The Argument
Although this study builds on elements of the views outlined, it departs from existing explanations by focusing on the importance of institutional design for political behavior, and particularly on the incentives that institutions create for the content and character of bargaining between political elites. What the Kenyan case shows is the critical importance of variables that have nothing to do with the persistence of kinship-based systems of social and economic organization—or with choice of economic policy per se.
Anyang' Nyong'o argues persuasively that in the case of Kenya, "it was largely the disintegration of the nationalist coalition that enabled a strong authoritarian president to emerge." He suggests that "a section of the nationalist coalition favoured this rise, seeing in it an opportunity to have access to states apparatuses and thereby acquire avenues for capital accumulation and personal enrichment." The dominant faction maintained the party when it suited its interests to do so, and then allowed it to "atrophy to deny any other organized faction . . . from using it politically to attain its objectives within the bounds of law."[42] This element of the coalition was able to exploit these opportunities and to become the core of a new bourgeoisie. The willingness to cede power to a single party in order to gain short-term economic benefits ultimately worked to the disadvantage of this new group. It began to lose control over the presidency and access to the state.
Anyang' Nyong'o's explanation stops short of a full account, however. It does not distinguish between the authoritarianisms (plural) that have governed Kenyan society during the first thirty years of the country's independence. It notes quite correctly that elements of the nationalist coalition used state resources to accumulate resources and to move themselves into the upper reaches of the petite bourgeoisie or the matajiri ("wealthy") class, dividing into factions as they competed for access to opportunities to acquire new assets. But Anyang' Nyong'o offers few explanations for the difficulty these groups experienced in creating a unified front and in opposing the president when he trampled on their interests. Why could they not control the authoritarian system whose creation they had sanctioned?
Answering this question requires an excursion into the ways in which political institutions shape interest-group formation and behavior. I propose that the shift from single-party dominance to the party-state, as it is defined here, depends on the concurrence of three main conditions.
1. In countries where ethnic, regional, and economic divisions coincide, as they do in most of sub-Saharan Africa, a shift from single-party dominance to a party-state is likely unless electoral rules or informal, extraparliamentary institutions force elites to bargain across boundaries in their efforts to secure winning coalitions.
The tendency of political elites to make unlimited claims in the public arena varies with the perception that other elites' demands are efforts to propel the interests of a single community ahead of those of others on a continuing basis. Where demands for economic redistribution to disadvantaged regions or for improvement of the competitiveness of a particular sector of the economy appear to be demands for communal advantage, the "rules of the game" Zolberg identifies are shattered and the likelihood that those in power will seek to limit opposition through use of the party as a vehicle for social control rises. Only where formal or informal institutions of government—electoral rules, for example—force bargaining across ethnic, regional, and economic boundaries will compromise appear individually as well as collectively rational.
It is important to distinguish ethnic and religious lines of division as understood in this proposition from the lineage-based social organization that figures prominently in the "weak states, strong societies" argument. In most parts of sub-Saharan Africa, the larger ethnic divisions that have become politically salient in recent years are new forms of
organization—political machines that employ ethnic idioms designed to draw members of different lineages and clan units into a political bloc, whether or not their members speak the same language and share the same systems of social organization.
These ethnic welfare societies or machines are often springboards for political careers. Where members occupy the same positions in the economy or a specific geographic area, the languages of class and community often intermingle, giving demands a cast of exclusivity. Where members occupy different regions or different positions in the economy, candidates may try to appeal to shared community membership as a way of distracting voters from the economic differences between them, creating a kind of "smoke screen."[43] Whether these kinds of bases and divisions promote political division along communal lines depends entirely on the incentives for bargaining a party system creates.
2. The shift to a party-state occurs when factions within the dominant party cannot constitute strong opposition to ascendant factions, whose members attempt to curtail political association or other civil liberties, because of the difficulty of organizing constituents around issues that generate no divisible benefits but improve collective welfare generally. Most civil-liberties issues are of this type. It is especially difficult for members of opposition factions to organize against the fusion of government and party in most African settings (a) because single-party or single-party-dominant systems decrease the likelihood that citizens will cast their votes for politicians who promise general policy changes rather than discrete projects, and (b) either because there are few people who have a disproportionate stake in these questions or there are few who can bear the exceptional costs entailed in organizing citizens in support of changes in government. Where one or more of the opposition factions have significant sources of money and compensate lawyers or public-interest groups for their efforts in fighting changes in government-party relations in court or in the international arena, the shift from single-party dominance to a party-state is likely to take place more slowly and possibly less completely than in countries where opposition factions are fragmented and financially weak.[44]
There are three parts to this observation. First, the argument centers on the difference between "maintaining the competitiveness of politics" as a campaign issue in contrast to pork-barrel politics, or demands that deliver divisible benefits. Some kinds of policy choices generate rewards for the particular individuals who lobby for them. For example, pro-
grams to build dispensaries or hospitals, irrigation schemes or market centers, sewer systems or access roads can be designed to benefit those who sought their passage in the legislature or who voted for candidates who incorporated these projects into their platforms. Because the benefits of these policies are divisible, it may pay members of Parliament to organize voting blocs to secure the required policy changes—to lobby their fellow representatives or political elites—provided the costs of the individual effort required are lower than the present discounted value of the possible gain. One might call these "type 1" issues.
Not all policy issues have this character, however. Many kinds of political reforms have the character of public goods; that is, once enacted, they benefit large numbers of people without regard for the level of individual contribution to the effort to secure their passage and implementation. They permit "free riding." Because of this characteristic, individual voters and the politicians who claim to speak for them have relatively little incentive to invest their efforts in overcoming opposition and in pushing reforms through the legislature. Specifically, because it is difficult for those who have expended time and money or accepted exceptional risks to exclude those who did not contribute from the benefits of reform and to procure the benefits for themselves and their constituents, special incentive is required to impel a politician to stand up for such measures. In the absence of such incentives, each individual will simply wait for another to move first to make the investment required to generate the reform—and in many cases nothing will happen. This kind of behavior may occur even when the anticipated individual benefits generated by the policy or reform exceed the costs of participation. One might call these kinds of policies and reforms "type 2" issues.
Under most circumstances, efforts to defend the level of public contestation in a political system are efforts to pursue "type 2" issues. Creation or protection of the legal ability to associate and to speak at public gatherings, to travel within and outside of the country, and so on, are public goods in the technical sense of that term.
Differences in institutional settings account for much of the variation in the willingness of political actors to accept the burdens of general reforms that affect the entire country and that potentially create large numbers of winners but a few powerful losers. That is, the structures of formal and informal representative systems can create special incentives for the provision of such policies and help overcome the "free-rider problem" political economists so often see as an obstacle to organization.
Elected elites can be assumed to act so as to preserve or extend their access to resources they can use for "constituency service" or local projects. These increase candidate visibility with constituents. In competitive political systems, where parties may alternate in power, political elites have an incentive to defend their "political space," their freedom to speak, associate, campaign, and otherwise contest public policy vigilantly. If they allow their ability to participate to be curbed, they stand to lose control over resources, either in the short run or when their party loses power.
The character of the party system makes a difference for political outcomes in three ways. First, competitive party systems help provide incentives for politicians to bear the costs of organizing reform platforms. Parties are different from factions in that they have independent organizational bases and personnel to carry out administrative tasks. The party leadership has a stake in the continued existence of the organization. That stake defines an interest in maintaining the "political space" the system of representation accords the opposition. Furthermore, the leadership can turn around and demand that candidates carrying the party banner contribute to the effort to pursue the watchdog and reform activities implied by that interest. That is, they can demand that candidates claiming affiliation with the party take time out from the pursuit of purely local interests or even attenuate local demands and devote themselves to more general, "type 2" issues as a quid pro quo of membership,[45] as V. O. Key noted in his study of the "single-party states" of the Old South in the United States.
Second, competitive party systems and single-party systems differ in the kinds of costs they create for those who try to organize to increase the level of public contestation within a political system. Although it is the free-rider problem created by general political reform as a "type 2" good that is central to the analysis offered here, it is also the case that single-party systems change the cost structure associated with organizing reform. In single-party systems, members of factions must tread lightly in defending their political space for fear of losing favor with the party leadership; absent a credible threat to defect to an opposition party, a candidate is likely to find that access to patronage during election periods diminishes; or the leadership may launch proceedings to expel a fractious candidate from the party altogether. Although faction members have an incentive to defend their political space in single-party systems, they face risks in doing so that candidates in competitive party systems do not encounter.
Third, the factions that proliferate in most single-party systems also provide a poor base for pursuit of policy changes at the national level in another respect.[46] In some one-party systems, a dominant faction is able to secure the patronage to construct and operate a political machine. This analysis is often used to understand African single-party systems.[47] Not all African single-party systems operate in this way, however. Many display high degrees of factional division—to which leaders often point as evidence of pluralistic competition. Factional division is especially common where mobilization across political divisions is difficult as a result of legal restrictions, geographic constraints, or significant differences in economic activities, and where the public treasury is so poor that patronage resources are few.[48] Indeed, it could be argued that such systems are increasingly the dominant form of one-party system in Africa. Where there is a high degree of factional competition and no clear and enduring organizational base for any subgroup, candidates are less likely to take a stand on a reform issue than they are simply to break with one ad hoc group and move to another. They have little incentive to demand changes in the positions held by other faction members. To do so would take investment of time, if not of money, and the creation from scratch of channels for articulating and discussing demands. Unless there is some additional factor that limits their ability to move between factions or otherwise encourages politicians to acquire a stake in the policy positions promulgated by one group and in the defense of the group's ability to participate, it will prove extremely difficult to constitute any sort of "watchdog" activities to make sure that those in power respect the desire of others to participate.[49] In general, "exit" will be preferred to "voice."
The observation made in point two also suggests that there is an endogenous element to the explanation of the shift to a party-state from a single-party system. That is, when a country moves from a multi-party system to a single-party system, it sets in motion a process in which, under the socioeconomic conditions that prevail in most sub-Saharan cases, the reduced incentives for political elites to support high levels of public contestation lead to passage of legislation that further weakens the ability of politicians to defend their political space. Whatever the pattern of factionalism that evolves within the party—a two-way split in which one faction monopolizes patronage resources or a high degree of fragmentation—once legislation that reduces opportunities for contesting policies goes into effect, the fusion of party functions with the administrative responsibilities of the state becomes increasingly difficult
to resist. Increasing political "departicipation" and disfranchisement are an inevitable result of the incentive pattern created.[50]
In this way, the creation of single-party states can contribute to or precipitate the weakening and even disappearance of the social groups that are major claimants to state resources in other parts of the world. Although it does not necessarily affect the positions of "predatory elites," the creation of such a state may make it more difficult for the elements of a "civil society" to coalesce. It may thus contribute to what Thomas Callaghy has called sociopolitical "shredding" and may make it easier for a head of state to initiate and secure acceptance of legislation of whatever type or content.[51]
Political repression also makes resort to this form of party-state much more likely, however. If political scientists can speak of legislative and bureaucratic structures enacting their environment or shaping interest-group structure and behavior, it is also the case that the party-state shapes civil society (by destroying it) and creates its own rationale. When a government has already moved to proscribe some forms of participation, opposition goes underground and becomes more difficult to monitor, save by resorting to obtrusive methods of surveillance and control. The organizational structure of the party, with its network of village representatives and local cells, becomes much more efficient as an instrument for gathering information during these periods than does the administration, with its less extensive network. For example, it is possible to mobilize youth wings to listen in marketplaces for rumors of political meetings. Further, the party-state provides a vehicle for indoctrinating potential opposition supporters and for communicating to them the high cost of deviation from the party line.
3. Finally, this argument explains variations between the African cases and other countries and variations among the African cases as a consequence of differences in the array and economic power of different social groups. Where opposition faction leaders can bolster their bargaining power with the spokespeople for the "in" faction by drawing on private resources for campaigns and for patronage, or where they can defend themselves against persecution by acquiring allies in an international forum, the shift to a party-state is likely to occur more slowly and less completely than in other situations.
Thus, movement toward a party-state form of governance is less likely when there are extra-parliamentary interest groups that have independent financial bases or occupy critical positions in the economy
that they may translate into bargaining power. The rise of a party-state is more likely (a) when the absence of private business or professional opportunities endangers the ability of opposition politicians to bring court cases contesting restriction of their political space, and (b) when either the absence of a free press or the absence of access to international ties by opposition figures makes it difficult to bring pressure on governments that restrict political space despite judicial decisions against them.
The construction of a party-state requires modifications to a country's constitution and usually to the constitution of the dominant political party as well. Although the modifications are likely to take place without open amendment of the documents concerned, these steps can be challenged by opposition leaders as long as these men and women have independent economic bases and so are not vulnerable to efforts by the incumbents to retaliate in economic terms, and as long as they can take their appeals to an independent judiciary, similarly protected. Furthermore, judicial decisions and requests must be binding, so that incumbents are forced to pay attention to them. In most cases, that assumes either a free local press to publicize violations or ties between opposition leaders and an international community capable of denying the incumbents critical resources unless they adhere to the letter and spirit of rulings.
To protest a restriction of civil liberties in court is to risk losing, and where the defendants, the government's managers, control most economic opportunities as well as civil service jobs, to lose is to lose big. Even where associational life has largely disappeared because of political repression, a public-spirited individual might be willing to challenge an administrative ruling or a law that blocked pursuit of personal livelihood. After all, in every society there are individuals who generate "public goods" for reasons of their own. If defeat in court would jeopardize the ability of an individual to provide for a household or ruin the future life chances of children, the probability of protest through the judicial system is much lower. Where the state plays a substantial role as an investor and manager of enterprises, or where it maintains substantial power over private employment through extensive licensing regulations, a country's governors can wield devastating power over plaintiffs. The size of the "private sector," the degree of concentration of holdings within that sector, and the level of economic diversification can thus strongly influence the willingness of citizens to press judicial challenges. That should come as little surprise. Historically, democratic political structures have always arisen in countries with "capitalist" eco-
nomic systems, or, more accurately, with substantial private economic activity.[52] (Not all capitalist systems or mixed economies are accompanied by democratic systems of representation, however.) African countries vary greatly in the degree to which they permit private economic activity without state intervention.
The existence of groups in civil society with economic "gatekeeping" powers, or the ability to diminish the state's capacity to function if a head of state tries to circumscribe political space, is important too. Only protest by independent domestic interest groups or critical elements of the international community that has clear material consequences is likely to discourage a head of state from passing new legislation to evade court rulings in cases individual plaintiffs have brought. Barrington Moore's analysis of the different political paths taken by industrializing countries makes clear the importance for "democratic outcomes" of groups in civil society, usually a commercial bourgeoisie or independent aristocracy with sufficient economic clout to be able to bargain with government leaders and counter the exercise of military or police powers.[53]
Although Kenya is shorter of land than most other sub-Saharan countries, it has generally offered greater latitude to private-sector economic activity than have most. That a party-state was in the making there in the 1980s is one of the puzzles this study seeks to solve.