Ethnic group institutions and electoral clientelism

Abstract

Individual electoral clientelism involves the allocation of handouts to voters around elections. Why is this strategy common in some contexts but not in others? This article demonstrates that ethnic group institutional structure helps to explain this variation. Where ethnic groups are organized hierarchically and have centralized leadership, politicians leverage this infrastructure to mobilize voters wholesale. Where they are not, politicians forge linkages directly with voters, resulting in more electoral clientelism. I provide evidence from a set of African countries, where there is variation in the social structure of ethnic and religious groups. I show that electoral clientelism is more widespread in countries where ethnic groups have a decentralized organization. An individual-level analysis of electoral clientelism in 15 African countries further shows that members of decentralized groups are most likely to receive electoral handouts. The findings contribute to the comparative literature on clientelism and highlight how the organizational structure of intermediaries can shape strategies of clientelism.

Keywords

clientelism elections ethnic politics traditional leaders vote buying

Individual electoral clientelism––often called “vote buying”––is prevalent in low-income countries around the world. Recent debates have centered on how electoral clientelism might be effective where ballot secrecy is protected (Nichter, 2008; Stokes 2005); when political parties will use electoral clientelism to buy votes (Stokes, 2005), buy turnout (Nichter, 2008), or buy voter abstention (Cox and Kousser, 1981; Gans Morse et al., 2014); and the voter targeting strategies of clientelistic parties (Nichter, 2008; Stokes, 2005). This article focuses on a different question: Why is individual electoral clientelism common in some contexts but not in others?

Indeed, individual electoral clientelism is not widespread in all low-income democracies. For example, Ghana and Kenya are countries with similar levels of development, urbanization, and formal political institutions. Yet in Ghana, about 7% of adults received an electoral handout during the 2008 elections, while in Kenya, about 40% received a handout ahead of the 2007 elections.

There also exists substantial variation in the prevalence of individual electoral clientelism, even among countries where others forms of clientelism are prevalent. For example, electoral clientelism is not, relatively speaking, prevalent during political campaigns in Senegal, a country where other forms of clientelism have been documented (Beck, 2008; Boone, 2003). This variation highlights that clientelism, where politicians trade private and particularistic benefits in exchange for political support, can take different forms in different contexts. In some settings, politicians target resources directly to voters at election time. In others, politicians trade benefits with local elites, in exchange for blocs of votes. In others, politicians exchange collective goods with groups of voters. This variation presents a puzzle. Why is individual electoral clientelism prevalent in some environments but not in others? And under what conditions does clientelism take the form of individual rather than collective targeting?

I argue that part of the answer lies in ethnic group institutional structure. Ethnic group institutions are important because they shape the options that politicians have at their disposal to secure votes and mobilize voters. Where ethnic groups are organized hierarchically and have centralized leadership, politicians will leverage this infrastructure to secure votes and mobilize voters wholesale. Where ethnic groups are organized less hierarchically and centralized leadership is not present, by contrast, politicians will be more likely to forge linkages directly with voters, resulting in more individual electoral clientelism. Thus, politicians turn to electoral clientelism in conditions where they lack other more cost effective infrastructure with which to work.

I test this argument with data from a set of African countries, where there is significant across and within-country variation in the organizational structure of ethnic and religious groups. First, I provide cross-national evidence and show that electoral clientelism is more prevalent in countries where ethnic groups lack centralized and hierarchical structure. Second, I code the structure of ethnic groups in 15 African countries and merge that information with data on individual electoral clientelism collected by the Afrobarometer. The data set includes 35,000 survey respondents who are members of 261 ethnic groups. I find that members of less-centralized ethnic groups are substantially more likely to be targeted with electoral clientelism, a relationship that holds after controlling for individual, ethnic group, and country-level variables. Finally, I conduct a more detailed analysis of Senegal and Uganda, two countries with significant within-country variation in ethnic and religious group structure. Together, the results highlight how ethnic group organization helps to determine the extent to which parties invest in individual electoral clientelism.

This article makes several contributions. First, it offers an argument about where individual electoral clientelism should be most prevalent that is distinct from those in the existing literature. Existing research shows that electoral clientelism is common where voters are poor (Brusco et al., 2004; Stokes et al., 2013), ballot secrecy is not protected (Lehoucq, 2007; Stokes, 2005), parties are organized like political machines (Stokes, 2005), and formal institutions incentivize the cultivation of a personal vote (Hicken, 2007). The goal of this article is not to rule out these arguments or to argue that ethnic group institutional structure is the only driver of individual electoral clientelism. The objective is instead to show that ethnic group institutional structure is consequential and should be considered in comparative analysis. Second, the argument builds upon recent research about how ethnic group institutions shape ethnic politics in African contexts (Koter, 2013), developing and extending the logic to individual electoral clientelism. Third, the article contributes to the literature on clientelism and vote brokers or intermediaries (Camp, 2017; Stokes et al., 2013; Szwarcberg, 2012). While this literature generally focuses on political party activists, this article highlights how the institutional structure of ethnic intermediaries can shape strategies of clientelism.

Ethnic group institutional structure and electoral strategies

Individual electoral clientelism involves the allocation of money or gifts to voters at election time (Gans Morse et al., 2014). I use the terms individual electoral clientelism and electoral handouts interchangeably. Individual electoral clientelism is a subtype of the broader phenomena of clientelism, which characterizes relationships of instrumental, reciprocal, and ongoing exchange between actors of unequal social, economic, or political status (Scott, 1969). Individual electoral clientelism can be distinguished from other forms of clientelism that occur outside of the electoral period or do not involve the direct targeting of individual voters.

This article argues that ethnic group institutional structure helps to explain variation in the prevalence of electoral clientelism across and within African states. The starting point for the argument is the observation that African political parties tend to be better organized in urban areas (Bratton, 2008). Yet in democratic elections, parties and politicians must win the votes from an electorate that remains largely rural.

In determining how to mobilize and persuade voters in rural areas, political elites face a landscape that is marked by substantial diversity. The structure of ethnic and religious groups varies greatly across the continent, heterogeneity with roots in the precolonial period. One important dimension on which these groups differ is the extent to which they are organized in a centralized and hierarchical manner.

Some ethnic groups––fragmented groups (Gennaioli and Rainer, 2007) which include Kenya’s Kikuyu and Nigeria’s Igbo, have historically been organized in a decentralized fashion (Tignor, 1971). In some cases, the concept of a common group identity did not emerge until the colonial and even independence periods (Southall, 1970). In many of these communities, authority was not vested in the hands of an individual chief––communities were often governed by councils of elders––and individual villages or homesteads did not recognize any authority above the community level (Boone, 2003).

Other ethnic groups––centralized groups—are organized more hierarchically with more centralized authority. The Ashanti, for example, had an empire that covered much of contemporary Ghana and which had many characteristics of a hierarchical bureaucratic state (Wilks, 1966). The Wolof kingdoms of precolonial Senegal were centralized and hierarchical (Boone, 2003). The Yoruba of Nigeria and Baganda of Uganda had similarly structured kingdoms prior to the arrival of British colonizers.

There is evidence that this variation has persisted into the contemporary period and is consequential for contemporary politics. Koter (2013) illustrates how this variation shapes ethnic politics in the contemporary democratic period. Baldwin (2014) shows that African presidents have in recent years granted greater power over the allocation of land to ethnic traditional leaders in order to buy their political support. Importantly, such power is only ceded when the ethnic groups have centralized organization.

This article argues that this variation can also help to explain the prevalence of electoral clientelism. Where ethnic or religious groups are characterized by centralized and hierarchical institutions, politicians should invest more effort and resources in forging linkages with the leaders of these groups in an effort to secure a bloc of votes wholesale. Where centralized and hierarchical institutions are not present, politicians will bypass local intermediaries and invest more resources and effort interacting directly to the voters, which results in more electoral clientelism. Importantly, the argument is not that political parties or politicians will never provide handouts directly to members of more centralized groups or that they will never seek to work with the group leaders of less-centralized groups. Rather, the argument is that, given budget constraints, their relative investments in individual electoral clientelism will be greater with members of more fragmented groups.

Centralized institutional structure facilitates wholesale electoral strategies for at least three reasons. Each relates to either the relative electoral effectiveness of working through local intermediaries or the relative effectiveness of targeting resources to groups versus individuals. First, in some contexts, leaders of centralized groups have the power to deliver votes using implicit or explicit physical or economic coercion. For example, intermediaries can use their monopoly control over valued resources––jobs, land, social services, and so on––to control and deliver votes (Medina and Stokes, 2007). Thus, if politicians deliver resources or benefits to group leaders, leaders of centralized groups will be better able to deliver the votes of their group as a bloc. This form of wholesale clientelism has been prevalent throughout history. In ancient Rome, elites would seek to buy the favor of tribal elites in exchange for a bloc of electoral support (Yakobson, 1999). In Chile, rural landowners delivered the votes of their employees prior to the adoption of stricter ballot secrecy laws (Baland and Robinson, 2008). In 19th century Germany, rural political elites, religious leaders, and employers exerted substantial control over voters, taking advantage of the absence of the Australian Ballot (Anderson, 1993).

The second set of mechanisms is less coercive. Leaders of centralized groups may be better able to help voters coordinate their voters around specific candidates. Strong local intermediaries may have more credibility and legitimacy with voters than do political party elites (Keefer and Vlaicu, 2008). An endorsement may therefore be effective because voters trust these leaders more than they do politicians. Alternatively, voters may interpret endorsements from group leaders as a sign that the leader and the politician will work together to provide public goods to their communities (Baldwin, 2015). Thus, politicians can work through centralized group leaders to convince voters that they will deliver the public goods and resources that they desire.

Third, centralized groups may have greater coordinating capacity, which makes the provision of collective goods more efficient for vote-seeking politicians (Gottlieb and Larreguy, 2016). This third mechanism thus emphasizes a reason why the provision of collective goods, such as local public goods, may be more electorally effective with centralized groups. If centralized groups are better able to coordinate their votes around a specific candidate, politicians can expect a greater electoral return for the provision of collective rather than individual benefits to the group.

Working through the leaders of ethnic groups becomes less efficient where social and political institutions are decentralized. One reason is that group leaders have the ability to control, coordinate, or influence fewer voters as the level of hierarchy decreases. Additionally, where institutions are decentralized and egalitarian, local leaders may have less control over voters, which makes it more difficult for them to deliver voters in a bloc. Tignor speaks to this issue in historical work on fragmented groups in Kenya and Nigeria, noting that such groups “were decentralized and fragmented into a number of autonomous communities administered by councils…the influence of these men was not hereditary or authoritarian” (Tignor 1971: 341). These egalitarian structures meant that no single individual or group exerted economic or social control over other community members or had monopoly control over crucial resources. Boone also highlights the challenges associated with working through weaker local leaders to secure votes. Writing of the more decentralized and egalitarian groups of Senegal, political leaders after independence simply “could not find interlocutors who were trusted enough, powerful enough, to package local populations into secure and acquiescent electoral blocs” (Boone 2003: 99). Local intermediaries in decentralized groups thus have both fewer voters to deliver and less power to deliver them.

In summary, there are three reasons why wholesale electoral strategies are likely to be more effective with centralized groups. The redistributive and normative implications of these wholesale strategies are distinct. It is beyond the scope of this article to empirically examine variation these different strategies. However, in the conclusion, I emphasize that an important area of future research will be to identity the conditions in which certain types of wholesale strategies will be selected over others.

This discussion produces several hypotheses that I test in this article:

H1: Individual electoral clientelism will be more prevalent in countries where fragmented groups are more populous.

H2: Individual electoral clientelism will be more prevalent among members of fragmented groups.

I note that ethnic and religious leaders are not the only intermediaries that can be used to mobilize votes. Central government actors may, for example, create subnational administrative units or build a political party organization. While the construction of new intermediaries is potentially important, I do not focus on them here for two reasons. First, despite efforts to displace traditional leaders in a number of African countries, traditional leaders continue to play an important role in much of the continent (Logan, 2013). Moreover, the formal power of traditional leaders appears to have increased since the wave of democratization that occurred in Africa in the early 1990s (Baldwin, 2015). Second, the creation of new intermediaries should only attenuate the effects of ethnic group structure.

Cross-national evidence

This section shows that electoral clientelism is most prevalent in countries where ethnic institutional structure is least centralized and hierarchical. To measure ethnic group institutional structure at the country level, I use data compiled by Gennaioli and Rainer (2007) on the proportion of the population in each African country that belongs to a centralized ethnic group. Gennaioli and Rainer (2007) use data from Murdock’s Ethnographic Atlas to code African ethnic groups as either centralized or fragmented. The Murdock data compile information from anthropological studies to form a data set of social, political, and economic characteristics of ethnic groups around the world.¹ The Murdock data include a measure of ethnic group “jurisdictional hierarchy beyond local community,” which captures each group’s level of centralized political authority. The most decentralized groups have no political authority beyond the local community and are coded as 1. Petty chiefdoms, which have one level of political authority above the local community, are coded as 2. More centralized groups have two levels (“larger chiefdoms”) and three levels (“states”) have two and three levels of political authority above the local community. Ghana’s Ashanti ethnic groups are coded as a “larger chiefdom” (3), while groups such as South Africa’s Zulu and Nigeria’s Yoruba are coded as once being “states” (4).

Centralized groups include larger chiefdoms and states (codes 3 and 4), while fragmented groups include petty chiefdoms and those ethnic groups with no political hierarchy above the local community (codes 1 and 2). The measure of focus is the proportion of individuals in the country that belong to a centralized ethnic group.

While the Murdock data capture information about the organization of ethnic groups prior to, or at the beginning of, colonial rule, they are advantageous in a study of contemporary clientelism for a number of reasons. First, the measure permits a degree of comparability across countries that helps to facilitate comparative analysis. Second, prior research provides evidence that the differences captured by the Murdock coding relate meaningfully to contemporary outcomes. Baldwin (2014) shows that presidents are likely to cede the power to allocate land to ethnic traditional leaders from other ethnic communities in order to buy their political support. Importantly, the study shows that such power is only ceded when the ethnic groups have centralized organization as measured by the Murdock data. Thus, the differences captured by the Murdock measures appear to be meaningful in contemporary politics.

Electoral clientelism data come from Afrobarometer’s round 3 and 5 surveys. The survey question on clientelism was not included in other Afrobarometer rounds. Round 3 was conducted in 2005 in 18 countries. Round 5 was conducted in 2012 in 34 countries. I use data from all countries where the Afrobarometer was conducted in round 3 and/or round 5 and where Gennaioli and Rainer (2007) have estimated the proportion of the population that is a member of a centralized ethnic group. The data include information on 27 countries and contain 43 unique country-round observations. I use the survey question that asks: “And during the [most recent] elections, how often [if ever] did a candidate or someone from a political party offer you something, like food or a gift, in return for your vote?” Respondents could answer “never,” “once or twice,” “a few times,” or “often.” I generate country level estimates of electoral clientelism that measure the proportion of respondents in each country that report being offered a bribe or a gift once or twice, a few times, or often. ²

Figure 1 displays the relation between the political centralization of ethnic groups and the prevalence of electoral clientelism in a country. The figure makes clear there is a great deal of variation within Africa in the extent to which ethnic groups have centralized structure. Countries such as Kenya and Mali are home to ethnic groups that are almost entirely decentralized. The residents of countries such as Botswana, Lesotho, and Malawi come almost exclusively from centralized groups.

Figure 1.

Social structure and electoral clientelism across Africa. Note: The vertical axis is the share of respondents in each country who report being offered a bribe or gift from a politician in their country’s most recent election. I use data from Afrobarometer round 3 and 5, the two rounds that include the question about electoral clientelism. The horizontal axis captures the share of the country’s population that belongs to an ethnic group with centralized ethnic institutions as estimated by Gennaioli and Rainer (2007) using Murdock’s Ethnographic Atlas (Murdock, 1967). The sample includes all countries where Murdock data are available and where Afrobarometer conducted surveys in round 3 and/or 5. Countries included twice are those where surveys were conducted in both rounds.

The figure highlights the strong correlation between social structure and individual electoral clientelism. Voters are targeted with electoral handouts at very high rates in countries where a relatively small share of the population belongs to a member of a centralized ethnic group, while electoral clientelism is relatively rare in countries where most voters are members of groups with centralized and hierarchical organization.

Online Appendix 1 presents Ordinary Least Squares (OLS) regression results which show that the relationship displayed in Figure 1 is robust to analyses that include country-level controls, including level of democracy, corruption, GDP per capita, urbanization, poverty, and colonial history.

Individual-level evidence

This section shows that individuals who are members of less-centralized ethnic groups are more likely to be targeted with individual electoral clientelism. I analyze individual survey data from 15 African countries. I selected these countries because they are countries where Afrobarometer conducted surveys in rounds 3 and 5, where the Afrobarometer survey included a question about ethnicity, and where I was able to match the Murdock data to a significant proportion of the population.³ The data set includes information about roughly 35,000 individuals, who are members of 261 ethnic groups.⁴

About 11% are members of ethnic groups that are coded as a 1 on the Murdock measure, 23% are members of ethnic groups that are coded as a 2, 33% are members of ethnic groups that are coded as a 3, and 33% are members of ethnic groups that are coded as a 4. About 34% of the samples are thus members of fragmented groups, while 66% of the samples are members of centralized groups.

The main dependent variable is a dichotomous measure of individual electoral clientelism. I code all individuals reporting that they received an offer of an electoral handout “once or twice,” “a few times,” or “often” as having experience with electoral clientelism (1) and all others as not (0).

I control for individual, ethnic group, and country-level factors that may be related to both ethnic group structure and patterns of clientelism. Further details about how the measures were constructed are in Online Appendix 2. At the individual level, I control for gender, age, residence in a rural area, level of education, and poverty using a lived poverty index (Mattes, 2008), where higher values capture a higher level of poverty (see Online Appendix 2). I also control for the intensity of individuals’ attachment to the ethnic identity (see Online Appendix 2), where higher values indicate a stronger relative attachment to the ethnic identity. The models also include a number of ethnic group specific controls, including the group’s share of the country population and whether a member of the ethnic group was the incumbent president. I include a control for each country’s level of democracy (POLITY score) and GDP per capita (logged). Models also include dummy variables for British or French colonial heritage. These variables capture historical differences between countries with different colonizers as well as many differences in contemporary political institutions.

Results

Table 1 presents the results. The unit of analysis is the individual and the dependent variable is the dichotomous indicator of individual electoral clientelism. Each model is a logistic regression, and standard errors are clustered by ethnic group survey round. I focus on two independent variables: a dichotomous variable indicating whether the individual is a member of a centralized (1) or fragmented group (0); and a continuous measure of centralization, running from 1 (least centralized) to 4 (most centralized).

Table 1.

Ethnic group structure and individual electoral clientelism in 15 African countries.

	(1)	(2)	(3)	(4)	(5)	(6)
Centralized ethnic group	−0.39*** (0.14)		−0.31** (0.14)		−0.28** (0.13)
Centralization (continuous)		−0.12* (0.07)		−0.09 (0.07)		−0.14** (0.06)
Poverty (individual)			0.19*** (0.03)	0.20*** (0.03)	0.13*** (0.04)	0.13*** (0.04)
Primary ed. only			−0.03 (0.06)	−0.04 (0.06)	−0.14** (0.06)	−0.15*** (0.06)
Secondary ed. only			−0.05 (0.05)	−0.06 (0.05)	−0.07 (0.04)	−0.07 (0.04)
Rural			0.16*** (0.06)	0.17*** (0.06)	0.07 (0.05)	0.06 (0.05)
Male			0.20*** (0.03)	0.20*** (0.03)	0.20*** (0.03)	0.20*** (0.03)
Age			−0.01*** (0.00)	−0.01*** (0.00)	−0.01*** (0.00)	−0.01*** (0.00)
Ethnic versus national identity			0.05** (0.02)	0.05** (0.02)	0.02 (0.02)	0.02 (0.02)
Coethnic with president			−0.01 (0.25)	−0.03 (0.26)	−0.05 (0.20)	−0.07 (0.21)
Ethnic group pop. share			−0.32 (0.58)	−0.44 (0.58)	−0.38 (0.42)	−0.43 (0.43)
POLITY democracy score					−0.08*** (0.02)	−0.08*** (0.02)
GDP per capita (log)					−0.21** (0.10)	−0.21** (0.10)
Former French colony					1.08*** (0.20)	1.13*** (0.20)
Former British colony					1.60*** (0.25)	1.68*** (0.25)
Round 5	−0.11 (0.15)	−0.12 (0.16)	−0.01 (0.18)	0.02 (0.18)	−0.01 (0.16)	0.01 (0.16)
Constant	−1.09*** (0.12)	−0.98*** (0.21)	−1.39*** (0.16)	−1.32*** (0.25)	−0.33 (0.83)	−0.19 (0.83)
No. of individuals	35,253	35,253	34,295	34,295	34,295	34,295
No. of ethnic groups	261	261	261	261	261	261
No. of countries	15	15	15	15	15	15

Note: Logistic regression analyses. Namibia is omitted reference category for colonial history. Robust standard errors, clustered by ethnic group––survey round in parentheses.

***p < 0.01; **p < 0.05; *p < 0.1.

Columns 1 and 2 present results of analyses that only include a survey-round control. In column 1, the effect of centralization is negative and significant. Members of fragmented ethnic groups are about 6.1 percentage points more likely to be targeted with individual electoral clientelism, a 36% increase over the mean in centralized groups. In column 2, the coefficient on the continuous centralization measure is significant at the p < 0.1 level and is also negative. These patterns are robust to the introduction of individual and ethnic group–specific controls into the models (columns 3 and 4). Here, the dichotomous measure remains significant and the magnitude of the effect goes down marginally with the introduction of controls. The continuous measure is not significant in column 4, although the magnitude and direction of the effect remain as expected. Columns 5 and 6 introduce country-level controls. The coefficients in both columns are significant and the magnitudes of the effects are not diminished by the introduction of these controls.⁵

Senegal and Uganda

This section presents further individual-level evidence by examining the cases of Senegal and Uganda in more detail. I analyze these cases for three reasons. First, in both countries, there is significant within-country variation in the organizational structure of ethnic and religious groups, and there is evidence that this variation is relevant in contemporary politics (Boone, 2003; Green and Bandyopadhyay, 2012). These are, therefore, cases where the existing literature validates meaningful differences in the organization of ethnic and religious groups, providing greater confidence in the centralization measures analyzed. Second, the extent of within-country variation in each case permits a within-country analysis. The advantage of within-country analysis is that it holds fixed a number of hard to measure historical, institutional, and cultural covariates that might be related to social structure and electoral clientelism. I can therefore analyze how social structure relates to campaign strategies adopted by politicians operating in the same political system. Third, in each case, I am able to analyze additional data in order to test further implications of the argument.

Before discussing the cases in more detail, I note an important difference between them: Senegal is a democracy,⁶ while Uganda is not.⁷ Nevertheless, I expect the argument to hold in both countries for three reasons. First, although Uganda is not a democracy, there is evidence that votes are consequential in the country, especially in elections for the legislature and local government. For example, opposition parties control a substantial number of elected positions in local governments in Uganda (Raffler, 2016). There is also evidence that MPs are more responsive to their constituents in electorally competitive electoral districts (Grossman and Michelitch, 2016), which suggests that electoral incentives do shape elite behavior. Second, some observers of Ugandan politics believed that President Museveni could have lost the 2011 election, and that this drove substantial campaign spending by Museveni and his party during the campaign (Conroy-Krutz and Logan, 2012; Izama and Wilkerson 2011). Conroy-Krutz and Logan (2012: 627) highlight that “state resources purchased everything from extensive advertising to public goods, with reported attempts to buy the political support of powerful local brokers and individual voters through cash disbursements” (emphasis added). Third, the literature suggests that autocrats often have a rational for engaging in clientelism, even if the result of an election is not in doubt. For example, clientelism can help autocrats win elections by large margins, which can have a number of indirect effects for the regime (Simpser, 2013), including deterring challenges and protest from elites and everyday citizens (Magaloni, 2006; Simpser, 2013). Clientelism can also mobilize voter turnout in order to signal the legitimacy of an election to domestic and international audiences (Gonzalez-Ocantos, et al., 2015). Thus, for both electoral and non-electoral reasons, there can be incentives for clientelism in non-democratic contexts such as Uganda, and this article tests the idea that ethnic group institutions shape relative investments in wholesale versus individual clientelistic strategies.

Cases

Uganda’s ethnic groups exhibit substantial variation with respect to their political centralization. Several of the country’s largest ethnic groups, including the Baganda, Banyankole, and Banyoro, had centralized, hierarchical, and relatively powerful kingdoms prior to the arrival of the British. In contrast, groups such as the Acholi, Bakiga, and Iteso, on the other hand, are described by Murdock as either petty chiefdoms or groups with no centralized authority above the community.

Importantly, many of the structures of centralized groups have persisted into or been resurrected in the contemporary period. For example, after the Kingdom of Buganda was reinstated in 1993 after being abolished in 1967 during the dictatorial rule of Milton Obote, the king established a number of legislative and administrative structures (Englebert, 2002).

In Uganda, electoral clientelism is prevalent during elections. The distribution of money and material goods to potential voters was especially intense during the country’s February 2011 presidential and parliamentary elections (Conroy-Krutz and Logan, 2012). Yoweri Museveni, in power since 1986 and running on the ticket of his National Resistance Movement (NRM) party, won the elections with 68% of the vote. Handouts to voters were reportedly widespread during the elections, with 25% of respondents in one post-election survey reporting that they were offered a handout during the campaign and 43% reporting that they observed parties distributing handouts (Conroy-Krutz and Logan, 2012).

Senegal is a former French colony. Although the country has been labeled a clientelistic democracy (Beck, 2008), rates of electoral handouts in the country are relatively low (Figure 1). Senegal is home to ethnic groups with a diversity of ethnic institutions. In the Wolof areas, for example, institutional structures were extremely centralized and hierarchical prior to colonial rule (Boone, 2003). The ethnic institutions of other large groups such as the Diola and Sereer contrast substantially with those of the Wolof. Of the Diola, for example, Boone notes that “village communities are largely independent” and “with a few exceptions, recognized no common authority” (Boone, 2003: 100). Even within villages, which are composed of extended family groupings, there is no “unified, village-wide political structure” (Boone, 2003: 101).

In Senegal, religious group organization has been important in the post-independence era. Indeed, colonial rule transformed traditional social structure and institutions in Senegal, opening the door for increased power and influence for the emerging Sufi Muslim Brotherhoods. In the 1880s, the French defeated the Wolof militarily, thus subjugating the traditional aristocracy and essentially transforming Wolof chiefs into administrative agents of the colonial state, a transformation that fundamentally shifted their power and legitimacy. In Wolof areas, the Mouride Brotherhood took over the status and many of the former functions of the Wolof chiefs. Mouride leaders also benefited from the expansion of groundnut production in Wolof areas during the colonial and post-independence eras, using their power to extract labor and agricultural surplus from their followers (Cruise O’Brien, 1971).

The Mouride are not the only brotherhood that emerged during colonial rule. Both the Qadiriyya and Tijaniyya brotherhoods also spread rapidly during the 19th century. The Tijaniyya currently has the largest membership in Senegal. While these brotherhoods also have a formal hierarchy, Cruise O’Brien (1971) argues that the Mouride’s degree of centralization, organization, and hierarchy made them distinct from their counterparts. Of the Mouride, he notes, “the leaders of the Mouride brotherhood, having authority over a relatively centralized structure, and one which had a reasonably effective hierarchical system, were able to deliver more assured votes than any other religious leaders…” (Cruise O’Brien, 1971: 262). In contrast, with respect to the Qadiriyya and Tijaniyya, he argues that “the real structure was extremely decentralized and the operative organizational unit was usually the local logde…” (Cruise O’Brien, 1971: 26). Thus, while the majority of Senegalese belong to Sufi Muslim Brotherhoods, there is variation with respect to how important the centralized and hierarchical structures of these brotherhoods are.

Results

Table 2 presents results from the analysis of the Ugandan data (AB rounds 3 and 5). The unit of analysis is the individual survey respondent and the centralization measure of focus is the continuous measuring running from 1 (least centralized) to 4 (most centralized). All analyses also include the individual-level control variables described above. Column 1 presents the bivariate relationship, showing that members of centralized ethnic groups are significantly less likely to experience electoral handouts. The estimates suggest that the least centralized groups are about 8 percentage points more likely to experience electoral clientelism than are the most centralized groups. The results in column 2 show that the effect is about the same magnitude in rural areas. Columns 3 and 4 show that the results are robust to the inclusion of controls. According to column 3, a one-unit increase in the centralization measure is associated with a 4.4 percentage point decrease in the probability of receiving an electoral handout (95% confidence interval from –6.7 to –2.1).

Table 2.

Ethnic group structure and individual electoral clientelism in Uganda.

	(1) Full sample	(2) Rural	(3) Full sample	(4) Rural
Ethnic group centralization (continuous)	−0.117*** (0.038)	−0.100** (0.043)	−0.111** (0.049)	−0.145*** (0.055)
Poverty index			0.121*** (0.041)	0.101** (0.047)
Residence in rural area			0.029 (0.092)
Age			0.006** (0.003)	0.004 (0.003)
Male			0.067 (0.072)	0.054 (0.082)
No formal education			−0.164 (0.161)	−0.016 (0.189)
Only primary education			−0.027 (0.118)	0.063 (0.152)
Only secondary education			−0.113 (0.115)	−0.044 (0.156)
Ethnic group size			−0.511 (0.332)	−0.180 (0.379)
Coethnic with president			0.583*** (0.131)	0.567*** (0.144)
Ethnic versus national identity			−0.081** (0.037)	−0.114*** (0.042)
Constant	−0.214** (0.091)	−0.218** (0.099)	−0.322 (0.211)	−0.153 (0.250)
Observations	3586	2747	3527	2703

Note: Robust standard errors in parentheses. Each model is a logistic regression model in which the dependent variable takes a value of 1 if the respondent was offered a bribe or gift during the 2011 (AB round 5) or 2001 (AB round 3) elections.

***p < 0.01; **p < 0.05; *p < 0.1.

Table 3 presents results on electoral handouts in Senegal. In Senegal, the relevant intermediaries are religious. I therefore code each individual in the Senegal Afrobarometer data with respect to their religious affiliation. About 27% report membership in the Mouride brotherhood, which I classify as centralized and hierarchical. The remainder belong to the Qadiriyya (6%), Tijaniyya (48%), or other religions (catholic, protestant, etc.), each of which I classify as lacking centralization.

Table 3.

Religious group structure and individual electoral clientelism in Senegal.

	(1) Full sample	(2) Rural	(3) Full sample	(4) Rural
Mouride	−0.252 (0.167)	−0.320 (0.222)	−0.141 (0.172)	−0.144 (0.230)
Poverty index			0.337*** (0.076)	0.335*** (0.095)
Residence in rural area			−0.196 (0.163)
Age			−0.009 (0.005)	−0.002 (0.008)
Male			0.162 (0.157)	0.084 (0.213)
No formal education			0.080 (0.315)	−0.076 (0.601)
Only primary education			0.237 (0.314)	0.011 (0.617)
Only secondary education			0.093 (0.307)	−0.037 (0.642)
Religious group size			−0.410 (0.500)	−0.637 (0.649)
Coethnic with president			−0.471*** (0.172)	−0.673*** (0.246)
Ethnic versus national identity			−0.064 (0.062)	0.043 (0.076)
Constant	−2.215*** (0.080)	−2.215*** (0.110)	−2.168*** (0.400)	−2.529*** (0.695)
Observations	2400	1320	2283	1261

***p < 0.01; **p < 0.05; * p < 0.1.

The results show that members of the Mouride Brotherhood are less likely to experience electoral clientelism than are members of other groups. According to the coefficient in column 1, which narrowly misses statistical significance (p = 0.13), Mouride are about 2.1 percentage points less likely to experience electoral clientelism (95% CI from –4.9 to 0.6). This is a substantively large effect, as only 9% of Senegalese received a handout. Column 2 shows that the magnitude of the effect is larger in rural areas, although the coefficient cannot be statistically distinguished from 0. Columns 3 and 4 introduce control variables, which have no impact on the magnitude of the coefficients on the Mouride variable. According to model 3, Mouride are about 2 percentage points less likely to receive electoral handouts (95% CI from –4.8 to 0.8).

I also code the structure of each ethnic group in Senegal using the four-point centralization measure. The results are presented in Table 4. The estimate in column 1 narrowly misses conventional statistical significance cutoffs (p = 0.11) and is qualitatively consistent with expectations. The result in column 2 shows that the magnitude of the effect is larger in rural areas. This estimate suggests that a one-unit increase in the centralization measure is associated with a 3 percentage point decrease in the probability of experiencing electoral clientelism (95% CI from –6 to 0.001). This is a substantively important effect size given that about 9% of the full sample reports experience with electoral clientelism. When the controls are introduced in columns 3 and 4, the coefficients are no longer statistically significant, although they are qualitatively consistent with expectations. According to model 4, a one-unit increase in the centralization measure corresponds with a 2.5 percentage point decrease in the probability of receiving an electoral handout, holding all other variables at their means (95% CI from –6.2 to 1.2). This is a large effect size given the overall low levels of electoral clientelism in Senegal.

Table 4.

Ethnic group structure and individual electoral clientelism in Senegal.

	(1) Full sample	(2) Rural	(3) Full sample	(4) Rural
Ethnic group centralization (continuous)	−0.221 (0.139)	−0.361** (0.183)	−0.141 (0.174)	−0.288 (0.228)
Poverty index			0.325*** (0.078)	0.332*** (0.097)
Residence in rural area			−0.192 (0.167)
Age			−0.008 (0.005)	−0.001 (0.008)
Male			0.162 (0.160)	0.082 (0.217)
No formal education			0.123 (0.335)	−0.155 (0.605)
Only primary education			0.260 (0.333)	−0.021 (0.624)
Only secondary education			0.115 (0.327)	−0.162 (0.653)
Ethnic group size			0.030 (0.659)	0.578 (0.902)
Coethnic with president			−0.510*** (0.193)	−0.813*** (0.272)
Ethnic versus national identity			−0.075 (0.064)	0.043 (0.077)
Constant	−1.715*** (0.365)	−1.376*** (0.470)	−1.971*** (0.536)	−2.115*** (0.818)
Observations	2321	1293	2212	1236

***p < 0.01; **p < 0.05; * p < 0.1.

In sum, while the Senegal results should be interpreted with some caution given the imprecision of some of the estimates, the patterns are consistent with the notion that religious and ethnic group organization has an impact on patterns of electoral clientelism.

Additional within-country evidence

I focus the within-country analysis on Senegal and Uganda, countries with significant within-country variation that has been validated in other studies. However, the data analyzed in the “Individual-level evidence” section also permit within-country analyses in a larger number of countries. These results are presented in Online Appendix 3. The results are statistically significant, and in the expected direction, in Malawi and Namibia. The results are qualitatively consistent with expectations in a number of other countries, including Kenya, Zambia, and Mali. In a few cases, such as Lesotho, the coefficients are very close to 0, though in such countries, there is limited internal variation in ethnic group structure to leverage.

In two cases, Benin and Ghana, the results run contrary to expectations: members of centralized groups are more likely to receive handouts. Although I can only speculate, one reason for this pattern in Benin could be the relative lack of trust in traditional leaders in that country, which limits their usefulness as potential intermediaries (Koter 2013). In Ghana, the results may differ because the two major political parties are well institutionalized and have a strong organization of grassroots activists (Osei 2012), which may impact how handouts are targeted. Further research is required to identify why the theory does not appear to explain patterns of clientelism in these cases.

Testing other implications

This section provides additional evidence that is consistent with the argument by focusing on the theory’s implications for outcomes among members of centralized groups. I use the survey data available to test two additional hypotheses. First, if the argument is correct, leaders of more centralized groups should have a greater influence on the political decisions of group members than do the leaders of less-centralized groups. Second, if leaders of centralized groups are better able to control the votes of group members, indicators of vote monitoring should be higher among members of centralized groups. Together, tests of these two hypotheses address the expectation that wholesale electoral strategies are more prevalent among centralized groups.

First, I examine the influence of traditional leaders on the political behavior of group members. To do so, I analyze an Afrobarometer survey question that was asked in Senegal (round 3). The question reads: “Of the following individuals, which of them are likely to influence your political choices? Religious leader.” Response options are: “not at all” (0), “a little” (1), “somewhat” (2), and “a lot” (3). I use this four-point response as the dependent variable. About 20% of the sample reports that a religious leader has at least a little influence on their political decisions. A similar question was not asked in Uganda.

The mean of the influence variable is 0.78 among the Mouride, the centralized religious group, and 0.39 among non-Mouride. Table 5 presents regression results, which show that members of the centralized group are significantly more likely to report that religious leaders influence their political decisions, even after controlling for a range of covariates.

Table 5.

Religious group structure and the political influence of religious leaders in Senegal.

	(1)	(2)
Mouride	0.38*** (0.07)	0.33*** (0.07)
Rural		0.25*** (0.08)
Male		−0.22*** (0.06)
Age		0.00** (0.00)
Poverty		−0.10*** (0.03)
No formal education		0.49*** (0.08)
Primary education		0.32*** (0.08)
Secondary education		0.34*** (0.07)
Constant	0.39*** (0.03)	0.11 (0.07)
Observations	1167	1130
R ²	0.03	0.08

Note: Robust standard errors in parentheses.

***p < 0.01; **p < 0.05; *p < 0.1.

The second set of analyses focuses on whether individuals believe that the secrecy of the ballot can be violated. The capacity of group leaders to control and deliver votes depends in part on their ability to monitor the vote choices of group members. If politicians are pursuing a strategy of wholesale clientelism, members of these groups should be less likely to believe that their vote choice is secret.

The round 4 survey conducted in Senegal and the round 4 and 5 surveys conducted in Uganda include a question about the secrecy of the ballot. The question reads as follows: “How likely do you think it is that powerful people can find out how you voted, even though there is supposed to be a secret ballot in this country?” Answers range from 0 (not at all likely) to 3 (very likely). I use this four-point scale as the dependent variable.

Table 6 presents the results, which show that members of centralized ethnic and religious groups are more likely to believe that ballot secrecy can be violated. Columns 1 and 2 present the evidence from Uganda. In each specification, the coefficient on ethnic group centralization is positive and significant. The Senegal results in columns 3 and 4 are qualitatively similar, although the coefficients are not statistically significant.

Table 6.

Religious and ethnic group structure and perceptions of ballot secrecy.

	(1) Uganda	(2) Uganda	(3) Senegal	(4) Senegal
Centralized ethnic group	0.04** (0.02)	0.07*** (0.02)
Mouride			0.06 (0.07)	0.08 (0.07)
Residence in rural area		0.06 (0.04)		−0.31*** (0.06)
Male		−0.02 (0.03)		0.08 (0.05)
Age		0.00 (0.00)		−0.00 (0.00)
Poverty index		0.07*** (0.02)		0.07** (0.03)
No formal education		−0.16*** (0.06)		−0.19 (0.13)
Only primary education		−0.11** (0.05)		−0.18 (0.13)
Only secondary education		−0.12** (0.05)		−0.04 (0.14)
Round 5 dummy		−0.28*** (0.03)
Constant	0.76*** (0.04)	0.82*** (0.08)	0.49*** (0.03)	0.62*** (0.13)
Observations	3536	3492	1092	1071
R ²	0.00	0.03	0.00	0.04

Note: Robust standard errors in parentheses. Dependent variable runs from 0 to 3, with higher values indicating a greater belief in the likelihood that powerful people can discover the vote.

***p < 0.01; **p < 0.05; *p < 0.1.

In sum, members of centralized religious and ethnic groups are more likely to report that group leaders influence their voting decisions and are more likely to believe that the secrecy of the ballot can be violated by powerful people in their community. These patterns provide further support for the argument.

Conclusion

This article has been motivated by two inter-related puzzles: first, that electoral clientelism is not prevalent in all low-income democracies or communities within low-income democracies; and second, that the prevalence of electoral clientelism can vary, even across contexts where other forms of clientelism are common.

I argue that ethnic group institutional structure can help to explain this variation. Where groups are structured hierarchically with centralized leadership, the expectation is that politicians will pursue more wholesale electoral strategies, either working through group leaders or seeking to trade collective goods for blocs of votes. In the absence of such structure, politicians are more likely to target their efforts directly to voters themselves, which results in larger investments in electoral clientelism. Evidence from across and within a set of African countries is consistent with this logic: electoral clientelism is more common in African countries where ethnic and religious groups are not organized in a centralized and hierarchical manner; and members of less-centralized and hierarchical groups are more likely to experience electoral clientelism.

This article suggests several directions for future research. First, this article has identified a correlation between ethnic group structure and clientelism that is consistent with the theory. Future research could build more empirical support for the mechanisms implied by the argument through in-depth qualitative research examining electoral strategies among groups with different types of organization. Second, although the results are generally consistent with the argument, there are specific countries where the findings run contrary to expectations, and further research is required to identify why the theory does not appear to explain patterns of clientelism in these cases. The cases of Benin and Ghana suggest that trust in traditional leaders and the organizational development of political parties may be important conditions to consider. Third, this article leaves open the question of how leaders of centralized groups deliver votes. We might consider two categories of mechanisms: one in which intermediaries leverage discretionary power over resources and decision-making to coerce voters into supporting their preferred candidate; and a second in which they play a more benign coordinating role in guiding group members toward the candidate or party that will best serve group interests. The theoretical and normative implications of these two mechanisms are distinct, and so it will be important for future research to identify the conditions in which the more coercive or more coordinating mechanisms are likely to emerge.

This study contributes to several literatures. First, it contributes to the literature on electoral clientelism by advancing a distinct argument about the types of conditions in which electoral clientelism is likely to be most prevalent. Second, the article builds upon recent work demonstrating the role of ethnic group institutions in shaping electoral strategies in Africa. I extend this literature to the realm of electoral clientelism, showing how ethnic group institutional structure shapes investments in handouts at election time. Third, the findings contribute to a growing body of research emphasizing the importance of intermediaries, or brokers, in shaping strategies of clientelism (Camp, 2017; Stokes et al., 2013; Szwarcberg, 2012). For example, Stokes et al. (2013) focus on the strategic interaction between politicians and party brokers, and show how this interaction shapes who gets targeted with clientelist transfers. This article shows how the organization of a different type of intermediary––leaders of ethnic and religious groups rather than party activists––shapes the character of clientelism.

Finally, although I have focused on the structure of ethnic and religious groups in Africa, the argument that group organizational structure is likely to shape strategies of clientelism may extend beyond Africa and beyond contexts where politically salient groups are defined along ethnic or religious lines. Testing this argument in other contexts will be an important area for future research.

Supplemental material

Supplemental Material, Kramon_revision_Supplementary_Material - Ethnic group institutions and electoral clientelism

Supplemental Material, Kramon_revision_Supplementary_Material for Ethnic group institutions and electoral clientelism by Eric Kramon in Party Politics

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Supplemental material

Supplementary material for this article is available online.

Notes

References

Anderson

(1993) Voter, junker, landrat, priest: The old authorities and the new franchise in imperial Germany. American Historical Review 98(5): 1448–1474.

Baland

Robinson

(2008) Land and power: Theory and evidence from Chile. American Economic Review 98(5): 1737–1765.

Baldwin

(2014) When politicians cede control of resources: Land, chiefs, and coalition building in Africa. Comparative Politics 46(3): 253–271.

Baldwin

(2015) The Paradox of Traditional Leaders in Africa. New York: Cambridge University Press.

Beck

(2008) Brokering Democracy in Africa: The Rise of Clientelist Democracy in Senegal. London: Palgrave Macmillan.

Boone

(2003) Political Topographies of the African State: Territorial Authority and Institutional Choice. New York: Cambridge University Press.

Bratton

(2008) Vote buying and violence in Nigerian election campaigns. Electoral Studies 27(4): 621–632.

Brusco

Nazareno

Stokes

(2004) Vote buying in Argentina. Latin American Research Review 39(2): 66–88.

Camp

(2017) Cultivating effective brokers: A party leader’s dilemma. British Journal of Political Science 47(3): 521–543.

10.

Conroy-Krutz

Logan

(2012) Museveni and the 2011 Ugandan election: Did the money matter? Journal of Modern African Studies 50(4): 625–655.

11.

Cox

Kousser

(1981) Turnout and rural corruption: New York as a test case. American Journal of Political Science 25(4): 646–663.

12.

Cruise O’Brien

(1971) The Mourides of Senegal: The Political and Economic Organization of An Islamic Brotherhood. Oxford: Oxford University Press.

13.

Englebert

(2002) Born-again Buganda or the limits of traditional resurgence in Africa. Journal of Modern African Studies 40(3): 345–368.

14.

Gans Morse

Mazzuca

Nichter

(2014) Varieties of clientelism: Machine politics during elections. American Journal of Political Science 58(2): 415–432.

15.

Gennaioli

Rainer

(2007) The modern impact of precolonial centralization in Africa. Journal of Economic Growth 12(3): 185–234.

16.

Green

Bandyopadhyay

(2012) Pre-Colonial Political Centralization and Contemporary Development in Uganda. Afrobarometer Working Paper No. 141. Available at http://afrobarometer.org/sites/default/files/publications/Working%20paper/afropaper141.pdf

17.

Gonzalez-Ocantos

de Jonge

Melendez

. (2012) Vote buying and social desirability bias: Experimental evidence from Nicaragua. American Journal of Political Science 56(1): 202–217.

18.

Gonzalez-Ocantos

de Jonge

Nickerson

(2015) Legitimacy buying: The dynamics of clientelism in the face of legitimacy challenges. Comparative Political Studies 48(9): 1127–1158.

19.

Gottlieb

Larreguy

(2016) An Informational Theory of Electoral Targeting in Young Clientelistic Democracies: Evidence from Senegal. Mimeo, 7 October.

20.

Grossman

Michelitch

(2016) Information Dissemination, Competitive Pressure, and Politician Performance between Elections: A Field Experiment in Uganda. Center for the Study of Democratic Institutions Working Paper Series 3-2016.

21.

Hicken

(2007). How do rules and institutions encourage vote buying? In: Schaffer

(ed) Elections for Sale: The Causes and Consequences of Vote Buying. London: Lynne Rienner Publishers.

22.

Izama

Wilkerson

(2011) Uganda: Museveni’s triumph and weakness. Journal of Democracy 22(3): 64–78.

23.

Keefer

Vlaicu

(2008) Democracy, credibility, and clientelism. Journal of Law, Economics, and Organization 24(2): 371–406.

24.

Kramon

(2016) Where is vote buying effective? Evidence from a list experiment in Kenya. Electoral Studies 44: 397–408.

25.

Koter

(2013) King makers: Local leaders and ethnic politics in Africa. World Politics 65(2): 187–232.

26.

Lehoucq

(2007) When does a market for votes emerge? In: Schaffer

(ed) Elections for Sale: The Causes and Consequences of Vote Buying, London: Lynne Rienner Publishers, pp. 33–45.

27.

Logan

(2013) The roots of resilience: Exploring popular support for African traditional authorities. African Affairs 112(448): 353–376.

28.

Magaloni

(2006) Voting for Autocracy: Hegemonic Party Survival and its Demise in Mexico. New York: Cambridge University Press.

29.

Mattes

(2008) The Material and Political Bases of Lived Poverty in Africa: Insights from the Afrobarometer. Afrobarometer Working Paper No. 98. Available at http://afrobarometer.org/sites/default/files/publications/Working%20paper/AfropaperNo98.pdf

30.

Medina

Stokes

(2007) Monopoly and monitoring: An approach to political clientelism. In: Kitschelt

Wilkinson

(eds) Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, New York: Cambridge University Press, pp. 68–83.

31.

Michalopoulos

Papaioannou

(2013) Pre-colonial ethnic institutions and contemporary African development. Econometrica 81(1): 113–152.

32.

Murdock

(1967) Ethnographic Atlas. Pittsburgh: University of Pittsburgh Press.

33.

Nichter

(2008) Vote buying or turnout buying? Machine politics and the secret ballot. American Political Science Review 102(01): 19–31.

34.

Osei

(2012) Party-Voter Linkage in Africa: Ghana and Senegal in Comparative Perspective. Berlin, Germany: Springer.

35.

Raffler

(2016) Does Political Oversight of the Bureaucracy Increase Accountability? Field Experimental Evidence from an Electoral Autocracy. Mimeo: Princeton University.

36.

Simpser

(2013) Why Governments and Parties Manipulate Elections. New York: Cambridge University Press.

37.

Scott

(1969) Corruption, machine politics, and political change. American Political Science Review 63(4): 1142–1158.

38.

Southall

(1970) The illusion of tribe. Journal of Asian and African Studies 5(1/2): 28–51.

39.

Stokes

(2005) Perverse accountability: A formal model of machine politics with evidence from Argentina. American Political Science Review 99(03): 315–325.

40.

Stokes

Dunning

Nazareno

. (2013) Brokers, Voters, and Clientelism: The Puzzle of Distributive Politics. New York: Cambridge University Press.

41.

Szwarcberg

(2012) Uncertainty, political clientelism, and voter turnout in Latin America: Why parties conduct rallies in Argentina. Comparative Politics 45(1): 88–106.

42.

Tignor

(1971) Colonial chiefs in chiefless societies. Journal of Modern African Studies 9(3): 339–359.

43.

Wilks

(1966) Aspects of bureaucratization in Ashanti in the nineteenth century. The Journal of African History 7(02): 215–232.

44.

Yakobson

(1999) Elections and Electioneering in Rome: A Study in the Political System of the Late Republic. Stuttgart: Franz Steiner Verlag.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB