James R. Scarritt Professor Emeritus Department of Political Science University of Colorado at Boulder

Transcription

1 WORKING PAPER The Explanation of Change and Continunity in Zambian Presidential Selection and Power: Leaders Strategic Interaction, Institutions, and Social Structure James R. Scarritt April 2012 Institutions Program Working Paper Series: INST

2 THE EXPLANATION OF CHANGE AND CONTINUITY IN ZAMBIAN PRESIDENTIAL SELECTION AND POWER: LEADERS STRATEGIC INTERACTION, INSTITUTIONS, AND SOCIAL STRUCTURE James R. Scarritt Professor Emeritus Department of Political Science University of Colorado at Boulder Paper presented on the panel Statehood and Governance: Challenges in Sub-Saharan Africa, International Studies Association, San Diego, April 2-5, 2012 with subsequent minor revisions 0

3 THE EXPLANATION OF CHANGE AND CONTINUITY IN ZAMBIAN PRESIDENTIAL SELECTION AND POWER: LEADERS STRATEGIC INTERACTION, INSTITUTIONS, AND SOCIAL STRUCTURE Electoral and party politics in all countries involves strategic interaction that is enabled and constrained by institutions and social structure. However, the nature of that interaction and its enablement and constraint vary extensively among countries and often within countries over time. This paper and the inprogress book manuscript on which it is based describe and explain the changes and continuities in strategic interaction within electoral and party politics in Zambia between 1948 and 2012 and their enablement and constraint by institutions and social structure over the seven time periods into which the sixty-four years are divided. Many analysts have focused on long term institutional and structural continuity and change, but far fewer have focused on how continuity and change in strategic interaction relates to them. Like many other African countries, Zambia attained independence after a relatively short struggle by African nationalists against the colonial power. This struggle involved the creation of some new institutions but little change in social structure and virtually no firm institutional guidelines for strategic interaction. It is therefore pertinent to ask to what extent such guidelines have subsequently evolved and how they are enabled or constrained by changes and continuities in institutions and social structure. Has change in strategic interaction occurred within institutional and structural continuity, and has there been continuity in strategic interaction when institutions and structures have changed? To answer these questions, a theoretical framework involving an integrated constraint/ enablement-agency and information-energy based linkage of strategic interaction, institutions, and social structure is first summarized. In the book a number of hypotheses about electoral and party politics in Zambia derived from this framework will be tested on a population of approximately 150 strategic interactions (the exact number will be determined when the coding is completed) spanning the seven time periods. To accommodate space limitations, the focus in this paper will be on six hypotheses about eighteen important and well-documented strategic interactions over the selection of, and exercise of power in electoral and party politics by, presidents of political parties and the nation, again spanning all 1

4 time periods. These hypotheses explain why strategic interactions of this type have changed significantly in periods of relative continuity in institutions and social structures, and why the most recent changes have been, on balance, positive for democracy because of learning by leaders that has partially overcome path dependence. Other strategic interactions in electoral and party politics in Zambia that partially overlap with presidential competition and one another, but are not analyzed here, include those over electoral turnout and legislative electoral results (including by-elections), party organization and internal conflict, relations among parties, party-government relations, the political implications of constitutional change, ethnopolitics, class/fraction politics, party-union relations, party-church relations, and partymedia relations. 1 The separation of strategic interactions over presidential selection and exercise of power from these others in this paper means that the analysis presented here does not attain maximum richness, but that is unavoidable. It will be argued in the book that recent changes in some but not all of the other strategic interactions in Zambian electoral and party politics have also been, on balance, positive for democracy. Major strategic interactions among leaders, usually also involving organized groups and members of the public, have been sufficiently well-documented through a combination of the scholarly literature and events data that I have gathered 2 to serve as comparable cases. The episodes that constitute cases of strategic interaction are distinguished by different actors, issues, and contexts (including time periods), although some of them overlap in time within the same period. Time Periods in Zambian Electoral and Party Politics The Zambian analysis begins in 1948, when Africans in the former British colony of Northern Rhodesia first officially became politicians and inaugurated national level politics involving a majority of the population because the Federation of African Welfare Societies officially changed its name to the African National Congress (ANC). It ends in 2012, when the final draft of this paper is being written. The sixtyfour years are divided into seven time periods: the colonial period and the nationalist movement ( ), the multiparty First Republic ( ), the one party Second Republic ( ), and the multiparty Third Republic under Presidents Frederick Chiluba ( ), Levi Mwanawasa ( ), 2

5 Rupiah Banda ( ), and Michael Sata ( so far). As in most historical narratives (Falleti and Lynch 2009: , ), these periods are somewhat arbitrary since many changes are not confined to one period and there are significant continuities across these periods. Yet the differences among periods in structurally and institutionally constrained strategic interaction are sufficiently great to justify their separation. The first three time periods are demarcated from one another and the fourth period by regime changes, while the last four periods are demarcated by different presidents holding office without regime change. The latter demarcation is justified by differences in the ways Presidents have used their great powers. Causal Mechanisms: Information-Energy and Constraint/Enablement-Agency Linkages, Path Dependence and Leadership Zambian leaders are agents who interact strategically with one another, foreign donors/ investors/ diplomats, societal groups and members of the public, often based on conflicting preferences, over electoral and party politics. These interactions are enabled and constrained by causally significant institutions and social structure. Agency and constraint/enablement involve both information and energy, and information and energy are both relevant to strategic interactions by agents and the enablement/constraint of these interactions by institutions and structural properties/relationships. These linkages are general causal mechanisms (Johnson 2002: ; Hedstrom and Sewdberg 1998) that operate at the strategic interaction, institutional, and structural levels of analysis and link them together. More specific causal mechanisms combining aspects of the two types of linkages such as path dependence and leadership are also important and are included in the hypotheses presented below. Information-energy linkages refer to interaction between what are often called ideas and material factors. Shared ideas or cultural information (Kiser and Bauldry 2005: 178-9) need to be translated into institutional rules and individual beliefs, and additionally tied to economic, political and cleavage based resources (energy) and the agency of many individuals, groups, and organizations to be effective, while the energy of agency, resources and power needs to be guided by beliefs, shared ideas and institutional 3

6 rules to be effective (see also Campbell 2004: , 123; Parsons 1960: 174; Sewell 1992: 13). Information-energy linkages are found within societies, formal organizations, and more informal groups, as well as in patterns of strategic interaction among individuals (Ackerman and Parsons 1966: 30-40; Parsons , 1966: 105-6). In this mechanism regulative information flows in one direction (from culture to cleavage/stratification to politics to the economy within societies, from values to norms to organizational/group structure to role behavior, and from beliefs to preferences to energized, resourcebased behavior within individuals engaged in strategic interaction) and facilitative energy or conditioning flows in the opposite direction in each case. This paper defines strategic interaction terms of a wide version of thick rational choice theory (Opp 1999). This version accepts the core assumptions of rational choice, but thick rational choice theory posits some additional description of agent preferences and beliefs (Ferejohn 1991: 282), while wide rational choice theory posits that these preferences may be non egoistic, that actors may not be fully informed, and that intangible and perceived as well as tangible and objective constraints affect behavior. Except for its greater emphasis on the causal significance of institutions and social structures, this approach is very similar to Bates and his collaborators (Bates, de Figueiredo, and Weingast 1998: 628; Bates et. al. 1998) integration of interpretive and rationalist accounts through analytic narratives, in which it is stated that rational choice explanation requires detailed knowledge of the values of individuals, of the expectations that individuals have of each others actions and reactions, and of the ways in which these expectations have been shaped by history. The crucial assumption in this framework about strategic interaction is that individual and group actors exercising agency are simultaneously engaged in making strategic choices to maximize their perceived interests (preferences) and following normative rules to which they are committed as part of their self-understanding; they are preference maximizers and rule followers simultaneously. In social action, human agents make strategic or allocative choices while simultaneously enacting (ontologically) prior understandings about the nature of the strategic situation in which they find themselves, the characteristics or identities of the players 4

7 (including themselves), and the common understandings or expectations as to how the game will be played (Ferejohn 1991:285). Hypothesis 1. Strategic interactions among Zambian leaders (and between them and other actors) over electoral and party politics have usually occurred in an ecology of games (which also includes interactions over institutions and policies) in which outcomes of every strategic interaction (game) and the relations among these outcomes are uncertain and are often unintended. Thus the focus of most actors has frequently been on winner-take-all. Nash equilibria have been unstable as losers have very quickly sought to reverse their defeats. These tendencies have been reduced since Prominent institutionalists conceive of institutions as linking strategic interaction and structural properties (Campbell 2004; Grief and Laitin 2004; March and Olsen 1989: 159; 2006: 4-17; Boone 2003: 1-19) and agree that institutions should be analyzed through a complex mix of rational actor, institutional and various structural perspectives, which includes conflict among institutions that is based in part on strategic interactions and conflicting values and beliefs. Thus institutions do not always create equilibria, but they often constrain conflictual strategic interaction and enable temporary coordination. Organizations, groups and roles--the participation of categories of individuals in interactions that are to some extent influenced by norms, beliefs, and preferences (Biddle 1986)--are parts of institutions, as are most norms. They are separate from both strategic interaction and structural properties and relationships at the societal level. But all institutions, especially one as powerful and stable as the Zambian presidency, are affected by the agency of many different strategic interactions over time that reinforce or change their characteristics, and are constrained by social structure (properties/relationships). Hypothesis 2. Strategic interactions provide the agency to sustain, and are in turn enabled and constrained by, political institutions that concentrate power in the Presidency (diffused from other African countries) at the expense of the legislature, judiciary, local authorities, social groups, and the general population. A strong presidency was adopted at independence because it was widely perceived as necessary for national system persistence and it still is viewed this way, so it is in equilibrium. A plurality single member district electoral system for the legislature (diffused from Britain) was also adopted at independence, and commitment to these institutions has been constant in spite of criticisms of all presidents (and opposition leaders who seek to replace them) and the probable benefits of a more proportionate electoral system for reducing winner take all strategic interaction (Burnell 2001: 142). Social structure consists of properties of societies and relationships among these properties. It also includes the society s position in the global economic and state systems. It enables and constrains 5

8 strategic interactions and institutions. Social structure usually changes very slowly, and is thus often considered as providing parameters for analyses of strategic interaction and institutions. Some aspects of Zambian social structure have changed over the long time period analyzed in this paper in ways that affect this analysis, so structural properties and relationships are considered as quasi-parameters (Grief and Laitin 2004: ). Zambian social structure contains widely held values and beliefs including nationalism, democracy, development, and economic inequality/status consciousness; a cleavage structure that includes race, ethnicity, formation of a dominant class and other nascent classes, religion, and organized groups representing one side of some of these cleavages; and an unevenly developed economy dependent on mineral exports and-until recently-heavily burdened by debt. Social structure also includes a society s position in the global system of sovereign states and international governmental and nongovernmental organizations. Values and beliefs provide information, the economy provides energy (resources and behavior), and cleavages, domestic politics and the global system provide both. The process of interaction over time between the two general causal mechanisms presented above can be illustrated by the way in which commonly held beliefs and values of Zambian leaders, mediated by institutions, enable and constrain their preference-based strategic interactions, but are also maintained or changed by these interactions and institutions. Most values are exogenous to institutions, and thus are cultural aspects of structure. The modal forms of beliefs, values and norms can be treated as ideational structural properties if they have significant effects (Inglehart and Welzel 2005: 231-9; Welzel and Inglehart 2007). They affect behaviour in conjunction with other factors because they are known to be relatively widely shared and relatively intensely held, and going against them is often sanctioned (Biddle 1986: 79; Kowert and Legro 1996; Lane and Ersson 2005: 273-7; Parsons 1960: 174-6), but also because they may be internalized and thus followed unconsciously (Abdelal et. al., 2009: 21-2). A number of authors (Barry 1978; Hechter 1987: 4-8, 20-25; 1992; Johnson 2003: ) have challenged the causal efficacy of values and norms, often citing Swidler (1986) who sees them as part of a cultural tool-kit, and their causal significance not in defining ends of action, but in providing cultural components that are used to construct strategies of action (273), but since she acknowledges that strategies of action are 6

9 cultural products, her position is actually similar to the one advocated here. Johnson accepts Swidler s definition of culture, but he says (2002: 243) that world view and rationality are reciprocally constraining and includes values and norms in world views. The critics of the causal power of values and norms focus on their independent rather than interactive effects, and do not conceive of informationenergy linkage as a general causal mechanism. Values norms, and beliefs constrain and enable strategic interaction, and norms are what make organizations also institutions. To a considerable degree, utility and interests are culturally defined. On the other hand, values, norms, beliefs, and identities do not exist in a vacuum, but rather must be interpreted and sustained by individual and group behavior in institutions and strategic interactions that are primarily in accordance with them. Hypothesis 3. Leaders values and beliefs in nationalism and democracy, fostered by education, urbanization, economic development, Christianity, and external conflicts, provide the most powerful explanation of why the prevailing pattern of enabled and constrained strategic interaction has not resulted in less nationalism and democracy, more intense cleavage-based conflict and violence, and greater corruption than has occurred in post-independence Zambia. Leaders values of inequality/status consciousness have the opposite effects (Scarritt 1971, 2010). Combining information and energy within and among strategic interaction, institutions, and social structure leads to the following hypotheses relevant to presidential selection and power exercising processes: Hypothesis 4. Nationalism has interacted with the strong presidency and plurality electoral institutions and the structure of politicized ethnic groups, which I call ethnopolitical groups (Scarritt 2010, 2006), to focus (weakly institutionalized) political parties on presidential elections involving broad multiethnic coalitions. Hypothesis 5.Path dependence, based more on self-reinforcement than positive returns (Page 2006: 88; Grief and Laitin 2004: ), explains why continuities in the Zambian presidential selection and power exercising processes (and other aspects of electoral and party politics) can be explained by previous hypotheses, with the path originating in the rise of African nationalism and continuing during the intervening years. Hypothesis 6.Leadership, especially but not exclusively by presidents--which includes articulating values and beliefs and learning from past strategic interaction/institutions/structure--and mobilization--which includes articulation of demands and support, protest, and organizing--explain why strategic interactions over presidential selection and exercise of power have been more favorable to democracy since 2001, although this change has been limited by path dependence. 7

10 Strategic Interaction over the Presidency of the Nationalist Movement in the Colonial/Nationalist Period, : Kaunda vs. Nkumbula 1.Kaunda vs. Nkumbula. Harry Mwaanga Nkumbula was elected president of the African National Congress (ANC) of what was then Northern Rhodesia in 1951, replacing Godwin Mbikusita Lewanika. Under his leadership, ANC became more militant, focusing on opposition to the creation of the Federation of Rhodesia and Nyasaland (Larmer 2011: 26). Nkumbula and most politically aware Africans in Northern Rhodesia correctly viewed Federation as a scheme to expand the white settler domination of Southern Rhodesia to Northern Rhodesia and Nyasaland and thus to avoid structural changes favoring Africans in those territories. After the anti-federation campaign failed and the Federation was established in 1953, ANC focused on boycotting white-owned shops that discriminated against African customers. Nkumbula s critics within the movement began to think that he was neither sufficiently militant nor organizationally effective to lead ANC s campaign for independence in the near future; they also criticized his personal habits, especially his heavy drinking. 3 Their doubts were significantly increased when Nkumbula agreed that ANC would participate in the 1959 elections under the Benson Constitution (named for Governor Arthur Benson) under which Africans were allocated a small minority of seats in the Legislative Council (Mulford 1967: ). While both Nkumbula and his critics were nationalists and were fighting for greater democracy, their strategic interaction involved mobilizing support from their respective ethnopolitical groups, the easiest enabling structural cleavage to activate. 4 When Nkumbula was challenged by his critics for the ANC presidency in 1958, he packed the meeting of the General Conference with members of his Tonga group, who had long been his strongest political and financial supporters (Macola 2010: 30-42, 78-83, 89-90, ). Although Nkumbula s critics came from all other ethnopolitical groups, their popular support and organization were strongest among the Bemba, especially those living in the economically vital Copperbelt. Their leader was Simon Kapwepwe. When Nkumbula s critics saw that they could not gain control of ANC by ousting Nkumbula, they formed a new party, the Zambia ANC, which was soon transformed into the United National 8

11 Independence Party (UNIP). Kenneth Kaunda, who had been a loyal Secretary General under Nkumbula (Kaunda 1962: ), and who was considered to be above ethnopolitical conflict because his parents were foreign born, was chosen to lead this party (Mulford 1967: ; Roberts 1974: 246). Complex institutional rules of the 1962 restricted franchise election allowed a limited ANC electoral pact with the predominantly white party to prevent UNIP from achieving a majority of elected Legislative Council members, and ANC s decision to form a coalition government with UNIP temporarily strengthened Nkumbula s position (Mulford 1964). But UNIP s overwhelming victory in the universal franchise election in 1964, a few months before independence, substantially weakened Nkumbula and put Kaunda close to a winner take all position. Strategic Interaction over the Presidency in the First Republic, : Kaunda vs. Nkumbula and Kaunda vs. Kapwepwe 2.Kaunda vs. Nkumbula The assumption of presidential powers by Kaunda at independence, the size of his parliamentary majority, and the continuing widespread perception of Nkumbula s weaknesses that had led to the ANC-UNIP split combined to convince Kaunda and UNIP that ANC could be eliminated in the next election and they could thus move to a full winner take all position while preserving democracy and increasing national unity. This proposed strategy underestimated both the strength of Nkumbula s support among the Tonga and the weakness of UNIP s support among the Lozi ethnopolitical group, especially after the UNIP government: 1) abrogated the Barotseland Agreement which gave their province a degree of autonomy, and 2) prevented them from working in the South African mines, their best source of income. These actions led to the formation of the Lozi-based National Party (NP), whose members joined ANC after the NP was banned for violent activities, and thus to a doubling of ANC s parliamentary seats in the 1968 election (Macola 2010: ; Mwangilwa 1982: ). Talks about merging UNIP and ANC were held between 1964 and 1969, but nothing came of them. Nkumbula successfully confronted several challenges to his leadership from within ANC. 9

12 3.Kaunda vs. Kapwepwe Conflict within UNIP increased sharply during this period. At its General Conference in 1967 the party held competitive elections for its Central Committee for the first time. A coalition of Bemba and Tonga candidates narrowly defeated a rival coalition of Easterners and Lozi for all of the positions for which they opposed one another. Most importantly, Kapwepwe replaced an Easterner as party (and thus national) vice-president, and a Tonga replaced a Lozi as secretary general. 5 This is one of two examples in Zambia s electoral history of a minimum winning ethnopolitical coalition that explicitly excluded other groups, and it had significant negative consequences for Kapwepwe and his heretofore close relationship with Kaunda (Mwanakatwe 2003: ). Kaunda attempted to maintain peace and unity within the party with only limited success, first by accepting the election results, then by changing the party constitution and appointing a new Central Committee in 1969, and at one point in 1968 by briefly resigning as President (Hall 1969: ). In the process he came to see himself as more indispensable to UNIP and the nation. In 1970 Kapwepwe resigned as Vice-President, probably after consultation with Kaunda (Pettman 1974: 57), and in 1971 he resigned from UNIP to lead the newly formed United Progressive Party (UPP). This party attracted a number of Bemba UNIP members, especially on the Copperbelt, who had long felt that Kapwepwe was the rightful leader of UNIP and the country. It also attracted people from all provinces who were disappointed with the fruits of independence. UPP was subjected to several forms of repression by the government, but Kapwepwe nevertheless won a legislative seat in the byelections of December 1971 ( Gertzel et.al. 1972; Mwangilwa 1986: 151). UPP and ANC agreed to cooperate but not to merge, and could have defeated UNIP in the 1973 elections. How effective UPP- ANC cooperation would have been, especially regarding selection of a presidential candidate, will never be known because Kaunda and UNIP decided in early 1972 to ban UPP and to institute a one-party state at the beginning of Kapwepwe and Nkumbula unsuccessfully asked the courts to block this move, so Kaunda and UNIP achieved a short term winner take all outcome. 10

13 Strategic Interaction over the Presidency in the Second Republic, : Kaunda vs. Kapwepwe and Nkumbula; Kaunda vs. Coup Plotters; Kaunda vs. Chiluba 4. Kaunda vs. Kapwepwe and Nkumbula The one-party state was supposed to foster democracy and nationalism; it was called one-party participatory democracy and was designed to eliminate or reduce ethnopolitical conflict by being all inclusive. Nkumbula and many former ANC leaders very quickly joined UNIP but were not permitted to seize control of that party s organization in Tonga areas. Kapwepwe and former UPP leaders were not asked and did not ask to join until 1977, and then were not permitted to seize control of UNIP s organization in Bemba areas, although pro-upp underground activity continued within UNIP (Larmer 2011: 95-99, ). Both former opposition leaders announced their intention to contest the party/national presidency in the 1978 elections, but party rules were quickly changed to prevent them from doing so. Kaunda again scored a winner take all victory, but continued to meet with both opponents and urge them to work within UNIP. Kapwepwe died suddenly in early 1980, while he was involved in planning the attempted coup discussed below. Nkumbula was awarded Zambia s highest honor in late 1982 shortly before his death in Kaunda vs. coup plotters In the early 1970s Zambia entered a period of economic decline that lasted about thirty years. Kaunda used his increasing presidential powers to prevent effective challenges to his Presidency until It was clear after 1978 that electoral challenges from within UNIP over the presidency were impossible. But in a declining economy his authoritarianism, although more enlightened than that of most African presidents, became increasingly unpopular. His ideology of Humanism and the Leadership Code that was enacted to implement this ideology were particularly disliked by educated and relatively wealthy Zambians because of their increasing class consciousness, inegalitarian values and falling incomes. Thus strategic interaction could now be enabled by a class cleavage among Zambian Africans as well as an ethnopolitical one. Some members of this class planned the first serious coup attempt in 1980, probably under the leadership of Valentine Musakanya, although Godfrey Miyanda was the military leader. The 11

14 plotters contacted Kapwepwe and Zambia Congress of Trade Unions (ZCTU) president Frederick Chiluba, and both cautioned against the use of violence. Kapwepwe was willing to become President after the coup if violence was minimized, but he died before the coup was attempted; Chiluba was unwilling to order union demonstrations to provide the rationale for the coup. Kaunda was aware that a coup plot existed, and the plotters were arrested before they could act (Larmer 2011: ; Mwanakatwe 2003: ). Another serious but unsuccessful coup attempt, led by General Christon Tembo and reputedly supported by a number of top level politicians and businessmen, occurred in late Little has been written about this attempted coup; Tembo was the only person tried in connection with it, and he was pardoned (Mwanakatwe 1994: ). A third coup was attempted by Lieutenant Mwamba Luchembe, acting primarily on his own, in June In spite of Kaunda s greatly reduced popularity by then, and the popular celebrations when it appeared to have been successful, most of the military opposed this coup and thus it too was unsuccessful. I believe Zambian leaders knew that change was coming, as described in the next section, and most wanted it to come democratically. 6.Kaunda vs. Chiluba In 1991, after twenty-seven years in power, Kaunda became the loser in winner take all strategic interaction, and initially accepted this outcome gracefully. The winner who initially took all was Frederick Chiluba, but his victory was quickly challenged in a number of strategic interactions. The Movement for Multiparty Democracy (MMD) was founded in mid 1990 and became a political party in early Chiluba was easily selected party president because of his union experience and support, and his supposedly strong support for democracy, nationalism, and free markets, although being a Bemba was also relevant. Kaunda s and UNIP s initial reaction to protests over the economy and the one-party state was to schedule a referendum on returning to a multiparty system. But after MMD was formed and demonstrated its high level of support, the referendum was abandoned because its outcome was clearly going to favor multiple parties. After considerable conflict over a new constitution, the MMD won the 1991 election overwhelmingly and Chiluba became President. Kaunda s initial gracious acceptance of 12

15 defeat was widely praised (Kabwe 1997), and led to the first suggestions that he could be useful as father of the nation since he was retiring from politics. Strategic Interaction over the Presidency in the Third Republic: The Chiluba Years, : Cbiluba vs. Kaunda; Chiluba vs. Mwanawasa; Chiluba vs. Mwila and anti-third term MMD leaders; Chiluba vs. Sata 7.Chiluba vs. Kaunda Because of Chiluba s insecurity about Kaunda s possible return to power and challenges from within the MMD, the return to a multiparty system did not eliminate the winner take all-uncertain and unstable outcomes nature of strategic interactions over Zambian presidential selection and use of power (van Donge 1995). Chiluba s first-hand experience with Kaunda s authoritarian rule caused him to hold democratic and capitalist values, but he did not maintain them when confronted by the challenges of strategic interaction with UNIP, supporters and dissidents in MMD, and international donors. Within a few months of his election, Chiluba fired a number of members of his cabinet and others resigned, with mutual accusations of corruption and incompetence. These dissidents did not challenge Chiluba s position within MMD because it was still very strong, but most of them joined the newly formed National Party (NP) and later other parties. Meanwhile, in 1993 a document advocating civil disobedience called the Zero Option Plan was revealed to have been circulating within UNIP. Chiluba declared a state of emergency and detained Kaunda and others briefly. After overcoming resistance in UNIP, Kaunda announced his decision to re-enter politics in Chiluba and MMD then amended the constitution to require that only those with two Zambian parents could be president and to establish a two term limit for them. UNIP decided to boycott the 1996 elections in response to this anti-kaunda measure. Kaunda and others were detained again in 1997 for allegedly being behind a very unsuccessful coup attempt by a few drunken officers. Upon his release Kaunda temporarily retired from politics, came back in 1999, and finally retired in 2000 (Smutanyi 2005: 81). Chiluba appeared to be winner take all against Kaunda once again. 13

16 8.Chiluba vs.mwanawasa The MMD founders who stood against Chiluba in the 1990 party election refused to serve as Vice- President under him. The opponents of Chiluba s selection sought a vice-presidential candidate who would serve under Chiluba, be acceptable to him, and balance the ticket. That candidate was Levi Mwanawasa, a lawyer who had served briefly as Kaunda s Solicitor General and had been fired from that position. Mwanawasa had been elected as chair of the MMD legal affairs committee in absentia. He may have been reluctant to run as party and national Vice-President, but seems to have easily been persuaded (Malupenga 2009: 39-51). Shortly after the general election he was seriously injured in an automobile accident and was unable to fulfill his duties effectively for a while. He felt that Chiluba and some cabinet members especially Michael Sata, who attacked Mwanawasa in harsh terms that earned him a rebuke from Chiluba (Malupenga 2011: 68-69) were corrupt and did not respect his Vice-Presidential role, and that Chiluba tolerated too much corruption. Thus he resigned as Vice-President in 1994, although he remained a member of MMD. He unsuccessfully challenged Chiluba for the party presidency in 1995 as a symbolic gesture, but got very little support allegedly because Chiluba bribed many delegates (Malupenga 2009: )--and he then officially retired from politics. 9. Chiluba vs. Mwila and anti-third term MMD leaders In early 2000 Ben Mwila, Chiluba s uncle (as defined in Zambia) announced that he would run for the MMD Presidential nomination in 2001 in defiance of Chiluba s order against declaring candidacy at that time, allegedly made to prevent intra-party conflict. Mwila was expelled from the party and harassed in a number of ways. It soon became obvious that Chiluba had issued the no declaration of candidacy rule because he wanted to amend the constitution and run for a third term. His campaign to do so was managed by MMD Secretary General Michael Sata, and was successful in getting him elected to another term as party president. Rakner (2003: 114) correctly states that the campaign was then met with a massive, and well-coordinated, response from the Zambian public. But this response was led by a number of ex-mmd leaders who opposed the third term because they wanted to become President, and this desire had led to their expulsion from MMD. The most prominent members of this group were ex- 14

17 generals Christon Tembo and Godfrey Miyanda, who became the Presidential candidates of other parties, as did Kaunda s son Tilyenji. In May 2000 Chiluba announced that he would not seek a third term. In order to avoid losing a winner take all contest after abandoning his third term quest, Chiluba had to control the selection of his successor. Given the conflicts that occurred between the two presidents in the early 1990s and after Mwanawasa s election (see below), observers have wondered why Chiluba chose Mwanawasa as his successor. Perhaps Chiluba thought he could dominate Mwanawasa and rule from off stage, but Mwanawasa s withdrawal from politics in 1994, when as Vice-President he first quarreled with Chiluba over corruption by ministers, does not support the idea of his manipulability. This withdrawal did mean that Mwanawasa was not one of the opponents whom Chiluba feared in the third term controversy, and that the former s honest image might help the party electorally. This and Mwanawasa s ethnopolitical similarity to Mazoka (see below), who was mounting a broadbased challenge to MMD, probably offer a better explanation (Simutanyi 2005: 76). 10. Chiluba vs. Sata Michael Sata expected to be nominated as the MMD Presidential candidate as a reward for organizing the third term campaign, and was furious when his old enemy Mwanawasa was nominated. He was expelled from the party for saying (correctly) that the method of Mwanawasa s selection went against party rules, and ran as the candidate of a new party, the Patriotic Front (PF). He later said he did not really support the third term, but was simply doing his party job (Rakner 2003: 114). Chiluba obviously opposed Sata s candidacy in 2001, supported it in 2006, and opposed it in 2012, when Sata won the Presidency and Chiluba died. Conclusion on These ten cases of strategic interaction over presidential selection and exercise of power support Hypothesis 1-5 presented above. Strategic interactions over presidential selection and use of power resulted in winner take all outcomes after periods of uncertainty, but most of these outcomes were short lived. Strong presidential power remained an equilibrium outcome. In spite of strategic interactions that 15

18 appeared to undermine nationalism and democracy, adherence to them as values remained constant. The dominant roles of presidential elections and ethnopolitical cleavages in party politics is well illustrated, although they cannot be fully explained without examining other aspects of electoral and party politics. This path lasted through four of the time periods analyzed in this paper that involved three major institutional changes, and the cases show that later interactions evolved in part from earlier ones, although not through increasing returns. Strategic Interaction in the Third Republic: The Mwanawasa Years, ; Mwanawasa vs. Mazoka and others; Mwanawasa vs. Chiluba; Mwanawasa vs. Kavindele and Mumba; Hichilema vs. Sikota; Mwanawasa vs. Sata and Hichilema Mwanawasa s strategic interactions were substantially although not entirely different from those of his predecessors, as the following cases illustrate. Those relevant to his relations with his predecessors and successors and Presidential selection and power generally involved: 1) building the broadest possible ethnopolitical coalition as a form of nationalism, incorporating all groups and sub-groups within them, in accord with his own mixed ancestry. (Critics say he made ethnopolitical identities more important to the entire population rather than just a tool in strategic interaction among leaders.); 2) declaring that he would not pick a successor but would let the party do that democratically; 3) maintaining the powers of the presidency at the expense of developing a strong leadership team that might be too critical of his policy choices, while making these powers seem less oppressive to many Zambians; and 4) taking an anticorruption stance, often without implementing it very effectively. He benefitted from economic growth, which was mainly due to increased Chinese and Indian demand for copper, but did not aspire to change other structural properties of Zambian society. He distanced his strategic interactions and policies from those of Chiluba by using the term new deal MMD, but rather than promising significant institutional change, he let these questions be answered through a constitutional review process that would effect change only after he was out of office. In 2003 he supported Kaunda being given Zambia s highest honor, and he subsequently involved Kaunda in a number of father of the nation tasks. 16

19 While Kaunda s and Chiluba s strategic interactions generated both strong support and strong opposition from other leaders, Mwanawasa s resulted in more ambivalence but a more positive response on balance. One leader interviewed in 2007, representative of the social democratic intellectuals who left MMD under Chiluba but returned under Mwanawasa, characterized him as a decent citizen who has made mistakes. Another member of this group said Mwanawasa is handicapped by the lack of an ideology, but he is a good man. The Post, a newspaper that frequently criticized Mwanawasa, editorialized on July 10, 2008 shortly after his stroke that President Mwanawasa in his awkward and sometimes clumsy way seems to have laid a reasonable foundation for Zambia to make progress in all areas of human endeavor. He has been determined to fight some very dangerous vices and practices at great personal and political peril. Writing in 2011, Bauer and Taylor (53-54, 85-86) praise his legacy: Zambia under the administrations of Mwanawasa and Banda is more democratic certainly more competitive than under past regimes, while noting some autocratic tendencies. In terms of the framework presented above, I suggest that Mwanawasa s strategic choices, based on firmer value commitments than Chiluba s and carried out more effectively through strategic interaction, represent significant learning and improved leadership in a situation of long term path dependence. Real world leaders without radical visions tend to learn incrementally. 11.Mwanawasa vs. Mazoka and others In order to get elected President Mwanawasa had to defeat four MMD defectors (Sata, Mwila, Tembo, and Miyanda), UNIP, and five other candidates. By far the greatest challenge came from Anderson Mazoka of the United Party for National Development (UPND). After getting no response to his offer to lead a united opposition to the MMD, Mazoka founded the UPND in 1998, and generated substantial support from Tonga, Lozi and Northwestern Province groups, and more limited support elsewhere. One of the people whose support he sought right away was Levi Mwanawasa, saying the two of them were the only ones who could save Zambia and Mwanawasa could be his successor as President. Mwanawasa was also approached by the Forum for Democracy and Development (FDD) that eventually nominated 17

20 Tembo. In both cases, he declined to get involved (Malupenga 2009: ). He remained loyal to the MMD, and perhaps reluctantly accepted their nomination, telling them that: I left government because I was not happy with the tenets that we were exhibiting. As far as I am concerned, when I come back, it will be to continue with the crusade which I had left. So its not too late. You can change your decision and pick someone else (Malupenga 2009: 135). The votes received by Mwanawasa and Mazoka were so close that it will never be known who really won, but the former was declared the winner. Mazoka and other losing candidates petitioned the courts over the results, and it was not until 2004 that the Supreme Court decided in favor of Mwanawasa. Mazoka considered this a winner take all outcome, possibly because of his ill health, but Mwanawasa was determined not to treat it as such. Sata finished in seventh place with relatively few votes. 12.Mwanawasa vs. Chiluba Chiluba remained party president for a while after Meanawasa was elected national President, hoping to retain some control of the government, but resigned that post in order to get his pension benefits. Mwanawasa removed a number of Chiluba appointees from the Cabinet, posts in the MMD, the civil service, and the military. In July 2002 Mwanawasa asked Parliament to remove Chiluba s immunity from prosecution for corruption, detailing incidences of Chiluba s corruption, and Parliament did so. A London High Court judge found Chiluba and others guilty of misappropriating forty million dollars from the Zambian government, but Chiluba and his lawyers were able to delay registering that finding in Zambia. Mwanawasa offered to pardon Chiluba if he repaid 75% of what he had stolen. Although these actions led to conflict with Chiluba supporters in MMD, Mwanawasa kept control of the party by firing his most vocal opponents, some of whom were expelled from the party. Since Chiluba retained significant influence in Zambian politics, this was not a winner take all outcome. 13.Mwanawasa vs. Kavindele and Mumba In mid 2003 Nevers Mumba, former leader of a small opposition party, joined MMD and was appointed Vice-President by Mwanawasa. He replaced Enoch Kavindele who had been initially appointed by 18

21 Chiluba, and had announced that he would be Mwanawasa s successor, accusing him of corruption. A motion to impeach Mwanawasa over the selection of Mumba failed badly. Just over a year later Mumba was fired for claiming that the Congolese government was funding the opposition. In return, he announced that he would challenge Mwanawasa for MMD president and national presidential candidate at the party s convention, and called for investigation of him in published comments in violation of party rules. He was expelled for doing so. At the convention Mwanawasa handily defeated Kavindele, who had planned to run for party vice-president until that election was cancelled. This was a winner take all outcome. 14.Hichilema vs. Sikota In May 2006 Anderson Mazoka died after a long illness. Although he knew of the seriousness of his illness, Mazoka had not named a successor, for which different explanations have been offered. Sakwiba Sikota, a Lozi, had been acting president of the party during Mazoka's illness, and apparently assumed that he would succeed him. But some Tonga activists insisted that the party president must be a Tonga, and Haikinde Hichilema, a. wealthy businessman who had donated money to the party but had not otherwise been involved in it, was elected leader by more than 70% of the vote at the party's general assembly in July. Sikota and his supporters left the UPND to form the United Liberal Party (ULP) a few days later. This return to winner take all interaction and exclusionary ethnopolitics in the new millennium surprised many Zambians and foreign observers. 15. Mwanawasa vs. Sata and Hichilema After intense maneuvering by other aspiring presidential candidates for the 2006 Presidential elections, it became clear that there were three viable contestants: Mwanawasa, Sata, and Hichilema. Sata stood out among ex-mmd pro-chiluba Bemba candidates and posed the most serious challenge to Mwanawasa by adopting populist economic policies that strongly appealed to the urban population, especially the poor. Although Sata s harshest criticisms were directed against foreign investors especially mainland Chinese--his injection of class issues into the campaign was a significant change in Presidential selection 19

22 politics (Larmer and Fraser 2007). Conflict between Sata and Mwanawasa increased in the months before and after the election was won by Mwanawasa. Sata was arrested several times on a variety of charges, but was never convicted. After the election he led large demonstrations protesting the results, and tried to use PF controlled local councils to cut local taxes, increase council powers, and review foreign investors contracts, calling this PF-MMD power sharing. Mwanawasa said this was interfering with national government powers, which was treasonous, but Sata and the PF soon offered Mwanawasa conditional support, and Sata soon faced rebellion from some of his MPs. Mwanawasa won the election because he compensated for the loss of Bemba support to Sata with strong rural support in the Eastern, Central, Western, and Northwestern provinces, at the expense of UPND, FDD and UNIP. He appointed Rupiah Banda, a UNIP member until 2001, as Vice-President. In April 2008 Sata had a near-fatal heart attack, and Mwanawasa had him evacuated to South Africa for life-saving treatment. Sata resolved to reconcile with Mwanawasa as a way of thanking him for saving his life. The road to reconciliation was not without problems (Malupenga 2011: 198). Both of them celebrated this agreement as a major positive step for state nationalism and democracy and against bitter ethnopolitical strategic interaction, while giving very different interpretations of it. Sata said that it meant that Mwanawasa had agreed to consult him on matters that he considered very important, while Mwanawasa denied it meant that he had chosen Sata as his replacement and emphasized that the MMD should defeat Sata and retake PF parliamentary and local government seats in the next election. Sata naturally denied this interpretation, but both leaders said it would take time for their agreement to bear fruit. Mwanawasa s stroke and death a few months later prevented a test of this prediction, but reconciliation improved both of their images, giving symbolic support to Mwanawasa s more democratic strategic interaction and helping to broaden Sata s political appeal (Larmer2011:262-63; Economist Intelligence Unit June 2008: 8). Mwanawasa recalled that Hichilema came to see him in early 2006, saying that he was not a politician and therefore should not have had trouble getting government contracts because of his alleged 20

23 membership in UNDP, and that if he ever joined politics it would be with the MMD because he liked Mwanawasa s policies. Mwanawasa was very unhappy a few months later when Hichilema when contesting for the UPND leadership said he was a founder of that party. He admitted that at their first meeting he thought Hichilema had great political potential, but denied the rumor that he had offered him the vice-presidency of MMD (Malupenga 2011: ). Some UPND members proposed a coalition with MMD after their party did poorly in the election due to internal conflict and Mwanawasa s outreach, but Hichilema did not accept this proposal. Strategic Interaction in the Third Republic: The Banda Years, : Banda vs. Magande; Banda vs. Sata and Hichilema 16.Banda vs. Magande Mwanawasa announced that he would not choose a successor, but said that he hoped the party would pick someone younger. Since Banda was more than ten years older, this seemed to rule him out. After Mwanawasa s death Banda became Acting President until a new President was elected in a by-election to finish the term. Within the MMD competition quickly narrowed to Banda--who had the advantages of incumbency, wealthy supporters, and apparent commitment to step down in 2011 (Cheeseman and Hinfelaar 2009: 59-61)--and Finance Minister Ng andu Magande. Mrs. Mwanawasa, who had presidential ambitions herself, strongly backed Magande, saying he was Mwanawasa s choice. Banda s supporters vehemently denied this, pointing out that Mwanawasa had publicly indicated he had no choice. They pointed out that Banda had delivered Eastern support to MMD, while Magande had not delivered Tonga support. Banda was the victor in the National Executive Committee (NEC) by 63 votes to 11. Magande was expelled from MMD in 2010 for trying to challenge Banda again, clearly a winner take all outcome. 17.Banda vs. Sata and Hichilema Banda said he would follow Mwanawasa s policies in governing. Immediately after Banda s selection as MMD candidate, Sata declared he would protect Chinese investors in Zambia, saying they had responded 21