The Perils of Semi-Presidentialism? The Collapse of Cohabitation and the Design of Executive Power in Sri Lanka


Asanga Welikala

Edinburgh Law School

The government of national unity, formed in 2015 promising major reforms in a new constitution, has now all but ended. Only a fragment of the promised reforms was enacted, and there is now no hope for any further. With only stagnation between crises in prospect, the weaknesses of the compromise reforms made to the semi-presidential system in 2015 are increasingly exposed.

The presidential and parliamentary elections of January and August 2015, fought and won almost entirely on issues of good governance and constitutional reforms, could have been a pivotal moment in Sri Lanka’s democratic development. They mandated a new constitution to replace the hyper-presidential 1978 Constitution that would abolish the executive presidency, devolve further powers to the provinces, and introduce a second chamber, a mixed member proportional electoral system, a new bill of rights, and comprehensive constitutional review. The elections were won by the widest coalition of political parties and civil society groups ever arrayed against a sitting president, and by a pluralistic democratic majority painstakingly constructed from across ethnic and religious divides. The government formed in 2015 therefore had an air of historic purpose behind it, not only because of its special reform mandate, but also because of the symbolic importance of the sense of nation-building unity that inspired the majority of Sri Lankans who voted for it.

By late 2017, however, the momentum for reform had run out of steam, with the old political culture of party antagonisms and self-interests reasserting itself. President Sirisena and Prime Minister Wickremesinghe proved incapable of rising to the occasion in implementing their joint mandate. In the context of the breakdown of their political relationship, the one constitutional amendment their government was able to enact in the immediate aftermath of the regime change – the Nineteenth Amendment in April 2015 – now takes centre stage.

The main change of the Nineteenth Amendment was to recalibrate the constitutional relationship between the president and prime minister, and by extension that between government and parliament. The initial proposal in 2015 was expansive. By the simple but significant expedient of requiring the president to act always on the advice of the prime minister, it would have transformed the constitution from a semi-presidential to a parliamentary model. However, in the process of negotiation and compromise between parliamentary parties including differences within the new governing coalition, the reach of the ‘advice clause’ was significantly diminished, although the prime minister’s mandatory advice has become more important in some ways, and his overall constitutional position within the executive was also strengthened in other ways. For defenders of semi-presidentialism, the Nineteenth Amendment represents an appropriate institutional balance by removing the more egregious features of the 1978 Constitution. But this appears to be an over-sanguine assessment. 

After the Nineteenth Amendment (the key provisions are Articles 30-33A and 42-48), the president is directly elected for a term of five years, with a two-term limit (both term limits and other important restraints on presidential power had been removed by the Eighteenth Amendment in 2010). The president is the head of state and government and commander-in-chief, and is responsible, although not answerable, to parliament. The president is a member and the head of the cabinet of ministers, which directs and controls the government. Cabinet ministers (other than the president) are collectively responsible and answerable to parliament. The president appoints as prime minister the member of parliament who is most likely to command the confidence of parliament. The president may consult the prime minister, where he considers it necessary, in determining the number of cabinet and other ministries, in assigning functions to them, and reallocating such functions. However, the president can only act on the advice of the prime minister when appointing and dismissing cabinet and other ministers. The prime minister cannot be removed by the president and continues in office throughout the period the cabinet continues to function. The cabinet is dissolved if the prime minister ceases to hold office by death, resignation, or ceasing to be a member of parliament. The cabinet is also dissolved if the government loses a vote of no confidence, or fails to get its statement of government policy or the annual appropriations bill passed.

Although the classification might be debatable, these changes have plausibly been characterised as converting the 1978 Constitution from a ‘president-parliamentary’ to a ‘premier-presidential’ model of semi-presidentialism. After the Nineteenth Amendment, the cabinet, as a matter of law, is exclusively responsible to parliament and not simultaneously to the president. The prime minister cannot be dismissed by the president and this can only be done by parliament withdrawing confidence from the government as a whole. Since he remains the head of the cabinet and government, and can reassign functions between them without prime ministerial advice, ministers may wish to maintain cordial relations with the president, but that is not a legal requirement. And of course, the president may well accrue greater de facto power in a situation in which his party also enjoys a parliamentary majority.

Such a model can be defended as a system of power-sharing, which is especially important in the context of a constitutional transition as envisaged in 2015. A model of consensus politics by invoking the ancient precedent of the Lichchavis was also how the current prime minister (then the leader of the opposition) articulated the rationale of the emerging coalition in late 2014. Mentioned with approval in the Buddhist Pali canon, the Lichchavi kingdom in Nepal (400-750 CE) is often cited as an ancient South Asian precedent for a form of government that reflected characteristics of federalism, power-sharing, and limited government. 

However, such well-meaning analyses lose sight of two important factors. The first is that the Nineteenth Amendment framework is not the product of a properly thought-through design, but the accidental upshot of an ill-disciplined and hasty process of political horse-trading. Amorphous language, inconsistent promises, and a general lack of transparency and truthfulness between parties within the winning coalition and beyond, led to a major division between those who interpreted the mandate as being for the abolition of the executive presidency and those who wanted only limited changes to it. Instead of constructing a solid political consensus in favour of abolition as the majority within the coalition desired – or at least on a principled compromise – before legislation was prepared, the government allowed a free-for-all in the debate and vote, with the result that the coherence of the constitutional amendment bill was gnawed out of existence without any control. That is why the advice clause is retained in some areas but not in others, and there is no rational explanation for why this is so. As a purely contingent framework, therefore, it could work only so long as the relationship between Sirisena and Wickremesinghe held up. 

The second factor is that no amount of institutional re-engineering, or popular mandates for consensus and cohabitation, seem able to change the fundamentally antagonistic culture of politics between the two main party groupings. The current unravelling of the ruling coalition closely follows the failing trajectory of the previous attempt at cohabitation in 2001-4. By the structure of incentives and disincentives that animates this culture, and in particular the gap in the accountability loop in between elections, the professional political class prefers a winner-takes-all logic in relation to power, and enjoys sufficient autonomy from the electorate to maintain it.

These two factors deprive the Nineteenth Amendment’s premier-presidential framework of any justification based on principle or pragmatism. It is not a model capable of delivering a meaningful constitutional transition through power-sharing; it merely institutionalises dysfunction by empowering obstructionists and crippling reformers. Moreover, rather than try to institutionally transform political culture overnight, the more realistic approach might be to harness this culture to the better ends of reform by the greater political accountability and diffusion of power necessarily implied by a parliamentary system. If it is accepted that presidentialism is not a system of government consistent with constitutional democracy in the context of Sri Lanka’s feeble political culture and its pluralistic social structure, then it is clear that the country needs a fully parliamentary model of constitutional state based on Commonwealth values.

Dr Asanga Welikala is a Lecturer in Public Law at Edinburgh Law School and the Director of the Edinburgh Centre for Constitutional Law.