Han-Hui Hsieh Department of Political Science, National Taiwan University

Reshaping Power Structures, Shifting Government Systems: The De Facto Semi-Presidentialism in Kenya and Zimbabwe A New Paradigm for Resolving Post-Election Controversies and Democratic Consolidation Han-Hui Hsieh Department of Political Science, National Taiwan University This paper addresses the changing politics in Kenya and Zimbabwe after their controversial presidential elections, and argues that the current semi-presidential system resulting from the power sharing agreements better fosters democratic consolidation compared with presidentialism. This paper defines the current form of government in Kenya and Zimbabwe as de facto semi-presidential, de jure presidential, which is a situation sui generis. The primary discovery here is that Kenya and Zimbabwe are unique to all the existing three major categories of semi-presidential states in that their way entering semi-presidentialism is not derived from the historical background or government structure inherited from the old regime, hence is a new entry mode. This study applies a semi-presidential analytical framework comprising the upstream (entry mode), midstream (system operation and proper subtypes), and downstream (institutional sustainability or transformation) of the system to demonstrate how such form of government is a functional mechanism to settle disputes and better foster and consolidate democracy. The paper further hopes to establish an argument that semi-presidentialism can be a useful means to settle political disputes in divided post-election African states. As such, the prospect of this paper is to seek relevant cases to further expand the validity of its proposed argument and thereby making it an applicable African paradigm for democratic consolidation. Introduction The reason why this essay puts Kenya and Zimbabwe in comparison is that there are many similarities between the two cases in terms of recent political development, especially in the background and process of their transition from the original presidential system into semi-presidentialism. First of all, the mutual background of their entry into semi-presidentialism is that they both underwent controversial presidential election results that later triggered a series of civil unrest and political crises that finally led to the power-sharing agreements that brought out the transition. In the presidential elections of Kenya and Zimbabwe, opposition parties of both cases accused the incumbent presidents Mwai Kibaki and Robert Mugabe, both of whom fell behind in polls, of manipulating elections result and engaging in fraudulence. Protests against Kibaki and Mugabe soon escalated to riots and political crises that

drew international attention and intermediation. In both cases agreements were reached to settle their political turmoil in the same manner, in which the incumbent president, in order to stay in power, was willing to render part of his authority to the newly established office of prime minister and invite his former rival to sit in that position. Furthermore, the opposition party that held parliament majority was also included in a coalition government, in which the president did not possess unilateral power to dismiss the prime minister and cabinet ministers. Though the president s power was reduced in the power-sharing agreement, yet he could still appoint vice president, the prime minister and his deputies, as well as cabinet ministers, and that his party still occupied more important positions in the government. Therefore, such situation in Kenya and Zimbabwe meets the requirements of a semi-presidential regime, namely a president directly elected who possesses considerable powers, and an independent position of prime minister whose government is responsible to the parliament, according to Maurice Duverger (1980: 165-187). It signifies that Kenya and Zimbabwe after the enactment of their respective power-sharing agreements now both enter semi-presidentialism. In this essay, the author adopts a semi-presidential analytical framework to examine the constitutional development and form of government in Kenya and Zimbabwe by discussing the upstream, midstream, and downstream of the semi-presidential system in the two cases. 1 A Unique Entry Mode into Semi-Presidentialism After the previous brief introduction of what elements led to the semi-presidential system currently in practice and the background of the political transition in Kenya and Zimbabwe, the content of this section, is about how Kenya and Zimbabwe entered semi-presidentialism, namely the upstream study of semi-presidential system in these two cases. The main argument is that the two countries underwent similar political turmoil resulted from controversial presidential elections as their background, and later with an effort to resolve the political crises, both Kenya and Zimbabwe adopted the same solution of establishing the position of prime minister to make the presidents share power with their rivals, which results in a design of semi-presidential system. Kenya Kenya was originally a presidential republic, which was explicated in its constitution. The article 4 of the Constitution of Kenya states that There shall be a President of Kenya, who shall be the Head of State and article 23 further states that 1 Such analytical framework to examine the functioning of semi-presidential system is introduced by Dr. Yu-Shan Wu in his essay Study of Semi-Presidentialism: A Holistic Approach in 2008.

The executive authority of the Government of Kenya shall vest in the President. Therefore the president of Kenya is not only the head of state, but also the head of government. The cabinet consists of the president, vice president, and cabinet ministers, without mentioning the position of prime minister in the original Chapter II of the constitution, which deals with the Executive, until the National Accord and Reconciliation Act that settled the political crisis in 2008. The president also has the power to appoint and to remove ministers according to article 16 of the constitution. From the above articles we can observe that the constitutional design of Kenya s government system is actually a presidential one. On December 27, 2007, Kenya held its general elections comprising the presidential and parliamentary elections. The incumbent president Mwai Kibaki faced the challenge from opposition leader Raila Odinga, and the pre-election opinion polls as well as the election-day exit polls all showed that Odinga was the leading candidate. 2 Odinga announced his winning of the election before the official election result was promulgated, 3 however the Electoral Commission of Kenya (ECK) declared Kibaki the winner with 4,578,034 votes, or 46.4% supporting rate, leading Odinga s 4,392,553 votes, or 44.1% of support. 4 The difference of the results between previous prediction and the official outcome triggered controversy. The opposition camp claimed that the results were manipulated by the ruling president Kibaki, 5 and accused that the ECK was controlled by the government, therefore refused to accept the results. This situation subsequently led to a series of riots and later expanded to a countrywide civil unrest across Kenya. Ethnic violence also erupted between the major tribes, especially the Kikuyus and the Luos, which Kibaki and Odinga belongs to respectively, and the two tribes accused each other of ethnic cleansing or genocide. In the unrest, it was reported that around 1,500 people died, 6 nearly 250,000 people were forced to flee home, 7 and the United Nations indicated 2 According to the exit poll conducted on the election day by the International Republican Institute, Odinga had the support rate of 46.4%, leading Kibaki s 40.3%. See: http://www.iri.org/africa/kenya/pdfs/2008%20august%2014%20kenyan%20election%20day%20poll,%20december%2027,%202007.pdf 3 The primary results released by the Electoral Commission also showed that Odinga was the leading candidate. See: Wallis, Daniel and Katie Nguyen. 2007. Opposition Leads Kenyan Election Count, Reuters, December 28, 2007. http://www.reuters.com/article/idusl277107920071228 4 The government spokesperson s office announce the result with slightly different numbers of Kibaki winning by 4,584,712 votes (47%) over Odinga s 4,352,993 votes (44%). See: http://www.communication.go.ke/elections/default.asp 5 The reports from the U.S. and the European Union considered that there was fraudulence and that the election result was manipulated in Kenya s presidential election. See: Kenya Election Was Rigged, U.S. Envoy Says, CTV News, January 7, 2008. http://www.ctv.ca/servlet/articlenews/story/ctvnews/20080106/kenya_opposition_080107?s_name= &no_ads 6 See: Kenya Police Ran Death Squads, BBC News, February 25, 2009. http://news.bbc.co.uk/2/hi/africa/7909523.stm 7 See: Kenya Opposition Cancels Protests, BBC News, January 7, 2008.

that there were some 600,000 displaced people within Kenya. 8 The series of unrest derived from the controversial presidential election results finally led to international concerns and intervention. Former U.N. secretary-general Kofi Annan mediated negotiations between the two sides to resolve the political crisis. Odinga, convincing that himself had won the presidency already, once insisted to hold another round of election and establish a new electoral commission, yet subsequently agreed to accept a power-sharing arrangement to reduce Kibaki s presidential power meanwhile creating the office of prime minister for himself. Both camps of Kibaki and Odinga finally reached a power-sharing agreement on February 28, 2008, which was later passed by the parliament to become the National Accord and Reconciliation Act. President Kibaki s party, the Party of National Unity (PNU), and the supporters of Odinga, the Orange Democratic Movement (ODM), together formed a coalition government, and the 40 cabinet positions were to be distributed equally among the two parties. According to the agreement, Kibaki would remain as the president while Odinga should become the prime minister. Furthermore, each camp also appointed one deputy prime minister. 9 The coalition government took office on April 17. In the election of the National Assembly of Kenya which took place alongside the presidential election, president Kibaki s party the PNU only won 43 seats, meanwhile the opposition ODM led by Odinga gained 99 out of 207 seats, thus becoming the largest party within the parliament. From the above-mentioned arrangement of the National Accord and Reconciliation Act, the office of the prime minister was reestablished in the government of Kenya, and the president yielded part of his power to the prime minister, which, according to Maurice Duverger, has met the requirements of what is called a semi-presidential regime, including a president directly elected by the people, a president with considerable power, and a prime minster responsible to the parliament, therefore the author considers the current form of government in Kenya a semi-presidential system. Zimbabwe Similar to Kenya, the form of government originally designed in the constitution of Zimbabwe was also a presidential system. The constitution states that There shall be a President who shall be Head of State and Head of Government in its article 27. http://news.bbc.co.uk/2/hi/africa/7174670.stm 8 See: U.N.: 600,000 Displaced in Kenya Unrest, CBS News, February 11, 2008. http://www.cbsnews.com/stories/2008/02/11/world/main3815702.shtml 9 Of the two deputy prime ministers, Musalia Mudavadi represents the ODM, while Uhuru Kenyatta from the Kenyan African National Union (KANU) represents the PNU.

Article 31D states that the president appoints ministers and deputy ministers, and article 31G states that the cabinet consists of the president, vice president, and ministers. From the articles on the executive powers in the constitution of Zimbabwe, we can observe that there is no such position of a prime minister, and according to the statement that the president is both head of state and the government, such design is clearly a presidential system. As what was previously arranged in Kenya, Zimbabwe held its presidential and parliamentary elections on March 29, 2008, and the incumbent president Robert Mugabe faced the challenge from opposition leader Morgan Tsvangirai. In the first round, Tsvangirai received 1,195,562 votes, or 47.9% of support, versus Mugabe s 1,079,730 votes or 43.2% supporting rate. 10 Since no candidate secured more than 50% of the total votes, there should be a second round election to decide the winner. However, there were controversies in the process and vote counts of the election, and Tsvangirai s camp claimed that their leader already won the presidency in the first round, and accused Mugabe s camp of conducting fraudulent behaviors. What followed the first round of election was violence between the two camps, which caused international concerns. After the escalation of violence, Tsvangirai announced that he would withdraw from the second round of election just days before the run-off, citing his supporters risked being killed. Subsequently, Mugabe became the only active candidate in the campaign and won his reelection with an overwhelming majority of 85.5% support. Meanwhile in the parliamentary election, Mugabe s Zimbabwe African National Union-Patriotic Front (ZANU-PF) gained 99 seats in the House of Assembly, while Tsvangira s Movement for Democratic Change (MDC) held 100 seats, thereby becoming the largest party in the parliament by its one-seat margin over ZANU-PF. 11 Concerning the post-election violence across Zimbabwe, 12 the Southern African Development Community (SADC) held an emergency conference to discuss the matter. There were also questions on the fairness of this election among the international community, requesting the African Union or the United Nations to mediate and investigate. Also the international society notified Zimbabwe that unless president Mugabe negotiated with the opposition to resolve the political crisis, otherwise there would not be international assistance to alleviate its hyperinflation. Thabo Mbeki, then president of South Africa, represented the SADC as a mediator, and ZANU-PF and MDC then settled on a power-sharing agreement. In the 10 See: Dzirutwe, MacDonald. 2008. Zim Heads for Run-Off, Independent Online, May 2, 2008. http://www.int.iol.co.za/index.php?set_id=1&click_id=3045&art_id=nw20080502152757558c887809 11 If adding the 10 seats of the MDC faction led by Arthur Mutambara, together the MDC holds 110 seats in the 210-seat parliament, which is a majority. 12 It was reported that about 200 died, 200,000 forced to flee home. See: Q&A: Zimbabwe Crisis, BBC News, January 30, 2009. http://news.bbc.co.uk/2/hi/africa/7293810.stm

Zimbabwean Power-Sharing Agreement, Mugabe still remained the president, while Tsvangirai would hold the newly established office of prime minister. According to that agreement, the executive power was shared between the president, prime minister, and the cabinet. Among the 31 ministerial positions, ZANU-PF acquired 15 while the MDC led by Tsvangirai held 13, and another smaller faction of MDC led by Arthur Mutambara gained 3. The two factions of MDC each appointed one deputy prime minister. The coalition government took office on February 11, 2009, and Zimbabwe under the power-sharing agreement also qualified the conditions of semi-presidential system, namely a president elected popularly with considerable power, and a prime minister responsible to the parliament. A New Entry Mode into Semi-Presidentialism After examining the process of entering semi-presidentialism in the cases of Kenya and Zimbabwe, it is clear that such process does not fit into the existing paths experienced by any previously or currently established semi-presidential republic. Among all the semi-presidential states, we can roughly categorize them into three major groups, namely Western European, Post-Leninist, and Post-Colonial, according to how they entered semi-presidential system (Wu, 2008: 8-10). The two cases of Kenya and Zimbabwe, if categorized according to their historical background, should have been considered to belong to the post-colonial group. However, as they did not adopt semi-presidential system in the first place after gaining independence, 13 therefore their initial choices of form of government were different from the other post-colonial semi-presidential states that mostly imitated their former colonizers political system when designing new constitutions, which subsequently led to the fact that when they turned into semi-presidential system, their entry mode was also different from those other countries and thus creating a new category of their own. Here the author would like to point out a primary discovery of this study as briefly mentioned previously, that Kenya and Zimbabwe s entering into semi-presidentialism is a unique way that deserves our further studies and search of similar new cases. Such new entry mode is different from all existing semi-presidential states in that it is not derived from the historical background or government structure of the ancien regime, but is triggered by similar political accidents that end up with a power-sharing agreement to form a de facto 13 Kenya gained independence in 1963 when it became the Dominion of Kenya with the British Monarch as its head of state, and there was a prime minister as head of government. In 1964 Kenya became a republic and the office of prime minister was abolished, while the office of president as both head of state and government was created, which was first held by Jomo Kenyatta, who was also the prime minister of the Dominion of Kenya. Zimbabwe gained its independence in 1980, and initially there was a weaker ceremonial presidency as well as the office of prime minister which occupied more executive powers held by Robert Mugabe. In 1987 Mugabe became the president and the position of prime minister was abolished when the president became both head of state and government.

semi-presidential system that leads to a transition in government system. Both the incumbent presidents of Kenya and Zimbabwe gave up part of their executive power in order to resolve the political crises that affected their legitimacy to rule, and their choice of reestablishing the position of prime minister might be a reference to the founding period of their countries history when there were still prime ministers, though such office was soon replaced by a powerful presidency. But considering the fact that the two presidents resorted to such mechanism in order to resolve political crises instead of for the purpose of reviving the old regime, therefore when compared with the entry mode of the established Western European, post-leninist, and post-colonial semi-presidential states, the cases of Kenya and Zimbabwe still provide a unique new entry mode. In the established Western European semi-presidential states, because of their parliamentarian traditions, and their consideration of creating a president with popular support as their head of state when designing their constitution, this group of states set up a mechanism to elect their presidents directly by the people. And the parliamentarianism combined with a president directly elected became a semi-presidential system. The post-leninist semi-presidential states inherited the dual leadership structure from their previous authoritarian regime design, in which the party general-secretary also head of state set the general direction of the country, while the government, which in theory should be responsible to the parliament, led by the premier would implement the policies decided by the head of state. After their democratization, such system design with two positions in the leadership continues, thus contributing to the trend of semi-presidential system among many newly-independent post-communist states. Lastly, the post-colonial semi-presidential states, of which most were former colonies of France or Portugal, both semi-presidential states, inherited such system from their former rulers. When these states gained independence, their constitutional design followed their colonists system, and the French Fifth Republic and Portugal after democratization both adopted a semi-presidential form of government, therefore making part of the newly independent former colonies semi-presidential. However, among the aforementioned entry modes, we discover that Kenya and Zimbabwe fit into none of them, hence the development of the two cases is described by the author as a unique new entry mode. Both Kenya and Zimbabwe adopted presidentialism after gaining independence, and there were no particular problems in the operation of the system or crisis that could break down their presidentialism, until there were overwhelming controversies over the results of presidential elections that subsequently led to civil unrest across these two countries. Under such circumstances, the incumbent presidents, on the one hand needed to deal with the domestic crisis and

international pressures, on the other wished to maintain their current position, therefore had to make certain adaptations and to accept power-sharing agreements. Although both Kibaki and Mugabe are popularly elected presidents according to constitutional regulation in Kenya and Zimbabwe, yet both now choose to share executive powers with their rivals to resolve the crisis, and agree to recreate the position of prime minister for opposition leaders Odinga and Tsvangirai. From the background and process that made Kenya and Zimbabwe enter into semi-presidentialism, we can observe the similarities between the two cases. To conclude such new entry mode into semi-presidentialism from the cases of Kenya and Zimbabwe, the author compresses the process into the following content: A powerful president who is directly elected and hence enjoys public support, when facing domestic challenge from the opposition, in order to deal with the political crisis and to maintain his own presidency, is willing to yield part of presidential power to a newly created prime minister, and the president together with the prime minister share the power to appoint cabinet ministers and the executive power of the government, thus such system is equipped with all the elements of semi-presidentialism. The Operation of the De Facto Semi-Presidentialism In this section, which could be considered the midstream semi-presidential system, a thorough study of the system operation of Kenya and Zimbabwe s semi-presidentialism will be provided through the analysis of the articles regarding executive powers and government operation in the two countries power-sharing agreements that made them semi-presidential states. Furthermore, the proper allocation of subtypes of the two countries semi-presidential system will be specified after examining the extent of presidential power in deciding ministerial positions and the triangular relationship between the president, the prime minister, and the parliament. Last but not the least, here we can find another striking similarity between the cases of Kenya and Zimbabwe in that their semi-presidential system currently in practice is actually stipulated in additional documents that provisionally operate alongside or amend the regulations regarding the system of government in the original constitution where there was a design of presidential system. Thus, such arrangement is defined as de facto semi-presidentialism, de jure presidentialism by the author, which is a unique situation in the study of semi-presidentialism. Kenya Under the original constitution of Kenya, the president is the head of state. The article regarding presidential powers includes article 16, which states that the president appoints ministers from the members of the national assembly, and the

office of a minister shall be vacant if the president so directs. Article 19 bestows upon the president the aforementioned powers regarding the assistant ministers. And according to these articles, one can see that the constitution renders the president full discretion on the appointment and removal of cabinet ministers, therefore the president is also the head of government. However, according to the National Accord and Reconciliatory Act that reduced the powers of the president, the original article 23 of the constitution which granted the executive authority of the government to the president has been weakened in that the power-sharing agreement transferred a significant amount of authority to the prime minister. The article 5 of the Act states that members of the cabinet include the president and vice president, the prime minister and deputy prime ministers, and the ministers. Article 3 states that the president nominates the prime minister and his deputies. Article 4 grants the prime minister the authority to coordinate and supervise the execution of the functions and affairs of the government of Kenya including those of ministries, thus we can clearly see that the prime minister leads the government. Moreover, the prime minister has to be a member of national assembly and also the leader of the largest party or coalition in the parliament. The original regulation in the constitution that the president can appoint or remove any minister does not extend to the prime minister and deputy prime ministers, and these positions shall be vacant only when the holders of office resign or cease to be member of the assembly, when the parliament has no confidence in the prime minister, or when the coalition is dissolved. In other words, in order to become prime minister, the occupier must meet certain qualifications relating to parliamentary support, and the office holder won t be unilaterally removed from position just because the president so directs, also such qualification that makes a person prime minister will not be affected or be deprived by the president, hence the office of the prime minister is in fact independent of the president. Moreover, regarding the composition of the cabinet, since the president only appoints but is not able to remove the prime minister and his deputies, and that the prime minister is responsible to the parliament because of the clause regarding no confidence, and that according to article 4 of the Act that the composition of the coalition government shall reflect the relative parliamentary strengths of the respective parties as well as take into account the principle of portfolio balance, one can clearly see that the president has lost complete control over the cabinet and that the series of articles regarding the cabinet of Kenya have met the principle of a parliamentarian system of government, which is then combined with a directly elected president to form a semi-presidential system. Furthermore, the power of the president to unilaterally remove ministers from their offices originally stated in the article 16 of the constitution has also been

deprived according to the article 4 of the Act that states the removal of any minister nominated by a parliamentary party of the coalition shall be made only after prior consultation and concurrence with the leader of that party, since in a coalition government, the appointment and removal of any member cannot be decided only by the president who is merely a member of one of the many parties that make up the coalition, especially under the circumstance that the current president Kibaki s party, the PNU, is not even the largest party in the parliament. Looking back to the regulations regarding the distribution of power between the president and the prime minister stipulated in the Reconciliatory Act, if the president wishes to remove the prime minister or any of the deputy prime ministers, it could only be done through the no confidence vote from the national assembly, but considering the fact that president Kibaki s PNU only has 43 seats in the parliament, which is roughly one-fifth of the total seats and not even half of the strength of prime minster Odinga s ODM, let alone controlling the votes required to pass a motion of no confidence, therefore it would be extremely difficult for Kibaki to remove Odinga from his position as head of government through this way. Aside from the above mentioned situation, the Act also states that the composition and allocation of cabinet minister portfolios must be consulted between all parliamentary parties that make up the coalition, thus the exercise of executive authority in leading the coalition government by the prime minister is in a sense independent from the influence of the president, and the prime minister and his government only have to be responsible to the parliament. Zimbabwe The constitution of Zimbabwe originally states that the president is head of state and head of government in its article 27, and that the executive authority of Zimbabwe shall vest in the president according to article 31H. Other sections in the article 31 further state that the president appoints his vice presidents, ministers and deputy ministers, and that these offices shall become vacant if the president moves them from office. Therefore, the original design of the form of government of Zimbabwe is clearly a presidential system. However, the Zimbabwean Power-Sharing Agreement has a new series of regulations regarding the framework for a new government in its article 20, which states that the executive authority of the government shall vest in and be shared among the president, the prime minister, and the cabinet. Also according to this article, the cabinet shall take decisions by consensus and take collective responsibility for all cabinet decisions. And from the above content one can observe that the original presidential power to fully control the cabinet is reduced. Article 20.1.3 of the Power-Sharing Agreement listed the president s authority,

including chairing the cabinet, exercising executive authority (shared with the prime minister), chairing the National Security Council, and appointing prime minister and his deputies as well as ministers and deputy ministers. Although the president keeps his power of appointing cabinet members, yet such power is restricted by related regulations in the agreement. For instance, the president has to consult with the vice presidents and the prime minister as well as deputy prime ministers to allocate ministerial portfolios, and that the ministers and deputy ministers may be relieved of their duties only after consultation among the leaders of all the political parties participating in the coalition government. Furthermore, the president must also consult with the prime minister the affairs such as appointing service/executive commissions in terms of the constitution and the dissolution of parliament, meanwhile, the president also makes key appointments after consultation with the prime minister. These articles show that the president yields part of his control over the government to the prime minister according to the regulation of the power-sharing agreement, and that the president enjoys no unilateral power in appointing or removing cabinet ministers, which also demonstrates the principle of the semi-presidential system. Regarding the power exercised by the prime minister, aside from the cabinet, the power-sharing agreement also establishes a new institution named Council of Ministers to ensure that the prime minister can properly execute his office and oversee the function of the government. Though article 20.1.4 of the power-sharing agreement states that the prime minister is only the deputy chairperson of the cabinet while the president is the chairperson, yet the same article also makes clear that the prime minister chairs the council of ministers, which literally consists of all cabinet ministers. And the primary function of the council of ministers is to assess the implementation of cabinet decisions, and also to report to the cabinet. From the content of the power-sharing agreement of Zimbabwe, one can observe that there is overlap between the authorities of the president and the prime minister in that the agreement states that both exercise executive authority, and they each chairs the cabinet and the council of ministers respectively, therefore the question is that how is the system with dual leadership actually works, and that who should be the one that leads the government? In order to clarify this point, the author would like to mention a similar arrangement in the semi-presidential system in the Republic of China, better known as Taiwan. First of all, from the previous articles regarding the authorities of the president and prime minister of Zimbabwe, one can observe that the cabinet headed by the president is the supreme policy-making organ in the government of Zimbabwe, while the council of ministers is the institution to implement the decisions. Such scenario resembles the division of responsibilities between the President of the

Republic of China and the Premier. The President of the Republic of China is the head of state who decides the general policy direction of the country, and the Premier is the head of government, heading the Executive Yuan, which is highest administrative organ of the state, to implement the policies decided by the President. When compared with the system of an enterprise, one can put the president of Zimbabwe in the position of the chairman of the board of directors, and that the cabinet is the board which acts as the highest decision-making body of the company, and the council of ministers of Zimbabwe serves as the executive division of the company to implement all the decisions made by the board of directors, thus its leader the prime minister shall fit into the place of the CEO. The Subtype of the Semi-Presidential System in Kenya and Zimbabwe After reviewing the relevant articles of the power-sharing agreements regulating the distribution of power between president and prime minister in Kenya and Zimbabwe, this section will focus on the actual operation of the semi-presidential system in the two cases. An analysis on the triangular relation between the president, prime minister, and the parliament will be provided in order to examine the extent of control over the cabinet by the president, i.e. the seats and portfolios that the president and his party can secure, under the circumstance that the president and parliament majority belong to different parties, namely under incongruence, in order to observe the system operation and to allocate the proper subcategories within the semi-presidentialism of these two countries. When it comes to distinguishing subcategories within semi-presidential system, the academic community has various propositions and opinions. For example, Shugart and Carey (1992) proposed two main categories of semi-presidential system, namely president-parliamentarism and premier-presidentialism. The primary distinction between them is that in the former the president holds more political dominance in that he possesses unilateral power to appoint and dismiss his cabinet ministers, including the prime minister, while in the latter the president does not have such authority but needs to consult and decide with the prime minister who enjoys the support of parliament majority. Therefore, according to the situation described regarding the operation of semi-presidential system in Kenya and Zimbabwe, it should be considered premier-presidentialism. However, a more detailed classification of semi-presidentialism is to look into the variable of whether the president can control the government via his nomination and appointment of the prime minister, and how important the portfolios his party can manage to obtain. In different scenarios, there are five possible conditions according to Dr. Yu-Shan Wu (2008: 21-24). A further note on such classification is that here we only need to

take into consideration the situation when the president and parliament majority belong to different parties or factions (incongruence), because when they are in the same party (congruence) and thus share somewhat similar political interests, the president will appoint a person from his party as the prime minister to form the government and to follow his directions, which means that all the five conditions will produce the same result. Therefore, it is meaningful to discuss an effective and detailed classification of semi-presidentialism only under incongruence, i.e. when there is division between the presidency and the parliament so that we can observe the differentiation of the results, meaning the president s ability to control over the cabinet, among all five subcategories. According to the classification that distinguishes five subcategories in the actual operation of semi-presidential system, if the president can always appoint his favorite prime minister to the office regardless of whether his party possesses majority in the parliament or not, such subcategory is named presidential supremacy. When the president and parliament majority are incongruent, if the president, though nominating a prime minister from the opposition party, still manages to control part of the cabinet positions, such situation is defined as a compromise mode. Then if the president appoints a prime minister from his party when in congruence, but must appoint the leader of the majority party in the parliament as his prime minister under incongruence, which is based on the spirit of parliamentarianism, this mode is called alternation/cohabitation. And when the president yields the power to control the cabinet to the parliament and the executive authority is in the cabinet led by the prime minister regardless of under congruence or incongruence, such situation is a quasi-parliamentarism. Last but not the least, if under incongruence, the president and parliament majority compete against each other for the power to form a government, it is called the collision subtype (Wu, 2008: 21-24). Reviewing the current situation within Kenya and Zimbabwe s coalition governments, the variable that we are looking into is the president s control over cabinet positions under incongruence. What we can observe is that in these two originally presidential states, when the presidents are facing parliament majority different from their own parties, both choose to share their executive powers with the opposition and consequently appoint opposition leaders as their newly created prime ministers. Meanwhile, both presidents do not enjoy the liberty to unilaterally dismiss the prime minister according to the design of the power-sharing agreements, therefore such arrangement fits into the definition of premier-presidentialism. Under the circumstances that the president appoints opposition leader as prime minister under incongruence, the next matter to be further discussed is the extent of the president s control over the cabinet, namely the number of positions and the importance of the

portfolios that the president s party can gain. In Kenya, president Kibaki s party the PNU and prime minister Odinga s party the ODM each holds half of the total 40 ministers and one deputy prime minister in the cabinet, however in reality the PNU and its allies still manage to occupy the most important positions, including minister of foreign affairs, minister of defense, minister of finance, and minister of home affairs. Therefore, though the number of ministers that PNU and ODM each possesses is the same, yet Kibaki s PNU, which is minority in the parliament, obtains more important positions and thus the president s party to some extent still dominates the cabinet, hence the author considers the proper subtype of Kenya s semi-presidential system belongs to the compromise mode. A similar situation yet more favorable to the president can be observed in the coalition government of Zimbabwe, when Mugabe s ZANU-PF not only secures the presidency and two vice presidents, but also gains 15 out of the 31 ministers. Though it is not the majority, yet ZANU-PF still has two more positions than the MDC s Tsvangirai faction in the government, thus becoming the party with the most strength in the cabinet. ZANU-PF also controls the more important ministries of foreign affairs, defense, and it also has one of the two co-ministers of home affairs along with Tsvangirai s MDC. In Zimbabwe, the president s party ZANU-PF, while has fewer seats in the parliament than the prime minister s ODM, still occupies the larger and more important share of the cabinet positions, hence Zimbabwe s president enjoys a higher control over the cabinet than Kenya s president under incongruence. However, because the position of prime minister still belongs to the opposition party, which is the majority in the parliament, the subtype of Zimbabwe s semi-presidential system is also a compromise mode like that of Kenya.

The Future of Semi-Presidentialism in Kenya and Zimbabwe: Institutional Sustainability or Exit The downstream of semi-presidentialism studies leads us to the discussion of the institutional sustainability or possible shift of the system and its subtype, as well as the further prospects on democratic consolidation especially in the cases of nascent/ unstable democracies. Therefore, aiming at extending the study of these two new cases of semi-presidential regimes and providing a broader and more complete analysis on the development of this issue, the author would hence address the prospects of how institutional sustainability of the semi-presidential system could influence democratic consolidation. As specified in the previous content, the current government system of Kenya and Zimbabwe under the power-sharing agreements is semi-presidential in that the president is popularly elected who holds considerable powers, and that the prime minister is responsible to the parliament. However, a point to notice is that these power-sharing agreements are merely provisional documents to ease the political crises following controversial presidential election results. For instance, in the case of Kenya, the National Accord and Reconciliatory Act is an agreement between the PNU and the ODM which is later sent to the National Assembly to be approved by the parliament as law, but this bill will cease to be effective according to its article 8 under the circumstances that the current parliament or the coalition is dissolved, or a new constitution is enacted, whichever is earlier. The Zimbabwean Power-Sharing Agreement has a legal status even weaker than that of Kenya in that it is merely an agreement signed by all parties taking part in the coalition under the witness of the international society. Although some of its provisions were incorporated in the 19 th constitutional amendment of Zimbabwe in 2009, however, most content regarding the operation of the semi-presidential system, including the decrease of presidential

authority and the powers of the reestablished office of prime minister, was not included in the constitution or the amendment. Given the above mentioned condition, we can see that while the operation of the political system in Kenya and Zimbabwe is semi-presidentialism, however the design according to their original constitution articles is actually still a presidential system, for those arrangements that make the two states semi-presidential are not specified in the constitution itself, but rather provisional legal documents that amend the constitutions, and that the presidents remain both head of state and head of the government as clearly stated in the articles of their original constitutions. Therefore, the author defines the form of government currently in practice in Kenya and Zimbabwe as de facto semi-presidentialism, de jure presidentialism. And in order to fully complete the transition from presidentialism to de jure semi-presidentialism, the two countries must undergo a process of constitutional amendment or enactment of a new constitution. However, in the new constitution of Kenya proposed in 2009 and approved in referendum in 2010, we see that the office of prime minister is abolished, and that the president is once again both head of state and government according to its article 131. The cabinet consists of the president, the deputy president, the attorney general, and cabinet secretaries, and the president chairs cabinet meetings. But the primary difference of the president s power between the original and the new constitution is that the new constitution states that the president appoints cabinet secretaries with the approval of the National Assembly in the articles 132 and 152, which somehow restricts the president s power for he does not possess the unilateral authority to form the government but needs the support of the parliament. From the new Kenyan constitutional design, we can observe that Kenya transfers from the current de facto semi-presidentialism to de jure and de facto presidentialism, therefore the downstream of Kenya s semi-presidential system evolvement has regressed back to presidentialism, which means the end of semi-presidentialism once the new constitution comes into force and the office of prime minister ceases to exist after the first general elections according to the new regulation. 14 And the prospects of Zimbabwe s turning into de jure semi-presidentialism still depend on whether it will have a new constitution incorporating the semi-presidential design currently in practice. According to the Zimbabwean Power-Sharing Agreement, in article 6 the parties agree that a new constitution should be drafted, and in article 24 the parties agree to include specified parts of the draft constitution made at Kariba on 14 According to the sixth schedule of Kenya s latest constitution, which states the extension of application of provisions of the former constitution, the provisions of the former constitution concerning the executive, and the National Accord and Reconciliatory Act 2008 shall continue to operate until the first general elections held under the new constitution.

September 30, 2007, which is also known as the Kariba draft, in the constitutional amendment that implements the Power-Sharing Agreement. However, after the constitution was amended by the amendment no. 19 in February of 2009, we still see the article 27 which states that the president is head of state and government remains unchanged, also, there are no related articles concerning the office of prime minister, but only on acting prime minister in article 31CA. Furthermore, in the Kariba draft, the form of government is still designed as a presidential system, and the specific parts to be included in the constitutional amendment, namely chapters 4 and 13, deal with citizenship and independent commissions respectively, and do not have anything to do with the design of the form of government. According to that document, the president is head of state and government, and he and the cabinet chaired by him possess the executive authority of Zimbabwe. Also the president enjoys unilateral power to appoint and to remove any minister, according to articles 93, 95, and 98. Therefore, from the design of the Kariba draft constitution, we can observe that Zimbabwe will still have a de jure and de facto presidential system if its new constitution adopts such design. Can Semi-Presidentialism Help Consolidate Democracy? Either from the newly promulgated constitution of Kenya or the proposed Kariba draft constitution of Zimbabwe, we see that both Kenya and Zimbabwe return or could be returning to presidential system, and according to the author s point of view, such situation is not a positive development and could possibly undermine the fragile democracy that was finally somewhat restored after a series of political turmoil following the controversial presidential elections in these countries. Therefore, what the author would like to propose regarding the analysis on the cases of Kenya and Zimbabwe is an argument that the institutional sustainability of semi-presidentialism is vital to democratic consolidation in the two countries when compared with the practice of presidential system. Here the author seeks to bring out a proper logic to support the argument that the hope for democratic consolidation in Kenya and Zimbabwe lies in the sustainability of the current semi-presidential system in that such mechanism prevents the emergence of a powerful, if not dictatorial, president, and thus the semi-presidential design of power sharing enables the prime ministers and their governments supported by parliament majority to check the power of the presidents, who in the two cases examined here were heavily criticized for undermining democracy in their countries, and that there were accusations that they won their current presidencies by fraudulent behaviors in the elections, which resulted in the domestic riots and civil unrest that led to political crises later settled by the power-sharing agreements.

With an effort to support the abovementioned argument, the author chooses to apply empirical data, instead of reiterating the various academic literatures currently existing that propose powerful presidential system could be more likely to undermine democracy, while compared with that situation, in a semi-presidential system when the presidential powers are reduced and elements of a parliamentarianism are incorporated into the system, democracy has better chances to sustain and to consolidate. Here the author applies empirical data of Kenya and Zimbabwe s performance on Democracy Index before and after entering semi-presidentialism. Table 1 Kenya s Performance in Democracy Index 2006-2011 Year Ranking Regime Type Overall Score Score 1 Score 2 Score 3 Score 4 Score 5 2006 101 hybrid 5.08 4.33 4.29 5.56 6.25 5.00 2008 103 hybrid 4.79 3.50 4.29 5.56 5.63 5.00 2010 101 hybrid 4.71 3.92 4.29 4.44 5.63 5.29 2011 103 hybrid 4.71 3.92 4.29 4.44 5.63 5.29 Table 2 Zimbabwe s Performance in Democracy Index 2006-2011 Year Ranking Regime Type Overall Score Score 1 Score 2 Score 3 Score 4 Score 5 2006 147 authoritarian 2.62 0.17 0.79 3.89 5.63 2.65 2008 148 authoritarian 2.53 0.00 0.79 3.89 5.63 2.35 2010 146 authoritarian 2.64 0.00 1.29 3.33 5.63 2.94 2011 147 authoritarian 2.68 0.00 1.29 3.89 5.00 3.24 Score 1: electoral process and pluralism Score 2: functioning of government Score 3: political participation Score 4: political culture Score 5: civil liberties Full democracies: scores of 8-10, Flawed democracies: scores of 6-7.9, Hybrid regimes: scores of 4-5.9 Authoritarian regimes: scores of 0-3.9 From the above two tables, the overall trend of democratic performance in Kenya and Zimbabwe shows that immediately after entering semi-presidentialism, the scores of democracy index in the two cases both declined. However, after semi-presidential system came into effect in Kenya for a longer period of time, we can