Thursday, October 7

Hans Peter Schmitz

Post-Doctoral Fellow

The Human Rights Program

This paper was originally presented under the title "Structure and Agency in the Study of Political Regime Change: Kenya and Uganda in Comparative Perspective" at the Annual Meeting of the American Political Science Association, Atlanta, September 2-5, 1999.

Abstract

This paper discusses the role of international norms in influencing recent processes of regime change in two East African countries, Kenya and Uganda. In the first part, I present a review of the pertinent literature and conclude with an argument for a more systematic integration of international factors into explanations of regime change in general, and non-material issues such as norms and ideas in particular. On the agency side, attention is mainly focused on the role of materially weak networks of non-governmental organizations active in the area of human rights and democratization. In the second part I use the comparison of two case studies to show under what conditions these norms and their entrepreneurs matter for domestic political change. For this purpose, I distinguish between a phase leading up to the initiation of regime change and the subsequent period of political transition. I conclude that current processes of regime change are best understood as conflicts between domestic and international institutions and their respective entrepreneurs and/or beneficiaries.

A combination of domestic and international factors accounts for significant differences in transition paths. After 1991/92, Kenya moved along the path of liberalization and increased competition, while the Ugandan transition after 1986 shows a preference for popular inclusiveness. The comparison across the two cases highlights the crucial importance of appropriate strategies for each phase of the transition process. Both on the domestic and international level, the shift from delegitimizing an authoritarian regime towards creating and legitimizing a new political order requires the development of new and different skills for governmental and non-governmental actors alike. While NGOs and civil society at large generally profit from such a shift, this does not necessarily translate into an automatic increase of influence on the course of events.

The subject of this paper is regime change understood as the modification of broadly defined political institutions and practices governing domestic politics. Democratization is a possible outcome of such a process if the modification of political institutions and practices reflects convergence with a particular set of liberal ideas and norms. During the last 25 years the interest of political scientists in such processes has steadily grown as the subject matter experienced a number of significant theoretical and empirical innovations. The democratic peace argument put forward in the international relations community led to a growing theoretical and normative interest in the conditions for democratic rule. Empirically, the number of suitable cases sharply increased when democratization spread from Southern Europe in the 1970s to Latin America in the 1980s and finally reached Eastern Europe and the other continents in the late 1980s and early 1990s. More recently, some have diagnosed an ebbing of the wave. While the empirical record of the third wave (Samuel Huntington) is mixed, its profound effects on the way the academic community studies democratization are undeniable. The new cases of democratization led during the 1980s to a paradigm shift away from structuralist explanations towards more contingency-driven and agency-based approaches. Challenges to the modernization school or cultural explanations of democratic change highlighted cases of regime change under structurally unfavorable conditions and called for greater attention towards the role of political actors, institutions, and contingencies.

While the regime changes in Southern Europe and Latin America opened in the 1980s a new intellectual terrain, this literature received a major boost as a result of the 1989/90 events in the Soviet Union and Eastern Europe. The third wave now affected many countries of the world posed not only a serious puzzle for structuralist approaches, but challenged also the separation of international and domestic politics, both in theory and practice. Concepts such as contagion or diffusion were now introduced into the debate to account for the new phenomenon. However, the reference to the international realm remained largely rhetorical. In light of the new research agenda within comparative politics and the abundance of cases, an additional set of international factors looked to most scholars much less attractive. Such a view was compatible with disciplinary boundaries and the academic training of many scholars.

This position further solidified when much of the initial clamor for fundamental political change in many parts of the world was lost in the mid-1990s. The temporal and directional uniformity implied in the picture of the third wave gave way to the reemergence of regional and individual diversity. Some countries completed the transition within weeks or months and can even be considered as consolidated democracies today. In other cases, the process has been much slower, stopped or was even reversed. In the worst cases, countries returned to authoritarian rule or even plunged into civil war. As a result, the emerging global picture with regard to democratization seemed to vindicate the regional and domestic bias within the discipline.

This study contributes to this literature by presenting a comparative evaluation of the recent regime changes in Kenya and Uganda. The paper follows the recent emphasis within the democratization literature on the role of political actors and institutions in bringing about and shaping regime change. It adds to this literature a specific concern for international human rights institutions and actors as potential sources of change. I argue that the international institutionalization of human rights norms represents an opportunity structure for international and domestic human rights actors. They use those norms and institutions to challenge the prevalent norm of state sovereignty and succeed in delegitimizing human rights violating regimes.

The paper evaluates the role of those international factors, first, in causing regime change and, second, in shaping the subsequent transition path. Variation in the results expressed by the mode of initial transition is explained for the onset of regime change by the dominant form of neopatrimonial rule. Whereas cohesive clientelist networks solidified in Kenya over almost 20 years, the Amin dictatorship and subsequent civil war destroyed the bases for such a system in Uganda. The dominance of positive integration by resource allocation in Kenya contrasted with a prevalent logic of negative integration expressed by violence and indiscriminate mass killings in Uganda. When transnational human rights mobilization affected both countries, the international-domestic interaction produced in Uganda a revolution from below (1986) and in Kenya limited reforms from above (1990/1991).

Variation in subsequent transition paths is mainly accounted for by the named initial mode of transition that affects the international and domestic mobilization. When the Kenyan government chose a path of limited electoral liberalization (extending competition) international human rights pressure remained strong and continued to challenge governmental repression. By contrast, the new Ugandan government initially emphasized a popularist transition path (extending inclusiveness) and engaged in elaborate constitutional reforms instead of allowing for multi-party electoral competition. A decline in human rights activism explains why this reform process was delayed and the government was able to extend its rule for more than 13 years. Rather than limiting institutions to the role of constraints upon actor's choices, this argument highlights the enabling role of institutions and identifies the potential conflict between norms embedded in different institutional settings as a source of social and political change. Hence, the cases generally confirm Dahl's claim that the sequence of transition (1) and the specific way of inaugurating a new regime (2) affect subsequent transition paths. However, the Ugandan case study tentatively questions his pessimism with regard to revolutions from below and cases where inclusiveness precedes the extension of (electoral) competition.

The paper will proceed in four steps. The following section contains a brief discussion of theoretical and conceptual questions. It follows a short introduction into the cases by way of presenting the explanandum, the extent and direction of regime change in Kenya and Uganda. In the subsequent main empirical part, the paper will evaluate those changes comparatively and present an explanation. In the final section, the paper will address alternative explanations and assess the contribution of this study to the literature on democratic transitions.

In this section the paper proceeds to clarify a number of theoretical and conceptual issues pertaining to the main subject of this paper, the comparative evaluation of regime change in two East African countries. Based on an initial critique of existing approaches to regime change, the second part of this section will present the conceptual tools that are suggested to supplement the current state of the art. The argumentation starts out with a rejection of a materially biased structuralism (modernization theory) and moves on to a critique of the agency-based literature on democratic transitions. In order to solve some of the remaining puzzles the paper suggests to supplement an agency-centered approach with an institutionalist perspective that puts the interaction between existing domestic arrangements and international institutions at the center of analysis.

Understanding the Effects of Modernization

Surprisingly enough, it has taken the social sciences several decades to fully understand the meaning of Seymour M. Lipset's original claim that "the more well to do a nation, the greater the chance that it will sustain democracy" . In the heydays of the modernization theory, many researchers and practitioners derived from this perspective a almost absolute dominance of economic over political development. Although the modernization paradigm was faced with many empirical and theoretical challenges from the start, it became for decades the dominant mode of conceptualizing the relationship between the economic and the political sphere in developing nations. During this time, the focus of research was almost exclusively on identifying the most 'efficient' linkages (e.g. urbanization, education, mass media, the development of a middle class etc.) promoting democratic change. More recently, a number of researchers have expressed increasing doubts both with regard to the interpretation of Lipset's and others empirical work and the causal path underlying claims based on a modernization perspective.

With regard to the causal path linking economic and political development, the number of studies reversing causality and claiming a positive effect of democracy on economic development slowly increased during the 1980s and 1990s. Przeworski/Limongi argued that this trend was a natural result of the global ideological changes in the late 1980s . For the purpose here it is sufficient to conclude that the original confidence about the force of economic development as a structural determinant of political change has disappeared. This is complemented by more recent quantitative studies attempting to reproduce the original claim of the modernization paradigm. Londregan/Poole maintain "a small but statistically significant effect" of increases in income . Vanhanen concluded from his work that economic development "is only an intervening variable that correlates positively with democratization because various power resources are usually more widely distributed at higher levels than lower levels of socio-economic development" . Arat maintained that "on the basis of these findings it can be concluded that increasing levels of economic development do not necessarily lead to higher levels of democracy, even for the less developed countries" .

Based on these more recent results investigating the relationship of economic and political development, Przeworski/Limongi (1997) suggest a more subtle reading of Lipset's original claim. While both authors accept that evidence about the generally positive correlation of economic development and democracy exists, they reject claims that economic development automatically 'produces' democratization. Hence, democracy is more likely to survive in richer nations , but the accumulated evidence does not support the claim that its emergence can simply be linked to increasing GNP and income levels. Essentially, such structural accounts of highly contingent processes attain a level of theoretical parsimony at the price of decoupling cause and effect. Hence, I conclude with Sørensen that there is "no clear answer to the trade-off question" because "countries move very fast between the regime categories" and "each time they make a stop in one of the categories, they lend their economic performance data, often covering a few years, to a different argument in these investigations" .

Critique of the Agency-centered View on Democratic Transitions

A rival agency-centered perspective on democratic transitions only entered the academic mainstream in the 1980s when for an increasing number of countries "the possibility of democracy appeared on the historical horizon" . Transitions from Authoritarian Rule represented the first major comparative effort to establish a perspective that questioned the imposing role of socio-economic and cultural prerequisites. However, the shift from the inevitability of authoritarianism towards the study of "possibilism" (Albert Hirschman) did not end in undue optimism. Many of the authors put much of their hopes in 'negotiated' pacts between elites that essentially sidelined the masses. Assumptions about human nature in general and the motivations of important agents for change were quite sober and demanded little normative commitment . This led to a somewhat curious tension between a more optimistic view on democratization under unfavorable structural conditions and a strong emphasis on the difficulties of successfully completing such a process. The challenge to the modernization literature remained profound as former prerequisites (independent variables) for democracy became now dependent variables.

Following Transitions to from Authoritarian Rule several distinct approaches within an emerging agency-based school developed. All of them emphasized the possibility of transitions from authoritarian rule simply based on (elite) actor's narrow choices instead of economic and cultural requisites. According to this line of thought, democracy with its intrinsic value of institutionalized uncertainty and protection against arbitrary rule develops a 'natural pull' on political actors which face declining power resources and/or intensified competition for such resources. Instead of loosing power to another party in a zero-sum fashion, these actors can create abstract and formal institutions. As a result "political forces comply with present defeats because they believe that the institutional framework that organizes the democratic competition will permit them to advance their interests in the future" .

The shift from structure to agency opens whole new research agendas within the study of regime change, but it does so at the cost of neglecting the role of the agent's material and non-material environment. The choice for democratic institutions can hardly be formally modeled in rational choice language because even minimal requirements about the existence of information about other's choices and the general course of events can not be taken for granted. How do we account for the confidence into the power of not even existing democratic institutions, if agents are expected to be driven by nothing else but their self-interest? Why should an elite that has been in almost complete control of the political system suddenly opt for risky political maneuvering? How do we account for the emergence of this assumed power stalemate in the first place? In order to solve these puzzles about actor's preferences under conditions of political uncertainty, I suggest to look into their ideational and material environment as a source influencing their actions. The question is not, if actors engage in interest-maximizing behavior or not, but what kind of pre-existing cognitive frameworks inform the very process of interest formation in a situation of an highly contested and fluid process of regime change.

Domestic Institutions and International Opportunities

In order to solve some of the remaining puzzles within an agency-centered perspective, I will draw on literature which has sought to move the definition of the structural environment of actors away from an overly materialist and positionalist understanding. With regard to the perceived increasing gap between 'structure' and 'agency' in the study of democratic transitions, a growing number of authors have suggested emphasizing the role of an institutional meso-level. Karl/Schmitter labeled such an approach structural contingency and maintained that "even in the midst of tremendous uncertainty provoked by a regime transition the decisions made by various actors respond to, and are conditioned by, socio-economic structures and political institutions already present, or existing in the people's memories" . Such an understanding of institutions extends the definition to informal procedures and rules, while it still maintains a separation of 'institutions' and 'culture' in a more general sense .

Domestic Institutions, Path Dependency, and the Limits to Regime Change

For Africa, Michael Bratton and Nicolas van de Walle held that prior regime type proved to be the main explanatory variable for predicting the sustainability of democratic change . The authors used the concept of neopatrimonial rule to describe as a set of weakly institutionalized practices which nonetheless shape actor's interests and behavior. From their comparison of 42 cases on the African continent they conclude that socioeconomic and international factors are less relevant to transition processes than domestic institutions shaping actor's choices. Their politico-institutional approach classified different forms of neopatrimonial rule (modal regimes) in Africa in order to predict outcomes of transition processes. While the authors remained generally pessimistic about the prospects of democracy in Africa, they hold that countries with a highly authoritarian history (e.g. a military oligarchy) tend to democratize more quickly, but are likely to fall behind again in the long run. In contrast, regimes with some prior democratic experience (e.g. a competitive one-party system) will make more modest, but also more sustainable advancements towards democracy.

Such a perspective on regime change allows for a more nuanced recognition of the actor's environment and the role of domestic institutions in shaping processes of regime change. However, it still leaves open the issue of where the preference for change away from authoritarian rule originates. Similar to an agency-centered perspective, an institutionalist argument is hard pressed to explain why previously solid authoritarian arrangements should suddenly fall apart. However, the initiation of regime change and the mode of transition have crucial effects on the subsequent transition path and the results. Reflecting the general wisdom of the agency-centered literature Karl/Schmitter hypothesized that elite compromise is the most successful path to (restricted types of) democracy, followed by elite imposition. In contrast, transition processes driven by mass protests are less likely to be sustainable .

	Regime Change
	Initiation	Path	Result
Structuralism	Yes	No	No
Contingency/Agency	No	Yes	Yes
Institutionalism	No	Yes	Yes

So far, the review of the transition literature has revealed more or less suitable ingredients for an explanation of regime change in a given country. Efforts to dynamically link economic and political development have not yielded satisfactory results. Democracy is more likely to survive under conditions of material well-being, but its emergence can not be explained or predicted only by identifying increases in GNP levels. This failure to account for the initiation of regime change together with a general lack of interest in subsequent issues of transition path and result make a perspective of materially biased structuralism least attractive for the purpose of this study.

A look at the alternative agency literature reveals a whole array of hypotheses about the sustainability and the eventual results of transitions. However, this literature is mute on the crucial question of how the process starts in the first place. A brief review of institutionalist efforts to solve some of the puzzles in the explanation of political transitions has also added valuable insights about the role of existing institutions and the path dependency of transition processes. However, the overall domestic bias prevails and institutions are mainly analyzed in their capacity to restrict rather than enable behavior. What is usually explored in all these efforts are the limits of democratic transitions, rather than the opportunities that represent the basic reason why we can study such social phenomena at all.

International Institutions as Opportunities

In order to get from the sphere of path dependency and institutional as well as socio-economic and cultural restrictions to the sphere of political opportunities, it is necessary to identify potential sites where knowledge about principled alternatives is produced and promoted. I argue here that international institutions and transnationally active actors represent an increasingly dense network interacting with the domestic realm of a given country. As such international institutions represent a "world polity", a process by which "certain rules become taken for granted while eliminating alternatives" . Within this international social context, this study focuses on the role of a specific sub-set of institutions and actors active in the area of human rights and democratization. The effects of these institutions and actors have been mainly studied by sociological institutionalists and authors interested in the growing role of transnational advocacy networks in international relations.

The Constitutive Role of the International System

Sociological institutionalism claims that the international social system and its embedded norms represent a powerful force influencing actor's preferences and behavior. This culture "creates and legitimates the social entities that are seen as actors" . Consequently, change in the system and its parts is not primarily driven by internal functional needs but in correspondence with existing external cultural norms . It is not the modernizing force of internal socio-economic development that creates sufficient preconditions for bureaucratic rationalization, but the existence of such norms in the international system. "Culturally and historically contingent beliefs about what constitutes a 'civilized' state (...) exert a far greater influence on basic institutional practices than do material structural conditions,..." . In contrast to a neorealist emphasis on material structures dominating international relations, such a perspective stresses the role of immaterial structures such as norms understood as shared understandings of appropriate behavior. The latter represent the social system upon which states exist and, thus, constitute the identities and interests of its members .

Meyer et al. further argued that the emergence and success of the state system itself has created the preconditions for more rapid processes of ideational diffusion between nation-states. Nation-states are understood as "constructions of a common wider culture, rather than as self-directed actors responding rationally to internal and external contingencies" . While traditional societies displayed a wide variety of dissimilar purposes, modern nation-states are formally equal entities and share rationalized identities and purposes. Hence, formal similarity is the basis of comparison between two or more entities in the system. Rather than arguing that similar material conditions lead to similar outcomes, this perspective assumed similar underlying ideational conditions that influence all units of the system. Nation-states will continue to be exposed to pressures for adaptation to world cultural examples even when they seem unable or refuse to live up to those requirements. Their identity, purpose and legitimization are intrinsically linked to the models they have originally based their existence on and other international and domestic actors can use these models to pressure for change. "If a particular regime rhetorically resists world models, local actors can rely on legitimacy myths (democracy, freedom, equality) and the ready support of activist external groups to oppose the regime" .

However, it is this latter issue that is usually neglected within a sociological institutionalist perspective. Sociological institutionalists are biased towards the homogenizing effects of a wider 'world culture' and are hardly equipped to identify variation with regard to the interaction of international norms and domestic practice. They fail to specify under what conditions the ubiquitous presence of such a culture translates into a significant force on the domestic level. "Rapid global changes across dissimilar units suggest structure-level rather than agent-level causes. They do not, however, prove them. One also needs to specify the mechanism of change and show the common source of the new preference and behavior" .

Norms have to become part of a socialization process driven by norm entrepreneurs. In some cases, international organizations can fulfil a role of norm teachers as they represent values widely shared in the international community of states. However, in many cases such favorable conditions for socialization do not exist. In particular, human rights norms often contradict strong domestic norms or interests. In such a case, non-governmental transnational advocacy coalitions have increasingly taken up the task of diffusing international norms into a given domestic context.

Transnational Advocacy Networks as Carriers of Norms

Holding state actors accountable to their words and written international treaties and conventions does not necessarily require the mobilization of matching material power capabilities, but can make use of the normative framework states are part of as a result of there mere existence in an international society. In contrast to institutions, norms are more specific in their prescriptions and reflect a shared expectation about appropriate behavior . Hence, social institutions such as sovereignty may be understood as a set of related norms which do not simply intervene between the interests and the behavior of actors, but constitute their very identity. Norms within this set "legitimize goals and thus define actors' interests" .

Transnational NGO networks add a particular set of agency and strategies to an international norm which largely evolved as a result of multilateral inter-state negotiations. NGOs in this area emerged as a result of a perceived (widening) gap between those prescriptions and the human rights situation in many countries of the world. While these norms slowly evolved with regard to their specificity and concreteness, their monitoring and enforcement remained underdeveloped .

In order to narrow the gap between international norm and domestic practice, human rights NGOs gather, provide and interpret knowledge about human rights situations and create alternative public arenas for debate on such issues. This creativity is necessitated by a strong normative protection states usually enjoy with regard to their internal sovereignty. Thus, in contrast to the majority of NGOs concerned with the delivery of services (e.g. health and education), transnational human rights NGOs are mainly interested in advocacy . They engage state actors on the basis of principled interpretation of human rights information and offer the domestic opposition a unique opportunity to circumvent an unresponsive and repressive state apparatus. After linking up to an international audience, individuals and groups on the domestic level can now transport their demands to the international arena and supplement pressure "from below" with pressure "from above" . In such a case, enabling international institutions and transnational networks are used to overcome severely constraining domestic structures.

Whereas initial research concentrated on target properties ('domestic structures') as an explanation for transnational impact , this debate becomes increasingly sophisticated in discussing the role of network characteristics and strategies (e.g. symbolic politics) as independent factors. This has moved the debate from a more conventional focus on the restraining role of political institutions to the enabling and constitutive role of institutions/norms and the subsequent creation of new meanings and identities. Current research on the role of transnational relations integrates rather than ignores state actors and raises the question how non-governmental agents are capable of transforming the social environment and identities of states. In the following, the paper will trace the effects of transnational human rights mobilization with regard to regime change in Kenya and Uganda.

The main subject of the study is the comparative evaluation of regime change in Kenya and Uganda during the last ten to fifteen years. For analytical purposes, this process is broken down into two distinct phases. The first captures the time prior to the initiation of regime change, in Kenya from the mid-1980s to 1990/91 and in Uganda from 1980 to early 1986. The second phase concerns the transition path that includes the period after the initiation of regime change until today. Even a superficial look at the evidence reveals significant variation among the cases. Whereas the initiation of regime change in Uganda was marked by a revolution from below, Kenya experienced reforms from above. Similarly, Uganda's transition path since the early 1990s reflects a strong preference for constitutional reforms within a popularist framework (extending inclusiveness), while the Kenyan path follows a more narrow electoral liberalization (extending competition).

	Positive Integration	Negative Integration
Kenya, 1978-89	High	Low
Uganda, 1981-85	Low	High

However, the crucial difference between both countries was not variation with respect to one-party rule, but the means and extent of violence used to secure control. President Milton Obote was unable or unwilling to prevent indiscriminate mass killings perpetrated by the military against the civilian population. In contrast, governmental repression in Kenya was, by and large, more circumscribed and focused on known representatives of the political opposition. Hence, in supplementing Bratton/van de Walle's conceptualization of neopatrimonial rule in terms of positive integration under a presidentialist system, I emphasize here the issue of negative integration represented by differences in the use of violence against alleged domestic opponents.

	Modal regimes of neopatrimonial rule/dominant form of governmental repression
	Prior to the initiation of regime change	After the initiation of regime change
Kenya	Positive integration: De jure one-party system with a dominant presidency and strong clientelist networks Negative integration: Repression of political dissent	Positive Integration: De jure Multiparty System with a dominant presidency and strong clientelist networks Negative Integration: Repression of political dissent combined with ethnic violence
Uganda	Positive Integration: De facto one-party system with weak clientelist networks; military uncontrolled Negative Integration: Indiscriminate mass killings of civilians	Positive Integration: 'Movement' system with elements of a competitive one party system Negative Integration: Repression of political dissent outside the Movement System

In Uganda, sustainable regime change set in when the National Resistance Movement (NRM) under the leadership of Yoweri Museveni won in early 1986 a prolonged civil war. For the first time since the first constitutional crisis in 1966 relative peace and security returned to most parts of Uganda. In Kenya, a process of political liberalization was set in motion in 1991/92, after the reintroduction of multipartyism.

The Ugandan leadership extended the initial four year ban on party activities to almost 15 years until 2001. With the introduction of a grassroots model of democracy, the participation of the Ugandan population in local affairs sharply increased. Following the military victory of the National Resistance Movement (NRM), the system of local resistance councils that had been partly established during the civil war was extended to the whole country. The Ugandan government embarked since 1986 on a lengthy constitutional reform process and simultaneously banned parties from participating in elections. Elections for national parliament were held for the first time again in 1992 and President Museveni stood in a nation-wide election only in 1996. The aggregation of political interests on the national level was now channeled through this movement system which replaced the previously dominant Western model of party competition. Within the logic of Robert Dahl's basic distinction between participatory and competitive aspects of democracy, the Ugandan government after 1986 chose the path of extending participation and inclusiveness rather than competition and contestation.

In contrast, Kenya represents a more 'classical' case of democratization. After heavy outside pressure, the authoritarian leadership agreed in late-1991 to the reintroduction of multipartyism. In late 1992 and 1997, multiparty elections twice returned the authoritarian leadership to power. While both elections cannot be labeled as free and fair, the 1997 contest was more democratic than the first round in 1992. However, continued repression of the opposition and a stubborn refusal of the government to enter serious talks about constitutional reforms accompanied the progress in the area of electoral democracy. In 1997, the government again tried to delay substantive political reforms until after the elections.

In using Bratton/van de Walle's concept of modal regimes in Africa it is now possible to visualize the significant differences with regard to the initiation of regime change and the transition path in Kenya and Uganda. These differences are to be explained in the following main empirical section.

Figure 1. Modal Regimes in Kenya and Uganda, 1980s and 1990s

High

MS

SO

1998

Competition CS

1982

1998

Uganda Kenya

MO 1985

1988

Low PS

Low High

Participation

The explanation of regime change in Kenya and Uganda can now be build on a model focusing on the interaction between international norms and their entrepreneurs and (pre-) existing domestic institutional arrangements. This main empirical section will present evidence on the role of transnational human rights mobilization prior (mobilization and regime change) and after the initiation of regime change (transition paths). In the selected cases, international and domestic mobilization was high prior to regime change. Afterwards, transnational mobilization decreased with regard to Uganda while it remained high in the Kenyan case.

	International Mobilization/NGO pressure
	Prior to the initiation of regime change	after the initiation of regime change
Kenya	High (1983-1989)	High (1990-1999)
Uganda	High (1974-1985)	Low (1986-1999)

Non-governmental organizations became active shortly after domestic patterns of systematic human rights abuses became visible in Uganda (mid 1970s) and Kenya (mid-1980s). Within a few years, human rights issues played a major role in the external relations of both countries. Domestically, the extent of human rights abuses under the Amin dictatorship (1971-1979) and Obote's second presidency (1980-1985) contrasted sharply with both the Kenyatta and the Moi era. The violent death of almost one million Ugandans during the period from 1972 to 1985 indicates the dominance of negative integration and the virtual destruction of neopatrimonial rule. In Kenya, this system with an emphasis on positive integration essentially survived the transition from Jomo Kenyatta to Daniel arap Moi in 1978. However, during the early 1980s Moi replaced supporters of the old regime at all levels of the government and the state-controlled economic sector and brought in his own ethnic coalition as the new main beneficiary. Subsequently, the conflict between a now marginalized and increasingly discontent Kikuyu and Luo elite and the new leadership led to an increase of government-sponsored human rights violations. However, the level of repression never reached the character of the dictatorships in Uganda during the 1970s and early 1980s.

Human rights conditions deteriorated sharply after army general Idi Amin had deposed elected president Milton Obote in January 1971. In their efforts to raise international attention about deteriorating human rights conditions in Uganda, Amnesty International and the International Commission of Jurists (ICJ) were initially keen to use mechanisms provided by the United Nations (UN) system. On 27 May 1974, the ICJ sent the first letter to the UN Secretary-General , who forwarded the application based on ECOSOC resolution 1503 to the Sub-Commission on the Protection of Minorities and the Prevention of Discrimination. One week later the report was released to the press and received world-wide publicity, especially in the British media. Idi Amin threatened to expel all British nationals, if the BBC would continue to report on these issues . Extra-judicial executions of Obote supporters and violent attacks on selected ethnic groups continued unabated. Consequently, the ICJ added three additional communications on 16 June, 26 July and 23 August 1974.

Despite the urgency of the situation, the UN Commission on Human Rights (CHR) decided in early 1975 to postpone its decision on the further treatment of the matter for one year. Some members of the UN Commission were irritated because the ICJ had published the information originally provided under a confidential procedure. On 1 October 1975, Amin declared in front of the United Nations General Assembly that Amnesty International was "fed on rumors and concoctions from discredited criminals and exiles" . Prior to the UN meeting, a former US ambassador to Uganda had urged President Ford in a letter on 3 September "to instruct the US delegation (...) to be absent when President Amin speaks to that body." Melady referred in his letter to "gross violations of human rights as distinguished from repression" . The US administration decided that only the Head of its mission to the UN would be absent.

In its 1976 session the members of the CHR decided to end the investigation without further action. The ICJ decided to prepare a second comprehensive report and sent this on 2 June 1976. Only when Idi Amin welcomed in mid-1976 the Palestinian hijackers of an Israeli passenger plane at Entebbe airport, diplomatic relations with Western countries faced a serious crisis. After an Israeli rescue operation , Britain broke off diplomatic relations with Uganda on 28 July 1976. Based on the second ICJ report on Uganda, the Sub-Commission recommended on 25 August 1976 resume the investigation and "make a through study of the human rights situation in Uganda, based on objective and reliably attested information." So far, similar studies had only been approved for South Africa, the Occupied Territories in the Middle East, and Chile. The CHR did not follow the recommendation but delayed the decision for another year until early 1977. Even when the Ugandan representative at the CHR meeting (and Attorney General) Godfrey Lule used his attendance in Geneva to flee into exile to London , this did not change the wait-and-see position of the UN.

At the very end of Amin's rule , there were only a few significant official responses to the human rights reports. Only in June 1977 the Commonwealth Head of States declared in a resolution that "cognizant of the accumulated evidence of sustained disregard for the sanctity of life and of massive violation of basic human rights in Uganda, it was the overwhelming view of Commonwealth leaders that these excesses were so gross as to warrant the world's concern and to evoke condemnation by the heads of governments in the strong and unequivocal terms" . As with all the other governmental actors, no immediate, let alone, effective actions to stop human rights violations in Uganda were taken.

The end to Amin's rule was mainly a result of internal divisions within his regime and the dictator's efforts to control those divisions by launching a military attack on neighboring Tanzania . In January 1979, Tanzanian troops and about 1,000 pro-Obote exiles invaded Uganda. Libyan and Palestinian reinforcement could not prevent the downfall of the Amin regime on 11 April 1979 . Despite the overall dismal gap between rhetoric and action, the transnational human rights movement succeeded in delegitimizing the Amin government as one of the worst human rights violating regimes. Human rights norms were now established competitors of the sovereignty norm when it came to international debates on Uganda.

Following two short-lived governments, Milton Obote returned to the presidency after his party Ugandan People's Congress (UPC) had been declared the winner of rigged multiparty elections held on 10 December 1980 . The anti-Amin coalition had already dissolved prior to the elections and some of the defeated political leaders, including Yoweri Museveni, refused to accept the election results and took up arms against the new government. Within a few months after the elections, the human rights situation in Uganda worsened again. The regular army was fighting several rebel groups and extended its repression to ever larger parts of the country. As a result, every successful rebel operation caused more repression, while each atrocity committed by the army led to an increase of popular support of the rebels.

Yoweri Museveni's National Resistance Movement (NRM) soon turned out to be the most successful challenger to Obote, not the least, because the group had developed a political platform prominently featuring basic human rights. These principles were applied to the areas controlled by the NRM and also enforced against members of the own rebel group. While official donor governments were caught in the conflict between the norm of state sovereignty and basic human rights, non-governmental organizations in Western countries soon openly supported Museveni and the NRM. In 1984/85, NRM leaders regularly traveled to European countries and met with non-governmental representatives, but increasingly also members of European parliaments or even concerned ministries. These growing linkages and the improving exchange of information between Ugandan rebels and Western partners further undermined Obote's position. However, similar to the Amin case, official governmental reactions to the human rights situation were inefficient and came too late . When it became apparent that Obote could not contain the rebellion, the military took over once again. Efforts by Kenyan president Daniel arap Moi to bring the new Ugandan leadership and the NRM rebels to open peace talks initially succeeded. However, the latter did not wait until a completion of these talks, but marched into Kampala on 26 January 1986 to seize power. Almost 20 years of violent repression from above ended in a revolution from below.

International attention for the human rights situation in Kenya increased as a consequence of the executive reactions to an unsuccessful coup d'etat against Moi on 1 August 1982. Since taking over the presidency in 1978, Moi had been eager to consolidate his initially weak position by expanding the security apparatus and concentrating executive powers in the Office of the President. In July 1980 all societal organizations with a "tribal bias" were banned and, thus, dissent was systematically suppressed. On 9 June 1982, the Kenyan parliament decided to turn Kenya into a de jure one party state. In the aftermath of the coup, several hundred dissidents and alleged members of the underground movement Mwakenya were arrested and kept in custody. In the following years, the universities as sites for the mobilization of the political opposition became a prime target for the security apparatus . The ruling party, Kenya African National Union (KANU) was revived and systematically used by the executive to perfect its control over society at large . By 1986, the churches were the only remaining significant societal actor with some autonomy from executive interference. In a situation of growing repression, information about deteriorating human rights conditions in Kenya increasingly disseminated through transnational advocacy networks. In September 1986, Norway became one of the first countries to grant political asylum to Kenyan dissidents, most notably Koigi wa Wamwere, a Kikuyu politician who turned into the "most important opinion leader in Kenyan affairs in Norway in the late 1980s" .

Norway became the first Western donor country that was profoundly affected by the activities of the emerging human rights network on Kenya. From there, long-held perceptions of Kenya changed in concentric circles starting in neighboring Scandinavian countries and moving to Continental Europe and the United States. For the time being, the main voice on the domestic level were the National Council of Churches of Kenya (NCCK) followed by a number of undeterred individual lawyers organized in Law Society of Kenya (LSK). In contrast to other societal organizations, which were streamlined under KANU rule during the same period, the NCCK and the LSK had always exceptionally strong relations to the international arena and mobilized these connections to defend their domestic independence and mobilize against rising authoritarianism. In late 1986, the NCCK published a protest letter against the government that had been signed by 1,200 pastors .

Pressure "from above" build up when Daniel arap Moi announced several state visits to Europe and the United States in 1987. The Kenyan government unwillingly contributed to the success of this campaign by arresting prominent opposition members shortly before Moi's departure to the United States in March. One of them, Gibson Kamau Kuria, a Oxford-educated defense lawyer in the MwaKenya trials, had filed a lawsuit against the security organs accusing them of torturing the suspects. In anticipation of the likely consequences, Kuria provided Blaine Harden, the Washington Post correspondent in Nairobi, with the compiled evidence. On 26 February 1987, Kuria disappeared. On 12 March, one day before Moi met with President Reagan in Washington, the government announced his arrest under the Preservation of Public Security Act and accused him of "disrespect of the President." The US State Department spokesperson, Charles Redman, declared that "the allegations of torture, apparently supported by signed affidavits from those in Kenya who claim to have been tortured, raise serious questions of human rights abuses" (citations in: Africa Contemporary Record, ACR, Vol. XIX, B 334).

The day after the talks between Moi and Reagan, the Washington Post subtitled a picture of both politicians on the front page with ‘Police Torture is Charged in Kenya’. Kuria's fate immediately turned into a cause célèbre (ACR, Vol. XX, B 324) for the transnational network. Moi cancelled his planned visit to New York and a meeting with the Secretary General of the United Nations and flew directly to less hostile Great Britain. In July, Amnesty International followed up on the Kuria affair and published the report Kenya: Torture, Political Detention and Unfair Trials . Pressure by human rights groups continued throughout the rest of the year and peaked again in fall when Moi planned to travel to a number of European countries. Contrary to his original plans, Moi only visited Finland and Romania, but decided to skip Sweden and Norway because of the negative press coverage prior to his visit . Upon his arrival back to Nairobi, he called the press reports "malicious and baseless" and argued that South Africa was a much better target than Kenya for the Norwegian and Swedish press. He said that this situation "had probably made envious outsiders become hostile to Kenya" (Foreign Broadcast Information Service, FBIS-WEU-87-172, 4 September 1987, p. 8).

Notwithstanding this flat rejection of outside criticism, the international mobilization slowly changed the domestic power game. In September 1987, the publication of the first issue of the government-critical Nairobi Law Monthly indicated that outside attention forced the government to be more tolerant towards dissent. Ethnicity remained a strong factor in this ensuing confrontation, because many members of the now side-lined Kikuyu or Luo elites realized that mobilization based on human rights issues resonated much better on the international level than directly playing the 'ethnic card'.

Although the transnational human rights mobilization seriously damaged Kenya's image as a reliable Western ally, donor governments took no concerted action. Only Norway and Sweden temporarily halted their bilateral aid in 1987, without any significant impact. To the contrary, in August 1988 Moi once again increased executive powers and pushed new constitutional amendments through parliament. The period for detention without charge or trial was extended from 24 hours to 14 days and the tenure for all public servants, including judges, was removed (25. Amendment). Members of the transnational human rights network answered these moves with intensified mobilization. The Robert F. Kennedy Memorial Center for Human Rights honored Gibson Kamau Kuria in March 1989 with its human rights award. A delegation of the organization visited Nairobi for the award ceremony as well as a meeting with President Moi .

As a result of the increasing international sensitivity, mobilization now occurred almost instantaneously whenever Kenyan security organs were caught or assumed to be involved in committing human rights abuses. The situation became uncontrollable for the Kenyan government in 1989/1990 when two prominent Kenyans died under mysterious circumstances and isolated domestic protests were supplemented by more frequent street demonstrations for political reforms. In February 1990, the Foreign Minister Robert Ouko was found murdered and his body burned not far from his farm. In August, the government-critical Anglican Bishop Alexander Muge died in a questionable car accident. Both events and the reactions of the government elite outraged the domestic and international public. In the case of the United States this development was further emphasized when the new Ambassador Smith Hempstone turned himself into a vocal supporter of the opposition shortly after he had taken up his appointment. Encouraged by the events in Eastern Europe, the domestic opposition now openly challenged the one-party rule in Kenya. On 3 May 1990, the former Cabinet Ministers Charles Rubia and Kenneth Matiba called a press conference to demand the introduction of multipartyism .

The opening created by the Rubia/Matiba attack led to intensified domestic mobilization for democratic change. On 4 July, the government sought to contain the upsurge of protests and detained Matiba, Rubia and others. Gibson Kamau Kuria took refuge at the US embassy and Smith Hempstone escorted him to the airport. Subsequent street demonstrations in the capital Nairobi and provincial centers culminated on Saba Saba (7 July 1990) when at least 29 civilians were killed by the police. The international response came almost instantaneously and worldwide protests against the violence sprung up. The International Bar Association canceled its biannual meeting with more than 3,000 participants to be held in Nairobi in September citing the general insecurity in the country.

As a result of the increasing isolation, Moi's resistance against outside pressure even increased. When the Norwegian Ambassador Niels Dahl attended a trial against Koigi wa Wamwere, the inner circle around Moi took this as another flagrant instance of an intrusion in domestic affairs of Kenya and severed diplomatic relations on 22 October 1990. The outside relationships of the Kenyan government were now openly strained, as representatives from Western donor countries had to react to the diplomatic standoff between Kenya and Norway. For the first time in October 1990, the US Congress attached human rights issues to foreign aid appropriations for Kenya. During this crucial period, the transnational human rights movement kept up its pressure. On 30 July 1991, Human Rights Watch came out with the first comprehensive (432 pages) human rights report on Kenya . While the report contained little new information, it re-emphasized that the human rights violations in Kenya were the result of systematic and intentional government policy rather than isolated acts of individuals.

On 2 August 1991, the opposition created a broad coalition called the Forum for the Restoration of Democracy (FORD) which was "inspired by Civil Forum in East Germany and Czechoslovakia" . The opposition was now strong enough to bypass the government's ban on political parties and represented a loose coalition of clergy, lawyers, and other individual dissidents. While the NCCK and the LSK had been the main institutional bodies backing opposition dissent during the 1980s, the formation of FORD was a breakthrough in establishing a direct challenge to KANU and Moi in the political arena. Initially formed as a pressure group, FORD's leadership was an ethnically broad coalition which included Oginga Odinga (Luo), Martin Shikuku (Luhya), Masinde Muliro (Luhya), Philip Gachoka (Kikuyu), George Nthenge (Kamba), and the Muslim Ahmed Salim Bahmariz from the Coast region. Many of the human rights activists such as the former LSK chairman, Paul Muite, and the editor of the Nairobi Law Monthly, Gitobu Imanyara, also joined FORD.

The global changes and the Kenyan independence history simultaneously inspired the formation of FORD. In contrast to prior unsuccessful attempts of institutionalizing political opposition, FORD mobilized both international support but also revived the historic Kikuyu-Luo alliance. "The FORD of 1990 echoed the KANU coalition of the 1960s, not only in its leadership but also in its middle class and urban base and its successful mobilization of rural support from its 'ethnic areas.' FORD opposed a KANU that resembled the 1960 KADU coalition of minority groups led by largely middle- and upper class leaders whose base remained rural and who were unable to attract an urban following" . However, internal conflicts about the future role of FORD and the leadership soon emerged. While one section of FORD advocated the immediate transformation of the movement into a political party, a second group wanted to keep FORD as a united opposition movement to press for reforms.

Efforts by Western ambassadors to facilitate secret negotiations between FORD and the government on the issue of a transitional government failed by mid-November 1991. When security forces arrested many members of FORD over night, the donor community suspended aid flows on 26 November. "Kenya, the long-time favorite of the West, was being treated as one of Africa's pariah regimes" . Four days after the donor decision, on 30 November, Hempstone and the visiting US Deputy Assistant Secretary of State for African Affairs, Bob Houdek, met with President Moi and Foreign Minister Ayah. Houdek said that he had been instructed to pursue two essential issues. First, he wanted to know a precise date when the opposition could hold its first legal public meeting. Second, he asked Moi to announce publicly elections with non-KANU candidates. Moi flatly rejected both demands and complained instead about the alleged misconduct of donors and embassy personnel. He asked the US to "detach itself from the dissidents and follow diplomatic conventions" . On 2 December, however, Moi announced the end of the one-party era in Kenya. As a first sign of these reforms controlled from above, FORD was officially registered as a party on 31 December.

Transnational advocacy networks played a major role in delegitimizing authoritarian rule in Kenya and Uganda. In both cases, the principled transnational human rights mobilization was a crucial factor in initiating regime change. The preexisting form of neopatrimonial rule explains the difference with regard to the initial mode of transition — in the Ugandan case revolution from below and in the Kenyan case limited reforms from above. In Kenya, a lack of transnational mobilization would have left the government with greater international and domestic legitimacy. Without the highly publicized human rights abuses and the interpretative frames provided by transnational human rights actors, the donor community would not have been united in making Kenya an exemplary case for good governance. While the principled transnational mobilization accounts for the fact that regime change was initiated, the specific form of neopatrimonial rule in Kenya determined the mode of transition. A largely intact clientelist system dominated by a strong presidency allowed the ruling elite to essentially survive the pressure.

In Uganda, a lack of mobilization would have seriously diminished the strength of the NRM and the relatively clear-cut elite turnover in 1986 would have been very unlikely. The difference between both countries was not so much the variation in one-party rule at one point or another, but the widespread violent mode of conflict resolution on the domestic level between 1972 and 1985. In Uganda, the institutionalization of terror replaced existing neopatrimonial institutions necessary to positively integrate society.

The second part of the study shifts attention from the initiation of regime change to diverging transition paths. For the subsequent reform process and the international mobilization, the differences in the mode of initiating these reforms matter greatly. While this phase ended for Uganda in early 1986 with a complete change of guards, the Kenyan government as the main target of mobilization survived. The international influences identified as causally relevant for the onset of the regime change continue to matter but the mode of transition now feeds back on to the international level. Strategies used to delegitimize an authoritarian regime must now be substituted or supplemented by efforts directed at the preparation of a legitimate new domestic political order.

The military victory of the NRM under Yoweri Museveni brought in a new government which emphasized human rights issues not only in rhetoric and developed a distinct Ugandan version of popular democracy. Although other rebel groups remained active and new one's emerged, the human rights conditions improved markedly. Press freedom, freedom of expression and assembly, and other basic human rights returned to the country, with only one exception. The NRM government identified political parties as the main scapegoats for the protracted civil war and outlawed public activities of all parties. Instead, the new government instituted a grassroots system of so-called resistance councils which was held together by the broad-based National Resistance Movement. In Robert Dahl's words, an emphasis on popular integration and inclusiveness clearly took precedence over the extension of electoral competition. Transnational human rights mobilization fell sharply as Amnesty International remained the only significant group reporting on Ugandan human rights issues after 1986. Donors contend themselves with the fact that the NRM brought peace and stability back to Uganda and began to generously increase their aid budgets directed at the country. The term of Museveni's transitional government was set to expire after four years in early 1990.

When the deadline of his government approached in 1989, the security situation in the main parts of the country had further improved, but most of the announced reforms were still unfinished or not even started. During this year, the NRM government held the first popular elections on the grassroots level and initiated a constitutional reform process. With a view on the remaining reform agenda, the government extended its rule for another five years. Only a few domestic and international voices protested this step. At the same time, increasing reports about human rights abuses committed by the regular army during counterinsurgency operations at the Sudanese border prompted Amnesty International to publish three extensive reports between 1989 and 1992. After four years of relative societal openness, the civil society sector remained weak and no significant groups had emerged yet which could possibly check a dominant executive and its 'movement' power base. Despite this consolidation of NRM rule, its grassroots ideas failed to take root and the NRM slowly returned to a more representational model of democracy. This shift became most apparent when elections for a Constitutional Assembly (CA) were held in early 1994 and the NRM was able to defend its status as dominant domestic political force.

The ban on party activities had the opposite effect: Instead of withering away, party politics as an idea re-emerged as a main focus of domestic public debates. Despite (or rather because of) the convincing victory of the NRM in the CA elections, the political system was now sharply divided between 'movementists' and 'multipartyists'. While the 'movementists' dominated the CA and were able to extend the movement rule for a second time until the year 2000, the domestic legitimacy of the movement as a whole decreased despite the near successful completion of its widely accepted reform projects. The new constitution now provided that a national referendum in 2000 would at that point decide about the introduction of multipartyism. At the same time, the constitution contained a significantly extended Bill of Rights as well as provisions for an independent human rights commission charged with overseeing the domestic conduct of security forces .

While parts of the donor community expressed some concern about the "no-party-rule", the general support for Museveni remained unquestioned. Instead, regional crises added to the pressure on the movement system. After a sharp decline in rebel activities in 1991/92, new groups emerged in the Northern and after 1997 also in the Western part of the country. In the North, the Lord Resistance Army (LRA) under the leadership of Joseph Kony killed hundreds of villagers and abducted more than 8,000 children between 1994 and 1998 . In order to end these rebellions, the Ugandan government involved itself into regional conflicts, first by supporting rebels in Southern Sudan, later by sponsoring the (successful) 1994 rebellion in Rwanda, and finally by supporting Kabila's rebel movement in Zaire. However, in the long run, all these direct or indirect military engagements led to increasing rather than decreasing threats to domestic stability.

Under the conditions of growing regional conflicts, Museveni stood in 1996 for the first time in an election for the presidency. Two weeks later, the members of the national parliament were also for the first time directly elected. In both elections, Museveni and the 'movementists' within the NRM prevailed once more by a large margin. However, the electoral victory did hardly strengthen the domestic position of the executive and the movement. After the more or less successful completion of the reform projects, the apparent rejection of the grassroots model of democracy, and the failure to end the armed rebellions in the North and West, the NRM was in danger of loosing its position as only credible political force. Although this could have created additional space for civil society actors, the domestic political arena remained dominated by the conflict between 'movementists' and 'multipartyists'. Instead of new political forces in and outside of parliament, entrenched movement leaders and the tainted old party leaders maintained their strong-holds on both sides of the divide. Only increasing activism within a strengthened parliament showed an increase of institutionalized domestic dissent.

When the government finally formalized the movement system in early 1997 and awarded every Ugandan automatic membership, this decidedly failed to increase the waning integrative force of the movement system. Despite convincing electoral victories and strong outside material support, the sacrosanct refusal of the NRM to allow party activities and its non-compromising attitude towards the various rebel groups came increasingly under pressure. The political opposition was now more successful in exploiting the gap between a rhetorical commitment to individual rights and the reality of banned party activities. While a number of domestic NGOs in the field of human rights and democratization had finally established themselves , they essentially ignored the contagious issue of multipartyism and remained heavily donor-dependent and urban-biased.

At around the same time, international attention towards the region focused on the situation in Zaire and the ongoing human rights abuses committed by the LRA . Leaders of Western governments, international organizations such as UNICEF alongside with members of the transnational human rights movement began to target these non-governmental human rights abuses. The plight of the abducted children made it to the headlines of international media and leading US government representatives such as Secretary of State Madeleine Albright (December 1997) and President Clinton as well as his wife Hilary (March 1998) spoke strongly on the issue while they visited Uganda. The coincidence between human rights mobilization and US strategic interests in the region (e.g. Sudan and former Zaire) made this a all too welcome topic.

This international reinforcement came at a time when the NRM and Museveni were under increasing domestic (normative) and regional (military) pressures. Combined with the fact that the 'movement' integrated in the late 1990s only by definition and increasingly failed to represent a inclusive national body, this led to a domestic blockage where the 'multiparty' challenge appeared stronger than the legitimacy of its supporters would one expect. The old movement guard used the regional threats to defend its anachronistic position and even occasionally increased the use of repression. As a result, the opposition forces gained ground simply by reminding the leadership of its rhetorical promises contained in its original political program and expressed in the 1995 constitution. It is doubtful that the currently dominant political forces are capable of leading the upcoming transition from the movement system to a new, more democratic political order.

In contrast to Uganda, regime change in Kenya after 1991/92 took the path of extending electoral competition while the inclusiveness of the political system as a whole decreased. Also in contrast to Uganda, a mushrooming and strengthening of domestic civil society groups and a powerful re-occupation of societal space accompanied regime change. As a result of this increased activism, transnational exchanges also further intensified . For the first time, domestic human rights NGOs such as the Kenya Human Rights Commission (KHRC), Release Political Prisoners (RPP), and even a broad NGO coalition emerged and were capable of resisting governmental strategies of suppression or cooptation . Increasing donor budgets for civil society groups, but also the shift of attention within the ruling elites facilitated this mushrooming. In 1991/92, Moi and his aides had to make hard choices and decided to focus on securing a KANU victory at the upcoming General elections.

Part of this strategy were violent attacks of 'ethnic cleansing' conveniently hidden under demands for a return to the original post-independence 'majimbo' (federalism) constitution (see fn. 15). In the end, about 2,000 Kenyans were killed and several hundred thousands displaced. A parliamentary investigation and many other independent observers concluded that a number of high level KANU representatives were prominently involved in instigating those outbreaks of violence . Other measures giving KANU an clear advantage at the polls were either already institutionalized or were taken during the period of voter registration. Over the decades, the executive had systematically expanded the number of constituencies in support of KANU, while traditional opposition areas remained unchanged. Hence, on Election Day the opposition had to gain three to four times more votes in order to get the same amount of seats in parliament . However, the most important reason why the opposition failed to remove Moi from power were the internal divisions and an inability to agree on a single challenger. While Moi succeeded in holding his ethnic coalition together, FORD split into several parties during 1992.

Following the electoral defeat, the political opposition had lost much of its momentum and almost disappeared. A total of 14 opposition MPs returned to KANU in 1993 . The donor community also returned to business as usual and ended the aid freeze. The KANU leadership was now able to reject demands for further democratization with even greater confidence than in 1989. The key for further domestic mobilization was now back in the hands of the strengthened societal groups (and their transnational partners), that now began to form new post-election policy platforms and organizations. In 1995/96, these groups renewed their efforts to force KANU into a meaningful constitutional reform process. The experience of the 1992 elections provided ample evidence that elections were almost meaningless as long as the incumbent government felt no formal and informal limits to its efforts in securing a favorable election result long before Election Day.

In spring 1997, efforts to unite the opposition again finally bore fruit. From a workshop in Nakuru a coalition of parties and NGOs emerged under the leadership of the newly created National Convention for Constitutional Reform. Its executive body, the National Convention Executive Committee (NCEC) became the main force behind the pressure for reforms prior to the upcoming multiparty elections in late-1997. Internationally well connected members of human rights NGOs such as Gibson Kamau Kuria, Kivuta Kibwana, and Willy Mutunga dominated this body. Initially, Moi and KANU flatly rejected all demands for reforms and hoped to enter the electoral race under the same conditions (apart from an even more split opposition) as in 1992. To prevent such an outcome, the NCEC called for regular street demonstrations in favor of substantive political reforms. When history repeated itself, and the police killed another twelve demonstrators in one such incident on Saba Saba 1997, international and domestic outrage peaked. Moi and KANU began to come close again to the situation of 1990/91.

On 15 July, Moi invited religious leaders to talk on the question of political reforms. The following day he made the first major concession and announced that public rallies no longer had to be licensed by the local administration. Furthermore, KANU offered talks about political reforms to the parliamentary opposition, but refused to include non-elected members of the opposition (such as the NCEC). The confrontation culminated by the end of August when Moi accused the NCEC of preparing a "civilian coup". Moderate sections on both sides of the divide now led an effort to break the blockage and enter negotiations about minimal constitutional reforms prior to the upcoming elections. Within only two-month time, an Inter-Party Parliamentary Group (IPPG) agreed on several amendments to the constitution that weakened the powers of the executive. The government agreed to register all parties and process all applications for radio and TV licenses as soon as possible. These changes were enacted in early November, about one and a half months before the elections (Daily Nation, 8 November 1997).

The December 1997 elections were "measurably fairer" than five years earlier, but the "institutional bias" in favor of KANU remained significant . Nonetheless, taken together the opposition succeeded to gain almost a matching number of seats in parliament. Again, the election period was marred by 'ethnic violence' against alleged members of the political opposition. After the elections, Moi and KANU insisted that the constitutional reform process should be exclusively in the hands of elected parliamentarians. Only after continued protests and protracted negotiations, KANU agreed, in principle, to the inclusion of civil society groups in the reform process. It was agreed that the central 25-member body would be made up of 13 MPs (from all parties; KANU agreed to take five seats) as well as 12 representatives from religious and civil society organizations. When the deadline for the nomination of the representatives approached, KANU reneged on its original agreement and demanded more seats. Since then, the reform process is stalled.

Transition paths in Kenya and Uganda varied widely after the initiation of regime change, ranging from a proactive policy of increasing inclusiveness to a reluctant acceptance of increasing (party) competition. During the same period, transnational and domestic mobilization levels as well as organizational capacities of NGOs took also very different directions. In a counter-intuitive way, both increased in the Kenyan case after 1991/92 under hardly changing government restrictions, while they hardly changed or even decreased in the Ugandan case despite a general societal openness. In both cases, the existing civil society actors are heavily donor-dependent and urban-biased.

While domestic factors became now more prominent in Kenya, continued transnational mobilization between 1992 and 1998 was crucial in keeping up the pressure for constitutional reforms. Those sections of civil society that had originally appealed in the 1980s for international support, were now the main forces behind a coalition for profound political reforms. However, their domestic legitimacy remains weak and their principled opposition was mixed with ethnic and personal power interests. A long-term reliance on outside support as a substitute for extending domestic legitimacy beyond one's own ethnic group has potentially detrimental effects on a democratization process. Hence, the counter-mobilization by KANU representatives based on a revived majimbo-agenda can not be reduced to an abuse of federalist ideas, but raises a number of valid questions about the future role of different ethnicities in Kenyan politics which were conveniently ignored by the opposition (and some of the outside observers).

In contrast to the Kenyan situation, mobilization decreased sharply in Uganda after 1986 and was almost insignificant between 1992 and 1995/96. A domestic network of non-governmental human rights groups only slowly emerged during the 1990s. As Susan Dicklitch has shown in a comprehensive study of the Ugandan civil society , the NGO sector generally thrived after 1986, but hardly fulfilled the (often-heard) promise of grassroots empowerment and a proactive role in the process of democratization. Most recently, regional developments in the Great Lakes region even increased the number of threats to the fragile transition process in Uganda. A new transnational human rights mobilization coincided with strategic interests of the United States in the region and further strengthened hardliners on both sides of the political divide.

For different reasons, the domestic situation in both countries is currently characterized by a reform blockage. In the Kenyan case, the advocacy network is still on alert and the domestic opposition is strong enough to push forward the reform process. Less promising is the situation Uganda. Even after thirteen years of relative peace and stability the state of the civil society sector as a possible stakeholder in the democratization process is dismal. It can not be expected that independent forces in the societal realm will contribute positively to completion of this reform phase ending with the 2000 referendum. Moreover, if the current leadership will resort to increasing domestic repression, the hope rests again mainly with the transnational human rights movement that legitimized the NRM's ascendancy to power in the first place.

Global norms as part of an evolving international society influence processes of domestic political change. In the case of human rights, transnational advocacy groups play a central role in transporting these norms from the international to the domestic level. However, neither the norms nor their entrepreneurs determine outcomes. It is the interaction of norm promotion on the one hand, and (pre-) existing domestic institutional arrangements on the other hand, which produces discreet paths of regime change. The study advances the argument that international human rights norms and their promotion by transnational actors play an increasingly important role for the initiation of regime change. Pre-existing institutional and power configuration matter, but it is the effective use of international norms as an alternative set of institutions, which forces authoritarian regimes into an initial opening.

Following this period of initial opening, the role of transnational human rights mobilization is likely to decrease, as domestic factors become more important. One of the main reasons for this development is the mandate of many transnational human rights groups. While basic human rights are likely to be prominent issues prior to regime change, the attention is likely to shift towards the choice of adequate democratic institutions once regime change has set in. Amnesty International simply has no competence (and interest) in deciding about the merits of multipartyism or presidential versus parliamentary systems of domestic governance. Both on the domestic and international level, the shift from delegitimizing an authoritarian regime towards creating and legitimizing a new political order requires the development of new and different skills for governmental and non-governmental actors alike. While NGOs and civil society at large generally profit from such a shift, this does not necessarily translate into an automatic increase of influence on the course of events.

For two reasons international factors remain a significant category for analysis. First, the interaction of international and domestic factors prior to the initial opening represents an important source of explanation for the subsequent transition path. Second, and perhaps more importantly, in cases of protracted transitions over an extended period of time, the presence (or absence) of outside mobilization matters for the final outcome. If the authoritarian regime remains in power and political reforms prove to be little more than 'window-dressing', transnational human rights groups will continue to mobilize unabated (Kenya). If the authoritarian regime is replaced and/or reforms are implemented, this will lead to a decline of the initial transnational network and a shift of emphasis towards the reinforcement of the reform process (Uganda). In the selected cases, these differences have produced peculiar transition paths. In Robert Dahl's terms, Uganda emphasized the extension of participation, while Kenya focused on the issue of increasing competition.

The approach spelt out here is rather broad and tends to add a variety of factors rather than highlighting the prevalence of one explanation over another. Still, the emphasis on non-material issues and the role of transnational human rights mobilization allows for a discussion of alternative scenarios. One such alternative would claim that the overall material dependence of both countries ensures that their leadership will remain "on track" with regard to dominant Western liberal norms anyway. The work of the human rights networks would simply be epiphenomenal to the dominant Western powers. There are a number of arguments that question such a 'shortcut' solution. First, the process tracing shows that events do not unfold in a way that donors take the lead and human rights groups follow suit. To the contrary, it is the human rights mobilization which gets (parts of) Western governments to come into the picture and use some of their material leverage (esp. in the Kenyan case). Moreover, the donor community is often at loss about its preferences and internally divided. In contrast to NGOs, they prove to be weak in terms of consistency and follow-up measures. Governmental actors are also much more socialized into the 'negative sovereignty regime' and hardly equipped to cross the crucial domestic-international divide. As it turns out, the supposed strength of possessing material means turns into a profound weakness when it comes to promoting long-term processes of regime change from within.

Second, there is considerable variation with regard to the issues of initiating regime change and transition paths. If materially weak countries were only to follow the beaten path, we would be puzzled about the significant differences between the Kenyan and Ugandan transition experience. Apart from highlighting the strong role of domestic pre-conditions, the study here suggests that the extend of transnational mobilization is directly related to the degree of delegitimization an authoritarian regime faces, both on the international and domestic level. The distinct outcome of reforms from above (Kenya) and revolution from below (Uganda) is a direct result of the transnational challenge to different forms of authoritarian rule. In the subsequent period after the initial regime change, the variation in mobilization level partly accounts for the distinct paths taken by Ugandan and Kenyan elites.

Another alternative scenario would claim that it is suffice to look at the domestic factors spelled out in the thesis and ignore international influences altogether. This position questions the added value of bringing in a whole set of new empirical and theoretical issues which compromise an otherwise fairly parsimonious explanation. Moreover, regime change is in itself an extremely complex issue and, as such, theoretical parsimony should prevail over other possible orientations. In order to evaluate this proposition one will have to use a counterfactual and ask, what would have happened in Kenya and Uganda if those transnational linkages had existed. If they matter, as claimed here, it must be convincingly argued that the course of political change in both countries would have taken a different direction since the early to mid-1980s.

For Uganda, the paper argued that Yoweri Museveni and the NRM would not have emerged as the most successful among many rebel organizations if it had not been for its distinct human rights agenda which allowed the group to link up with international supporters. After regime change, mobilization diminished considerably, but the new government was clearly much more responsive to outside criticism. Overall, the transition process after 1986 became much more of a domestic affair because issues of choosing the appropriate political institutions go largely beyond the mandate of those organizations originally mobilizing for change in the late-1970s and early-1980s. Most recently, transnational human rights mobilization has even tended to support an increasingly reform-tired regime because it focused on violations committed by rebel groups in Northern and Western Uganda. Again, the mandate of the organizations involved often seems to prevent them from taking the larger regional picture into proper account. At the same time, the domestic civil society has remained remarkably weak, in particular when compared with neighboring Kenya.

For Kenya, the study argued that Moi would not have faced a united donor community in late-1991, if advocacy networks had not successfully changed the international image of a regime generally seen as a reliable Western ally and the only stable country in the region. Under these conditions, the effect of transnational mobilization is remarkable. Mobilization remained high after the initiation of regime change, because the Moi regime tried to simply modify means of repression rather than implement substantive changes. While the political opposition organized in the newly emerging parties by and large failed to press the government towards democratization, it was the domestic part of the original transnational human rights network (now organized in an expanding civil society) which successfully pushed for constitutional reforms in 1996/1997. Although these groups were not included in the negotiations and it is still unclear if they get access to the process of a comprehensive constitutional review, their activism accounts for the current state of the Kenyan transition process.

The mobilization protects and empowers the domestic opposition prior to regime change and, in a very basic way, can help to ensure the physical survival of dissidents that re-emerge during the period of a (protracted) transition path. On this issue, the comparison between Uganda in the 1970s and 1980s where these elites were systematically killed and Kenya where such protective measures were much more effective is particularly telling. This supports the overall claim, that transition processes are highly contingent and that their interpretation requires an acute awareness of actor's choices and strategies. Henceforth, the paper maintains that the story told in the empirical section and the can not be fully understood without reference to the existing ideational framework of international society in general and the human rights mobilization carried out by non-governmental advocacy networks.

However, this does not mean that transnational mobilization always produces desirable domestic results. In the Kenyan case, an economically strong Kikuyu elite has effectively monopolized such new channels of voicing dissent and, as a result, transnational mobilization against the Moi regime (and its supporting ethnic coalition) has tended to undermine compromise between both groups. Moi's position even hardened because he perceived of these activities as an intervention into the internal affairs of his country and accused domestic dissidents of being agents of a 'new colonialism'. On the other hand, the outside support tended also to harden the critics of Moi, who thought that they were now strong enough to remove Moi and had no need to compromise with the old regime. Considering the prevalence of ethnic voting patterns, this situation can not be solved by a simple rotation of elites, but requires a lasting compromise on the issue of power sharing based on a negotiated 'ethnic formula'.

This paper has sought to make a case for the study of international human rights norms and their entrepreneurs' prior and after domestic regime change. Although there is growing interest in analyzing the role of those institutions and actors, the comparison of two cases does not allow yet for a conclusive evaluation. As argued above, it is the role, strategies, and techniques of those norm promoters that seem to make a crucial difference for the effectiveness of a given norm. One possible way forward would be to include more cases and develop more fine-tuned instruments to actually measure and compare the extend of mobilization and the subsequent results. While it is always desirable to check on other and more cases, it is doubtful that a quantification strategy will yield meaningful results. Initial efforts to quantify mobilization levels with regard to Kenya and Uganda has not led to additional insights about the role of transnational human rights activism. It is more likely that a more systematic effort in incorporating results from the social movement literature (esp. regarding movement strategies and tactics) will bring this research further ahead. This idea is further supported by the fact, that the literature on social movements has recently developed a growing interest in the role of the international context of their main subject of study.