The Millet Mills
In 2016, two local governments in rural Senegal had a total of seventeen millet mills to distribute. The first local government had been granted eleven mills by a national government program.Footnote 1 When choosing from among the many villages in the local government, Mamadou Dia, the former mayor who had brokered the program, carefully designated the mills for the villages of his political supporters and extended family.Footnote 2 Because the mills arrived after Dia’s defeat in the 2014 local elections, however, his successor, Abdoulaye Ka, quickly put his own mark on the project. Upon taking office, Ka dissolved the women’s committees that Dia had formed to manage the mills’ profits and recreated them with more politically favorable members. In a political coup for himself, Ka was able to change one mill’s destination to his own supporters because the recipient village had been ambiguously listed in official paperwork. As Dia grumbled to me later, Ka had really “played the situation.”Footnote 3 As a result of these distributional choices, villages with less than 100 residents received mills while women in villages three times that size remained with no choice but to pound their families’ daily millet by hand.
In the second case, the local government purchased six mills and delivered them to six different villages, none of which had a functioning mill previously. Recipients reported that their villages were chosen because their female residents spent endless hours each day pounding millet and because every year their villages paid the entirety of their local taxes. Neighboring villages did not contest the decision or its logic. Rather, despite the fact that the current mayor had won with a much narrower vote share than in the first case, local elites noted with pride that their community possessed a powerful sense of solidarity that guided such distributive decision-making. As the imam of the local government’s capital village mused one evening, “the best politics are the politics where everyone gets their share.”Footnote 4 Across the local government’s many villages, no one had much to say about the affair.
The seventeen mills represent a fraction of the services delivered each year by local governments in rural Senegal. Since 1996, the country’s democratically elected local governments have been tasked with delivering basic social service investments, ranging from building primary schools to organizing yearly youth soccer tournaments to purchasing millet mills. Yet as these illustrations suggest, local governance has produced stark differences in how goods are delivered and, by extension, who benefits from government largess. Why do some local governments deliver goods broadly across their populations while others target even the most basic of services to meet narrow political ends?
The answer to this question does not only concern Senegal. Similar scenes unfold across Africa – and the Global South more broadly – every year. Over the past quarter century, donors have urged most sub-Saharan African states to adopt similar governance reforms, rendering decentralization among “the most significant facets of state restructuring in Africa since independence.”Footnote 5 As an institutional reform, decentralization has become the bearer of most good things desired by the international development community: it is argued to promote democracy and political participation, to increase the voices of women and minorities, and to improve efficiency and equity in basic service access while stimulating bottom-up development and economic growth in the process.Footnote 6
The varied fates of the seventeen millet mills should give advocates of decentralization pause. The example of the first local government suggests clearly that the decentralization of public goods delivery leaves basic investments vulnerable to local political pressures. After all, local officials spoke frankly about using the mills for their own political ambitions; when asked how he decided which villages would receive the eleven mills he had to distribute, the former mayor Dia responded clearly “I chose the villages that are with me.”Footnote 7 Dia’s candor calls into question the promise that decentralization would help enact the preferences of the local community, rendering government more accountable and responsive.Footnote 8 But such cases risk obscuring the second case, where the local government more or less did what scholars and donors would have hoped: they allocated scarce goods according to equity and need.
Herein lies the first of two questions motivating this book: why are some communities able to come together to improve their collective lot while others are not? I examine the distributional political logics adopted by local governments to advance a novel answer to this question: local decision-makers are better able to cooperate around basic service delivery when their formal jurisdictional boundaries overlap with informal social institutions, by which I mean norms of appropriate comportment in the public sphere demarcated by group boundaries. I refer to this spatial overlap as institutional congruence. When local government boundaries pool together villages that share preexisting social institutions, the preferences of local elites (such as local government officials, village chiefs, and other notables) shift toward group goals, dampening potentially explosive political situations, curtailing opportunism, and enabling elites to adopt longer-run time horizons. This is the dynamic that explains the six mills delivered by the second local government in 2016. The most significant lesson from the story of the seventeen millet mills is not one that fulfills our more pessimistic expectations about distributional politics therefore, but rather in the comparative banality of cases where such dynamics are absent.
If the social institutions that I argue are critical for reorienting elite interests are not evenly distributed across space, then why do some local governments possess robust cross-village norms while others do not? This is the book’s second motivating question. If we want to understand why elites in the second local government did not pursue their own private gain, we must situate it in a much longer history than the tale of the millet mills alone. This local government falls in the heart of the precolonial state of Cayor, where petty disputes over a few millet mills cannot be allowed to undermine centuries of cohabitation and cooperation between villages. In areas that fall in the footprints of the country’s precolonial polities, elite incentives are more prosocial toward their neighboring villages because they inherited robust cross-village social institutions that are repurposed and redeployed following institutional reform.
This book thus argues that historical structures of political order can remain politically consequential long after the formal institutions themselves have disappeared. Even if the precolonial past does not at first glance appear to have much of anything to do with contemporary local governance, the following chapters document how the advent of decentralization in the 1990s granted new venues for much older social repertoires, forcefully bringing the weight of the past into the sphere of formal politics. In brief, this book is about such differences – between areas defined by institutional congruence, where local elite comportment is constrained by shared social institutions, and communities where such congruence is absent, between areas that were home to precolonial kingdoms and those that were not. These differences, I argue, drive profound variation in the politics of representation and redistribution at the grassroots in rural West Africa today.
Redistributive Dilemmas in Rural Africa
African governments have made dramatic investments in basic social services since the turn of the millennium. Even in the wake of the 2008 financial crisis, social spending increased countercyclically in many countries.Footnote 9 For its part, Senegal increased education spending from 3.2 percent of its gross domestic product in 2000, to 7.1 percent by 2015.Footnote 10 The material outcomes of these budgetary changes are highly visible as new clinics, roads, and schools pop up across the landscape.
The academic consequence of this expansion in public goods delivery has been a resurgent interest in distributive politics on the continent. Though research has long recognized the political calculations behind public policy and redistribution, publications on service delivery on the continent have increased nearly tenfold since the early 1990s.Footnote 11 The vast majority of this work, however, focuses on the political logics of the central state, meaning that we possess relatively less theoretical and empirical insight into local redistributive decision-making.Footnote 12 In general, local governance in Africa remains understudied compared to the experiences of other regions.Footnote 13 This is a notable oversight given the fact that African citizens report their most frequent government contact at the local level, echoing Wibbels’ (Reference Baldwin, Raffler, Rodden and Wibbels2019, 15) observation that in much of the Global South local governments “are often the only governments that materially impact the lives of citizens.”Footnote 14
More insight into the nature of local distributional politics might be found in the proliferation of work on decentralization over the past quarter century. Yet this body of work tends to fall into a dichotomy between an overly macro-focus on the institutional environment of reform on the one hand and a micro-focus on community-specific outcomes on the other. The former focuses on failures of institutional design where the animating assumption is that the benefits of decentralization will be unleashed once the institutions are “right.”Footnote 15 An alternative version of this macro-perspective emphasizes the politics surrounding state implementation of decentralization reforms, rendering differences in local government performance a function of the political ambitions of the central government, be it manifested through politicized implementation, uneven fiscal transfers, or other forms of favoritism.Footnote 16
At the micro level, decentralization has largely been studied as a question of elite capture.Footnote 17 This literature, which focuses on how elites “control, shape, or manipulate decision-making processes or institutions in ways that serve their self-interests and priorities,” suggests that when power is concentrated in the hands of a few, elites will collude, using their role as intermediaries to advance their own interests and divert resources to themselves.Footnote 18 In this book, I conceptualize local elites as holders of local social status within their villages, the base social unit in rural Senegal. This definition captures both customary authorities, such as village chiefs, and lineage-based status hierarchies, such as those holding notable or high-caste status, as well as religious leaders and individuals who obtain status via their association with the state, such as elected officials. Theories of elite capture have generated substantial pessimism about the fate of local redistribution, but they offer little theoretical leverage on why these dynamics would be more active in one local government, but not another.Footnote 19 If local elites systematically pursue their own interests at the expense of the collective good, then absent institutional safeguards, decentralization should produce equally poor development outcomes everywhere.
Taken together, neither the recent literature on distributive politics nor work on decentralization easily explains the varying distributional decisions adopted by Senegal’s local governments that I document in the following chapters. As I show, we see variation in local government performance within a shared macro-institutional environment, indicating that we must look beyond questions of institutional design. My research also reveals that local political cleavages only rarely parallel national ones and are more clearly shaped by deep-seated social histories at the grassroots than the nature of political connections to the center, suggesting that distributional politics follow distinctly local political dynamics and are not reducible to the ambitions of the central government.
At the same time, my findings challenge the idea that local governance is reducible to the question of when local elites are able to control local positions of authority because this ignores a prior question of when local elites perceive themselves to have shared interests. Recent scholarly and policy-oriented attention to elites has often focused on identifying specific actors capable of moderating interventions or, alternatively, those whose authority must be checked.Footnote 20 I seek to trouble this emerging dichotomy of local elites as either a peril or a promise for development and democratization by focusing on the nature of intra-elite relationships as the variable of interest. By emphasizing elites as self-interested individuals, we have lost sight of the ways in which elite authority is interdependent and imbricated in broader social webs that are as, if not more, consequential than individual motivations.Footnote 21 In other words, even if elites may be able to capture local government positions everywhere, I do not predict this to generate universally good or bad outcomes because what matters is not their positions as elites alone but the broader social networks that shape the social and political strategies available to them.
In this way, I contribute to our understanding of local governance by drawing attention to the political negotiations that take place between local elites, shifting the unit of analysis downward from macro-level work focused on the central state and up from the micro-focus of work on elite capture, to look squarely at the local state itself. Because democratic decentralization pools numerous villages into one administrative unit, local elected officials are obliged to negotiate across villages when deciding how to distribute scarce projects and resources. Certainly, local elites seek to maintain and reinforce their social status within their villages, pursuing local government office in order to capture state resources as one means of doing this. Nonetheless, when local elites meet within the local state, we can neither assume that they share a common interest to collude nor that they are inherently rivalrous. Rather, their political incentives emerge from their social worlds and are intimately influenced by historical legacies that long pre-date the modern advent of decentralized governance. It is only by recognizing the social underpinnings of local political life, that I argue we can understand distributional politics following decentralization.
The Argument
I advance a theory of institutional congruence to explain subnational variation in local government performance following decentralization. Specifically, I posit that local redistributive politics is a function of the degree of spatial overlap between the formal institutional extent of the local state, created and reformed from above, and local social institutions, by which I mean norms of appropriate behavior in the public sphere demarcated by group boundaries, inherited from the precolonial past. Despite the fact that West Africa’s precolonial kingdoms are long defunct, I show that they left a legacy of robust, cross-village social institutions that continue to define local social life. When social institutions embed the vast majority of local elites from across a local government’s many villages, elites are more likely to factor the interests of others into their own preferences. This is true even though today’s local government boundaries are much smaller than the territorial extent of precolonial kingdoms; institutional congruence does not necessitate the perfect overlap of physical boundaries as long as local elites are mutually embedded within shared social institutions. Though comparable social institutions exist in historically acephalous zones, or those that lacked a centralized precolonial political authority, they remain delineated by village or clan boundaries and do not map onto the arena of the local state.
Why would such an overlap alter elite behavior? In cases of high congruence, the presence of shared social institutions encourages local actors to engage in more prosocial behavior toward group members. Here, a localized sense of community elongates time horizons and leads local elites to “encapsulate” the interests of their neighbors, enabling long-term cooperation.Footnote 22 In contrast, elites’ political and social worlds are discordant in cases of low congruence, minimizing the risks that local elites face when they pursue individual interests and diminishing any rewards they may receive from acting with an eye toward the collective good.Footnote 23 This generates the expectation that representation and redistribution under decentralization will be broad and equalizing in cases of high congruence, but contentious and targeted in areas where it is low.
Social institutions shape the decision-making of local elites through two complementary mechanisms. The first mechanism is the presence of a shared social identity endowed by a precolonial state that ties local elites into a sense of shared fate. When actors invoke a shared basis of social identification, they delineate in- and out-group members, reinforcing the meaning of their mutual membership through shared narratives. As I document, the precolonial past forms a powerful identity that local actors can lay shared claim to, even in the face of other forms of diversity such as religion and ethnicity. The second mechanism arises from the social network ties among local elites. Communities that fall in the footprint of precolonial polities exhibit denser social ties across the many villages of the local state than in incongruent communities both because generations of cohabitation have fostered enduring ties around intermarriage and intervillage assistance, and because the ability to claim descent from the precolonial polity, no matter how romanticized, generates a mutually reinforcing justification for elite status across villages. Denser social network ties not only valorize shared social identities and social institutions, they also circulate information about actor behavior, enabling rewards and sanctions for those who uphold or transgress group norms.
Together, the identity and network mechanisms help explain both why social institutions have persisted to the present and why they continue to influence elite behavior: broad social identities create focal points for local action while dense social networks generate incentives to abide by locally understood social norms of behavior. Like all informal institutions, therefore, social institutions shape behavior through sanctions and rewards. What makes social institutions distinct is the degree to which actors internalize these expectations for both their own social comportment and that of their social relations.
In this way, I present a novel explanation for why local governments make strikingly different distributional choices when delivering basic public goods. Local elected officials face structurally distinct political and social incentives in cases like the second community with which this book opened, where the boundaries of the local state fall within the territory of the precolonial state of Cayor. Here, cross-village social institutions enable intra-elite cooperation by disincentivizing political conflict and fostering a sense of “we-ness.”Footnote 24 Where such cross-village institutions are absent, however, as seen in the first case, politics pervades even the most basic of local government investments, skewing allocative decisions away from the neediest to the most politically useful or politically expedient.
The argument is not that areas home to precolonial states somehow produce more benevolent leaders, but rather that elites in these areas are particularly sensitive to perceived violations of local norms because they face distinct social risks in addition to those they face at the ballot box. In this way, social institutions simultaneously impose standards for political behavior and create broader webs of obligation for local elites, rendering the local politics of representation and redistribution endogenous to local social structure. Though shared identities and cross-village social networks are not necessarily unique to areas that were home to precolonial states, the presence of both mechanisms in these zones jointly reinforces the value of social institutions to elites from across a local government’s many villages with powerful effects on actor behavior.
Why would the precolonial past generate such persistent social legacies? I locate the continued relevance of the precolonial era in the durability of rural social hierarchies throughout the West African Sahel, where social status is largely shaped by one’s family lineage. Despite the onslaught of colonialism, the French colonial state stopped short of fundamentally altering village social hierarchies, meaning that most village chieftaincies and other markers of social status within villages have persisted by and large intact within families. Because the ability to claim descent from a zone’s founding families bestows particular social and political authority, local elites repeat and valorize local historical narratives justifying their right to local authority as a means to legitimate and shore up their own status. Precolonial identities are a particularly potent means to do so, but by invoking this historical right to authority, elites simultaneously validate the status of others making identical claims. The very act of claiming social status in these zones is therefore collective, linking local elites into broader cross-village networks of reciprocal recognition of elite status.
Social institutions that develop over generations of cohabitation thus come to mutually embed local elites in their behavioral stipulations by virtue of their ties to local status hierarchies. Yet in contrast to other prominent arguments about historical legacies, my theory stipulates that we should not expect the precolonial past to always matter for redistributive politics even if these social institutions are reproduced over time. Rather, it is only following decentralization reforms, when allocative authority is transferred downward, that the dynamics of institutional congruence should arise because it is only at this moment that social institutions map onto the scale at which distributional decisions are made.
Points of Departure
My argument holds theoretical and empirical implications for four prominent debates in the study of the African state and political economy of development more broadly. In order, I detail below how the argument speaks to (a) the value of local narratives as a lens into actors’ political strategies; (b) the social identities we prioritize in the study of African politics; (c) prospects for state-building on the continent; and (d) our understanding of how historical legacies shape contemporary development outcomes.
Local Narratives, Local Political Strategies
This book joins a growing interest in redistributive politics in developing countries. The motivating assumption of much of this work is that politicians are instrumental and vote-maximizing leading studies of service delivery in Africa to view public goods delivery as a function of the electoral calculi of politicians who target supporters along partisan or ethnic lines in an effort to earn votes.Footnote 25 I do not dispute the assumption that politicians pursue political power above all else. Like all politicians, the local elected officials I speak with in rural West Africa are sensitive to the demands of political competition. Rural politicians almost always run for reelection and cite delivering goods to voters as the way to get reelected. As one mayor cautioned, “never forget that tomorrow they will have to reelect you.”Footnote 26
But focusing on short-term electoral motives alone risks leaving critical dimensions of local political behavior unexplained. During interviews, local politicians in rural Senegal voice broader and more interdependent preferences toward redistribution than those suggested by dominant theories of distributive politics. For example, they frequently detail at great length how they take into account prominent family lineages and social relations, balancing political decisions, such as who gets their name on electoral lists, across villages. One mayor in Fatick Region stressed the role of “social alliances” in his community, which he could not ignore because while such social ties may not always manifest themselves electorally, they were critical for local political success. In this case, the mayor’s social relations and the obligations they engendered had led him to invest heavily in villages that had not voted for him and which he did not expect to do so in the future.Footnote 27
While it is easy to dismiss these claims as face-saving stories told to an outsider, I suggest we should take them seriously as a lens into how local actors understand and interpret their political worlds. Listening to how local elites root their explanations of their political lives in the specific histories of their community helps us understand the beliefs they bring with them to their social and political interactions because such collective memory sheds unique light on the “history of play” between individuals.Footnote 28 To this end, the rhetoric that individuals deploy when explaining their behavior is a particularly powerful vantage point into the collective beliefs and norms that they use to formulate and justify their actions.Footnote 29 Narratives of local political life thus reveal what individuals view as “ordinary and right,” generating a valuable form of data for understanding local action.Footnote 30 That the very telling and retelling of local narratives should be studied as an explicitly political act has recently been argued by Klaus (Reference Klaus2020), who documents the ways in which narratives surrounding land ownership in rural Kenya delineate in- and out-group members to create or reaffirm the cultural meaning of land within communities. I suggest that a similar process is at play in rural Senegal: when local elites narrate their local political worlds, they reveal and reinforce their understandings of community membership and the rights and obligations this entails.Footnote 31
Ultimately, it is less critical that the narratives I recount in the following chapters are strictly historically accurate because I am more interested in how actors believe their social worlds operate than in the absolute truth. Though I take care to cross-validate the degree to which narratives of local history and political life are collective across a local government and though I draw on multiple methodologies to test the ability of these narratives to explain redistributive patterns, I am indifferent to whether or not they document a community’s “true” history. To the extent that the actors I interviewed recount their communities’ histories from their social worlds, they shape how local social and political objectives are understood and, by extension, the appropriate policy to pursue them.
Accordingly, I theorize how elites pursue power through their social relations. When local elites share collective narratives about their community, they also share norms about appropriate political comportment, including not only what is expected of them but also what others will tolerate and what will enhance their reputation and their social and political influence. How elites negotiate, in other words, varies as a function of the social institutions they are mutually embedded in and not their narrow political interests alone. This is a less radical reimagining of political strategies that it may at first imply, but it does nuance considerably the assumption that elites can pursue purely electoral strategies. The reality of local politics in the developing world is that many rural elites live within the communities they serve, engaging in iterated, daily interactions with one another that extend their time horizons.Footnote 32 The effects of this become clear in the narratives of local political life heard in many Senegalese communities where key political actors – representatives, patrons, chiefs, and brokers alike – speak openly about how their social and political lives are entangled, limiting their ability to pursue individual or electoral gains alone explicitly because their political strategies are shaped by a broader set of political, economic, and social interests.Footnote 33 Listening to local narratives, I argue, offers an unparalleled lens into these dynamics.
Rethinking Forms of Identity Politics in Africa
By studying how local elites understand and negotiate their political worlds from the ground up, I am able to make two interventions in a long-standing interest in identity politics in Africa. I begin by pushing back at the assumption that ethnicity is the most obvious political cleavage for mobilizing political support in the region. Ethnicity has been granted a particular privilege by students of African politics, where it is viewed as a core source of political preferences and where ethnic ties are considered the prevalent channels of clientelism and service delivery.Footnote 34 From the advent of ethnic or hometown associations in the colonial era to elected politicians in the present, patrons with access to the state are conceptualized as conduits for services and privileges that are channeled downward to co-ethnic clients.Footnote 35 In the aggregate, ethnic diversity is associated with slower economic growth and worse development outcomes.Footnote 36 Indeed, the benefit of ethnic homogeneity has been so widely accepted that much of the ensuing debate focuses on adjudicating between potential mechanisms.Footnote 37
The influence of this literature is seen clearly in the recent explosion of research on service delivery in the region, which takes ethnic favoritism as a default hypothesis. But while African politicians are frequently assumed to find the most viable path to electoral victory in ethnic coalition building, the evidence for this tends to be strongest in Anglophone cases and some scholars have suggested that co-ethnicity may not yield the benefits that have long been ascribed to it.Footnote 38 Critically, even if ethnicity is a key political cleavage nationally, it may not operate similarly at a local level, a reality Nugent (Reference Nugent2010, 50) identifies in The Gambia, where ethnicity’s “salience remains weak in the internal politics of most rural communities” despite being politicized nationally.Footnote 39
In questioning the dominance of ethnicity in the study of African politics, I do not deny that ethnicity is a relevant social category for Senegalese. Individuals describe themselves along ethnic lines and they invoke ethnic stereotypes and categories, but it has never mapped neatly onto national political cleavages.Footnote 40 In areas home to precolonial states, ethnicity is rarely mentioned by respondents other than in simple descriptive terms, though such claims do come out strongly in many historically acephalous, or stateless, areas of the country. In the following chapters, I test the ethnicity hypothesis in numerous ways, but like others before me I find little evidence that ethnicity is a useful predictor of Senegalese political cleavages or distributive politics in the country.Footnote 41 Instead, place-specific identities dominate local political life. Among these, a sense of descent from the precolonial past holds special currency, but throughout the country local histories of settlement order local claims to political resources. These identities may intersect with ethnicity, but they are not reducible to it.
This leads me to make a second, more novel intervention about the prevailing tendency to deduce politically relevant identities from observed political behavior.Footnote 42 Despite growing recognition that political identities rooted in the past can be consequential decades if not centuries later, I challenge the assumption that they are always politically relevant.Footnote 43 My argument, which documents how social identities can persist at the grassroots with minimal political effect only to forcefully reenter political life decades later, offers a framework for understanding why identities may generate distinct behavior at some moments of time but not at others.Footnote 44 We analytically limit ourselves when studying institutional change when we assume that identities are exogenous or static variables. In contrast, it is exactly in these moments that we might expect shifts in which identities individuals prioritize.Footnote 45 As I show, previously latent identities can become politically relevant once they become congruent with newly introduced formal institutions, revealing why the spatial interaction between identities and institutional boundaries deserves our explicit attention.
By taking seriously the social identities that rural Senegalese themselves utilize in conversation, I reinforce Cramer Walsh’s argument that researchers should not ignore how individuals describe their own political identities since these are “touchstones for individuals’ understandings of the political world” (Reference Cramer Walsh2003, 176). If, as I argue, old identities can be politicized in the relatively short term following institutional reform, then assuming that rural Senegalese see themselves in ethnic terms, say as Wolofs or Serers, rather than as descendants of the Kingdom of Saloum, means we risk never understanding local politics at all. In this way, my findings caution against the tendency to discuss what we can most easily measure at the risk of ignoring consequential identities that are less amenable to quantification.
Revisiting the Question of State-Building in Rural Africa
Scholars have long diagnosed the ills of the African state. It is either too heavy-handed, suppressing a dynamic society or conversely it is too embedded within society itself, propping up its weak legitimacy by reinforcing blurred boundaries between the public and private realms.Footnote 46 Among our most foundational questions about the African state is why African leaders have been unable or unwilling to expand the authority of the state outward across their territories. Why, nearly three generations of scholars have asked, have we failed to see the emergence of political bargains between the center and its citizens?
Recall that decentralization was proposed by many as a means to address these various “pathologies” by encouraging social-contracting. Accordingly, my argument generates a distinct set of implications for the fate of state-building on the continent. I present empirical evidence that precolonial histories position some communities to do better under decentralization than others in a way that is independent of institutional design or the political ambitions of the center. This means that forces often conceptualized as impediments to the state-building project – precolonial, indigenous social organizations – can actually entrench the state locally when the local state is sufficiently able to graft onto a preconceived understanding of community.
The more radical implication is that the very reforms intended to “fix” the African state risk heightening inequalities in state strength as some communities attach themselves to the local state, slowly building capacity and commitment to the democratic project, while elsewhere others disengage from local governments which they see as unresponsive and foreign. The argument holds mixed lessons, therefore, for long-standing questions about the relationship between state and society in the region. The state is ultimately stronger, I find, when it is embedded within or legitimated by society, but only in particular spatial constellations. Like Boone (Reference Boone2003b) or Herbst (Reference Herbst2000) before me, I document great territorial unevenness driven by structural, historical legacies, but I root the source of this variation in purely local dynamics.
Critically, the production of these inequalities is ongoing, which would indicate that even if we hold the central state’s relative strength or weakness constant in the short term, organic properties of social structures at the grassroots are capable of generating variable degrees of state-building subnationally. This refocuses our attention to the local level even when we want to understand the center because state-building ultimately takes place at a granular level within communities. In lieu of studying how the state looks downward, I emphasize how local actors negotiate reforms initiated by the center, integrating them into their own political worlds. This, in turn, has implications for national outcomes, from the nature of party strength to the distribution of economic opportunity.
Decompressing Historical Legacies
In recent years, it has become near gospel that “history matters” in the study of why contemporary development outcomes vary across and within countries.Footnote 47 Our renewed interest in historical legacies assumes a distinct temporality: the past influences future sequences of events by raising the costs of deviating from a given pathway through the development of routines, beliefs, or networks, to name but a few common causes.Footnote 48 Yet despite the embrace of path dependence by students of political economy, the study of institutional legacies has become surprisingly apolitical and, in its own way, atemporal. This project offers a corrective on both fronts.
I illustrate this by referencing the most relevant debate to my own argument: the growing interest in the long-term legacies of precolonial centralization. This literature is dominated by a series of cross- and subnational studies that demonstrate how early experiences with state formation correlate with a number of contemporary outcomes, including greater economic growth, rule of law, and tendencies toward authoritarianism.Footnote 49 Much of this work relies on the assumption that a historical cause, once unleashed, persists to the present in a uniform fashion; for example, that group-specific attributes, like ethnic norms, explain why historically centralized regions perform better over the long term than others.Footnote 50 Michalopoulos and Papaioannou’s (Reference Michalopoulos and Papaioannou2015) prominent paper concludes in this vein that the robust correlation between precolonial centralization and better public goods access today is the result of “ethnic-specific attributes,” which they argue raises an important question as to how “ethnic institutional and cultural traits shape economic performance.” Arguments of this nature are not inherently at odds with my own: the persistence of cultural attributes could plausibly explain why some local governments perform better under decentralization than others. Problematically, however, these arguments assume that group attributes are homogenous across space, equally internalized by members, and persistent across time, and empirical claims that I test and find no evidence for in Chapters 4 and 6.
This literature bolsters my contention that precolonial history continues to matter in the present, but the explanations it offers stop short of helping us understand the “causal channels” through which historical factors course.Footnote 51 By estimating a cause and effect with decades, if not centuries, in between, scholars “compress” history and lose insight into how the temporal evolution of our independent and dependent variables alike indicates more and less plausible hypotheses.Footnote 52 One consequence is that historical legacies are only rarely tested against each other; yet as I show in Chapter 7, different legacies may face distinct prospects for persistence. Despite the powerful influence of the colonial state on subnational public goods access in the twentieth century, for instance, this effect has largely faded by the twenty-first.
This is particularly striking because we know that institutions, which are never distributionally neutral, are inherently subject to political contestation, rendering historical interactions all the more consequential, both empirically and theoretically.Footnote 53 Apart from Mahoney’s (Reference Mahoney2010) study of colonialism in Latin America, which argues that the nature of precolonial political organization conditioned the forms of extractive regimes implemented by Spanish colonizers, much of this literature is strangely silent on the question of power at all. Like Mahoney, I focus on how the past generates distinct power dynamics over time, culminating in local elites’ deployment of narratives of the precolonial past that I document in subsequent chapters. I go on to “decompress” history in order to examine empirically whether precolonial legacies have always impacted service delivery, only to find no evidence of institutional congruence until the exogenous shock of Senegal’s 1996 reforms that devolved distributional authority to the local level. This demands a theory that is both rooted in precolonial political geography and that is logically consistent with the precolonial past only impacting the local level and only under certain formal institutional configurations. By looking at when the precolonial past can explain distributional patterns, I offer an explicitly temporal story: the long shadow of the precolonial past on contemporary development outcomes is at best a story of intermittent effects.
Empirical Strategy and Overview of the Book
The remainder of this book elaborates upon my theoretical and empirical claims that the precolonial past intimately shapes the nature of contemporary local governance in rural West Africa. In the following chapters, I employ a comparative, subnational analysis of one country’s experiences with formal institutional reform, leveraging the uneven geography of precolonial statehood to examine variation in my independent variable, the degree of institutional congruence following decentralization. Subnational research designs such as my own are particularly well-suited for uncovering the divergent consequences of political and economic reforms because they allow us to hold constant the nature of reform (in this case decentralization) as well as concerns about any number of unobservable characteristics that are expected to differ across but not within countries.Footnote 54 In the process, I am able to shed light on the “humanly important” variation that is unleashed by institutional reform, while avoiding the risk of theory-stretching by not assuming a priori that theories developed at the national level will translate neatly downward to subnational units.Footnote 55 In contrast, I demonstrate that where a local government falls spatially within a country offers more leverage than prominent theories of public goods delivery developed at higher levels of governance, which prove to be poorly suited to explain the variation I document in the following chapters.
The risks of this strategy are twofold. First, there is the question of cumulation and generalizability.Footnote 56 My findings might tell us a lot about local governance in Senegal, but is the impact of precolonial centralization limited to Senegal alone? In seeking to keep national-level institutions constant within one case, I risk inadvertently selecting a country where something inherently unique is happening. To address this concern, I expand my dataset to look at all of West Africa in Chapter 8, where I find consistent and robust results across the subregion. Second, subnational studies can pose a trade-off between gains in internal validity and our ability to ask enduring questions in the field.Footnote 57 But this is not necessarily so. The big, meaty questions that have long defined the study of comparative politics, such as order, democracy, or inequality, are alive and well in the chapters of this book. As I document, only by taking seriously local dynamics of redistribution can we see the important subnational variation in development and democratic consolidation emerging across West Africa following the introduction of decentralization.
Why Senegal?
Senegal is not the only African state whose territory was home to a heterogenous political landscape in the precolonial era. Yet Senegal provides a unique opportunity to explore the long-run legacies of precolonial statehood on contemporary distributive politics because it offers clear leverage on both the independent and dependent variables. First, Senegal was home to a dynamic microstate system prior to French colonization. On the eve of the final French push to conquer Senegal in the early 1880s, slightly under half of Senegal’s territory was under the control of a centralized political organization. Introduced in more detail in Chapter 2, the West African state system was populated by polities that were far more state-like than is often assumed, capable of enforcing property rights and of adapting to the changing whims of capitalist markets.Footnote 58 Importantly, Senegal’s precolonial states were smaller and less capable of capturing the colonial or postcolonial political arena, in contrast to well-studied precolonial heavyweights, such as the Ashanti in Ghana or Buganda in Uganda.
Senegal also offers clean leverage on the central dependent variable - the nature of local social service delivery – as well as the question of representation in the local state, the secondary outcome of interest. Though Senegal began its decentralization process in 1972, earlier than many in the region, it was only in 1996 that local government councils became fully elected by the popular vote and gained the right to independently allocate public services, most notably the placement of primary education and health facilities.Footnote 59 As a result, studying Senegal allows me to isolate the local logics of social service delivery that came with decentralization reforms in 1996 from those emanating from the central state in years prior. In essence, the 1996 reforms serve as an exogenous shock, revealing how local political dynamics are conditioned on antecedent conditions rooted in the precolonial past because, as will be shown, the effect of precolonial centralization only emerges when local actors are charged with making distributional decisions.
If anything, Senegal is a hard test for the argument that precolonial legacies persist to the present. Unlike the dominant outcome in British colonies, the French colonial administration’s policies of direct rule resulted in more consistent dismantling of precolonial hierarchies and chieftaincies. This means that any legacies of the precolonial past have to persist through some informal mechanism not tied to the formal institutional structure of the state. Moreover, the fact that Senegal’s decentralization project was top-down should bias against the risk that precolonial identities were instrumentally revived by politicians for the sake of obtaining subnational autonomy, such as has been seen with the creation of new states in Nigeria.Footnote 60 Together, these two factors suggest that the persistence of grassroots identities was far from overdetermined.
In choosing to investigate the question of how decentralization reforms interact with precolonial legacies in Senegal, I join a long tradition of looking to the country’s dynamic spatial variation in political form. The iconic study for political scientists is Boone’s (Reference Boone2003b) Political Topographies of the African State, where Senegal serves as a key case in Boone’s argument that state-building efforts are shaped by the relationship between the center and the periphery. Whether rural areas become allies or rivals for the center is determined for Boone by two axes: the presence or absence of social hierarchy, which grants rural areas bargaining power, and the degree of economic autonomy, which generates or suppresses incentives to collaborate with the state. Recent work has expanded these insights to examine variation in the ability of local Senegalese leaders to serve as brokers under the long-dominant Socialist Party (Parti socialiste or PS) as well as under the Alliance for the Republic (Alliance pour la république or APR), in power at the time of writing.Footnote 61 More recently, Honig (Reference Honig2017) suggests that the value of hierarchical customary authority extends to land markets as well, with chiefs’ resistance to or facilitation of state efforts to control authority over land generating subnational variation in state-building.
Collectively, however, this scholarship has kept its focus on relations with the center. I shift focus to look solely within local governments, revealing divergent outcomes in otherwise quite similar local governments. For example, Beck (Reference Beck2008) argues persuasively that Senegal’s Mouride Brotherhood, an influential Sufi, Islamic sect, possesses both high political and social autonomy, making them particularly influential political brokers with the central government. I show in the following chapters that predominantly Mouride communities that fall within the bounds of a precolonial state nonetheless engage in distinct redistributive strategies following decentralization compared to their counterparts in historically acephalous areas. One implication of my study for our understanding of Senegalese politics, therefore, is that the social structures that are consequential for center-periphery relations do not automatically correspond to those that are critical for determining the nature of political life at the local level.
Overview of the Book
I develop and test my theory through a nested empirical analysis that structures the following chapters. Specifically, I combine extensive qualitative data with quantitative tests of my theoretical predictions in order to iteratively develop, test, and refine my argument. I begin by detailing my theory of institutional congruence in Chapter 1. I argue that persistent forms of social cooperation at the grassroots are revitalized following institutional reform where communities have inherited robust social institutions stipulating appropriate social behavior from a precolonial polity. I elaborate on the theory’s dual mechanisms of shared social identification and dense cross-village network ties to illustrate how institutional congruence helps local elites navigate a two-level political game introduced by decentralization: local elected officials face pressure within their villages on a first level that may be more or less compatible with their incentives at the second level of the local state itself, where they must negotiate with other elites from other villages. My theory of institutional congruence offers leverage on how local elites resolve this unique redistributive dilemma by arguing that shared social institutions stretch across the many villages of a local government, elites find it easier to negotiate at the second level of the local state because these social institutions reorient them toward group-based goals.
Chapters 2 and 3 lay out three historical building blocks that are critical for my theoretical argument. Chapter 2 presents the core antecedent condition under study. Here I introduce my measurement of precolonial statehood and specify why these states left enduring legacies even in the face of substantial upheaval during the colonial encounter by detailing the mechanisms of persistence: the nature of village-based social hierarchies. Chapter 3 goes on to introduce Senegal’s decentralization reforms, the second historical condition for the argument, in depth, before turning to the third and final historical building block, the delimitation of Senegal’s subnational boundaries during the colonial and postcolonial era. Employing archival and interview data, I demonstrate that decentralization and boundary delimitation were largely top-down processes, suggesting that the emergence of institutional congruence was not driven by endogenous, bottom-up demand.
Subsequent chapters document and explain the empirical variation in local governments’ redistributive strategies through the use of five distinct data sources. Chapter 4 offers initial evidence for my core empirical claim that there is subnational divergence in the nature of local political life following decentralization. I draw on an original survey with more than 350 rural Senegalese political and traditional elites to develop insights into how local decision-makers themselves understand their sociopolitical worlds. The survey reveals remarkable diversity in how political cleavages are articulated across the country, but which nevertheless share structural similarities rooted in relative degrees of institutional congruence. From this data, I deduce the two mechanisms of group identities and social networks to explain how social institutions impact local politics under decentralization.
Chapter 5 employs an original, geocoded dataset of village-level primary education and basic health infrastructure across rural Senegal to test the effect of precolonial centralization on a village’s likelihood of receiving a public goods investment between 2002 and 2012. I find robust evidence that falling within the territory of a precolonial state increases a village’s chance of receiving local infrastructural investments from the local state. This result is robust to a number of alternative explanations and model specifications, affirming the argument that there is something different about how local governments deliver local public goods in formerly centralized areas, even when accounting for similar objective need.
While the quantitative dataset allows me to confirm that we are witnessing the emergence of subnational variation in patterns of public goods delivery, I return to the communities where this book opened in Chapter 6 to present model-testing case studies that help me explore the theory’s mechanisms in the third stage of analysis. Specifically, I follow a “typical” or on-lier case selection strategy from the statistical analysis in Chapter 5 to select cases that are similar in as many respects as possible apart from their exposure to a precolonial polity.Footnote 62 By pairing oral histories, in-depth interviews, and network analysis of local elite social ties, I trace how the presence of a shared social identity and dense network ties shape redistributive preferences in a “typical” case of institutional congruence, while their absence generates more biased forms of redistribution elsewhere. Collectively, Chapters 4 through 6 suggest that the nature of local governance following decentralization varies meaningfully and systematically in line with my theory.
One of the central claims of this study is that the impact of informal social institutions is contingent on the formal institutional environment they operate within. This argument is explored in Chapter 7, which looks at the historical trajectory of basic public goods investments in Senegal from the onset of colonial rule in 1880 to the present. By extending my quantitative dataset backward in time to the onset of French colonial rule through archival data and ministerial reports, I “decompress” history by taking both spatial and temporal processes seriously. In so doing, I find that the impact of precolonial centralization only appears in force following the 1996 decentralization reforms that transferred authority over public goods placement to local governments. Chapter 7 makes two contributions to the broader project: first, it isolates the 1996 decentralization reforms as an exogenous shock that facilitated cross-village social institutions to emerge as a key driver of subnational distributional politics. Second, it offers me the opportunity to disprove the possibility that precolonial legacies interact with or reflect what are ultimately colonial legacies. This fourth stage of analysis shows that the colonial past did matter for access to social services, but that colonial effects have largely faded by the 2000s.
A final empirical chapter assesses the generalizability of my empirical findings. Chapter 8 scales outward from Senegal, using data from the Demographic and Health Surveys (DHS) and the Afrobarometer to establish that we observe similar patterns in public opinion and development outcomes across decentralized West Africa in areas that were home to precolonial states, indicating broad, empirical tractability for the argument. While my theory is built around the specific legacies of precolonial statehood, I secondly move beyond Africa to show the analytic leverage of the theory’s twin mechanisms for Comparative Politics more broadly. The chapter concludes with a discussion of scope conditions for the argument. A conclusion to the book in its entirety follows.
Open access