The Nineteenth Amendment to the Constitution, enacted in April 2015, has considerably changed the constitutional framework of Sri Lanka. Some of the most crucial shifts have been in the nature of the executive system, with the change of the powers and the institutional relationships between the president, the cabinet, the prime minister, and the legislature. It is both useful and appropriate to consider the changes wrought by the Nineteenth Amendment through the concept of semi-presidentialism, in particular to see how it has changed the executive branch and the coordination between the executive and legislative powers. The Nineteenth Amendment has turned Sri Lanka from a ‘president-parliamentary’ system with sweeping presidential powers to a ‘premier-presidential’ system with less presidential powers.
In this regard it is important to first discuss the distinctive characteristics of semi-presidential systems, highlight their most frequently discussed problematic features, and juxtapose these to empirical cross-case evidence. Second, it is useful to build the Sri Lanka’s constitutional framework established by the Nineteenth Amendment into a comparative context with other historical and contemporary cases of semi-presidentialism. Finally, it is useful to consider what the theoretical insights and empirical evidence of the existing literature tell us about the future of the new form of semi-presidentialism in Sri Lanka. It is important to note that this chapter only covers the changes introduced by the Nineteenth Amendment that relate to the structure of the executive branch and executive-legislative coordination. Other important changes introduced by the Nineteenth Amendment are consequently out of the scope of this chapter.
The chapter is structured as follows. The first section provides a short background to the concept of semi-presidentialism and its definitions, as well as the distinctiveness of semi-presidential systems from both parliamentary and presidential ones. Section 2 discusses the most commonly raised criticisms of semi-presidentialism and juxtaposes these critical arguments against the findings of the existing cross-case comparative research. This section pays particular attention to the phenomena of cohabitation as well as divided and minority governments. The section ends with the discussion of the factors that determine formation and control of executive power in semi-presidential systems. In section 3 I briefly describe the changes that the Nineteenth Amendment has introduced to the structure of the executive power, as well as its relations with the legislative power. The final section provides a discussion of the constitutional amendment in light of the theoretical arguments and empirical findings in the existing literature.
Semi-presidentialism: Definitions and Constitutive Characteristics
Reflecting on the form of government of the French Fifth Republic, Maurice Duverger was the first to identify semi-presidentialism as a distinct form of government. The early conceptualisation by Duverger, which first emerged in 1970, was further elaborated in his 1980 article. He defined semi-presidentialism as a political system where there is a popularly elected president who possesses quite considerable powers and is faced with a prime minister and a cabinet that have executive power and who stay in power as long as they enjoy the support of the parliament. At the time Duverger identified seven countries meeting these criteria: Austria, Finland, France, Iceland, Ireland, Portugal, and the historical case of the German Weimar Republic.
Duverger’s definition has been criticised for its combination of institutional and behavioural attributes. The behavioural attribute of ‘possessing quite considerable power’ has been brought up as a source of conceptual ambiguity and empirical confusion. Overcoming the problems identified with the Duvergerian definition, Robert Elgie has suggested what is currently the most widely accepted definition. According to him semi-presidentialism is a form of government where a directly elected president is facing a prime minster and cabinet who are collectively responsible to the legislature. Unlike Duverger, Elgie’s definition has no behavioural attribute and is restricted solely to the institutional characteristics of semi-presidentialism.
Elgie’s definition has been crucial in several respects. The first and most important contribution of the purely institutional definition has been the emergence of semi-presidentialism as a distinct form of government, rather than a hybrid system shifting between that of parliamentary and presidential properties. Elgie’s definition has been useful in avoiding a lot of the confusion related to the arbitrary assignment of the label of semi-presidentialism based on the interpretation of behavioural elements of the definition in a number of separate cases. Instead the purely institutional attributes which are easily identified through a country’s constitution leave little room for debate whether a country is semi-presidential or not.
Despite its obvious advantages Elgie’s definition does not address another problem of semi-presidentialism – extreme heterogeneity of formal constitutional powers of actors and the often quite divergent actual behaviour they exhibit in these systems. This extreme variation poses considerable difficulties for cross-case comparative empirical analysis and causal inference. Several attempts have been made in accounting for the institutional and behavioural variation within semi-presidential systems through developing metrics of presidential power. Given such extreme variation of formal powers, what is it then that makes semi-presidentialism a distinct form of government?
The distinctness of semi-presidential systems becomes evident if we think of government systems as chains of delegation of political authority from voters to politicians and state institutions. This view has been elaborated early by Matthew Shugart and John Carey and further discussed by Shugart, and Shugart and David Samuels, as well as research that uses principal agent theory. Shugart suggested that origin and survival of the executive authority is a useful angle to look from when differentiating authority patterns in semi-presidential systems. In parliamentary systems both the origins and survival of the chief executive (prime minister) is fused with that of the legislature. The parliamentary majority is in this case the principal of the cabinet and prime minister, which are its agents. In presidential systems the origin and survival of the executive are separated from the legislature. President and legislature are elected separately by the people, and both are independent of each other for their survival. Semi-presidential systems are distinct in that one part of the dual executive, the president has its origin and survival separated from the parliament, while the other component, the prime minister and cabinet have both their origin and survival fused with the parliament.
Using the differentiation of origin and survival of the executive presidents, cabinets and prime ministers, and legislatures, Shugart and Carey have proposed the most widely used framework for differentiating semi-presidential systems in an analytically useful way. They identified two basic variations of semi-presidential systems: premier-presidential and president-parliamentary. At the root of this differentiation is the origin and survival, as well as the relative power of the directly elected president vis-à-vis the assembly. The most important differentiation concerns the formation and dissolution of the cabinet and the extent of legislative powers. In the president-parliamentary systems it is the president who forms and dissolves the government at his/her own discretion. Although parliamentary confidence is still required for the survival of the cabinet, it is the sole prerogative of the president to form the cabinet. The cabinet in this case remains accountable to both the president, who can dissolve the cabinet, and to the parliament to which the cabinet is accountable and can be dismissed through a no confidence vote. In premier-presidential systems the prime minister is formally chosen by the president, but the cabinet is solely dependent and accountable to the legislature. Once the president makes the prime ministerial appointment he/she has no control mechanism over the working of the cabinet. The sole right to dissolve the cabinet resides with the legislative majority. In terms of formal constitutional powers, it is the president who assumes control of the executive government in president-parliamentary systems, while in premier-presidential systems it is the prime minister.
Perils of Semi-Presidentialism? Arguments and Comparative Empirical Evidence
The effects and consequences of semi-presidentialism has been a subject of long and heated academic debates. In the context of highlighting the deficiencies of presidentialism for democratisation Juan Linz argued that semi-presidentialism exhibits the same problematic characteristics as presidentialism. Linz has particularly and recurrently stressed the tendency of presidential systems to introduce a zero-sum element in political competition, to encourage personalisation of politics, and to undermine political parties. Later Linz and Alfred Stepan developed an additional and since than a standard critique of the destabilising effect of cohabitation and intra-executive conflict especially in new and fragile democracies.
Cindy Skach presented a similar critique, arguing that the combination of shared power between prime minister and executive president and their unequal legitimacy and accountability create tensions in semi-presidential countries. Arend Lijphart joined the critics of semi-presidentialism by arguing that it offers only a slight improvement over presidentialism and its perils. The direct election of the president keeps the zero-sum nature of the political competition in place and encourages personalisation of politics along with weakening of institutions and political parties. Moreover, Lijphart argued that depending on the outcome of elections, semi-presidential systems can easily become ‘hyper-presidential’ with the presidents acquiring much more sweeping powers than any president in ‘pure’ presidential systems.
Other scholars have been more optimistic about the consequences of semi-presidentialism. For Giovanni Sartori, the presence of two independent executives is an advantage as it allows for ‘head shifting’ and greater institutional flexibility.Considering the record of semi-presidentialism in Western Europe, Gianfranco Pasquino praised semi-presidential systems along similar lines for their supposed greater government capability and institutional flexibility. Reflecting on the experience of post-communist Central and Eastern Europe and the former Soviet Union, François Frison-Roche comes to a similar conclusion, praising the flexibility of semi-presidentialism, its ability to adapt to various contexts, and as a system that best enables rapid transition from dictatorship to democracy. 
These reflections on the contribution of semi-presidentialism are however based on either single country studies, or are restricted to specific regions with particular historical and political trajectories, which may prevent a fuller assessment of the impact of semi-presidentialism. Large N comparative research on the impact of semi-presidentialism has produced much less categorical conclusions. In the work of Sophia Moestrup, and Jose Antonio Cheibub and Svitlana Chernykh, the comparison of parliamentary, semi-presidential, and presidential systems found no significant difference between semi-presidential and presidential regimes. Taeko Hiroi and Sawa Omori show on the other hand that semi-presidential systems outperform parliamentary systems in democratic sustenance, while Milan Svolik and Ko Maeda find that semi-presidential and presidential systems are more prone to regime termination. These empirical findings show that the arguments about the deadly consequences of semi-presidentialism are somewhat exaggerated, since there is no conclusive evidence on their negative consequences compared to other regime types.
The comparative research has produced much more convincing findings on the varying effects of different types of semi-presidentialism. Elgie and Petra Schleiter show that premier-presidential systems are more conducive to survival than president-parliamentary systems. Similarly premier-presidential systems exhibit higher democratic performance than president-parliamentary systems. Moestrup comes to similar conclusions when analysing the regime effects in young democracies. Her research shows that premier-presidential systems have better record of civil liberties and democratic survival.
Cohabitation and Divided Government
The case against semi-presidentialism usually revolves around two situations that are said to be threatening to democracy: cohabitation and divided or minority government. Cohabitation is the situation when the executive president and the prime minister come from opposing parties, and when the president’s party is not represented in the government. The fear is that this will lead to intra-executive conflicts and government deadlock until such time as elections give a single party control over both the legislature and executive presidency. Elgie however shows that cohabitation mostly occurs in full democracies and all of the instances of cohabitation in full democracies survive without democracy collapsing. Cohabitation is on the other hand rather rare in partial democracies. Elgie counts only three such instances Weimar Republic (1923-24; 1927), Sri Lanka (2001-2004) and Niger (1995). Out of these three, democracy collapsed only in Niger as a result of cohabitation, which is therefore cited as a classic example of the perils of cohabitation. Meanwhile, no full democracy ever collapsed because of cohabitation and intra-executive conflicts.
Given the characterisation of cohabitation as a perilous outcome of semi-presidentialism it is important to consider factors that give rise to situations of cohabitation. First, it has been shown that cohabitation is more common in premier-presidential than in president-parliamentary systems. Elgie and Iain McMenamin show that cohabitation arises mostly following elections and even more commonly, mid-term elections, when the majority in the parliament changes, forcing a cohabitation with the incumbent president and newly formed majority in the parliament. Cohabitation is likely when there is a split of electoral preferences, either because of non-concurrent legislative and presidential elections, or vote splitting as a result of different electoral logics of parliamentary and presidential elections. In addition to showing that cohabitation on its own has no really empirically founded negative consequences, Elgie also shows that the factors that could give rise to cohabitation do not cause democratic breakdown. Actors’ anticipation of forthcoming cohabitation does not push them toward unconstitutional means of resolving political conflicts and democracy does not collapse. Overall, cross-case as well as case study research shows that the alleged deadly consequences of cohabitation have been exaggerated.
Minority or divided governments have been brought up as another undesirable consequence of semi-presidentialism. Minority and divided governments emerge when neither the president’s party nor any other party, including those explicitly opposed to the president, can form a government with a sustainable majority in the legislature. The argument goes that these insecure majorities lead to shifting alliances and coalitions in the legislature, with recurring government deadlock and intra-executive conflicts. This in turn leads to the executive president’s frequent use of decree and emergency powers and potentially to more extreme unconstitutional measures. Elgie finds that five out of 17 partial democracies collapsed in the periods of divided or minority government. Even those that did not collapse experienced increased political tensions, presidents had to frequently resort to their decree powers, and in two of them, Russia and Madagascar, the system moved towards granting more powers to the president. Similar to the case of cohabitation, however, divided and minority government led to democratic breakdown only in partial democracies, and never in full democracies. This might point to an apparent conjunctural relationship. Perils of semi-presidentialism come to the fore in absence of the wider ‘safety net’ that comes with fully democratic regimes, such as an independent judiciary and other independent institutions, strong civil liberties, an independent media and civil society, and so on.
Cabinet Formation and Control
The discussion in the previous sections shows that the allegedly most problematic effects of semi-presidentialism emerge because of the conflicts between presidents and parliaments, most commonly over the formation and control of the cabinet of ministers. Since semi-presidential systems exhibit high diversity in their institutional and behavioural characteristics it is evident that there is no single pattern of government formation and control. In this context it is important to elaborate on which actors – presidents, prime ministers, parliaments or parliamentary parties – have the upper hand in cabinet formation and control. Under what circumstances does this or that actor get advantage over the others? What are the mechanisms of influence that each of these actors possesses and applies?
Several complementing theoretical accounts exist on which actors compete for formation and control over cabinet in semi-presidential systems. For Octavio Amorim Neto and Kaare Strøm cabinet formation is the outcome of a ‘tug of war’ between the prime minister and the executive president. For Oleh Protsyk on the other hand, cabinet formation and control is an outcome of a bilateral bargaining between the executive president and parliament which (depending on the precise constitutional design) have the powers to nominate, appoint, and/or dismiss the cabinet, the prime minister and individual cabinet members. Schleiter and Edwards Morgan-Jones offer a more integrative account. They look at the chain of power delegation in semi-presidential systems from the perspective of a principal-agent theory. Since semi-presidential systems assume a joint control over the government by parliamentary parties (through confidence-investiture procedures) and executive president (through at least one of the three: cabinet appointment, dismissal, legislative powers), the legislative parties and the executive president are the principals of the cabinet, which is their agent.
A long-standing debate in the semi-presidentialism research has been about the relative importance of formal constitutional powers in cabinet formation and control. Several authors have argued that formal constitutional powers have little importance on actual balance of powers in semi-presidential systems. Contrary to these claims, the analyses of Protsyk, Amorim Neto and Strøm, and Schleiter and Morgan-Jones, of different samples of semi-presidential systems show that formal constitutional powers do play a crucial role in determining the formation and control over executive government. However, these authors also argue that despite the importance of formal constitutional powers other factors can amend the constitutional power granted to the actors.
Exactly what institutional mechanisms matter in determining cabinet formation and control, once we take for granted the finding that formal constitutional powers are crucial for cabinet formation and control? For Amorim Neto and Strøm and Protsyk it is the power to nominate, confirm and dismiss cabinet members that give one or the other party bargaining advantages. The power to nominate is a crucial agenda-setting mechanism while the unilateral dismissal power is a powerful tool that structures the entire negotiation process. In addition Protsyk shows that symmetric dismissal powers that give both the president and the parliament the right to dismiss the newly nominated prime minister and cabinet members provide a powerful institutional incentive for compromise and moderation since both the president and parliamentary majority know that their first preference candidates could easily be dismissed. Schleiter and Morgan-Jones on the other hand argue that government formation and control emerge from both, the formal constitutional structure as well as the results of elections that grant the authority to form and control the government to the parliamentary parties and the president.
As already mentioned, despite emphasising the role of formal powers these authors also argue that several intervening factors introduce considerable changes to the formal powers. Both Protsyk and Schleiter and Morgan-Jones show that actors’ de facto powers increase when government formation is triggered and takes place immediately after elections. Recent elections give both parliaments and executive presidents incentives and legitimacy to be proactive in imposing their preferred government composition. These authors show that the ability of executive presidents to influence cabinet formation and exercise control increase with the increase in the fragmentation of the legislature. In fragmented legislatures the bargaining environment is complicated by the number of veto actors. In such circumstance the executive president takes up the role of broker among several parties and can bargain the president’s preferred cabinet composition irrespective of the formal constitutional powers. The Amorim Neto and Strøm study also shows that the electoral volatility of parties in parliament from one election to another increases the executive president’s powers. On the other hand, these authors show that cabinet’s fragmentation or the size of the governing coalition decrease the executive president’s ability to influence government formation, since the multiple parties in the coalition all demand a share of ministerial positions for themselves, restricting the president’s opportunities to nominate his own or non-partisan members.
Beyond parliament’s and the governing coalition’s fragmentation, the nature of the party system is also a crucial aspect. Protsyk shows that presidents in premier-presidential systems acquire important leverage when the party system has a clientelistic nature and is characterised by frequent factional instability and floor-crossing which give the president more room in manoeuvring and finding a suitable candidate for the PM’s position that would be in her/his interest. In addition, S.G. Kang shows that the president’s membership in a legislative party is an important factor in the executive-legislative power balance. The president’s de facto powers considerably increase when she/he is the leader or a member of a strong legislative party.
The Nineteenth Amendment: What Has Changed?
The Nineteenth Amendment introduced considerable changes to the Sri Lankan constitution, particularly the balance of powers between the executive president, prime minister/cabinet and parliament. The president is directly elected by the people for a term of five years. The president is the head of state, the commander-in-chief of the armed forces, and also the head of the executive. The president is elected for a term of five years renewable only once, while individuals who have already served in the post for two terms are ineligible to run for the post again. The Nineteenth Amendment introduced a provision which states that the president is ‘responsible to parliament’ for the exercise of his/her duties and functions, although it is not entirely clear what this ‘responsibility’ entails. The president can be impeached by parliament by a two-thirds majority.
The president appoints the prime minister from among the members of parliament at his/her own discretion. However, the prime ministerial nominee should be able to command the confidence of the parliament, essentially constraining the president’s discretion with the actual distribution of power in parliament. This is given an additional twist since the term of the parliament is essentially fixed. The president has no power to dissolve parliament unless in the last six months of the lifespan of a parliament and only if two-thirds of MPs vote in favour of the resolution. This means that the president cannot dissolve parliament and call new elections in the hope that new elections will yield a parliamentary majority commensurate with the president’s preferences.
The president’s role in the functioning of the cabinet remains considerable. The president is a member and the head of the cabinet of ministers and consults whenever he/she finds it necessary the prime minister on the determination of the number of ministries, ministers, and the assignment of subjects and functions to such ministries. On the advice of the prime minister the president appoints the ministers and deputy ministers. The cabinet is formally collectively responsible to the parliament, but also de facto to the president who can change its composition, assignments, and functions of the ministers. However, the president cannot dismiss the prime minister, and the entire cabinet. The president can dismiss cabinet ministers only on the advice of the prime minister. The cabinet in its whole can be dissolved only if the parliament passes a no confidence vote or rejects the statement of government policy. As such the president losses the highly powerful leverage of cabinet dismissal as the survival of the cabinet rests only with the parliament. It is here that the collective responsibility of the cabinet to the parliament is exercised. Consequently the command of the majority of the parliament by the prime minister is essential to the survival of the cabinet.
Through the changes introduced by the Nineteenth Amendment Sri Lanka has moved from the president-parliamentary to premier–presidential category of semi-presidential systems. The most important change has been in the origin and survival of the cabinet. Before the Nineteenth Amendment the president and legislature had separate origin and survival. The cabinet’s origin was with the president, while its survival was with the executive president, through dissolution power, and the legislature, through confidence vote. However, even in the case of no confidence vote the president had the discretion to reappoint the cabinet as he/she wished. This placed both the actual origin and survival of the cabinet with the president. With the Nineteenth Amendment the origin of the cabinet lies with the president mediated through parliamentary majority. However, and most importantly, the prime minister’s and the cabinet’s collective survival solely rest with the parliament. In terms of principal agent theory cabinet and the prime minister are now the agents of the parliament and through the parliament of legislative parties.
Adopting and measuring the extent of presidential powers based on the Shugart and Carey scale, the current Sri Lanka executive president scores 10 points compared to the 16 points it had in the Shugart and Carey research from 1992. Table 1 below shows the power of Sri Lanka’s president as measured on Shugart and Carey’s scale before and after Nineteenth Amendment. As compared to semi-presidential systems measured by Shugart and Carey, Sri Lankan president under Nineteenth Amendment has 10 points, which is below the average of 13 points of all semi-presidential regimes including both, president-parliamentary and premier-presidential systems as initially measured by Shugart and Carey. Adopting Steven Roper’s modification of Shugart and Carey’s scale for premier-presidential regimes shows that from the point of view of formal constitutional powers the Sri Lankan president is the second strongest among premier-presidential systems as originally measured by Roper. Only the president of Iceland with 16 aggregate points has more formal powers than the president of Sri Lanka.
Table 1: Powers of the Sri Lankan President before and after Nineteenth Amendment
Compared to the pre-Nineteenth Amendment the president has not gained any additional legislative powers, but retained the crucial right to propose an issue to a referendum – potentially a powerful tool of agenda-setting and a check on the powers of the parliament. Importantly the president’s sweeping non-legislative powers were somewhat constrained. Prior to the Nineteenth Amendment the president had full power to form and dissolve government, as well as the right to dissolve the parliament. With the Nineteenth Amendment the president still retains the nomination of the prime minister, subject to the candidate commanding support of the majority of the legislature. The president also appoints members of cabinet, but only on the advice of the prime minister, and has the right to change the composition of the cabinet, the functions and assignments of the cabinet ministers. These changes though must not alter the continuity of the cabinet and the continuity of its responsibilities to the parliament. On the other hand the president lost the power to dissolve the cabinet of ministers. The president has to an extent retained the right to dissolve the parliament. However, this dissolution power is severely constrained in time. The president can only dissolve the parliament four and a half years after the first sitting of the parliament.
As already discussed an executive president’s actual powers can be considerably different from formal powers depending on a number of factors. Existing research has identified the nature of the party system, electoral volatility, the degree of parliament’s as well as governing coalition’s fragmentation, and the occurrence of minority government, as some of the most important such factors. In this context it is crucial to consider the impact of the electoral system. The extent of fragmentation of the legislature, and the governing coalition(s), the shape of the party system and electoral volatility are all to a great extent influenced by the electoral formula. Additionally, the design of electoral districts, in particular of district magnitude, as well as the overall cleavage structure of the society are all of important consideration in this regard. This goes back to a limited, but important stream of research that has argued of the crucial importance of studying interactive effects of institutions rather than their separate, net effect on any given phenomenon.
The existing research by Schleiter and Morgan-Jones and Amorim Neto and Strøm shows that fragmented legislatures would increase the president’s role in government formation by giving him/her more space for manoeuvre and nomination of a prime minister closer to his/her preference. In this context a more permissive electoral system such as proportional representation systems could be expected to fragment the legislature and provide more leverage for the executive president in cabinet formation. Given Sri Lanka’s ethnic diversity, drawing electoral districts along the settlement patterns of ethnic groups will enable majoritarian systems to provide for quite extensive representation of ethnic groups and achieve a degree of fragmentation of the legislature. Legislative fragmentation will likely mean that no single party would be able to command absolute majority. Parties would need to enter into governing coalitions. The larger these coalitions, the more room would the executive president have in the appointment of the prime minister, but also less control over the appointment of ministers and other cabinet members. Legislative fragmentation and inability to form a stable majority by fewer parties could also be expected to lead to situations of minority government, again paving the way for more extensive powers of the executive president.
However, it is important to consider not only the mechanical degree of fragmentation but also the role of the executive president in the party system. As Kang shows in his research given president’s nomination power the executive president’s role will highly increase whenever he/she heads or is a member of a parliamentary party. Irrespective of the formal power of the president the post is likely to continue to occupy an important role for Sri Lankan political parties. In this sense it is unlikely that the position of the president would become apolitical and solely symbolic.
Subject to the changes of the electoral system, the party system can change considerably from the current mainly two-party structure. The party system could change if the new electoral system is very permissive, encouraging and rewarding possible fragmentation of the two major parties. The party system might also evolve from the traditional two-party system if minorities consolidate around one to two parties and actively engage in electoral politics so that the minority parties’ share of representation in the legislature approximate their share in the population. However, if we assume the party system does not change, then it is very likely that the post of the president will be occupied by one of the two powerful parties holding the parliamentary majority on their own or in coalition with other parties. This suggests that formal presidential powers aside, it is the presidential parties in the legislature that will be a crucial source of power for the presidential executive.
In case the party opposed to the president is holding a majority in the parliament the president’s power can be considerably restricted. In case a party opposed to the president is holding an inconclusive majority in the parliament there is likely to be a cohabitation or minority government. The peculiarity of these two situations in the current Sri Lankan system is that the term of the parliament is essentially fixed. There is no constitutional provision that would allow for dissolution of the parliament unless there is a strong consensus on dissolution within the legislative parties. On the one hand this locks all parties to four and a half years of tense intra-executive and inter-party relations and unstable government. On the other hand the fixed parliamentary mandate is a powerful incentive for more consensual relations that can incentivise parties to compromise.
 M. Duverger, ‘A New Political System Model: Semi-Presidential Government’ (1980) European Journal of Political Research 8(2): p.166.
 For detailed criticism of Duverger’s original definition see R. Elgie, ‘The Politics of Semi-Presidentialism’ in (Ed.) (1999) Semi-Presidentialism in Europe (Oxford: Oxford University Press): pp. 4-12.
 Ibid: p.12.
 These authors also show that heterogeneity is also high among presumably ‘pure’ parliamentary and presidential systems. See e.g., J.A. Cheibub, Z. Elkins & T. Ginsburg, ‘Beyond Presidentialism and Parliamentarism’ (2014) British Journal of Political Science: pp.1-30.
 Because of this extreme diversity especially in the powers of the executive presidents, Alan Siaroff has expressed doubts whether semi-presidentialism is a distinct and coherent form of government: see A. Siaroff, ‘Comparative presidencies: The inadequacy of the presidential, semi-presidential and parliamentary distinction’ (2003) European Journal of Political Research 42(3): p.307.
 L.K. Metcalf, ‘Measuring Presidential Power’ (2000) Comparative Political Studies 33(5): pp.660-685; Siaroff (2003).
 M.S. Shugart & J. Carey (1992) Presidents and Assemblies: Constitutional Design and Electoral Dynamics (Cambridge: Cambridge University Press).
 M.S. Shugart, ‘Semi-Presidential Systems: Dual Executive And Mixed Authority Patterns’ (2005) French Politics 3(3): pp.323–351.
 Shugart & Samuels (2010).
 P. Schleiter & E. Morgan-Jones, ‘Party government in Europe? Parliamentary and semi-presidential democracies compared’ (2009a) European Journal of Political Research 48:5, 665–693; P. Schleiter & E. Morgan-Jones, ‘Review Article: Citizens, Presidents and Assemblies: The Study of Semi Presidentialism beyond Duverger and Linz’ (2009b) British Journal of Political Science 39:4, 871; P. Schleiter & E. Morgan-Jones, ‘Who’s in Charge? Presidents, Assemblies, and the Political Control of Semipresidential Cabinets’ (2010) Comparative Political Studies 43(11): pp.1415-1441.
 Shugart (1992).
 Ibid: p.325.
 Schleiter & Morgan-Jones (2009b).
 Shugart (2005): p.325.
 Ibid: p.327.
 Shugart & Carey (1992).
 J. Linz, ‘Presidential Versus Parliamentary Democracy: Does It Make a Difference?’ in J. Linz & A. Valenzuela (Eds.) (1994) The Failure of Presidential Democracy (Johns Hopkins University Press); J. Linz, ‘Introduction’ in T. Ray (Ed.) (1997) Postcommunist Presidents (Cambridge University Press).
 J. Linz & A. Stepan (1996) Problems of Democratic Transition and Consolidation (Johns Hopkins University Press): pp.278–279.
 C. Skach (2005) Borrowing Constitutional Designs: Constitutional Law in Weimar Germany and the French Fifth Republic (Princeton University Press).
 A. Lijphart, ‘Constitutional Design for Divided Societies’ (2004) Journal of Democracy 15(2): pp.96-109, also available at: http://muse.jhu.edu/content/crossref/journals/journal_of_democracy/v015/15.2lijphart.html (last accessed 18th March 2016).
 G. Sartori (1997) Comparative Constitutional Engineering: An Inquiry Into Structures, Incentives, and Outcomes (New York University Press).
 G. Pasquino, ‘Semi-Presidentialism: A Political Model at Work’ (1997) European Journal of Political Research 31: pp.128-137.
 F. Frison-Roche, ‘Semi-Presidentialism in Post-Communist Context’ in R. Elgie & S. Moestrup (Eds.) (2007) Semi-Presidentialism Outside Europe (Abingdon: Routledge): pp. 56-77.
 S. Moestrup (2007a) ‘Semi-presidentialism in Niger: Gridlock and Democratic Breakdown – Learning from Past Mistakes’ and S. Moestrup (2007b) ‘Semi-presidentialism in Young Democracies: Help or Hindrance?’ in R. Elgie & S. Moestrup (Eds.) (2007) Semi-Presidentialism Outside Europe (Abingdon: Routledge): pp. 105-120 and pp. 30-55; J.A. Cheibub & S. Chernykh ‘Are Semi-Presidential Constitutions Bad for Democratic Performance?’ (2009) Constitutional Political Economy 20(3-4): pp.202-229.
 T. Hiroi & S. Omori, ‘Perils of Parliamentarism? Political Systems and the Stability of Democracy Revisited’ (2009) Democratization 16(3): pp.485-507.
 M. Svolik, ‘Authoritarian Reversals and Democratic Consolidation’ (2008) American Political Science Review 102(2): pp.153-168; K. Maeda, ‘Two Modes of Democratic Breakdown: A Competing Risks Analysis of Democratic Durability’ (2010) The Journal of Politics 72(4): pp.1129-1143.
 R. Elgie & P. Schleiter, ‘Variation in the Durability of Semi- Presidential Democracies’ in R. Elgie, S. Moestrup & Y.S. Wu (Eds.) (2011) Semi-Presidentialism and Democracy (Oxford: Oxford University Press): pp. 42-61; R. Elgie (2011) Semi-Presidentialism Sub-Types and Democratic Performance (Oxford: Oxford University Press): pp.43-49.
 Elgie (2011): pp.69–94.
 Moestrup (2007b).
 R. Elgie & I. McMenamin, ‘Explaining the onset of cohabitation under semi-presidentialism’ (2011) Political Studies 59(3): pp.618-619.
 R. Elgie, ‘The Perils of Semi-Presidentialism. Are They Exaggerated?’ (2008) Democratization 15(1): pp.49-66.
 According to M.S. Shugart & D.J. Samuels (2010) Presidents, Parties, and Prime Ministers: How the Separation of Powers Affects Party Organiztaion and Behavior (Cambridge: Cambridge University Press): pp.45-46, Sri Lanka 2001-2004 is the only case of cohabitation that existed in a president-parliamentary system. All other experiences of cohabitation come from premier-presidential systems.
 Elgie (2008): pp. 49-66.
 L. Kirschke, ‘Semipresidentialism and the Perils of Power-Sharing in Neopatrimonial States’ (2007) Comparative Political Studies 40(11): pp.1372-1394; Moestrup (2007a).
 Elgie (2008); R. Elgie, ‘Semi-presidentialism, Cohabitation and the Collapse of Electoral Democracies, 1990-2008’ (2010) Government and Opposition 45(1): pp.29-49; Elgie & McMenamin (2011).
 Elgie (2010).
 Elgie (2008).
 O. Amorim Neto & K. Strøm, ‘Breaking the Parliamentary Chain of Delegation: Presidents and Non-partisan Cabinet Members in European Democracies’ (2006) British Journal of Political Science 36(4): p.619.
 O. Protsyk, ‘Prime Ministers’ Identity in Semi-Presidential Regimes’ (2005) European Journal of Political Research 44(5): pp.721-748.
 Schleiter & Morgan-Jones (2010): pp. 1415–1441.
 J.A. Cheibub, ‘Making Presidential and Semi-Presidential Constitutions Work’ (2008) Texas Law Review 87: pp.1375, 1398-1401; Duverger (1980): pp.179-180; Linz & Stepan (1996): p.278; Siaroff (2003): p.303.
 Protsyk (2005); Amorim Neto & Strøm (2006); Schleiter & Morgan-Jones (2009a); Schleiter & Morgan-Jones (2009b); Schleiter & Morgan-Jones (2010).
 AmorimNeto and Strøm(2006); Protsyk (2005).
 Protsyk (2005).
 Schleiter & Morgan-Jones (2010).
 Protsyk (2005); Schleiter & Morgan-Jones (2010).
 Amorim Neto & Strøm(2006).
 Protsyk (2005).
 S.G. Kang, ‘The Influence of Presidential Heads of State on Government Formation in European Democracies: Empirical Evidence’ (2009) European Journal of Political Research 48(4): pp.543-572.
 When outlining the legislature-cabinet relationship the amendment states that the cabinet is “responsible” and “answerable” to the Parliament. The word answerable is omitted form the description of legislature-president coordination.
 For a detailed discussion of the scale see in Shugart and Carey (1992): Ch.8 at pp.148-154.
 S. D. Roper, ‘Are All Semi-presidential Regimes the Same? A Comparison of Premier-Presidential Regimes’ (2002) Comparative Politics 34(2): pp.253-272
 T.D. Sisk (1996) Power Sharing and International Mediation in Ethnic Conflicts (United States Institute of Peace); K. Belmont, S. Mainwaring & A. Reynolds, ‘Institutional Design, Conflict Management, and Democracy’ in A. Reynolds (Ed.) (2002) The Architecture of Democracy : Constitutional Design, Conflict Management, and Democracy (Oxford: Oxford University Press).
 Schleiter & Morgan-Jones (2009); Schleiter & Morgan-Jones (2010); Amorim Neto & Strøm (2006).
 Kang (2009).
 The Nineteenth Amendment requires that the resolution on dissolution of the legislature must have the support of two-thirds of MPs including those not present.