character of national politics, development trajectories, and patterns of policy competition. Around the world, economic inequality across subnational regions is strongly associated with core–periphery tensions, tensions between wealth-generating and lagging regions, problems of national integration (including the high political salience of ethnic and regional identities), and tensions arising from divergent regional policy preferences (Rogers Reference Rogers2016). In cross-national studies, underprovision of public goods, weak programmatic politics, the prevalence of accountability-eroding electoral clientelism, and civil conflict around questions of territorial dominance or autonomy are sociopolitical ills that have been attributed, at least in part, to high levels of spatial inequality. The relationship of “ethnicity” to these outcomes is variable and contingent, even within the context of one country.
In African countries, the lack of systematic and reliable empirical data at the subnational level has made it difficult to develop and test theories linking spatial inequalities and political outcomes. An earlier generation of qualitative political scientists and historians accorded considerable importance to the role of regional tensions in shaping politics in the 1950s through the 1970s, and these insights have fed into contemporary research that considers regional politics and questions of national integration.Footnote 2 So far, however, this type of analysis has not congealed into theories of structure and variation in national-level electoral geography, or of territorial dynamics in African politics.
From the 1990s onward, much scholarship on African politics has downplayed spatial inequalities and has seen politicized cultural heterogeneity as the cause of political division and of the prevalence of clientelism over substantive policy appeals in elections. Influential political sociologists and economists, such as Donald Horowitz (Reference Horowitz1985) and Easterly and Levine (Reference Easterly and Levine1997), along with a generation of scholars focused on elections and individual voting behavior in the multiparty era, identified ethnicity as an overwhelmingly determinant force in African politics and an ideological force that is orthogonal to – that is, that cuts across and neutralizes – programmatic economic interests and socioeconomic cleavages. Many of these scholars would grant that ethnicity produces a territorial or regional effect when coethnics are spatially clustered, but the spatial clustering itself is often portrayed as an effect of ethnicity, a prepolitical expression of ideological or cultural preference rooted in the distant past, and exogenous to politics. Institutions and economic geography are usually taken to be invariant (or controlled for) within and across countries. Two key shapers of structure and variation in political competition within and across nations – institutions and economic geography – are thus left out of the analysis.
This work leverages CPE theories of social cleavage, uneven development, and institutions to propose a theory of the sources and correlates of regional cleavage in national politics. I will argue that these give structure to electoral competition at the national level and shape the policy content of the national political agenda in many, perhaps most, African countries.
A theory of regional cleavages as the driver of national politics links the study of politics in Africa to classic works on social cleavage and political economy in other parts of the world. This opens the door to significant revisions of prevailing axioms in the study of African politics. Many inferences about national-level politics that are extrapolated directly from micro studies of electoral behavior ignore the ways in which factors that are unobserved in micro studies – including highly uneven patterns of economic development and strongly territorial political institutions – shape politics at the national level. I argue here that economic geography and political institutions play a significant role in shaping political interests and identities, producing cohesion and division in these over time, and in defining the mechanisms and dynamics by which local-level clientelism may (or may not) be subsumed within regional-level electoral cleavages. My analysis replaces common arguments about volatility and fluidity in national political alignments in Africa with theory and evidence of persistent cleavage structures over time. This more expansive theoretical framing sharpens our ability to draw political inferences from existing descriptions of ethnic and electoral politics and reveals the larger stakes in political and policy debates that have long been sidelined in political science studies of African countries.
In an influential study, Herbst (Reference Herbst2000) pointed to political divisions along geographic lines in African countries when he wrote of the stark cleavage between capital cities and rural hinterlands. Herbst depicted hinterlands as largely resistant to incorporation into the national fold, and to the “rural areas” as constituting a largely undifferentiated and recalcitrant periphery in most countries. Here, I invert Herbst’s model by locating dynamic drivers of politics in the regions. What Herbst refers to generically as the “hinterlands” are treated here as segmented and differentiated landscapes composed of regions marked by economic inequalities and power differentials, variable alignments vis-à-vis the center, distinctive production profiles and economic interests, and varying modes and degrees of integration into the national economy. Relations among such regions are potentially competitive, with tension arising from rival policy preferences and competing visions of state- and economy-building. National regimes are shown to be rooted to a very significant extent in regional strongholds, to achieve national predominance on the basis of predominantly rural electoral coalitions, and to be invested politically in the prosperity of regionally specific sectors of the economy (Rabinowitz Reference Rabinowitz2018). In many African countries, the persistent lines of social cleavage that are visible at the national level are regional in nature, often taking forms that are familiar to scholars of regional competition and cleavage in other parts of the world.
1.1 Inequality Structures in African Countries
This work focuses on territorial oppositions – the divergent interests and priorities of subnational regions in relation to, and in competition with, each other – that grow out of the process of building a national state and economy. It argues that institutions contribute to the structuring of these territorial differences and inequalities, and that institutions work to bring these inequalities to the fore in national political competition.
My point of departure is the dramatic extent of both interpersonal income inequality and regional economic disparities in most African countries. Not only are some of the world’s highest levels of interpersonal income inequality found in sub-Saharan African countries, but levels of spatial inequality among regions (provinces) in most African countries are also extremely high. Economic disparities across subnational regions in most African countries are higher than they are in textbook cases of high spatial inequality featured in the CPE literature – including Spain, the United States, Mexico, and Argentina.
These strong inequality patterns are captured in the scatterplot in Figure 1.1, adapted from Rogers (Reference Rogers2016). High inequality countries from around the world – including Brazil, Mexico, Malaysia, and the USA – are featured in orange, for comparison.
Figure 1.1 Interpersonal and spatial inequality in African countries, 2012 (with comparison to some other high inequality countries)
Notes and sources: The Y-axis is the coefficient of variation in adjusted nighttime luminosity across provinces (Admin1 regions), unweighted for population, in 2012 (Lessmann & Seidel Reference Lessmann and Seidel2015, Reference Lessmann and Seidel2017). The X-axis is Gini of interpersonal income inequality (Milanovic Reference Milanovic2014). See Appendix Figure A1.1 for the population weighted data.
The vertical axis captures interpersonal income inequality. Measured by national Gini coefficients (as shown), Africa’s levels of income inequality are among the highest in the world, even exceeding Latin America’s stratospheric levels by some measures.Footnote 3 By the Gini index of household consumption expenditure for 2008, Africa was the world’s most unequal macroregion, with a Gini of 67 compared to 50 for Latin America and the Caribbean.Footnote 4 These high levels of interpersonal inequality are a long-standing feature of economic structure in African countries and have been traced back to the 1950s, when the earliest data are available (Milanovic Reference Milanovic2014: 11).
The income Gini captures the familiar picture of African societies as polarized between small, wealthy elites in urban, formal sector employment (i.e., positions in the higher echelons of the state, and in top private and multinational firms) on the one hand, and large majorities trapped in low-productivity rural livelihoods (agriculture, pastoralism, etc.) and the informal sector. As van de Walle explained (Reference Van de Walle2009), the smallness and weakness of the middle class is a corollary of this bifurcation. National-level policy factors, including regressive taxation structure and very low levels of income redistribution, contribute to high levels of interpersonal inequality, but the main drivers lie on the production side. There is strong variation across African countries: Those with higher levels of economic development tend to score higher on the Gini index.Footnote 5 This is consistent with Kuznets’ prediction that development itself would increase interpersonal income inequality.
The horizontal axis in Figure 1.1 measures spatial inequality. The figure captures variation in levels of economic development across provinces, proxied by nighttime luminosity, a commonly used if coarse measure of subnational GDP. By this measure, levels of spatial inequality, or inequality across provinces, in sub-Saharan Africa countries also rank among the highest in the world (Table A.1).Footnote 6 Most African countries rank higher on this measure than country exemplars of very high spatial inequality featured in the CPE literature, including the United Kingdom, Spain, Indonesia, Argentina, and Mexico. Indeed, in a global sample of countries measuring spatial inequality, almost all African countries stand out with extreme scores on the spatial inequality dimension (along with a number of non-African countries, including Syria and Pakistan).
In African countries, national capitals do have stark advantages over the provinces in terms of levels of economic development, as the urban bias literature underscores. Yet these data and many alternative measures of spatial inequality show that there are also sharp differences across predominantly rural regions. In Tanzania in 2016, for example, average consumption per adult in Manyara region, the richest of the provinces, was almost three times that of Ruvuma and Kigoma regions.Footnote 7 Similarly, in Ghana, living standards in the central cocoa-producing regions are twice what they are in the poorest rural regions of the northern savanna, where livelihoods are centered on pastoralism and subsistence agriculture. In Kenya, early childhood mortality rates are almost four times higher in the western county of Homa Bay, lying on the lowland shores of Lake Victoria, than they are in Nyeri County in the agriculturally rich central highlands.Footnote 8 As Sahn and Stifel argued, in most countries, indicators of poverty “differ markedly between rural regions of almost every country” (2000: 593). Changes over time in rural well-being – net improvements and declines – also “differ dramatically across rural areas” and are often highly regionalized (Sahn & Stifel Reference Sahn and Stifel2000: 593). When countries are growing, some rural regions benefit, while others fall behind.
The existence of two forms of extreme socioeconomic inequality – interpersonal income inequality and spatial inequality across regions – complicates the inequality puzzle in African countries. Melissa Rogers argued that “these indicators represent distinct potential distributional conflicts within a nation” (2016: 27). In Rogers’ global sample of countries, interpersonal and spatial (interregional) inequality are weakly correlated. She found this to be the case for sub-Saharan Africa, as well. Some countries are indeed marked by levels of interpersonal income inequality that are far more extreme (by both world and African standards) than their levels of spatial inequality.Footnote 9 South Africa is the clearest example. Others exhibit the reverse combination, where spatial inequality appears far more extreme than income inequality. Mali and Ethiopia are exemplars of this pattern. Yet for most African countries, both income and spatial inequality are very high by world standards. The presence of both high income and high spatial inequality means that there are stark potential lines of socioeconomic cleavage and distributional conflict in most African countries. Will structural economic inequalities find expression in politics, and if so, which ones, and how?
So far, most studies of African electoral politics have discounted programmatic economic concerns as a driver of political competition, or argued explicitly that the ethnic cleavages visible at the national level are not systematically related to structural socioeconomic cleavages. This is puzzling for comparative political economists who expect inequalities rooted in the economy – be they along class, sectional, or sectoral lines – to stoke systemic distributive conflict.
1.2 Main Argument
CPE scholars argue that institutional structure is critical in shaping inequality’s political effects. A substantial line of theory predicts that, where strong regional inequalities are overlaid by strongly territorial institutions, “distributional conflict will exist primarily among territorial groups” (Rogers Reference Rogers2016: 2). Strongly territorial political institutions will work to accentuate the political salience of spatial inequality and to channel the politics of inequality into “a distribution game … across and within geographic districts” (Rogers Reference Rogers2016: 2; see also Rodden Reference Rodden2010; Beramendi Reference Beramendi2012).
In this book, I deploy this insight to identify and unlock the puzzle of inequality politics in African countries. In most African countries, spatial inequality is high, and structures of political representation, government administration, and resource allocation are strongly territorial. As CPE theory would predict, in most African countries, national competition can be described as “territorial politics in regionally divided countries.”
Figure 1.2 captures theoretical expectations about the relationship between high spatial inequality and strongly territorial institutions (Rogers Reference Rogers2016), and locates African countries on this conceptual map. My main argument is that most African countries fall into Cell 1 of Figure 1.2, where high spatial inequality and strongly territorial institutions coincide, and that territorial politics predominates. Yet there is cross-national variation in the relationship between institutional and inequality structures in African countries, as noted above, and this offers some additional analytic leverage.
Figure 1.2 Territorially divided states: Theoretical expectations and African exemplars
Notes: Territorially divided states are characterized by high spatial inequality and strongly territorial institutions that align with geographical economic disparities, as typified in Cell 1. “TZ” is mainland Tanzania.
Most African countries are located in Cell 1 in Figure 1.2. Rural populations predominate – in most countries, over 50% of the population is largely dependent on land-based livelihoods. In most, natural endowment is highly unevenly distributed across space, giving rise to strong sectoral–spatial differentiation. This unevenness is associated with spatial inequalities and different modes of integration into the national polity and economy. Strongly territorial institutions of administration, political representation, and land tenure (property rights in land) overlay regionalized patterns of economic development and magnify their political salience. Rules of electoral competition, including multiparty competition organized around territorially defined electoral constituencies, contribute to the political salience of geographically specific interests. In these countries, regionalism is expected to trump class politics. Examples are Kenya, Zambia, and Côte d’Ivoire.
In Cell 4 of Figure 1.2, where strongly centralizing institutions coincide with lower spatial inequality, the political salience of interpersonal income and class differences is expected to be stronger. In more industrialized and urbanized countries, the political salience of spatial inequality falls relative to income inequality. With nationally integrated labor markets and high levels of proletarianization, the salience of rich–poor polarization increases in national politics. Universalizing institutions associated with the development of mass- and class-based society – such as national parties and trade unions – give voice to class-like demands and policy responses. Administrative divisions cut across, rather than overlay and reinforce, the most extreme spatial disparities. In the sub-Saharan Africa context, South Africa is the exemplar. In South Africa, a strongly integrated national bureaucracy and an electoral system based on proportional representation have also promoted what Caramani (Reference Caramani2004) calls the “nationalization of politics.”Footnote 10 Under these conditions, class politics dilutes or trumps the politics of regionalism. In Botswana and Namibia, where levels of urbanization and of interpersonal income inequality are also very high, territorial cleavages also have lower political salience (see Baleyte et al. Reference Baleyte, Gethin, Govind and Piketty2020). In these counterfactual cases, regional tensions do not predominate; they are overwhelmed by other social cleavages.Footnote 11
The two cells on the opposite diagonal are theoretically generated types that are not associated with unambiguous empirical referents in the African context. Cell 2, where spatial inequality is low and national institutions are strongly territorial, is not expected in African countries. Under colonial rule, strongly territorial state institutions coevolved with highly uneven patterns of economic development and political control of African populations (see Chapter 3).
Cell 3 describes a configuration in which spatial inequality is high, but strongly nationalizing institutions mute or dilute the salience of territorial identities and interests, and provide few supports for regional interests to coalesce in the political arena. As suggested in Figure 1.2, Tanzania has features that pull it toward Cell 3, but it is not a pure type (as indicated by the tilde).Footnote 12 While there are clear patterns of regional inequality and Tanzania’s electoral rules encourage the political expression of territorial interests, key features of institutional structure have been nationalizing: These include centralized ruling party and line-ministry control over regional, district, and local administration; direct administration of rural localities; nationalizing (rather than region- or district-specific) land tenure institutions; and one-party rule between 1965 and 1995.Footnote 13 Regional tensions have decisively shaped the national trajectory, but in mainland Tanzania these have been attenuated for much of the last sixty years of independence.
This book argues that in most sub-Saharan Africa countries from the 1950s to today, including the post-1990 era multiparty politics, the salience of territorial politics has been high. I show that national politics in many and perhaps most African countries is structured along regional lines in ways that are critical to understanding the stakes of economic modernization and bureaucratic consolidation. Cell 1 captures the predominant tendency, describing not only Kenya but also most of the countries in the twelve-country study (see Section 1.4) that provides empirical foundations for the main arguments advanced across Chapters 2 through 7.Footnote 14 Most of the analysis concentrates on identifying structure and variation within the predominant pattern captured in Cell 1, identifying the institutional and sectoral–spatial features of African political economies that account for this, and identifying the effects of regionalism in political and policy dynamics at the national level. South Africa serves as the main counterfactual in an argument that is developed in the first part of Chapter 7.
This book develops the theory that regional differences and spatial inequalities, in interaction with state institutional structure, shape politics and policy in African countries, as they do in countries with high spatial inequality and strongly territorial state institutions in other parts of the world. These effects are not reducible to ethnic effects. Indeed, the same institutional factors have shaped both economic geography and ethnic identities. Territorial institutions play a strong role in defining the content of ethnic and regional identities, which identities find expression in politics, and the spatial scale at which they find expression in the national arena. Cleavages that find expression in national politics in contemporary African countries tend to be associated with regional identities and interests. These transcend the ethnic identities that are linked to precolonial indigenous languages and ancestral property rights, and colonial native authority units.Footnote 15
1.3 Theoretical Foundations
Unfolding processes of state-building and national economic development have created regional winners and losers. Regional actors have coalesced with the framework of regional administrative and political institutions and mobilized over time in defense of their interests. Since the 1940s if not before, moves toward economic integration and political centralization have stirred center–periphery tensions and other territorial oppositions, as predicted by Lipset and Rokkan (Reference Lipset, Rokkan, Martin Lipset and Rokkan1967) and other theorists of territorial politics. Forces of market expansion and agglomeration exert a pull in favor of already leading regions. These tend to be export-producing regions, as much trade theory would predict.Footnote 16 Postcolonial governments have both spurred on and regulated these processes. They have differentially allocated the costs and benefits of change over time, shaping and politicizing the pace and direction of national economic integration. Clientelism, pork barrelism, and low levels of redistributive spending are starkly visible in African countries, as is the case in other territorially divided countries around the world (Rogers Reference Rogers2016).
Theoretical tools from historical sociology, economics, and comparative political economy provide foundations for these arguments.
The “Territorial–Geographical Dimension” of Politics: Rokkan
A two-dimensional space describes alternative cleavage structures in classic studies of the political development of Europe, with the functional (income or class) axis of cleavage predominating over the territorial axis as the major source of differentiation in party system structure in Europe from the 1950s to the 2000s.Footnote 17 For the earlier period, however, a different configuration tended to prevail, and this is what interests us here.
Lipset and Rokkan (Reference Lipset, Rokkan, Martin Lipset and Rokkan1967) studied the ways in which patterns of urban–rural, core–periphery, and church–state cleavage were activated in national politics in modernizing Europe by the accelerating forces of urbanization and early stages of industrialization. Development of national economies and the growth of modern national bureaucracies under centralizing rulers are drivers of tensions between core and periphery, and across regions within peripheries. Lipset and Rokkan emphasized what they depicted as “territorial oppositions,” which they argued were linked to different regional economies, and the varying types of social organization (including but not reducible to cultural and social values) embedded in them. I adopt a similar theorization, emphasizing heterogeneity in regional economies and hierarchy among them, and arguing that tensions and competition arising from this find expression in political organization and economic interests, and ultimately, national politics.
Class politics and territorial politics are associated with very different forms of political competition and bargaining. Class politics revolves around social contracting to divide economic surpluses generated by growth (or the costs of decline), predominantly between two nonspatial but intrinsically interdependent parties: owners of capital and owners of labor power. Territorial politics by contrast focuses on relative advantage and disadvantage across geographic segments of the national polity. It is driven by concern not only with securing state spending “here” rather than “there,” as scholars of the electoral logics of distributive politics at the micro level have argued, but also with the evolution of relations between leading and lagging regions, adverse incorporation into national political economy of lagging regions or those on the sidelines of power, and caution vis-à-vis growth strategies that may channel advantages to already dominant regions. Rather than a politics around division of an economic surplus generated within firms, sectors, or the national economy, as is the case in class politics, territorial politics revolves around winners and losers created by spatial expansion and/or segmentation of markets, spatially targeted policy, the commodification and decommodification of access to immobile natural resources, and the spatial limits, autonomy, and powers of the local state vis-à-vis the center.
In the African context, the spatial–territorial dimension of politics focuses on distributional tensions among regions. The politics of clinics and boreholes may be an instantiation of this at the neighborhood or micro level. Yet when analysis scales up to the national level, regional coalitions are primary contenders, especially those regions that are most deeply integrated into the national economy, and when larger, more programmatic issues and tensions are at stake in the formulation of national policy.
Particular types of collective action problems around growth strategy and redistribution emerge at this level. Programmatic policies, both social and economic-developmental, often have a strong spatial bias. Indeed, these often involve spatially targeted state spending and development policies. Market-expanding policies are likely to reinforce the advantage of leading regions and may compound the adverse incorporation of lagging regions. The bearers of such interests are not only the cabinet-level ethnic brokers featured in studies of autocratic rulers’ “ethnic arithmetic,” but also political party builders, local and regional elites, and multitudes of ordinary citizens who are linked to regional economies through their occupations, livelihoods, place-based associational lives, local citizenship responsibilities and rights, and property holding. This includes farmers organized into producer associations such as cooperatives in sectors that are deeply incorporated into the national economy. They are sensitive to regionally targeted sectoral development and regulatory policies.
The argument that politics in many African countries can be understood as driven by distributional tensions that cleave the national body politic along regional lines was a leitmotif in African studies from the 1960s. This study argues that such strains have persisted over many decades and are strongly visible today in patterns of coalition and cleavage in electoral politics at the national level, as well as in cleavages around political issues that tap into regional tensions and divides.
Spatial Inequality: Rokkan Meets Krugman
The New Economic Geography (NEG) as pioneered and popularized by Krugman (Reference Krugman1991, Reference Krugman1998) helps explain where regional inequalities are likely to emerge, and how and why they are likely to deepen over time in the absence of profound change in technology or institutions. The Old Economic Geography assumed that, over time, regional growth rates and income levels would converge as capital chased lower labor costs, and as migrants moved to regions that offered the best wages for their skill-mix. The old school believed that regional economic disparities would erode over time. The New Economic Geography undercut this logic, demonstrating convincingly that economies of scale in production – theorized as a type of market imperfection – generated increasing returns for already dynamic regions. Density and agglomeration themselves have a positive effect on economic activity. Empirically, the New Economic Geography has won out, as regional convergence ended in most OECD countries around the 1970s, giving way to decades of growing regional inequality within the EU countries and the United States that would contribute to the political and social strains of the 2000s. “The great divergence” in regional growth patterns within the developed countries emerged as a new, politically urgent object of study in political economy and regional economics.
In Africa, too, the convergence process that 1950s and 1960s economists expected to see has failed to materialize. There is far less convergence not only between African countries and the West, but also across regions within African countries than was predicted in the heyday of mainstream development theory in the 1960s and 1970s. With a few notable exceptions, the relative rankings of regions within national economies in Africa over the last several decades (measured by regional GDP or standards of living) have proved to be remarkably stable. Analysts often describe uneven spatial distributions of income that were established in the first half of the twentieth century as entrenched (Moyo Reference Moyo2014: 11).Footnote 18 This reflects inter alia the limited structural transformation of African economies over the last several decades (limited industrialization in particular, and the persistence of large and growing rural populations). But even in other settings, agglomeration theory suggests that markets tend to concentrate growth in already leading regions, making the catch-up of lagging regions a very uncertain proposition (Scott & Storper Reference Scott and Storper2003). Regional economic convergence was unlikely to happen automatically, or evenly in time and space. Poverty traps are real.
Venables (Reference Venables2010: 470) suggested that focusing on agglomeration forces can support a core–periphery view of the national economy. This is the line of reasoning followed here. This suggests an economic model of persistent spatial inequalities that is highly complementary to the Lipset–Rokkan model of political and economic development.
Drawing on earlier theorists of uneven economic development, my analysis foregrounds the role of national economic integration in producing centripetal (agglomeration) pressures that tend to exacerbate uneven development and spatial inequality. In African countries, regional growth rates and endowments differ, creating hierarchies of regions defined by differential growth rates (and levels), and geographies of relative winners and losers. Large literatures in economic geography lead us to expect such unevenness to give rise to demands from some peripheries for alternative models of incorporation and state-building. Demands may take the form of calls for more local autonomy (and less central state intervention), more redistribution, or sector-specific economic policies that reflect particular regional interests. Core–periphery tensions and region-specific needs and claims on the state create the potential for regional or territorial politics in African countries. These would be expected to trump class cleavages in structuring politics and policy debates where the potential for class cleavages to form as a strong axis of competition is low.
In African countries, evidence of the cumulative advantages of already favored regions is starkest in the rise of capital cities. Expansion of the state itself has been a major driver of primary (or primate) city growth and agglomeration of economic activity, including tertiary sector activity and wage/salaried employment, in capital cities. However, the rise of primate cities has generally not been accompanied by structural change in national economies along the agrarian-to-industrial pattern observed in America and Europe in earlier centuries. In the Lipset–Rokkan model, urbanization and industrialization drove the “nationalization of politics” by eroding preexisting, regionalized and regionally differentiated social structures, and promoting the rise of a new, nonterritorial [territory-wide] national economy and class structure. Africa’s leading cities, by contrast, have been dubbed service or consumption cities. Their rise has not been driven by industrialization (although they tend to be sites in which national manufacturing industry is concentrated), or accompanied by absolute decline in rural population counts or, in most cases, the formation of social classes based on the capital–labor division associated with industrialization. (South Africa’s leading cities provide a stark exception to this rule.) The phenomenon of urban economic advantage is well recognized in the political science literature in Africa, along with some of its political correlates and effects. Less recognized is the fact emphasized here – growth and forces of agglomeration have also tended to favor the already leading predominantly rural regions.
Predominance of natural resource-based sectors and industries means that levels and location of economic activities are strongly influenced by natural endowment. Other factor-mobility-constraining “market imperfections” reinforce endowment-related disparities between leading and lagging regions. These are related most obviously to institutions and transport infrastructure (or connectivity more generally), but less visible impediments to population mobility are also important.Footnote 19 Uneven levels of incorporation into national markets mean that market-based “universalistic policies” are not actually space-neutral or spatially blind. Political dynamics, in turn, play a role in shaping economic structure over time.
Divergence of regional trajectories is a corollary of the above. When overall national growth rates are positive, different parts of a country will share very unequally in this prosperity. When times are bad, the burden is distributed unevenly. Divergence itself (relative gains/losses) as well as region-specific attributes shape the political preferences of region-based actors.
Institutions Matter: CPE Theorists
Theory in comparative political economy predicts distinctive political outcomes in countries that combine high levels of regional socioeconomic inequality with strongly territorial, or territorially fragmented, political systems (Beramendi Reference Beramendi2012; Rogers Reference Rogers2016).Footnote 20 Such political systems are defined not only by strongly territorial systems of political representation, but also by structures of public administration and bureaucracy that divide the polity into internally cohesive “state segments” with low cross-unit bureaucratic interdependency. In these settings, the structure of the national political system is expected to bring territorial oppositions to the fore, to amplify the political salience of interregional distributive issues and of spatial inequalities in national politics, and to highlight tensions around the terms of incorporation of regions into the national political economy. Canonical cases are the United States, Canada, Spain, Germany, and Argentina, where concerns with disparities in interregional resources and distribution of power are central to the logics of national politics.
Focusing on institutional determinants, Melissa Rogers (Reference Rogers2016) argues that in “territorially divided states” we should expect that (a) territorial institutions define collective actors in politics – including political identities and agendas; (b) institutional structure promotes and heightens the salience of distributive conflict across geographic units – “at the extreme, politics becomes conflict about distribution of resources among jurisdictions” (Beramendi Reference Beramendi2012: 40); and (c) political competition produces centrifugal forces: Diverse regional economic profiles generate distinctive and often divergent political preferences, including divergent preferences arising from cross-regional economic inequality itself.
Juxtaposed to these are countries with strongly centralizing and homogenizing institutional structures. These include strong proportional representation (PR) and national list voting and parliamentary systems that dilute the salience of geography in electoral dynamics; resource allocation through corporatist (functional) structures, and state territorial administration through functionally defined authorities (e.g., national social security administration or national education ministry), rather than territorially defined authorities, to check spatial nonuniformity in the supply of government. A corollary of these is strong party systems and nationalized parties, and civil society organizations such as strong labor unions, producer peak associations, and religious organizations of national scope that pull together cross-local political coalitions. France is often cited as a paradigmatic case.
In this study, I argue that the predicted effects of strongly territorial political institutions and high spatial inequality feature prominently in the politics of most African countries. In many African countries, strongly territorial political institutions structure the integration of citizens into national political life. These work to aggregate political preferences around territory in general and around region or district in particular, apportion resource access among territorially defined population groups, and create territorially defined channels by which citizens access top-down resource flows from the central state.
The most important institutions are ethnically defined territorial constituencies (former Native Authorities) which prevail in many parts of most countries, single-member district (SMD) electoral systems, and territorial administrative units at the provincial (Admin1) and district (Admin2) levels that divide national space into state segments that often differ strongly from each other in their economic and governing structures. Territorial administration creates “containers” within which policy is tailored for local circumstance and implemented, and within which parties and political entrepreneurs may build coalitions and alliances across multiple levels of the administrative hierarchy, from the microlevel, to the constituency, district, and regional level. State institutions and institutional practices often define separate and unequal systems of property holding (in land) for territorially defined subgroups of citizens, foster regional economic segmentation, and supply regionally differentiated government services, resource flows, and policy.Footnote 21 Correlates of these state features are weakly nationalized party systems (i.e., regionalized electoral mobilization) and the relative weakness of trans-local or cross-regional societal organizations and mechanisms of interest aggregation.
As Isham et al. (Reference Isham, Woolcock, Pritchett and Busby2005: 145, n. 5) note, a large class of models and theories in comparative political economy shows that political institutions are themselves conditioned by economic structure. Both factors shape the power and interests of different social groups and these, in turn, compete to shape and reshape political institutions. These endogeneities are strongly present in the case material examined here. The strong territoriality of governance structures in African countries coevolved with colonial economies that were dependent upon “enclaved” or regionally specific production of export commodities. In this study, I deal with this endogeneity head-on by showing how territorial institutional structures and uneven development coevolved in the British and French colonies of sub-Saharan Africa, producing a “regionalization by design” of both colonial economies and citizenries. The institutional structure of the postcolonial state itself is treated as an overtime effect of postcolonial politics, again conditioned by economic structure. As Rogers (Reference Rogers2016) argues, struggles over constitutional order and state design in regionally divided states with high spatial inequality can be expected to unfold in particular ways: that is, over the distribution of political authority and prerogative between regions, and between regions and the central state.
1.4 Main Arguments and Chapter Outline
In many, perhaps most, African countries, governmental institutional structures (administrative, economic, and representative) overlay patterns of spatial inequality and uneven economic development across predominantly rural regions. The central argument of this book is that, where this is the case, regional cleavages go far in structuring national electorates, and regionalism is a salient force in national-level politics. As Stephanie Rickard (Reference Rickard2018: 24) observed, “regional effects are strongest when institutions and socio-economic inequalities align.” Because these institutional structures and patterns of uneven development are traceable to state formation projects set in motion in the early twentieth century, I dub the dominant, overtime tendency Regionalism by Design. Counterfactual cases in Africa are countries where these forces are mitigated by more strongly nationalized economies (nationally integrated economies), administrative systems, and political systems.Footnote 22 These counterfactual conditions strongly mitigate or attenuate the political salience of regional inequality and economic differentiation. Class-like politics tends to trump territorial politics.
Institutions play a critical role in producing and reproducing regionally uneven development, and shaping the political expression of regional inequalities and difference. State institutional structure has gone far in defining the political identities of subnational collectivities (via ethnicization of state-delimited “tribal” territories under colonial rule).Footnote 23 At the same time, state-led economic development within the template of colonial- and postcolonial territorial administration has given rise to, and/or entrenched, differences in levels of development and sectoral profiles of subnational regions. Factors related to endowment and location combine with active state promotion of regional economic specialization, accentuating the sectoral economic profiles (e.g., export crop production, labor-exporting systems, pastoral zones) of subnational units delimited, roughly, by the boundaries of administrative districts or provinces. Spatial inequalities across such units have tended to be persistent, indeed entrenched, over the last fifty years. Economically leading regions of the late 1950s and early 1960s are, in most cases, economically leading regions in the 2010s.
Such persistence of spatial inequality and hierarchy, combined with the alignment of socioeconomic differences with formal (institutional) territorial divisions, has contributed mightily to making regional inequalities politically salient in national politics in African countries. Alignment of economic inequality with institutional structures of representation has worked to promote regional bloc voting in national politics and to structure patterns of competition and coalition formation across regional units. Policy competition across regions arises around policies that have distributive implications and regionally uneven effects. Regional tensions also surround choices over constitutional arrangements that distribute powers across territorial subunits of the state.
These arguments about the causes and effects of regional political cleavages and regionalism in African politics are developed across Chapters 2–7 of the book, following the schema in Figure 1.3.
Figure 1.3 Regional cleavages and national politics: Causes and effects
Notes: Schematic outline of main arguments and chapter structure.
Operationally, Chapters 4–6 adopt an electoral geography definition of “cleavage” by taking spatially distinct clusterings and discontinuities in electoral patterns as indicative of political cleavages in the electorate.Footnote 24 Using electoral geography methods and logics, I identify patterns of persistent regional bloc voting in presidential elections in a diverse sample of twelve African countries since the 1990s and take this as evidence of regional structure and cleavage in political competition at the national level.Footnote 25 The persistent electoral bloc is taken as the expression of a type of political coalition (or coalescence of interests) that developed since the late colonial period, generally at the provincial scale. Political competition between these blocs goes far in defining the structure of competition in national political arenas over time. This is a form of politics shaped in large part by the alignment of the territorial structure of the state with sectorally heterogeneous and economically unequal regions.
By leveraging analysis of economic inequality across these persistent electoral blocs in each of twelve countries, it is possible to address a series of questions about how spatial inequalities are manifest in national politics, and how they shape issues that are salient on the national political agenda. Why do persistent regional blocs emerge in some parts of a country, but not in others? What are the origins and bases of bloc cohesion and cleavage? How do these relate to ethnicity in politics? Are patterns stable over time? To what extent does economic hierarchy among blocs correspond to patterns of political dominance (and exclusion) in national politics? How do spatial inequalities and cleavages shape coalition-building strategies in national elections? Zooming out to encompass the secondary literature on a wider range of African countries, I ask: Is regionalism manifest in national policy agendas, and if so, how? Answers to these questions are developed over the course of the chapters outlined below.
Chapter Outline and Summary
Chapter 2 develops the conceptual and theoretical framework in five steps. The first situates this analytic and theoretical approach with respect to earlier work on ethnicity and region in African countries. Earlier work is marked by a strong tendency to look away from regional economic inequality as a political force in African countries, and has defaulted to theories centered on ethnicity, understood as a force orthogonal to programmatic economic policy interests and devoid of economic ideology. The second part of Chapter 2 grounds the concepts of “region” and regional inequality. In African countries, the sources of regional economic difference are found in the uneven distribution of natural endowments, regionally specific patterns of state intervention in the economy that date to the colonial period, strong spatial–sectoral differentiation, and institutional administrative structure. Regional inequality is a corollary of the sectoral heterogeneity of regions that make up the national economy, and of constraints on the mobility of factors of production that would, in a perfectly neoclassical world, lead to convergence in regional levels of economic development and incomes. Part three follows Lipset and Rokkan in theorizing the particular sources and nature of regional cleavages that arise in the course of state-building and national economic integration. Part four identifies institutions that contribute to the “regionalization” of national economies and politics in African countries. The last part lays out the main elements of an approach to the analysis of regionalism that is fit for African contexts.
Chapter 3 develops the spatial inequality and institutional arguments in historical context. I identify the colonial origins of the institutional structures and patterns of uneven economic development that create the template for political regionalism. The analysis shows how and why functional economic regions and administrative regions tended to align in the African colonies, defining patterns of regional difference and inequality in the early decades of the twentieth century that are often strongly visible today. This process also established the framework within which politically salient ethnic identities developed, and defined political (eventually, electoral) constituencies in strongly territorial terms. With the opening of political space for colony-wide politics in the 1940s, existing administrative and political structures worked to channel regional interests and ideologies of regional consciousness into the national political arena. A boundary persistence analysis underscores the large extent to which colonial territorial grids have been reproduced over time, defining the institutional playing field of politics in many African countries today. The final section of Chapter 3 (Section 3.6) argues for explaining institutional persistence (and changes) in terms of when and how African political leaders, social elites, farmers, landholders, and members of rural communities found advantage (or the avoidance of bads) in territorial institutions forged under colonialism. Regional cleavages, and the territorial institutions that help frame and reproduce them, that were in place in the 1950s and 1960s have structured patterns of national-level political competition in many African countries for many decades.
Chapter 4 argues that regional interests and tensions are manifest in regional bloc voting in the contemporary era of multiparty politics (1990s–2010s). It is coauthored with Juliette Crespin-Boucaud and draws on earlier work with Michael Wahman and other collaborators (Boone, Wahman, Kyburz, & Linke Reference Boone, Wahman, Kyburz and Linke2022). We present an electoral geography analysis of constituency-level voting in presidential elections in twelve countries over the course of 1990–2015 (44 elections).Footnote 26 Regional bloc voting is visible in all twelve. Across the twelve countries, we identity a total of thirty-six distinct, persistent regional-level electoral blocs. Leveraging a variety of data sources, we describe the economic attributes of the persistent electoral blocs and how they differ from parts of each country in which persistent clustering of regional bloc voting does not occur. These data consist of forty rounds of DHS surveys for geocoded data on education and ethnicity, nighttime luminosity data, historical maps of producer regions, and raster data describing population densities and contemporary crop production profiles. The analysis in Chapter 4 underpins two critical planks of this book’s argument. First, the prevalence of regionalized voting patterns across a diverse set of African countries supports the argument that regionalism is a distinct feature of political competition in many, perhaps most, African countries. Second, the economic profiles of the electoral blocs provide strong clues as to bloc etiology. We begin to see why persistent, multiethnic electoral blocs coalesce in some regions but not in others.
We find that most of the persistent electoral blocs arise in rural regions that are wealthier, better educated, more densely populated, and more deeply incorporated into the national economy than other rural areas. Most of the persistent blocs are specialized in high-value export crops (or traded food crops). Some have nonagricultural production profiles as labor-exporting or mining regions. Most coalesce within provincial-level or Admin1 regions that were defined in the colonial era, the boundaries of which have mostly been reproduced intact since then (sometimes with subdivisions). Rural regions in which persistent electoral blocs do not form (which we call the “non-bloc” regions), by contrast, tend to be characterized by lower levels of economic development, lower educational levels, lower population densities, and the lack of sectoral economic specialization. The evidence is consistent with the argument that state institutions work to channel politics arising from uneven economic development into the national political arena.
Section 4.7 of Chapter 4 identifies multiple micro- and mezo-level mechanisms that contribute to this outcome. They are related to interests, organizations, ideology, and the actions of political agents and coalition-builders. Cross-nationally, the observed regional bloc voting is not systematically related to the geographic distribution of ethnic identities. Blocs emerge in countries wherein ethnicity has low political salience (Mali, Tanzania), as well as in high ethnic salience countries like Kenya. Almost all persistent electoral blocs are multiethnic, even in Kenya where politics is highly ethnicized. Many of the blocs are nonethnic, and/or explicitly regional (e.g., Northern Nigeria). Conversely, the non-bloc areas of all countries are home to multi-constituency ethnic groups that do not present as persistent electoral blocs.
Chapter 5, “Bloc Hierarchies and National Winning Coalitions,” argues that regional economic hierarchies are strongly reflected in the structure of electoral dominance in national political arenas. Using the twelve countries examined in the preceding chapter, I show that the within-country distribution of spatial inequality produces hierarchies of electoral blocs and non-blocs, structuring political dynamics at the national level. In most countries, economically dominant blocs tend to dominate politically and to forge national coalitions with constituencies in non-bloc areas to win presidential elections. I argue that these patterns are isomorphic to patterns of coalition-building between economically leading and lagging regions in other parts of the world, and that they tend to reproduce regionalism in national politics.
Chapter 6 identifies patterns of bloc polarization, or “territorial oppositions” in national politics. I argue that axes of territorial cleavage arising between predominantly rural regions tend to take canonical forms associated with core–periphery politics in countries that are undergoing national economic integration and the growth of the central state. This chapter argues that stable axes of sectional competition, whereby leading regions square off against each other or against the periphery, are indeed visible in the electoral data and in persistent policy cleavages in countries in this study. In broad outlines, these often conform to models of territorial oppositions in national politics advanced by earlier scholars (Lipset & Rokkan Reference Lipset, Rokkan, Martin Lipset and Rokkan1967; Gourevitch Reference Gourevitch1979; Bayart Reference Bayart2013). Four countries – Kenya, Zambia, Malawi, and Uganda – serve as archetypes to show this.
Chapter 7 argues that regional competition in African countries finds expression in tensions, debates, and competition over policy. I argue that regional economic tensions tend to find expression in four persistently salient issue areas: (a) redistributive social policy, (b) regionally specific [sectorally specific] investment and regulatory policy, (c) land policy, where we see redistributive tensions and conflicts that arise in the building of national land markets, and (d) issues around state structure (the territorial division of powers and prerogatives). In most countries, regional cleavages trump class-like or interpersonal income inequalities as a driver of national-level contestation over issues of policy and collective choice. South Africa, where regional inequality is lower and nationalizing institutions are stronger, is an outlier: Redistributive social policy is more developed than it is anywhere else in Africa, and the issue of national land market integration is far less salient than in many African countries.Footnote 27
The conclusion argues that a theory of regional cleavages embeds the behavioral assumptions of microlevel, ethnicity-focused models in a broader spatial, institutional, and temporal frame. A national-level, theoretically grounded framework built on economic geography and institutions produces general findings about political dynamics in African countries that are close to what classical and mainstream treatments in the CP and CPE literatures would lead us to expect. Many debates within the literature on African politics can be refined or transcended by more systematic consideration of the dynamics of political scale, with the regional scale providing a theoretical and conceptual map that enables scholars to identify cleavage structures in national politics.
The challenges of territorial politics in regionally divided countries that confront most African countries today are increasingly prominent in non-African countries, including the postindustrial countries of the West. The United States and the United Kingdom are striking examples. This makes research on the politics of spatial inequality in African countries, including the present work, I hope, relevant to general understandings of how economic and spatial inequalities may evolve over time and heighten the challenges of national politics, and to calls for place-based economic development strategies in the name of national cohesion.
