Few scholars have taught us more about African voters, legislators, and legislatures than Joel Barkan. His numerous research projects, which took him from capital cities to the most rural locales, were underpinned by a continuous focus on political linkage: the informal and formal networks through which citizens and representatives are connected. Barkan was interested in the question of how citizens were connected to their governments in Africa’s new states—if, that is, they were. He was also interested in thinking through what enabled governments that typically had little presence in rural areas to mobilize their populations. In asking these questions he was implicitly seeking an answer to one of the most interesting puzzles of African politics: why did civilian one-party systems prove to be the most stable forms of government in the three decades following independence (Allen Reference Allen1995)? Although Barkan’s research typically focused on Kenya, and to a lesser extent Tanzania, his answers to these questions are also relevant to other single-party states such as Senegal and Zambia. Partly for this reason, his publications became fundamental building blocks of how we understand African politics, then and now.
This article argues that one of the key insights provided by Barkan’s work is the recognition that although single-party governments in Africa were authoritarian in many respects, they contained important democratic elements, and that these elements were critical to the maintenance of political stability. Put another way, Barkan understood that civilian one-party states featured aspects of “competitive-authoritarianism” long before the term became fashionable (Levitsky & Way Reference Levitsky and Way2002). As Henry Bienen put it in relation to Kenya (1969), political stability was generated not simply through authoritarian repression, but through the combination of participation and control.
More specifically, Barkan’s research demonstrated that the maintenance of one-party elections for constituency-based MPs operating in a first-past-the-post system created important lines of political linkage that ran all the way from voters to the executive via a series of intermediaries, most notably MPs, regional political brokers, and cabinet ministers. Of course, this highly personalized and hierarchical set of networks was better at sending messages down the pyramid than up. Furthermore, the impossibility of changing the party in power or the president meant that representation could only go so far. But this did not mean that representation was not important. Rather, it came to be expressed through an extremely localized form of politics. In particular, MPs were expected to focus on local issues of direct and immediate concern to the constituency, with voters demanding that their representatives link them to streams of revenue and services emanating from the political center.
In making this argument, Barkan was keen to stress that African MPs acted the way they did not because they were ignorant or did not understand any other way to “do politics,” but because they understood that this was the best way to get elected in the contexts within which they operated. His focus on the structural pressures placed on African legislators was symptomatic of his determination to treat African politicians and voters as knowledgeable and rational individuals, motivated by the same kinds of considerations that influence leaders and electorates in other continents. Indeed, one of the main reasons that Barkan’s work remains so relevant is that he refused to ghettoize or exoticize Africa.
One important aspect of this approach was that Barkan refused to accept the notion that the continent’s politics were so unique that they required the use of distinctive research methods. Instead, he sought to apply the same techniques that were becoming fashionable among political scientists researching North American and European politics in the 1970s, pioneering the use of survey data as a tool through which to understand African popular opinion. Another important aspect was that he took political institutions seriously, and recognized that the structure of the state—and the forms of political linkage—have profound implications for the responsiveness of the political system to the needs and demands of its citizens.
This article seeks to demonstrate the more nuanced appreciation of authoritarianism in Africa that emerges out of Barkan’s work, and the way in which his innovative research helped to lay the foundations for contemporary developments in African political science, which over the last decade has increasingly recognized the value of survey research and the significance of formal political institutions. It also argues that Barkan’s work holds the clue for understanding Kenya—and similar African states—today. The continued weakness of political parties, the strength of “Big Men,” and the pressure for the devolution of power are a direct consequence of the highly personalized and localized form of political linkage that his work did so much to bring to our attention.
Competitive-Authoritarianism in the African One-Party State
According to the influential formulation of Levitsky and Way, “competitive-authoritarianism” refers to a “particular form of hybrid regime” in which “formal democratic institutions are widely viewed as the principal means of obtaining and exercising political authority. Incumbents violate those rules so often and to such an extent, however, that the regime fails to meet conventional minimum standards for democracy” (2002:52). Levitsky and Way provide a number of examples to illustrate this point, including “Russia under Vladimir Putin, Ukraine under Leonid Kravchuk and Leonid Kuchma, Peru under Alberto Fujimori, and post-1995 Haiti, as well as Albania, Armenia, Ghana, Kenya, Malaysia, Mexico, and Zambia through much of the 1990s” (2002:52). As these examples make clear, Levitsky and Way see the rise of competitive-authoritarianism as the product of a particular historical moment, namely the end of the Cold War and the subsequent proliferation of a set of political systems in Africa and postcommunist Europe that were neither fully authoritarian nor fully democratic.
Joel Barkan began to use the term “competitive-authoritarian” only in his later work, following the reintroduction of multiparty politics on the continent in the 1990s (see Barkan Reference Barkan, Cooke and Stephen Morrison2009a). But a careful reading of his early work suggests that he had identified the emergence in Africa of a different kind of competitive-authoritarian state much earlier (and much earlier than is conventionally recognized) in the shape of the civilian one-party states that took hold in places such as Kenya, Senegal, Tanzania, and Zambia. In these countries, where popular nationalist parties often won multiparty elections in the colonial era and went on to assert their political hegemony in early postindependence polls before constructing one-party states, the stability of the political system rested on a complex combination of coercion, political participation, and constrained elite competition.
Of course, these states were not all competitive-authoritarian in the strict sense. Under Levitsky and Way’s formulation, competitive-authoritarian states were officially democratic but in reality authoritarian. By contrast, the states that Barkan looked at were officially one-party states but nonetheless contained elements of democracy. In this sense, the use of the competitive-authoritarian concept here stands the original definition on its head. As a result, it might be more accurate to swap Levitsky and Way’s terms around and categorize Africa’s civilian single-party systems as “authoritarian-competitive” states. Doing so makes it clear that some of their findings, such as the potential for political violence to be driven by controversial national elections, do not apply to the cases discussed in this article.
However, the framework that Levitsky and Way operationalize remains useful because competitive-authoritarian and authoritarian-competitive states exist on the same continuum—and often exhibit more similarities than they do differences. Although the legal barriers to opposition political parties in countries like Kenya and Zambia prevented national-level political competition, Footnote 1 in their first decade they were often rated as being just as open and democratic—and in some cases considerably more so—than their contemporary competitive-authoritarian counterparts (Cheeseman Reference Cheeseman2015). Indeed, when it comes to freedom of the press and the independence of the legislature, the gap between multiparty Uganda and Rwanda, and one-party Kenya and Zambia, is vanishingly thin. Moreover, Africa’s civilian one-party states exhibited many of the key elements of competitive-authoritarianism as described by Levitsky and Way, and it is only by understanding these elements that the stability of the one-party state, and its political legacy, can be understood.
Indeed, reviewing Barkan’s research on the sites of democratic contestation that are most important within the competitive-authoritarian framework—namely the electoral, legislative, and judicial arenas—reveals just how much contestation occurred within some African single-party systems. Both in the early and the later stages of his career, Barkan was keen to highlight the democratic reality and potential of African states. Rereading his work is also a worthwhile exercise because it demonstrates the capacity for this political competition to generate strong links between citizens and their representatives. In the discussion that follows, space restrictions require me to focus on the two arenas of contestation that received the most attention in Barkan’s own work: the electoral and legislative arenas.
The Electoral Arena
According to Levitsky and Way, “the first and most important arena of contestation is the electoral arena” (2002:55). Elections in the African one-party states did not offer voters a choice of parties (although electors in Zambia were allowed a “yes”/“no” vote on the ruling party’s presidential candidate), but this does not mean that one-party elections were not competitive. Footnote 2 Soon after independence, Hyden and Leys (Reference Hyden and Leys1972) documented the high levels of political competition in legislative elections in Kenya and Tanzania, with rates of turnover in excess of 50 percent. As Gertzel (Reference Gertzel1970) and others have shown, this high level of political turnover often led to the defeat of senior political figures, although rarely those particularly close to the president. In this way, voters were able to bloody the noses of leaders who were seen to have underperformed, with assistant ministers particularly vulnerable to defeat. Footnote 3
Much of Barkan’s early work focused on demonstrating the competitive features of one-party elections and the way in which the political linkages generated through these processes served to legitimize the governments of the “new states” in Africa. In taking on this task he was painfully aware of the failure of some of his contemporaries to recognize the complex roots of African political systems, and the rational political decisions of the agents who operated within them. For example, his 1976 paper “reassessing the conventional wisdom” argued passionately that researchers should recognize that the “existence of informed publics on the periphery of the new states is complemented by rational political behavior on the part of these publics, behavior which appears in turn to affect the behavior of those who purport to be political leaders” (1976:452).
There were two parts to this argument. The first was that single-party elections in Africa could not be understood simply as an ethnic census in which voters unthinkingly lined up behind the relevant “Big Man.” Rather, the fact that constituency elections for MPs often took place in fairly homogenous rural constituencies, and were conducted in the absence of a divisive contest between presidential candidates of rival ethnicities, meant that single-party elections were more likely than multiparty polls to focus voters’ minds on issues relating to the actual performance of their representatives (Barkan Reference Barkan, Smith and Musolf1979b; Posner 2005). The second part of the argument was that African voters were often well informed about their candidates and fully able to make up their own minds based on the evidence. Putting these two points together, he came to the conclusion that
the strong belief that MPs should communicate the views of their constituents to the central government and strive to obtain resources for their districts, when coupled with the relatively low incidence of no answers [to survey questions], suggests that the publics of peasant societies not only have a general knowledge of the political process . . . but also a well-defined conception of the roles their political leaders should play. (1976:453)
This “well defined conception” was particularly important, because it demonstrated the great demand for, and existence of, political linkages between citizens and the central government in countries such as Kenya. Indeed, a survey conducted by Barkan and John Okumu in thirteen rural Kenyan constituencies found that when voters were asked what activities they thought it was most important for their MPs to perform, the “overwhelming proportion of the respondents” named “activities which constitute linkages between the publics on the periphery of the political systems of the new states and the central government” (1976:453). More specifically, the three most common answers were that MPs should visit the district frequently (11%), obtain projects and benefits for the district (25%), and tell the government what people in the district want (29%).
These findings were important for number of reasons. For one thing, this evidence made it much harder for Afropessismists to suggest that high levels of illiteracy meant that the continent was doomed to authoritarian rule. Barkan’s work was also significant because it highlighted the way in which the bias in favor of linkage over nonlinkage activities—such as taking part in legislative debates (5%)—placed MPs under tremendous pressure from below. As his later research would explain, fulfilling local expectations had two aspects: playing a role in the coordination and funding of development efforts at the constituency level, and linking the constituency to sources of funds and services emanating from the political center. Competition among aspiring political leaders thus took the form of mobilizing funds for the construction of locally specified projects. As a result, elections quickly became referenda on the development performance of MPs, generating strong ties of local accountability (Barkan Reference Barkan, Smith and Musolf1979b).
Building on Barkan’s insights, my own research has shown that one of the reasons that this system proved to be so durable is related to the dynamic fusion of patronage and accountability (Cheeseman Reference Cheeseman2006a). Because MPs, at least in the first fifteen years of the one-party state, were forced to seek reelection in polls that were reasonably open, they could not afford to ignore the demands of voters. Voters, for their part, continued to engage in elections because MPs controlled substantial resources, and so the contests took on considerable relevance to constituents’ everyday lives. Taken together, these two mutually reinforcing factors gave rise to one of the continent’s more responsive political systems (Hyden & Leys 1971; Hornsby & Throup Reference Hornsby and Throup1992). At the same time, because elections mattered for the distribution of resources, constituents and representatives engaged in the electoral process in a serious way—although turnout tended to decline as a result of the growing manipulation of the polls in the 1980s. Footnote 4
In Kenya, the fusion of patronage resources and electoral competition had important developmental implications, because it drove the remarkable uptake of harambee (self-help) initiatives through which local communities could secure state funds for, say, a nurse or a teacher if they first grouped together to build a clinic or a school (Barkan & Holmquist Reference Barkan and Holmquist1989). Footnote 5 By allowing communities to specify their most pressing development needs, this system had the potential to respond to the demands of voters in a way rarely seen in authoritarian states—even if the reality did not always live up to the rhetoric (Widner Reference Widner1992). Of course, the system also provided President Jomo Kenyatta with a neat way of deflecting public pressure for public services away from the central government and onto the shoulders of individual MPs (Cheeseman Reference Cheeseman2006b).
The extent to which these kinds of linkage activities enabled ordinary Kenyans to feel that they had a stake in the system is perhaps best demonstrated by the impact that the rigging of party and general elections in the mid- to late 1980s had on the fortunes of the ruling party. Levitsky and Way note that “although incumbents may manipulate election results [in competitive-authoritarian systems], this often costs them dearly and can even bring them down” (2002:55). So it was in the Kenyan one-party state.
As David Throup has argued (1993), the decision of President Daniel arap Moi to introduce queue voting as a way of ridding himself of disloyal party leaders and MPs played a major role in the collapse of the single-party system. On the one hand, the violation of the secret ballot drew criticism from religious leaders and some members of the ruling party, undermining what had been one of the main pillars of the regime’s legitimacy. On the other hand, the blatant attempt to intimidate the members of the electorate by making them stand publicly behind their favored candidate, and hence expose themselves to potential retribution, backfired when many voters refused to comply. As a result, the subsequent announcement that some of the candidates with the shortest queues had won rendered the scope of the rigging clear for all to see, and represented an “emperor’s new clothes” moment for the ruling party (Widner Reference Widner1992). Shortly after the election, two of the leaders who had been the subject of election rigging, Charles Rubia and Kenneth Matiba, launched a campaign for the reintroduction of multiparty politics, which culminated in the repeal of Section 2A of the Constitution and the legalization of opposition parties.
The Legislative Arena
The flipside of Barkan’s work on elections and voters was his enduring interest in legislatures (see Mattes and Mozaffar, this issue). His work on political linkage had made it clear that MPs spent little time in Parliament not simply because they had little interest in national issues, but because their electorates expected them to be in the constituency delivering patronage and development. But this observation left three questions unanswered. First, who got elected and what did this mean about the composition of the legislature? Second, what did legislators do when they did spend time in the national assembly? Third, and relatedly, under what conditions was legislative strengthening—that is, the emergence of powerful and autonomous assemblies—possible in semiauthoritarian contexts?
Barkan’s answer to the first question was that the candidates most likely to win were those whose nonpolitical credentials suggested they would be successful at linkage activities (Barkan & Okumu Reference Barkan, Okumu and Barkan1979). As a result, there was a gradual change from one-party legislatures dominated by traditional leaders and teachers to parliaments that were composed largely of businessmen capable of using their wealth to undertake the kind of large-scale development activities that would impress constituents (Hornsby Reference Hornsby1989). One implication of this in single-party systems such as Kenya was that politics increasingly became the preserve of the wealthy and the well connected, excluding most of the ordinary citizens whose votes helped to put businessmen in power in the first place.
Barkan’s answer to the second question was that MPs could still have an important effect on distributional issues: the politics of who gets what, when, and how (Barkan Reference Barkan, Smith and Musolf1979b). More effective legislators were able to generate more roads, schools, and hospitals for their constituencies. As a result, inequalities began to emerge in the level of public services and infrastructure between those areas with better connected and more highly skilled political entrepreneurs and those who lacked political, economic, and social capital. In this sense, political linkage in some single-party systems was not so different from the situation in the U.S. House of Representatives, a point also made by Gertzel (Reference Gertzel1970).
The third question concerning legislative strengthening took up much of Barkan’s time in the later part of his career, culminating in the publication of one of the first books to focus exclusively on African parliaments, Legislative Power in Emerging African Democracies (2009a). In this book Barkan argued that parliamentary politics during the eras of one-party states and also under the compromised multiparty systems that succeeded them deserved greater scholarly attention. As Levitsky and Way put it, “legislatures tend to be relatively weak, but they occasionally become focal points of opposition activity” (2002:55–56). For example, although parliaments were rarely able to effectively check the power of the executive following the emergence of single-party systems, a degree of political contestation remained. During the “golden age” of the Kenyan Parliament in the 1960s, it was widely known that the ruling Kenya African National Union (KANU) party was split into two factions, “KANU A” and “KANU B.” Indeed, this division ran so deep that the dissenters sat on the opposition benches in the National Assembly and held their own caucus (Gertzel Reference Gertzel1970).
However, as the composition of the legislature changed, so did the potential for meaningful debate. On the one hand, the high costs of providing “development” for constituents meant that those who secured election tended to be financially indebted to businessmen and senior political figures, locking them into networks of patronage, and in many cases, corruption. In turn, this undermined the potential for “good governance” reforms (Barkan Reference Barkan2009b). On the other hand, voters who were frustrated with their marginalization from high political office could choose more radical representatives to travel to Nairobi and “speak truth to power.” In turn, the potential for grassroots concerns to be articulated through the election system led to the repeated reelection of figures such as Martin Shikuku, popularly known as the “people’s watchman.” Along with a group of more critical MPs that became known as the “Bearded Sisters,” Shikuku used his knowledge of the Standing Orders of Parliament to frustrate the executive and demand that key issues be debated (Cheeseman Reference Cheeseman2006a).
This sliver of legislative independence was important, because it sustained the idea that the National Assembly was supposed to be a vehicle for political contestation and scrutiny. It was no accident that following the reintroduction of multiparty politics, the Kenyan legislature emerged as one of the most vibrant on the continent, voting itself greater autonomy over its finances and organization (Barkan Reference Barkan2009b). This continuity suggests that the elements of competitive-authoritarianism that were present during the one-party era shaped the way in which multiparty competitive-authoritarian states evolved. Moreover, the popular memory of dissenting MPs, combined with the pressure on legislators to fulfill local expectations, came to play an important role in the evolution of the National Assembly.
Reflecting on developments in Kenya also led Barkan to argue that, under the right circumstances, the great financial pressure placed on legislators by voters could help to generate progressive coalitions of reformers and those he called “opportunists” (Barkan Reference Barkan2008:132). A good example of how such marriages of convenience work in practice can be found in legislators’ efforts to improve their own remuneration. As Barkan argued, “reformers recognize that better pay is an imperative step in professionalizing the legislature” because well-paid legislatures are more likely to be able to resist executive patronage, while opportunists “simply want more money and the possibility of fulfilling their constituents’ expectations” (2008:131). In this way, democracy can be advanced even in competitive-authoritarian contexts in which many of those in the legislature are not democrats (Cheeseman Reference Cheeseman2015). This conclusion nicely epitomizes both Barkan’s relentless spirit of realistic optimism and his awareness of the ways in which the political structures that he described could shape political outcomes.
Understanding Contemporary Africa
As demonstrated by the preceding discussion, the way in which politics operated under authoritarian rule is not just of historical interest: it helps to explain contemporary Africa. This is especially the case for the one-party states that Barkan studied, for three reasons. First, in Anglophone Africa the reintroduction of multiparty politics that occurred during the 1990s largely took place in the absence of far-reaching constitutional review. Thus the political systems that emerged in the 1970s and 1980s lived on into the multiparty period. In Kenya, for example, significant constitutional change was deferred until 2010, while in Zambia it has yet to take place. Second, in the rush to attract supporters in order to compete in hastily arranged national elections, many political entrepreneurs fell back on tried and tested methods of political mobilization. As a result, multiparty politics tended to reinforce, rather than to erode, previous forms of political linkage. Third, the kind of patron–client relations that emerged after independence shaped popular expectations of the appropriate behavior of legislators. This, in turn, acted as a constraint on political leaders in subsequent decades. Although the reintroduction of multipartyism led to fresh optimism and higher levels of engagement with the political system, what citizens demanded from their representatives remained largely the same.
The strength of these continuities demonstrates that for many countries the politics of the 1980s represents a helpful guide to the politics of the 1990s, and highlights the value of understanding the one-party state as a distinctive form of competitive-authoritarianism. As Levitsky and Way recognize (2010), referring to single-party systems in Africa as authoritarian, and the multiparty systems of the 1990s as some kind of partial democracy, obscures how similar they actually were. In reality, both are best understood as variants of the same competitive-authoritarian regime type: the multiparty systems of the 1990s simply allowed a greater degree of controlled competition at the national level. Indeed, it is important to keep in mind that for all the talk of political reform, the quality of democracy on the continent actually fell during the early 1990s, as presidents resorted to established authoritarian strategies in order to retain control of the political agenda (Bratton & van de Walle Reference Bratton and van de Walle1997). As a result, sitting presidents in Africa have won around 85 percent of the elections in which they have stood (Cheeseman Reference Cheeseman2010).
Of course, politics in these countries has also changed in important ways over the past two decades. While some states have become increasingly authoritarian, others have started to democratize. To what extent does Barkan’s early work help us make sense of the state of politics today? In 2009 Barkan categorized Burkina Faso, Burundi, the Central African Republic, Djibouti, Ethiopia, Gabon, Gambia, Togo, and Uganda as competitive-authoritarian states on the basis of the scores for political rights and civil liberties awarded to each country by Freedom House, an American think tank (Barkan Reference Barkan, Cooke and Stephen Morrison2009a). By this point, he believed that countries such as Kenya, which would have counted as a competitive-authoritarian state in the early 1990s according to his schema, had progressed to a more promising category that he entitled “aspiring democracies.” It is certainly true that, despite the shocking ethnic clashes that followed the flawed 2007 general elections, the political space available to the Kenyan media and opposition parties increased significantly after 2002, when KANU was finally defeated at the ballot box (Cheeseman Reference Cheeseman2008b). Thus, in the most basic sense, Barkan’s optimism about the prospects for democracy in certain African states has been borne out.
But what of the more fine-grained changes that have driven this uneven process of political liberalization and weakened the power of authoritarian leaders? To what extent does Barkan’s research on political linkage help contemporary students of Kenya, for example, understand the changes that have occurred to the structure of the government and its relationship to its people? Remarkably, Barkan’s work anticipated and helps to explain one of the most significant yet unheralded political trends in Kenya over the last forty years: the move toward decentralization. This process has gone through three main stages. First, in October 1980 President Daniel arap Moi announced that the government would “henceforth allocate its resources for rural development on a decentralized basis, to be more responsive” to its people (Barkan & Chege Reference Barkan and Chege1989:431). Moi argued that decentralization—by which he meant a shift of emphasis from the provincial level to the district level—was necessary to improve the efficiency and efficacy of service provision. Footnote 6 The change, he promised, would “create for the people and their chosen representative a whole new world of opportunity” (quoted in Barkan & Chege Reference Barkan and Chege1989:431).
While acknowledging that decentralization had the potential to bring economic benefits, Barkan, writing with Michael Chege, saw through Moi’s rhetorical justification for the new District Focus for Rural Development policy. Most significantly, he understood that the main driving force behind the reforms was not developmental, but political. One reason that Barkan was so quick to identify the deeper motivations at play was that they were rooted in issues that he had previously written about, namely the harambee system of development and the form of political linkage constructed under Kenyatta. The localization of politics under the one-party state had institutionalized a system in which political responsibility for development had been relegated to the constituency level. Shifting administrative responsibility from the province to the district brought the civil service into line with this vision. At the same time, the new policy was easy to communicate to Moi’s supporters because it resonated with his preindependence commitment to a “majimbo” (regionalist) constitution in which the country’s different ethnic groups would enjoy a degree of self-government.
In addition, Moi was desperate to break up the administrative and political networks that had grown so strong under Kenyatta’s rule. He was motivated to do so because a considerable proportion of Kenyatta’s fellow Kikuyu elite had sought to block Moi’s path to power, advocating a change of the constitution in a failed attempt to prevent the vice president from succeeding Kenyatta upon the latter’s death in 1978. Moi also feared the wealth and influence that these networks represented, which far exceeded his own. Footnote 7 As a result, Moi spent the next decade carefully restructuring the state in order to redistribute power and resources to his own supporters. As Barkan and Chege wrote in their evaluation of District Focus—published seven years after the policy was announced—“Moi’s populist mode of governance . . . has had its intended effect of circumventing the influence of the most senior politicians of the Kenyatta era, especially those from the Central Province” (1989:437–38). Footnote 8
The determination of Moi to use political reform as a vehicle through which to gain greater political control meant that in reality power was not devolved away from the center but was instead “deconcentrated.” In other words, rather than creating more autonomy for local operatives, the changes introduced through District Focus led to “the posting of greater numbers of more central personnel to an expanded number of field offices to exert greater control over development initiatives on the periphery” (Barkan & Chege Reference Barkan and Chege1989:433). Thus the state that emerged under Moi represented a significant political reorientation but precious little liberalization, as the reach of the central government was extended ever further. Barkan’s sharp understanding of the complex motivations underpinning decentralization did not lead him to underestimate its significance. He concluded, in fact, that the nature of political linkage in Kenya, which generated strong popular support for the localization of politics, meant that the “highly centralized system of the Kenyatta era . . . is gone for good” (Barkan & Chege Reference Barkan and Chege1989:452). This analysis, which was written prior to the reintroduction of elections in the 1990s, and a full quarter of a century before constitutional reform ushered in genuine decentralization in 2010, proved to be remarkably prophetic, as we shall see.
The second stage of decentralization occurred following the defeat of KANU by Mwai Kibaki’s National Rainbow Coalition (NARC) in 2002, more than a decade after Barkan and Chege’s analysis was published. The key change during this phase was the creation of a Constituency Development Fund (CDF) worth 2.5 percent of government revenue which gave all MPs new resources to be used for the development of their constituencies. Footnote 9 This was officially billed as a measure designed to break up the patronage networks that had sustained the KANU regime, because it went hand-in-hand with the introduction of new legislation that banned politicians from making personal donations at harambee meetings. It was also said to have the potential to enable a more equitable distribution of resources, as 75 percent of the total fund was to be allocated equally among all 210 constituencies, with the remaining 25 percent allocated according to constituency poverty levels. But in reality the CDF was not transformative; rather, it served to entrench the key features of the Kenyan system of political linkage that Barkan had identified in the 1970s (Cheeseman Reference Cheeseman, Whitfield and Mustapha2008a). Not only did it identify MPs as the agents of development, but it gave them even greater funds with which to perform this role and strengthen their patron–client ties in their constituencies. Given this, the CDF is best interpreted as an attempt by MPs to increase the resources available to them, while simultaneously responding to the pressure that they were regularly placed under by constituents. In this sense, it represented a rational response to the political context that Barkan identified in his early work.
The third main stage of decentralization occurred some eight years later in 2010, when the country voted to accept a new constitution that created forty-seven counties, each complete with its own governor and senator. It is too early to evaluate the results with certainty, but these new positions have been granted sufficient authority, resources, and constitutional protection to suggest that power has, for the first time in Kenya’s postcolonial history, been deconcentrated (Cheeseman, Lynch, & Willis Reference Cheeseman, Lynch and Willis2015). If this is true, then Barkan’s early predictions will turn out to be only partly true when it comes to this most recent development. Footnote 10 As discussed above, in his analysis of District Focus, he was decidedly pessimistic about the prospects for the meaningful decentralization of power in the Kenyan context. On this issue alone, then, Barkan may have underestimated the capacity of the centrifugal forces let loose by the particular form of political linkage practiced in Kenya to undermine the hegemony of the central government.
To be fair to Barkan, though, he identified greater potential for democratic reform than most commentators did in the dark days of the 1980s. Against the backdrop of a heavily repressive competitive-authoritarian regime, it seemed unfeasible that the government could be persuaded to create a further tier of political leaders capable of challenging central authority. Even after the flawed presidential election of 2008 gave way to a new mood of national reconciliation and a growing recognition of the value of reform, few researchers predicted that President Kibaki would be prepared to preside over such a dramatic reduction in the powers of the central government. Footnote 11 In hindsight, the ratification of the 2010 Constitution appears to have been the product of a rare combination of events that would have been unforeseeable just a few years before: the postelection violence in 2008, the power-sharing agreement that ended it, and heavy international engagement to encourage a new political system that would be better placed to manage interethnic tension (Kasfir Reference Kasfir, Förster and Koechlin2015). It therefore seems reasonable to conclude that the introduction of genuine devolution in Kenya was one of those events that are very hard to predict ahead of time, like the collapse of the Soviet Union or the Arab Spring—albeit on a much smaller scale. In turn, this serves as an important reminder that political science is much more successful at predicting and explaining continuity than change.
This feature of the discipline makes it all the more remarkable that—in contrast to most commentators brave enough to make predictions—Barkan was right much more than he was wrong. His analysis of the direction of the Kenyan legislature, for example, continues to look sound. In his book of 2009 he identified the Kenyan National Assembly as one of the most vibrant and promising legislative bodies on the continent. Part of the reason he held this view was that his early analysis had correctly identified the Kenyan political system as one of the more responsive precisely because the structure of political linkage placed great pressure on MPs to respond to the priorities of their constituents. Although significant problems of corruption and weak scrutiny remains, these judgments still hold true. Although Kenya is still traveling down what Kennedy Opalo has called the “long road to institutionalization,” Parliament now enjoys relatively formal independence from the executive (2014). Indeed, the 2010 Constitution strengthened Parliament in three key ways: It created a second chamber—the Senate—to represent the interests of the forty-seven newly created county governments; it removed ministers from the National Assembly so that they could not directly influence proceedings on the floor of the house; Footnote 12 and it granted Parliament new powers to vet presidential appointments to key positions. To date, the new Assembly and Senate have failed to make the most of these new opportunities due to the capacity of the executive to co-opt legislators and the poor discipline of opposition parties, but the potential for a much more effective and assertive legislature exists. Moreover, the growing number of political heavyweights who are first and foremost dependent on their constituencies and counties for their posts means that local grievances are increasingly likely to be translated into legislative concerns.
Conclusion: The Importance of Institutions and Institutional Reform
One of the most valuable legacies of Barkan’s research is the contribution that he made to our understanding of African political institutions and their implications. At a time when it was more common to argue that African politics was effectively “institutionless” (Chabal & Daloz Reference Chabal and Daloz1999), he took legislatures and electoral systems seriously. In doing so, he helped to lay the foundations for the current trend toward more institutionalist analyses of the continent. Much of the recent research on African legislatures (Chaisty, Cheeseman, & Power Reference Chaisty, Cheeseman and Power2014), legislators (Lindberg Reference Lindberg2010), and public opinion (Bratton, Mattes, & Gyimah-Boadi Reference Bratton, Mattes and Gyimah-Boadi2005) owes a great debt to the body of work that he produced.
Barkan’s work is also important because it demonstrates the democratic aspects of the continent’s one-party states, and the impact that these have had on the process of political liberalization. As I have argued elsewhere,
Variations in the institutional structure of authoritarian rule in the 1970s and 1980s . . . shaped the different pathways that countries took to multipartyism in the 1990s. . . . Moreover, competitive forms of authoritarian rule were more likely to have developed norms in favor of representative government and institutions capable of maintaining their independence from the executive. (2015:53–56)
Given this institutional background, an appreciation of the distinctive competitive-authoritarian elements of civilian one-party states is critical to an understanding of their democratic potential. Barkan understood this point better than perhaps any other Africanist.
Moreover, while the form of political linkage he described is particularly pronounced in Kenya, this is not just a Kenyan story. Recent studies in other countries have also found that voters first and foremost want their representatives to plug them into public services. For example, in 2013 Weghorst and Lindberg published the results of a Ghanaian survey that asked questions very similar to the ones that Barkan had posed some thirty years before, and their analysis came to very similar conclusions. As in Kenya, voters in Ghana mainly want their MPs to provide localized goods and services such as schools and health clinics (which Weghorst and Lindberg refer to as “collective goods”). The similarities between their analysis and Barkan’s publications some twenty years earlier suggest that the demand for political linkage, and the focus on the developmental performance of MPs during legislative contests, is not specific to Kenya but is a common feature of politics in states that employ first-past-the-post electoral systems.
The implications of Barkan’s research for how we understand contemporary debates around African politics are therefore profound. As Barkan argued, the types of political linkages that are constructed to connect citizens and governments shape the way in which political systems operate and the extent to which voters feel connected to the political system. This has significant implications when it comes to the costs and benefits of different types of constitutional design. Consider the choice of electoral system. It is often said that a system of proportional representation (PR) would benefit African states because by curtailing constituency-based elections it would reduce the pressure on MPs to meet the financial demands of constituents, ensure greater levels of ethnic inclusion, and, by enabling party leaders to decide the order in which MPs are ranked on the party’s list of candidates, promote internal party discipline (Barkan 1998).
However, if we follow Barkan’s analysis we can also see that introducing PR into countries that currently hold first-past-the-post (FPTP) constituency-based elections would also have negative side effects. For one thing, one of the most effective forms of linkage between the political center and the periphery would be undermined, which would almost certainly make the political system less responsive to local concerns. There is also a real possibility that the introduction of PR would make the emergence of reform coalitions interested in legislative strengthening less likely.
According to Barkan’s work, the creation of strong reform coalitions is most likely under two conditions: “when the ruling party and opposition parties hold nearly equal numbers of seats,” such that “a majority coalition can be formed by the opposition in alliance with a modest number of the ruling party’s backbenchers” (2008:133); and when the discipline of the ruling party is low, thus facilitating coordination across party lines. The implication of these findings is that the impact of PR on legislative strengthening is likely to be mixed. By reducing the magnitude of victory of the ruling party, proportional elections have the potential to promote reform coalitions. But to the extent that PR also strengthens the hand of party leaders over their own parties, it also makes intraparty rebellions, and the coordination of efforts across party lines, less likely. Given that proportional systems may also make it harder to give citizens a sense that they have a stake in the system, the benefits of PR over FPTP are questionable.
The structure of political linkage also has important implications for how civil society groups and donors should go about building democracy in Africa. For example, the pressures that MPs face from above and below to focus their energies on “linkage” issues mean that efforts to strengthen African legislatures by offering them training, workshops, and learning experiences in Western parliaments are likely to be unsuccessful. As Barkan recognized, a more fundamental change in the position of MPs within the wider political system is necessary before they can be expected to devote the majority of their time to dealing with national policy issues on the floor of the chamber.
Barkan’s insights into the nature and significance of political linkage in Africa can also help explain some of the political outcomes that are of most interest to contemporary political scientists. Take the current debate about the prevalence of neopatrimonialism on the continent. Footnote 13 Variations in the extent of patrimonialism can be explained, in part, by the forms of political linkage that were constructed in the late colonial and early postcolonial eras. For example, the highly localized and personalized linkage established by Jomo Kenyatta in Kenya promoted an extreme type of patron–client politics that entrenched neopatrimonial tendencies within the political system. By contrast, countries such as Zambia developed less pronounced versions of “Big Man” rule because they featured the construction of political linkage arrangements in which resources ran through the party committees and key interests groups—trade unions, in the Zambia case—rather than just individuals (Cheeseman Reference Cheeseman2006a). In this way, the study of political linkage has much to tell us about why, despite similar colonial experiences, not all African states are equally “neopatrimonial.”
The way in which Barkan conducted his research also has much to tell us about how we should approach the study of Africa. The body of work that he produced would not have been possible without his commitment to understanding the deeper political structures through which power and influence flow. He was one of a small number of foreign scholars who continued to believe that the study of African politics could benefit from methods and ideas developed in North America and Europe, and vice versa. It was precisely because he conceived of voters and legislators as rational actors whose decisions are shaped by political and incentive structures that he asked so many questions that remain at the heart of African studies, and stayed cautiously optimistic about the possibility for democratic change throughout his life. To stay true to his legacy we must insist on the value of the study of political institutions, and on the relevance of Africa for comparative politics more generally.
Acknowledgments
I would like to thank Nelson Kasfir, Ella Kusnetz, Gabrielle Lynch, Justin Willis, and three anonymous reviewers for their useful comments, which have significantly strengthened this article. Thanks also to those who came together to commemorate Joel Barkan’s life and discuss his scholarship at the annual meeting of the African Studies Association in November 2014. Some of their thoughts and ideas are reflected here, but any errors remaining are mine.
