Blog of the International Journal of Constitutional Law

Distinguishing Among Referenda (I-CONnect Column)

Aslı Bâli, UCLA School of Law

[Editor’s note: This is one of our biweekly I-CONnect columns. Columns, while scholarly in accordance with the tone of the blog and about the same length as a normal blog post, are a bit more “op-ed” in nature than standard posts. For more information about our four columnists for 2017, see here.]

Over the last two months, the role of the referendum as a tool of populist mobilization and democratic decline has weighed heavily on my mind. The Turkish referendum was the obvious catalyst, but the use of referenda is also a subset of the broader question of how democratic processes may be used to subvert democracy. And all of these concerns are related to the question of the role of comparative law in a time of “democratic decay,” as my fellow I-CONnect columnist, Tom Gerald Daly has framed it. A quick survey of recent ICONnect posts makes clear the widespread concern about the corrosive effect of populist strategies on democratic practices. For instance, the joint I-CONnect/Verfassungsblog mini-symposium on populism and constitutional courts draws attention to the degree to which populism may adopt the language of democracy and constitutionalism—as in Jan Werner Muller’s discussion of the Hungarian case—while eroding the substantive “minimum core” of democracy, including “longstanding commitments to free and fair elections, the separation of powers, basic human rights and government accountability,” as Rosalind Dixon argues.

These recent ICONnect pieces all engage important dimensions of the question of how populists harness the engine of democratic processes to undermine basic democratic commitments. How can we best understand this dynamic in the case of referenda? This question seems especially important in light of the remarkable number of high profile referenda that have taken place over the last year. Perhaps the most famous referendum of 2016 was the Brexit vote, but other examples include the constitutional referenda in Italy and Thailand, the Hungarian referendum over that country’s EU migrant quota and the referendum on the Colombian peace deal.  The Turkish constitutional referendum earlier this month may well have been a particularly acute instance of invoking the popular will to disable democratic restraints, but it is not the only one.

In a piece responding to the Brexit vote, Kenneth Rogoff famously characterized referenda as “Russian roulette for republics.” Rogoff’s argument was less a substantive critique of the use of a referendum as one tool for democratic governance than a procedural critique of how the British referendum was structured. Rogoff begins by noting that “the idea that somehow any decision reached anytime by majority rule is necessarily ‘democratic’ is a perversion of the term.” Yet there are many circumstances in which we endorse majoritarian decision-making as democratically permissible. Rogoff’s proposed procedural corrections seek to identify the conditions that distinguish legitimate deliberative decision rules from those that undermine democratic values. That is, he argues that the process by which a referendum is conducted is at least as important to its democratic pedigree as the substance of the question being put to the public. He suggests that a legitimate process requires mechanisms that provide checks and balances, protect minority interests and ensure the deliberative quality of public participation. In the case of the Brexit vote, he would have required an iterated vote over a long interval—such as two popular votes spaced over at least two years—and a subsequent confirming parliamentary vote requiring a supermajority of 60%.

Stephen Tierney’s book Constitutional Referendums is a sustained scholarly engagement that addresses a similar set of issues at a higher level of abstraction, namely the question of when and why constitutional referenda may be appropriate and how they should be conducted. As he wrote in his reply to Peter Oliver’s review of the book for ICONnect, he takes the position that “referendums are defensible for higher order, or constitutional, decision-making.” More specifically he notes that the importance of constitutional decisions and their impact on shaping citizen identities over time provide a “plausible argument for engaging citizens directly in exercises of constitutional authorship.” In other words, there is an appropriate democratic role for constitutional referenda under some circumstances. But, like Rogoff, Tierney is concerned that such exercises be structured in ways that promote deliberative participation and he also identifies specific risks associated with the use of constitutional referenda in deeply divided societies.

Tierney acknowledges longstanding criticisms of referenda that include: “elite influence, aggregative rather than deliberative decision-making and that [referenda] structurally favor majoritarianism.”[1] His proposed solutions include guidelines for multiplying opportunities for participation that could be designed to mitigate elite control over issue-framing, accord minorities a greater role and enhance public deliberation. Thus a referendum that meets the requirements of deliberative democracy, on this account, would be part of a broad process with a number of stages including fair provision of information to voters, engagement of civil society, regulation of campaign expenditure and access to media, and inclusion of counter-majoritarian devices requiring more than a simple majority in cases of divided societies. For Tierney, the criticisms of referendum democracy are “problems of practice rather than principle” that can be overcome, for the most part, by “good process design.” He comes out differently than Rogoff on his view of the legitimacy of the Brexit vote because he finds that the referendum was subject to a statutory scheme, rigorously regulated, with broad dissemination of information by all sides, and was not marked by voting irregularities or apparent breaches of funding or spending rules.

While Tierney and Rogoff may disagree on their assessment of a particular constitutional referendum in practice, they reflect a consensus that the democratic legitimacy of a referendum depends on a minimum set of procedural protections. By this standard, the claim that the Turkish referendum employed democratic means towards anti-democratic ends is simply incorrect. Both the means and the ends of the Turkish referendum were plainly anti-democratic. Substantively, there can be little doubt that the eighteen constitutional amendments that were voted on as a package in a Yes/No vote in Turkey present a threat to the “democratic minimum core” defined by Dixon and Landau.[2] As Ilayda Gunes has shown, what was at stake in the Turkish constitutional amendment proposal was “dissolving the very institutions that give democracy any proper vitality or legitimacy.” In brief, the amendments replace the parliamentary system with a highly centralized and partisan executive with vastly expanded powers to control judicial appointments, rule by decree, declare states of emergency, appoint vice-presidents and ministers without parliamentary approval and dissolve the parliament. Parliamentary powers to investigate or impeach the president are also significantly diminished. There are further aspects of the amendments that subvert both separation of powers and any check on the executive, but this list is sufficient to convey the damage to Turkey’s constitutional order. While this is not the first package of constitutional amendments to pass by referendum in Turkey, it is the first to be broadly denounced by the Venice Commission, which stated that the referendum “represents a dangerous step backwards in the constitutional democratic tradition of Turkey.”

Procedurally, however, the popular focus on discrediting the outcome has been on credible allegations of fraud and irregularities, including a last minute change by the High Election Council allowing unofficial ballots to be counted. The emphasis on outright ballot-stuffing threatens to occlude the prior reasons why the process by which the referendum was conducted, no less than the substance of the amendments themselves, represents an assault against democratic practice in Turkey. The referendum was conducted under a state of emergency, with over a dozen opposition parliamentarians and thousands of dissidents detained, tens of thousands of civil servants purged from their positions including judges, lawyers and public university professors, an unprecedented assault on press freedoms including the jailing of journalists, a de facto ban on campaigning for proponents of the “No” campaign in the referendum, and the flooding of print and broadcast media as well as public spaces by the government with the messaging of the “Yes” campaign. Government officials used their public platform to denounce “No” voters as traitors and incited a campaign of public and private intimidation and violence against opposition voters. The manufacturing of a crisis with the German and Dutch governments in order to both influence diaspora voters and produce a nationalist backlash in support of the “Yes” campaign at home projected the democratic deficit of the referendum well beyond Turkey’s borders.[3] In short, the referendum was deliberately conducted to exacerbate polarization, marginalize opposition, stymie public access to information and maximize an uneven playing field to the government’s advantage. The process by which the referendum was conducted undermined minority rights, civil liberties, rule of law, separation of powers and checks and balances in Turkey just as much as the substance of the amendments. Understanding the referendum as an instance of democratic decline by democratic means is to give the Turkish referendum—like many other recent referenda—undue credit. Neither the means nor the ends qualify as democratic.

Suggested citation: Aslı Bâli, Distinguishing Among Referenda, Int’l J. Const. L. Blog, Apr. 27, 2017, at:

[1] Stephen Tierney, Constitutional Referendums: The Theory and Practice of Republican Deliberation (OUP 2012), at 41.

[2] Rosalind Dixon and David Landau, ‘Transnational Constitutionalism and a Limited Doctrine of Unconstitutional Constitutional Amendment’ (2015) 13 International Journal of Constitutional Law 606.

[3] For an excellent overview of the circumstances in which the referendum was called, see Umit Cizre, ‘Fear and Loathing in Turkey,’ MERIP, April 26, 2017.


Leave a Reply

Your email address will not be published. Required fields are marked *