Perilous presidentialism or precarious power-sharing? Hybrid regime dynamics in Myanmar

ABSTRACT Although Myanmar's transition from military rule has been thoroughly studied, research into the evolving presidential system is lacking. This article maps Myanmar's unique hybrid form of presidentialism and characterises executive-legislative relations between 2012 and 2017. It not only examines institutions themselves but also explores how actors within them have acted and shifted strategically. It is argued that elite-behaviour determines how institutions work. Myanmar's 2008 constitution imposed by the military established a hybrid presidential system with a formally powerful president. The first phase under President Thein Sein saw competitive yet cooperative executive-legislative relations between President Thein Sein and the ambitious House Speaker Shwe Mann, both members of the former military regime. However, this power-sharing arrangement of former members of the military regime became increasingly fragile with the rise of Aung San Suu Kyi and her attempt to change the 2008 constitution. After the election victory of the National League for Democracy (NLD) and the passage of the state counsellor law, a de facto semi-presidential system was established – Burmese style. Since then, the frontline has been between the civilian and military parts of the administration.

Over the past three decades, the democratisation literature has been closely linked to the literature on institutional design. Nowhere has this been more visible than in the debate about the 'perils of presidentialism' and the 'virtues of parliamentary systems'. Since the publication of Juan Linz's seminal article in 1990, there have been several rounds of scholarship debating the effects of these regime types on democratic consolidation (Elgie, 2005).
This article shifts the debate to Myanmar, the largest country in territorial terms in mainland Southeast Asia. Starting in 2011, the long ruling military regime in Myanmar embarked on a liberalisation of the political system culminating in the November 2015 elections, which were won by the oppositional National League for Democracy (NLD) in a landslide. The reforms moved the country into the category of hybrid regimes (Egreteau, 2017a;Freedom House, 2017). Much has been written about the country's evolving institutional framework, such as the guardianship role of the military (Bünte, 2014;Huang, 2014), the emerging pattern of parliamentary politics (Egreteau, 2017b;Kean, 2012) or the lack of the rule of law (Cheesman, 2015). What has been lacking is a discussion of Myanmar's form of presidentialism, the president's relationship to an increasingly assertive legislature and the consequences for the country's democratisation. How did the country's institutional framework impact on stability and democratisation?
This article does not only look 'at the institutions themselves (formal and informal) but on how actors "within" these institutions act and, on how they strategically shift' (O'Donnell, 2006, p. 85). It argues that it is not so much the institutional character of Myanmar's presidential system that matters but how it has been shaped by the actors involved. The case of Myanmar reveals that elite-level behaviour determines how institutions work, more than the other way around. Myanmar's 2008 constitution, written by the military regime a decade ago, established a unique form of presidentialisma system that cannot be properly categorised within existing typologies of comparative politics. The military designed this system to enable the transition from direct military rule to a civilian government. Providing for a powerful executive and a weak legislature, Myanmar's presidential system could have easily evolved into a form of 'hyper-presidentialism'. Yet, the rivalry between reformist president Thein Sein and an ambitious House speaker Shwe Mann led to a competitive, yet cooperative form of executive-legislative relations (2011)(2012)(2013)(2014)(2015)(2016).
This elite arrangement has been increasingly challenged since the NLD joined the political process in 2012. After the NLD election victory, the power-sharing arrangement could only be sustained by inserting a semi-presidential element. The state counsellor law (discussed in more detail below) enabled Aung San Suu Kyi to take over the leadership of the state. She exercises much more authority than the president. Today, the frontline runs not between the president and parliament, but between the civilian and the military parts of the government. The NLD has attempted to appease the military and there has not been an open confrontation between the NLD and the military, although the latter is showing increasing discomfort with the ongoing liberalisation.
This article proceeds in three steps: first, it develops a typology based on the dominant paradigm of executive-legislative relations and revisits the central arguments of the perils of presidentialism debate. Second, it analyses the formal institutions of Myanmar's 2008 constitution, established by the outgoing military regime from a position of strength. Third, it looks at the power-sharing arrangements under President Thein Sein (2011)(2012)(2013)(2014)(2015)(2016) and President U. Htin Kyaw (since 2016).

Presidentialism, parliamentarism and semi-presidentialism
In order to prepare the ground for a comparative analysis, it is imperative to give a classification of executive-legislative relations. Since Walter Bagehot's ([1867] 2009) comparison of constitutional practice in the British and American political systems, the dichotomy between presidentialism and parliamentarism has been fundamental. The traditional typology allows us to distinguish between two conflicting principles of executive-legislative relations within presidentialism and parliamentarism: a 'transactional relationship' in the former and a 'hierarchical delegation' in the latter (Shugart, 2008). The transactional relationship of presidential systems derives from the strict separation of powers between the president (and his cabinet) and the parliament, and their separate source of legitimacy. Consequently, Giovanni Sartori (1997, p. 82) defines the direct or directlike election of the president as the 'defining condition' of presidential systems. He adds that neither the president nor his cabinet are appointed or can be dismissed by the parliament. For Stepan and Skach (1993, p. 22) the 'essence of pure presidentialism is mutual independence' between presidents and parliaments. The 'hierarchic relationship' in parliamentary systems refers to what Sartori (1997, p. 101) calls 'executive-legislative powersharing': governments in parliamentary systems are appointed, supported and may even be dismissed by parliamentary majorities. Despite the extraordinary variation in both presidential and parliamentary systems, Shugart (2008) claims that all forms of democratic design must trade off these two conflicting principles, which according to Stepan and Skach's (1993, p. 22) terminology function as 'analytically separable propensities'.
Building on this traditional understanding, contemporary studies have developed a more sophisticated typology, particularly since the introduction of semi-presidential regimes (Duverger, 1970). The latter regimes are often referred to as 'hybrid' (Shugart, 2008, p. 6) or mixed regimes (Cheibub, 2007), or worse, a 'bastard tribe' between parliamentarism and presidentialism (Cheibub, Elkins, & Ginsburg, 2014, p. 519). Since then, the dual structure of authority between a president as a head of state and a prime minister, who heads the government has become the epitome of semi-presidential systems (Sartori, 1997, p. 121). According to a widely accepted definition, semi-presidentialism consists of a 'popularly elected fixed term president existing alongside a prime minister and cabinet who are collectively responsible to parliament' (Elgie, 1999, p. 13). 1 Robert Elgie's definition (1999, p. 13) has a powerful simplicity that appears to fit well within the classic typology. It focuses on the formal provisions of the constitution and, thus, eliminates many ambiguous cases associated with other definitions. The general tripartite distinction between parliamentary, presidential and semi-presidential regimes faces little serious conceptual competition (Cheibub et al., 2014, p. 514;Siaroff, 2003).
Of perils and virtues: The impact of regime type on democratisation During the last three decades, we have witnessed several 'waves' of scholarship debating the effects of various regime types on democratic survival (Elgie, 2005(Elgie, , 2011. Driven by the institutional turn in political science, the research focus in the early 1990s has revolved around the possible 'perils' of presidentialist and the 'virtues' of parliamentary institutions. At the same time, a debate began about the effects of semi-presidentialism. While the first research wave had examined debated the effects of 'pure' regime types on democratic consolidation, the second research wave focused on the interplay of presidentialism and other institutions. Moreover, the dependent variables also broadened and included good governance, effectiveness and conflict management (Elgie, 2005).
The initial trigger of this debate was Linz's classic argument for the superiority of parliamentarism over presidentialism (Linz, 1990a(Linz, , 1990b(Linz, , 1994. Debating the suitability of diverse systems of government for the consolidation of third wave's democracies, Linz (1994, pp. 6-8) put forward a number of strong arguments against presidential systems. For instance, he lamented the 'dual legitimacy' of presidentialism, referring to the fact that both president and legislatures enjoy the support of 'the people'. Institutional conflicts have also been exacerbated by the winner-take-all nature of presidential elections, which leaves the losing group without any representation. Linz was also concerned with the rigidity of presidentialism and the lack of proper institutional mechanisms to overcome gridlock situations. In order to solve these, either the legislature may resort to impeaching the president, thus intensifying constitutional crisis, or the president may use his powers to govern over and above the legislature, thereby threatening the rule of law. In these crises, the military might be more willing to intervene. In short, presidentialism is far less conducive to democratic consolidation than parliamentarism. The latter provides more flexible solutions for young democracies, such as votes of no-confidence or the Premier's dissolution of legislatures. A number of critics raised objections, particularly due to Linz's case selection and implicit focus on Latin American presidentialism.
The research focus of the 'second wave' of presidentialism research shifted attention from the effect of 'pure' regime types to the effects of broader institutional configurations. Debating the fundamental institutional variables of regime types in conjunction with other institutional variablessuch as the election or party systemsscholars criticised the combination of fractionalised multi-party systems and presidentialism. Supporting Linz's general argument, Mainwaring (1993, p. 199) finds a correlation between stable presidential democracies and two-party systems and concludes that multipartism and presidentialism together make a very difficult combination. 2 According to him, multipartism increases the likelihood of gridlock situations and ideological polarisation, and it makes inter-party coalition building more conflictive and less likely. Shugart and Carey's work (1992, p. 156) supports this argument by looking at the legislative and non-legislative powers of the elected presidents. They found reasons to believe that more powerful presidents tend to be more problematic, particularly in which authority over cabinets is shared between assembly and president.
In their study of Latin American presidentialism Shugart and Mainwaring (1997, p. 395) argue that presidential systems vary so greatly in the powers accorded to the president, the types of party and electoral systems with which they are associated, and the socioeconomic and historical context in which they were created that these differences are likely to be as important as the oftassumed dichotomy between presidential and parliamentary systems. Sedelius and Linde (2017) Have shown in a large N-analysis of all regimes that parliamentarism generally does not perform better than the other regimes. Neither their measures of democracy nor the measures of government performance show significantly better records for parliamentary regimes. This indicates that a parliamentary constitution with an indirectly elected president does not necessarily go along with better political performance than a premier-presidential one with a popular elected but weak or medium weak president.

Presidentialism in Myanmar: Power-sharing on the military's terms
Myanmar is a special case for an analysis of executive-legislative relations. Institutions in the country are still young and bear the imprint of 23 years of direct military rule and 50 years of authoritarianism. The current institutional framework, the 'praetorian constitution' of 2008, drafted under military control from 1993-2008, reflects the military's view of how the civilianised rule should work once direct military rule is ended (Egreteau, 2017c, p. 122). It served as an instrument of power-sharing between the military and civilians. The military has been able to control the entire constitution-drafting process. It has drafted the basic principles of the constitution and dominated the discussions in the National Convention. Delegates were handpicked, oppositional and ethnic parties hard hardly any say in the drafting the new constitution.
Large parts of the 2008 constitutionparticularly the strong role of the president and the heavy influence of the military in politicswere long rejected by the NLD (Than, 2004). The military regime made the support for the 2008 constitution a precondition for participating in the 2010 elections, which were boycotted by the NLD and won by the military-sponsored Union Solidarity and Development Party (USDP). After the dissolution of the military junta in 2011, then President Thein Sein started a reconciliation process with NLD-leader Aung San Suu Kyi, paving the way for the NLD's reintegration into the political system. The NLD registered as a political party and competed in the by-elections of April 2012. In the course of the liberalisation, the NLD transformed from an anti-system to a 'transition-seeking party' willing to legitimise the political order by participating in institutions agreed upon by the authoritarians, but to the end of reforming them from within (Bünte, 2016, p. 379). The NLD consequently agreed temporarily to the powersharing institutions installed by the military a decade before while working toward revising them.
The following analyse Myanmar's presidential system based on the 2008 constitution. I argue that the system of checks and balances is determined much more by the competition of the most powerful actors than the institutional foundations laid down in the constitution. While the formal institutions provide for a powerful presidential system with a strong military, executive-legislative relations were initially quite evenly balanced by the rivalry between President Thein Sein and house speaker Shwe Mannboth highranking members of the former military regime. Moreover, despite far reaching reforms, the dominance of the military has remained untouched. Yet, since the inclusion of the NLD after the by-elections in May 2012, Aung San Suu Kyi's attempts to change the constitution and the NLD's election victory in November 2015, frictions between the military and the oppositional NLD have increased.

Classifying Myanmar's political system according to the 2008 constitution
Myanmar's 2008 constitution established a hybrid system between parliamentarism and presidentialism. It has some characteristics of the hierarchical delegation mode of parliamentary systems combined with the 'transactional relationship' of presidentialism. Moreover, it carries a number of provisions, which ensure military dominance in the political arena and are a sign of power-sharing on military terms.
Myanmar does not fulfil Sartori's 'defining condition' of pure presidentialism: the direct (popular) election of the president. Instead, an electoral college consisting of three groups elects the president: members of parliament (MPs) from the lower house, MPs from the upper house and the appointed members of the armed forces, which make up 25% of both houses. Each group proposes one candidate and the Electoral College choses the president. The losing candidates become vice-presidents (Constitution Myanmar 2008, Art. 60). With this route to the presidency, the military ensured that the candidate nominated by the military would become at least vice-presidentif not president (Clapp, 2017, p. 85).
While parliament plays a vital role in the selection of the president, her or his political survival is not necessarily tied to the majority party in parliament. In accordance with presidential systems in general, the president is supposed to serve for a fixed term -Myanmar's president serves for five years, with a chance of one time re-election. The president can be dismissed only by what Arend Lijphart calls the 'highly unusual and exceptional process of impeachment' (Lijphart, 1994, p. 92). However, the constitution does not link the president's impeachment to criminal activities only, such as high treason and unconstitutional behaviour. The impeachment process is also opened up to political criteria, such as 'misconduct' and 'inefficient discharge of duties assigned by law' (Art. 71)which has resonances of a parliamentary system. Admittedly, the hurdles for impeachment are high: 2/3s of elected parliamentarians need to support the impeachment motion. Yet, the wording of Art. 71 opens the impeachment process for the political manoeuvers of parliamentary majorities which are not satisfied with the president's performance.
Despite these ambiguous constitutional provisions, which resonate with parliamentarism, there are a number of provisions which move the political system into a presidential direction. First, the concentration of powers in the president's hands puts her/him in a very strong position as leader of the executive: all executive actions of the Union Government shall be taken as action in the name of, and are ultimately attributable to, the President. In this regard, the president is not accountable to the legislature (Art. 215). Second, the president can chose his cabinetparliament only needs to check the qualifications of the ministers involved, whereas the president decides on the composition and size of the cabinet. Ministers are solely responsible to the president and not to parliament (Art. 232 h). The president's authority, however, is limited in certain areas due to the heavy influence of the military. In relation to the Ministers for the Ministries of Defence, Home Affairs and Border Affairs, the president must choose from a list of candidates submitted to him by the Commander-in-Chief (Art. 232). The President must then submit the list of selected persons, along with the list of Defence Service personnel nominated by the commander-in-chief, to parliament for approval (Art. 232(c)).
Third, the separation of powers is granted in the provision that the president (and also her/his cabinet ministers) are not allowed to be MPs. They have to relinquish their mandate once elected (or appointed as ministers). Due to the military's mistrust of elected civilians, the military drafted constitution even promulgated that the president (and the ministers) must not undertake any political activities for their respective parties during the presidency. Fourth, the president also has some budgetary authority: he proposes the state budget, which is discussed (and revised) by parliament (Art. 218). Fifth, the president has certain legislative powers and ceremonial functions as well. He has a veto-right and can send laws back to the legislature for revisions, although the legislature can override his veto. His ceremonial functions include the ability to grant pardons or amnesties, start or end relationships with other states (Article 206), make a speech in both houses of parliament, call an emergency session of the legislature (Art. 211), put into effect an urgent law when the legislature is not assembled (Art. 212). Sixth, the president has the power to appoint and remove the Chief Ministers of the regions and states (Art. 261(a),(b)); Art. 263). This gives him a very powerful position and virtually the control of the administration of the regions and states.
The president also coordinates security policy and chairs the National Defence and Security Council (NDSC)an institution, which has extensive powers over defence and security issues. The council comprises 11 members: the president, the two vice president, two parliamentary speakers, the commander-in-chief, the vice-commander-in-chief and the ministers for defence, home affairs, border affairs and foreign affairs. This composition allows the military to overrule civilian politics, since military appointees outweigh civilian politicians (6:5). Finally, the president controls the appointment and removal of the judges of the Supreme Court (Art. 299 and 302) and the chief justices of the High Courts of the Regions and States (Art. 308(b). The president also selects a third of the judges of the Constitutional Tribunal (Art. 321).
We can conclude that Myanmar's 2008 constitution establishes a governmental system that cannot easily be classified according to the existing typology and thus has to be categorised as 'special case'. It is a hybrid between presidentialism and parliamentarism that provides for a powerful president who is elected indirectly by both houses of parliament and the military. However, the accountability mechanisms are blurred. This might not come as a surprise, as the constitution gives extensive powers to the armed forces which act as guardian of the political system and share power with civilian politicians in executive and legislative affairs. They are particularly influential in security and defence issues and remain insulated from civilian control. Since more than 75% vote is needed to change the constitution, with 25% of parliamentary seats the military has de-facto veto power over constitutional change. The system set up by the military resembles Indonesia's post-authoritarian system (1999)(2000)(2001)(2002)(2003)(2004), in which the president was also indirectly elected by both Houses of Parliament (the so called People's Consultative Assembly). 3 However, Indonesia's post-authoritarian system faced serious instabilities in 2000/2001, which resulted in the impeachment of President Abdurrahman Wahid and his removal from office in July 2001 (Mietzner, 2009). The next sections discuss whether Myanmar's presidential system has seen similar challenges resulting from the country's liberalisation and whether the power-sharing agreement based on the military's terms has endured.
Perilous presidentialism? Executive-legislative relations under Thein Sein (2011)(2012)(2013)(2014)(2015)(2016) Given the concentration of political powers in the executive and the weakness of the legislature and judiciary after 50 years of military rule, Myanmar's political system could have easily developed into a form of 'hyper-presidentialism'a system, in which a powerful president faces few effective checks from the legislature or the courts (Ackermann, Desierto, & Volosin, 2011). 4 However, to avoid a concentration of power and safeguard his own personal, familial or commercial interests after stepping down, the strongman of the former military junta Senior General Than Shwe handpicked key persons of the new administration himself (Callahan 2012, p. 122;Kyaw, 2012). Thein Sein, a loyal and long-time member of the junta, who had chaired the National Convention to draft the constitution and served as Prime Minister of the junta since 2007, became the first 'civilian' president. Four-star General Thura Shwe Mann, former joint chief of staff of the military and third highest member of the military hierarchy, was made speaker of the lower house.
Both were placed at the helm of the regime party, the USDP. 5 Reportedly, a number of senior members of the first cabinet were also handpicked by Senior General Than Shwe. Altogether, 29 of the 36 cabinet members were former military generals (Kyaw, 2012). Before going into retirement and dissolving the long-term junta, Than Shwe picked Min Aung Hlaing as commander in chief. All in all, the balancing of former members of the military regime prevented a concentration of power in the president's hands and made far reaching reforms, which might endanger the military's dominant position, highly unlikely (Callahan, 2012, p. 125).
Having the support of merely a handful of ministers in the beginning 6 , Thein Sein broadened his support base with the cabinet reshuffles in August 2012 and September 2013, included more reformers and sidelined some of the hardliners (Callahan, 2012;Min Zin, 2016, p. 119). There were six ministerial positions assigned to manage presidential affairs, to communicate with the legislature, to coordinate economic development, to look after ethnic affairs, to formulate policies for national projects and to coordinate with subnational governments (Kyaw & Farrelly, 2016, p. 26).
Thein Sein was able to use his presidential powers to start a reconciliation process in which he convinced Aung San Suu Kyi to join mainstream politics. He also began liberalising his regime which undermined some of the foundations of authoritarianism, although not the dominant position of the military (Bünte, 2014;Huang, 2015). The president had turned out to be reform-minded, seeking consensus on a number of contentious issues within the framework of the 2008 constitution. President Thein Sein established a number of committees to discuss protracted issues such as the issue of political prisoners, press liberalisation or religious violence (Kyaw, 2012;Robinson, 2013). His reliance on technocrats and his inner circle enabled him to dissociate himself from party politics and overcome the USDP's resistance. Citing the constitutional requirement that the president was not allowed to take part in party activities during his term in office (Art. 64), Thein Sein managed to distance himself from the more conservative elements in his party. For instance, the USDP did not want to hold the by-elections 2012 at all because of fears about strong popular support for the NLD. But the president overrode his party's objections to show the West that he was committed to reforms (Min Zin, 2016, p. 119).
Although dubbed as '15-minute parliament' in the beginning, the legislature under the leadership of Shwe Mann also started to support a number of important reform laws, such as the authorisation of labour unions, the Peaceful Assembly Law and laws related to foreign investment (Egreteau, 2017b, p. 33;Fink, 2015, pp. 327-329). The deeper reasons for this lie in the personal ambitions of house speaker Shwe Mann. He felt frustrated that he had not been named presidentthough he was higher in military rank than Thein Sein. He therefore put his energy into turning the legislature into a much more powerful and independent body than envisioned by the constitution (Fink, 2015, p. 336). Competing with the presidential power embodied by reformist president Thein Sein, the new legislature started to assert its autonomy with a non-partisan approach of oversight of the executive and with the inclusion of opposition figure Aung San Suu Kyi beginning in 2012, making Myanmar's parliament an important symbol of Myanmar's transformation (Kean, 2012, p. 65).
The USDP and the NLD did not act as government and opposition in Parliamentthere was no clear division between authoritarians and democrats; often both parties supported the same liberalisation initiatives. This resulted in an unusual lack of party divisions and verbal conflict (Kean, 2012). Shwe Mann also discouraged MPs from using the word 'opposition' in parliamentary debates. Nevertheless, the parliament was assertive under Shwe Mann's leadership. It started to hold the government accountable and voice the grievances of the citizenry (Fink, 2015;Kean, 2012). Parliament also played a crucial role in the peace-process that started under President Thein Sein in 2011, leading to a nationwide ceasefire agreement in October 2015 (Chit Win, 2016).
Moreover, in contrast to other Southeast Asian electoral authoritarian regimes like those in Malaysia, Singapore or even Cambodia (Case, 2011), Myanmar's legislature placed some checks on the powerful executive. These oversight mechanisms relied mainly on motions and questions raised during the plenary sessions. Even USDP-lawmakers started to question the government, in order not to be seen as obstructing reforms (Fink, 2015). Shwe Mann sometimes steered parliament towards a collision course with president Thein Sein, such as, for instance, in the impeachment of the judges of the Constitutional Tribunal in 2012 (who had decided parliamentary committees are not equal to government ministries) or in the introduction of a Development Fund for Parliamentarians, to which Thein Sein objected. Moreover, President Thein Sein established the National Human Rights Commission by decree, but when it came to approving the budget, the parliament refused until the Commission was established by law (Crouch, 2016, p. 226). In addition, Thein Sein did not hesitate to use his power to refer draft bills back to the Union Parliament with suggestions for reform, although under the constitution parliament has the final say on legislation, with laws coming into force 14 days from approval regardless of the president's opinion (Crouch, 2016, p. 226). This constitutional provision prevented extensive gridlock and prolonged conflicts. According to an interview with a member of the former National Convention, this provision has been introduced to find a middle way between presidentialism and parliamentarism and avoid excessive conflicts associated with the presidentialist system (interview with a member of the National Convention, Naypidaw, November 2015).
At this stage, Myanmar's hybrid form of presidentialism had not experienced any of the perils associated with other, 'purer' forms of presidentialism. Although there were conflicts between the executive and legislature, they did not escalate into long periods of stalemate or gridlock between the two sides. Moreover, conflicts were also rooted in the personal rivalry of the main actors involved. But, crucially, these rivalries did not endanger the military's power-sharing agreement established with the 2008 constitution.
This, however, changed with Aung San Suu Kyi's growing involvement in parliamentary affairs since 2012 and her attempts to change the 2008 constitution. For instance, in August 2015, when Shwe Mann openly sided with Aung San Suu Kyi and agreed to change the constitution, the Ministry of Home Affairs, under direct control of the military, sent some 400 police officers to remove Shwe Mann from the position as Chairman of the USDP (Min Zin, 2016). 7 This 'party coup' not only highlighted the military's dissatisfaction with the emerging alliance between the former general and the opposition leader, but also the deep divisions within the USDP party leadership. However, the conflict did not spill over into parliament where Shwe Mann remained house speaker. It is also a sign of a growing discomfort of the military with changes in the political order. The military had not only devised the new political order, it is also actively guarding what it once termed 'disciplined democracy'. To the surprise of many, military representatives supported President Thein Sein's liberalisation. The military did not block the release of political prisoners in 2011 and 2012, the liberalisation of the press or the relaxation of the freedoms to protest. It also allowed the elections in November 2015 to run smoothly. There were hardly any open and disruptive conflicts between the quasi-civilian government of Thein Sein and the military leadership. In parliament, the military bloc tried to act as moderators or arbitrators (Egreteau, 2017a). All this is hardly surprising, since President Thein Sein has been a high-ranking member of the former military junta and is one of the key architects of the transition to disciplined democracy (Huang, 2017). Moreover, Commander-in-Chief Min Aung Hlaing backed Thein Sein from the beginning (Min Zin, 2016, p. 177).
Yet, the resistance of the military towards a further liberalisation has grown since 2014: in June of that year military representatives blocked constitutional changes that would lower the 75% threshold for constitutional change to 70%, which would have potentially undermined the veto power of the military. Moreover, the military bloc also prevented changing Article 59 (f), which bars Aung San Suu Kyi from the presidency due to the nationality of her sons. In December, Tin Aye, chairman of the election commission and confidant of President Thein Sein, stated in a meeting with artists that the military would seise power in the event of political or ethnic turmoil in the country, which some of the opposition members understood as veiled threat to stop pushing for changes to the constitution (Fink, 2015, p. 350). In October 2014, the military block proposed in parliament an expanded role for the National Defense and Security Council (NDSC), a militarydominated 11-member body that holds wide-ranging powers, including the right to take over from the civilian government in a state of emergency. The military block proposed the NDSC should be able to dissolve parliament (New York Times, Return of the Military?, 17 November 2014). This revealed the military's increasingly unease about growing political freedoms.
To take the example of press freedom: the press is not allowed to criticise the military or write the negatively about military projects. If press outlets cross this line, they are approached by the military and told to watch their reporting (Bünte, 2017). The military has also been worried about protests since 2011, which have often been triggered by land-grabbing or military development projects (Bünte, 2018). Military representatives in parliament staged a protest in February 2016, when a NLD MP openly criticised alleged misconduct by local authorities involved in the Letpadaung copper minea controversial enterprise with heavy involvement by a military conglomerate. The symbolic and theatrical behaviour by military representatives bluntly shows that the armed force refusal to be criticised for its involvement in commercial activities, including land-grabbing and natural resource-related projects.
Semi-presidentialism -Burmese style: the state counsellor law The election landslide of the NLD in November 2015 challenged the elite consensus on the military's terms and brought about a major change in the configuration of government. Already days before the 2015 elections, Aung San Suu Kyi had declared that in the event of an NLD victory, she would be 'above the president.' After the election victory, the NLD chose the relatively unknown Htin Kyawa close aide of Aung San Suu Kyi and veteran of the pro-democracy struggleas 'proxy president'. A few days after Htin Kyaw's inauguration, the NLD-dominated parliament was able to push through the State Counsellor Law, specifically engineered to allow Aung San Suu Kyi to serve as de facto leader of the NLD-government. The State Counsellor Law created an additional advisory position for Aung San Suu Kyi. According to the law, the State Counsellor is accountable to the legislature. The State Counsellor can communicate with cabinet members, government departments, organisations, associations and individuals (The State Counsellor Law, 2016, ch. 4). The law is restricted to the tenure of the presidentthe position is consequently restricted to a term of five years. The state counsellor law is consequently a step into a 'semi-presidential direction' -Burmese style. 8 As such, it is a power-sharing deal par excellence. It reduces the number of veto-players to the military and forms a coalition of NLD and the military. Restricting the law to five years also stresses Aung San Suu Kyi's belief that she would be able to overcome the resistance of the military to change Art. 59(f) within a relatively short time. Despite this, the military representatives staged a boycott to demonstrate their dissatisfaction. One of the military legislators stated: 'As the legislature did not consider our proposed amendments, we refused to vote on the bill' (Global New Light of Myanmar, 6 April 2016, 3). At the end of the session, the bloc stood in silent protest and subsequently denounced the passage as 'democratic bullying' by the majority. Since the NLD appointed all members of the Constitutional Tribunal, the military likely felt a formal challenge would be unsuccessful and risked further demonstrating its legislative impotence (Bünte, 2017, p. 110).
Apart from the position of state counsellor Aung San Suu Kyi took up four ministerial positions: foreign affairs (to be included in the NDSC), minister of the president's office, electricity and energy, and education. The whole civilian administration was effectively under her controlaside from the three ministries remaining under the control of the military. Andrew Selth hinted at the centralisation of power in Aung San Suu Kyi's hands when he wrote: 'some foreign commentators have even labelled her a "democratic dictator in the making", who may precipitate the very crisis she is trying to avoid' (Nikkei Asian Review 29 March 2016).

Cohabitation 2016 and 2017
Since passing the state counsellor law Aung San Suu Kyi and the NLD have avoided serious confrontations with the military. They have not attempted to push the military out of politics by addressing the issue of constitutional reform. The government has also attempted not to threaten the core interests of the military. It has neither encroached on military's business nor reformed its grip on land or resource extraction (Bünte, 2017, p. 121). This is particularly remarkable, since the NLD won the 2015 elections based on the promise to create a transparent economy, modernise ailing state enterprises and return the land stolen the decade before. 9 At the same time, the government has avoided to convene the NDSC, although it represents the highest-level forum in security related matters. Aung San Suu Kyi even appointed her own security advisor. These explicit policy decisions aim at engaging the military with civilian institutions and improve civil-military relations (Kyaw & Farrelly, 2016).
However, the Rohingya-crisis in September 2017 demonstrated her powerlessness in security issues when she openly defended the military's brutal response to an attack by a militant Rohingya group. The brutal crackdown on Rohingya villageslabelled as textbook example of ethnic cleansing by Zeid Ra'ad al-Hussein, the United Nations high commissioner for human rightsled to an outcry of the international community. Aung San Suu Kyi's initial silence and then failure to criticise the military's operations are signs of her institutional weakness: security is the military's prerogative in the NLD-military coalition and she apparently did not dare to endanger the trust built up during the early phase of the cohabitation. 10 It also highlights that the military is solely responsible for certain policy fields. Apart from security issues and border affairs, the military has control over the Home Ministry and the General Administrative Department (GAD), which gives it considerable influence over the countrywide local administration. Furthermore, there has been growing militarisation of lower administration positions in civilian ministries resulting from military officers joining these departments (Bünte, 2017). All this highlights the precariousness of the civilian-military coalition. It demonstrates that in this 'cohabitation' arrangement, the frontline does not run between president and parliament, but between the civilian and military parts of the government.
Because of the huge election victory of the NLD, the role of the legislature has changed significantly. The NLD, which controls a majority of 59% of the seats of both houses of parliament, 11 has introduced strict party discipline. Unlike during the Thein Sein government, there have been no major conflicts between the executive and the legislature thus far (Kyaw & Farrelly, 2016, p. 35). Admittedly, it is still early days, but the concentration of power in Aung San Suu Kyi's hands and her control of the political party rule out major confrontation for the time being.

Conclusion
Despite the rigidity of the 2008 constitution, Myanmar's governmental system has been quite fluid since the beginning of the country's transition from military rule in 2011. The military-drafted constitution establishes a unique hybrid between presidentialism and parliamentarism that provides for a powerful president who is elected indirectly by both houses of parliament and the military. However, the accountability lines between parliament and president are blurred. This does not come as a surprise since the constitution given considerable powers to the armed forces which act as guardian of the political order and share power with civilian politicians in executive and legislative affairs. They also act as veto-power over constitutional change. Although the constitution vests power in the executive and could have easily led to the establishment of a 'hyper-presidential system', the rivalry between reformist president Thein Sein and house speaker Shwe Mann led to a more competitive, yet still cooperative form of executive-legislative relations (2011)(2012)(2013)(2014)(2015)(2016).
Yet, this elite arrangement has become increasingly precarious since the NLD joined the political process in 2012. The military has been particularly dissatisfied with the alliance of Thein Sein and Aung San Suu Kyi, attempts by the NLD to change the constitution and some of the liberal freedoms established under Thein Sein. Consequently, the military has actively tried to limit democratisation (Bünte, 2017). After the NLD election victory in 2015, the power-sharing arrangement could only be sustained by inserting a semi-presidential element. The state counsellor law establishes a semi-presidential system -Burmese Style. Since the president is not elected directly, making Aung San Suu Kyi state counsellor does not establish a pure semi-presidential system. Yet, this political system reflects the growing concentration of political power in Aung San Suu Kyi's hands. Since coming to power, there has been no open confrontation between the NLD and the military, since her government so far has avoided threatening the core interests of the military. The case of Myanmar has shown that it is not so much the formal institutional framework established by the constitution that matters. Institutions are determined by powerful actors and they do not really constrain elite-level behaviour. Thus, executive-legislative arenas were reshaped by the rivalries between President Thein Sein and the speaker, Shwe Mann. They were reconfigured again by the NLD's capture of the legislature and the installation of Aung San Suu Kyi as State Counsellorformally accountable to parliament yet informally the most powerful position in the civilian administration. The mode by which executive power is organised in Myanmar seem no longer to matter very much for the military has made it apparent that it is the main arbiter of power. The coalitions of actors involved, their manoeuvres and alliances might lead to harmony or conflict. In Myanmar, presidentialism has not been perilous. Rather, a precarious power-sharing arrangement has been put in place between a military that has been ruling for five decades, and a newly elected government. Notes 1. By responsible to the legislature, Elgie means that the legislature has the power to refuse to approve the cabinet in an investiture vote and/or to topple the cabinet through a vote of confidence/no confidence. If the final decision for this lies with the president, then Elgie speaks of presidentialism. 2. For the opposing view see Pereira and Melo (2012). 3. The constitutional reforms in 2004 moved the system to a Presidential System, in which the president was directly elected. 4. Observers feared in the beginning that Thein Sein could turn out to be a mere puppet of longterm junta chief Senior General Than Shwe. 5. Although Thein Sein was the Chairman of the part at the time of its formation, he had to step down on becoming President, as required by the Constitution. Vice-chairman Shwe Mann became Chairman. 6. Industry Minister Soe Thein, Railway Minister Aung Min and Zaw Min. 7. Another reason for the party coup was that Shwe Mann refused to place 150 military officers on the USDP party list for the November 2015 elections. Among those rejected were Thein Sein's and Min Aung Hlaing's supporters (Zin Min, 2016). 8. It would be only be a 'proper' semi-presidential system if the president were directly elected.
Since the system was not a 'pure' presidential system before the, it is not a pure 'semi-presidential system' with this adaptation either. 9. Land grabbing is a major problem in the country. The military has confiscated a huge amount of land to expand industrial zones, build factories or barracks. Cronies with close connections to the military have profited from their links to military generals. 10. A different reading sees her silence as compliance with ethno-nationalist Buddhist groups, who consider Rohingya illegal migrants. 11. The NLD won 80% of the contested seats. The military has a representation of 25%.