• 1 Introduction

  Direct democracy is considered an important instrument of democratic innovation and a panacea against the crisis in parliamentary and presidential systems. It has been argued that referendums strengthen political legitimacy ( Bowler and Donovan 2002), increase feelings of efficacy (Bowler and Donovan 2002; Mendelsohn and Cutler 2000; Qvortrup 2018) and mitigate conflicts (Collin 2014). Nevertheless, modern authoritarian leaders also use referendums as a tool to bolster their legitimacy.
  This analysis of two case studies, Turkey and Russia, contributes to academic research on modern authoritarian regimes (Levitsky and Way 2002; Puddington and Roylance 2017) as well as on electoral integrity (Schedler 2006; ). Numerous terms have been used to refer to modern authoritarian regimes including “semidemocracy”, “virtual democracy”, “pseudodemocracy”, “illiberal democracy”, “soft authoritarianism” and “electoral authoritarianism” (Levitsky and Way 2002: 51). Although characterised by authoritarian leadership, primarily by one individual, these regimes often adopt democratic institutions to maintain power and gain domestic and international legitimacy (Levitsky and Way 2002: 57). Their stability is reliant on their leaders’ ability to act effectively and legitimise their power on different levels including by establishing mechanisms of representation and participation that help justify their authoritarian practices. One such tactic is to “validate autocracy through the ballot” (Levitsky 2002: 10).
  In recent years, democratic innovation in numerous countries (such as Ireland and Chile) has seen constitutional referendums combining direct democracy with deliberative democracy (Welp and Soto 2020). This technique is also being deployed in some authoritarian countries. However, it has also been argued that modern authoritarian systems are using these deliberative instruments to develop deliberative authoritarianism (He and Warren 2011). This study investigates whether this deliberative constitution making is being used in authoritarian systems such as Russia and Turkey. If so, how is it being combined with direct democracy and what effect does it have on the integrity of the referendums? Therefore, our second research question focuses on the use of deliberative tools in the constitutional review processes of Turkey and Russia. We hypothesise that the integrity of referendums in these countries is often dubious. Modern authoritarian systems¹x In the following text, we use the terms ‘modern authoritarian regimes’ and ‘modern autocracies’ interchangeably. typically use repressive measures to demobilise civil society leading to political apathy and disengagement. However, authoritarian systems may also attempt to mobilise citizens and activate voters using propaganda and participatory instruments.
  We here analyse Turkey and Russia using a new research instrument that measures direct democratic integrity. This tool was first applied to Turkey in 2017 (Kersting and Grömping 2021 and, later, to the democratic system in Italy in 2020 (Kersting and Regalia 2023). This is the first comparative analysis article on applying this tool to two authoritarian systems. Both of these case study countries transitioned from more democratic systems to systems with very strong authoritarian presidents. Both used referendums to strengthen the incumbent’s power.
  Turkey and Russia are also both considered hybrid modern authoritarian regimes in which democratic institutions are regularly misused, subverted and instrumentalised for the purpose of electoral legitimisation (see Coppedge et al. 2018). Under the Erdogan government, particularly since the attempted coup of 2016, Turkey has been considered an electoral autocracy (Baghdady 2020). In the 1990s, Russia was characterised as a “defective democracy” regime, that is, it incorporated not only democratic but also authoritarian and anarchic elements (Mommsen 2019: 29). However, Russia has transformed under the presidencies of Vladimir Putin and has been a strictly “managed democracy” since 2000. As the term suggests, constitutional principles have been bent and democratic institutions and procedures manipulated to authoritarian ends (Mommsen 2019: 29). This study contributes towards filling the gap in research on how referendum integrity operates under undemocratic conditions. Considering the integrity and fairness of referendums for minorities and the opposition to be crucial, we have investigated the level of integrity of constitutional referendums in authoritarian systems.
  By comparing the integrity of referendums in Turkey (2017) and Russia (2020) based on the results of the new Direct Democracy Integrity (DDI) Index and its expert survey, the following study develops a theoretical framework for identifying the links between modern authoritarianism, referendums and deliberation. Section 2 presents the methodical and empirical approach used to assess the integrity of the referendum as an instrument of direct democracy. We compare the integrity of referendums in Turkey (2017) and Russia (2020) using the new DDI-Index and its expert survey. Section 3 compares the political context of the referendum in Russia to that in Turkey. Section 4 presents the results of the expert surveys for Russia and Turkey and provides further detail on some of the empirical data. Section 5 presents our conclusions.

• 2 Plebiscites and Referendums under Autocracies

  Modern authoritarianism has several defining characteristics: among them, an “illusion of pluralism that masks state control over key political institutions”, “state or oligarchic control over key elements of the national economy”, “legalized political repression” and “suppression of nongovernmental organizations (NGOs) that focus on human rights or political reform” (Puddington 2017: 6). In such regimes, democracy is essentially corrupted as the regime must control election outcomes for, as Puddington demonstrates,

  they need to hold votes to validate their rule, but they also recognise the risk involved, as elections remain a potent instrument of democratic renewal even in deeply troubled societies. (2017: 14)

  They are looking for, to use Max Weber’s term, “legitimate domination” in which the governed accept authoritarian rule because the leader comes “from an authoritative source” or because “they accept the reasons provided by the rulers” (He and Warren, 2011: 6).
  Despite the awareness of these issues, the role plebiscites and referendums play in autocracies remains largely overlooked in the literature (the prominent exceptions being Altman 2010; Qvortrup 2018). Mechanisms of direct democracy have been extensively deployed in authoritarian states since the revolution in late-eighteenth-century France (Altman 2011: 88). The most notorious cases took place in Nazi Germany and there have been more recent cases in Romania under Ceausescu and the Philippines under Marcos. Most of the votes in these regimes were marred by massive fraud, coercion and blatant propaganda. Qvortrup (2018: 15) suggest that autocracies use referendums to “intimidate and control the population” and “disorganize its potential opponents”. They also find that referendums are most often used in autocracies with high ethnic fractionalisation and ‘sultanistic’ regimes, while post-communist states are less likely to use plebiscites. According to other studies, party-based autocracies also rarely use plebiscites as they rely less on personal charismatic power and obtain political legitimacy through alternative channels (Monje and de la Cruz 2019).
  The integrity of a referendum as a direct democratic instrument highlights normative standards. Direct Democracy Integrity focuses on a free and fair referendum process that ensures equal universal suffrage, in particular, and protects minority rights by facilitating campaigns that enable free expression for both ruling and opposition parties, in addition to majorities and minorities. The concept is very closely related to electoral integrity, which has been the focus of international research projects since 2012 (see Norris et al. 2014). Election integrity is based on Article 2 (3) of the Universal Declaration of Human Rights (1948). The Electoral Integrity Project (EIP) has already demonstrated that electoral malpractices often occur on polling day (stuffed ballot boxes) and during the vote count (vote rigging). The question, now, is whether the level of integrity is compromised in any phase of a referendum. Until recently, research into and the analysis of plebiscites have been quite limited. Therefore, this study sheds the much needed light on how referendum integrity is affected under non-democratic conditions using Russia and Turkey as paradigmatic case studies.

  2.1 Authoritarian Deliberation

  Deliberation is a “mode of communication involving persuasion-based influence” (He and Warren 2011: 3). Essentially, increased use of deliberative practices can stabilise and strengthen an authoritarian regime or can serve as a tool for democratisation. Therefore, deliberation is not intrinsic to democracy. When combined with non-inclusive power, it becomes authoritarian deliberation. This hybrid form of deliberation does provide “space for people to discuss issues, and to engage in the give and take of reasons, to which decisions are then responsive” (He and Warren 2011: 12). However, authoritarian deliberation “is not in itself a decision-making procedure” as authoritarian control also implies control of the agenda, not only in terms of which “policies or issues are deliberated, but also the forums, levels of organization, timing, and duration” of the discussions (He and Warren 2011: 12). Thus, authoritarian regimes have a particularly complex relationship with the media. Essentially, Puddington (2017: 11) has highlighted how new authoritarian regimes target key media that are perceived as having the “greatest impact on public opinion” and, thereby, the greatest influence on people’s votes. By exerting “state or oligarchic control” over such media outlets, according to Puddington, those in power are able to control “information on certain political subjects and key sectors of the media, which are otherwise pluralistic, with high production values and entertaining content”.
  In He and Warren’s (2011) analysis, which focuses predominately on China, they identify ideal communication techniques based on the distribution of powers. The communication techniques include purely instrumental (i.e. the “use of communication to express preferences, without regard to the preferences of others”), strategic (i.e. the “use of communication to express preferences, with the aim of maximizing an agent’s preferences while taking into account the preferences of others”) and deliberative (i.e. the “use of communication to influence the preferences, positions, arguments, reasons, and justifications of others”). According to their study, deliberative authoritarianism is an uncommon type of rule “in which powers of decision are concentrated, but power holders enable communicative contexts that generate influence (responsiveness to claims, reasons, and empathy) among the participants” (He and Warren 2011: 12). Therefore, this strategy involves more than the solicitation of input associated with consultative authoritarianism, in which those in power “collect the preferences of those their decisions will affect and take those preferences into account… [in] their decision-making” (He and Warren 2011: 12). In contrast, in command authoritarianism (traditional), power holders “use communication solely to indicate the content of commands”.
  As seen in recent years, critical constitutional referendums have been held in (semi-) authoritarian systems including the Philippines (1977), Uganda (2000), Zimbabwe (2000), Comoros (2001, 2009), Congo-Brazzaville (2002), Madagascar (2007), Venezuela (2008), Crimea (2014) and Belarus (2022). In most of these cases, authoritarian leaders used these referendums to bolster their legitimacy. For example, in the referendum in Zimbabwe, the government implemented broad outreach programmes and deliberative democratic instruments that were characterised by strong propaganda in favour of the incumbent President Mugabe (see Kersting 2023).

• 3 Assessing the Integrity of Referendums

  Direct democratic instruments of participation, such as referendums (Qvortrup 2018; Setälä and Schiller 2009), are instruments of democracy that focus on the thematic topics and policies rather than on the election of personnel and candidates. There is also a wide range of different institutional settings, and this complicates the evaluation of direct democracy in referendums (Kaufmann et al. 2010;). In the analysis that follows, direct democracy and referendums will be used synonymously as umbrella terms. The typology of referendums is based on who initiates them and what topics they are allowed to deal with. They are either mandatory or initiated from a top-down or bottom-up perspective. Plebiscites are initiated from above, usually by the executive. Referendums can be binding or consultatively determined by constitutions or electoral law and usually involve specifically defined legal requirements such as quorums and time frames. Subjects may include a new constitution or particular constitutional or other legal issues. In the following, in addition to the umbrella terms mentioned earlier, we use the terms constitutional referendum and plebiscite when referring to the 2017 Turkish and 2020 Russian constitutional votes, respectively. Both included a broad constitutional review. The Russian national vote is considered a plebiscite because it did not follow the regulations for a constitutional referendum as defined by the electoral laws (see section 20 Russian federal law). In contrast, the Turkish referendum, which had to be held because it did not get a clear majority in Parliament, was indirectly initiated by the President.
  In this study, we have also used a referendum cycle model based on the electoral cycle framework (EC 2006; Norris 2014). Drawing on the work of the EIP (Norris et al. 2014), a theoretical framework was built to construct a new measurement instrument for use in expert surveys (see Kersting and Grömping 2021). The new empirical instrument, or the DDI-Index, was first tested in a pilot study on the Turkish constitutional referendum of 2017 and then on the Italian referendum of 2020 (see Kersting and Regalia 2023). Finally, the tool was applied to the Russian constitutional referendum of 2020.

  3.1 The Referendum Cycle and the DDI-Index

  The DDI-Index is based on the EIP. The EIP has been using its Perceptions of Electoral Integrity (PEI) Index to analyse all parliamentary and presidential elections globally since 2012 (Norris 2014; Norris and Grömping 2019). Both the PEI Index and the DDI-Index draw on experts’ assessments and a survey questionnaire. The newer DDI- Index was specifically built to assess the quality of referendums.
  The approach taken by the DDI- Index is based on the electoral cycle framework developed by the EIP (EC 2006; Norris 2013) and divides each referendum into four phases: pre-referendum, campaign, polling day and post-referendum (see Kersting and Grömping 2021, see Figure 1). The characteristics of a referendum are similar to those of an election in the polling day and the post-referendum phases. However, referendums are extremely different in the two pre-polling day phases. Referendums are initiated differently, involve different legal requirements and instruments and other actors become relevant. The DDI-Index assesses the quality of the referendum in relation to eleven dimensions. In the pre-referendum phase, the legal framework, referendum initiation process and voter registration are the significant dimensions. In the campaign phase, the campaign itself, the media coverage and campaign finance play an important role. On polling day, it is essential to evaluate the referendum procedures and the voting process. Finally, in the post-referendum phase, the vote count, post-referendum environment and electoral authorities are the key dimensions. Each dimension has a number of indicators and questions that must be considered (for further details and the list of questions used in the survey, see the Appendix in Kersting and Grömping 2021). While much media coverage has focused on irregularities such as ballot box stuffing, ‘ghost voting’ and vote count falsification, problems can also emerge through the formulation of the legal framework, the media’s campaign coverage or campaign financing. Thus, procedural integrity can be broken at any step of the referendum cycle.
  Specifically, regarding the pre-referendum phase, some factors are crucial to the procedural integrity of the referendum: the legality of the process, the existence of bias in favour of the status quo in how the referendum is conducted and the extent to which the executive uses its privileged position to dominate the agenda-setting process. The form of the referendum topic is also important. In this regard, “accurate, balanced, accessible, and relevant information about the referendum topic” and the “unambiguous wording of the referendum question itself” are all important safeguards that may help protect the integrity of the vote (Kersting and Grömping 2021: 7).
  As has been shown, combining deliberative democracy and direct democracy is becoming more common. Therefore, any use of deliberative democracy is also accounted for by the DDI- Index as an important feature of the pre-referendum phase.

  The Direct Democracy Integrity Cycle
  Source: Kersting and Grömping 2021.

  3.2 Methodological Instrument

  The discussion until now has focused on the procedural factors before, during and after polling day. In order to assess the integrity of referendums in a systematic, comparable and suitably fine-grained manner, the DDI-Index has been constructed by adopting and adapting the approach used by the EIP’s PEI Index. The DDI-Index’s survey questionnaire includes fifty-nine measures in the eleven sub-dimensions of referendum integrity covering the whole referendum cycle defined earlier. The eleven sequential sub-dimensions reflect the dimensions of the referendum cycle (see Figure 1). When completing the survey, experts were asked to indicate whether they agree with the statements about approximately sixty integrity items on a five-point Likert scale (strongly disagree – to strongly agree). Election expert is defined as a political scientist (or other social scientist in a related discipline) with demonstrated knowledge of the electoral process in a particular country (for example, through publications, membership of a relevant research group or network or university employment.) Survey invitations were sent to 230 experts on Turkey in mid-June 2017 and to 250 experts on Russia at the end of 2021, with two follow-up reminders one week apart each. We received forty-five responses on Turkey and fifty on Russia, representing a response rate of around 20% for Turkey and 18% for Russia. The results were controlled for inconsistencies and biased answers in the different pro and contra groups.

• 4 The ‘Constitutional Plebiscites’ in Russia, 2020 and in Turkey, 2017: Political Contexts

  In April 2017, 46 million Turkish people voted in the popular referendum, either in Turkey or from abroad, with a turnout of 85.3%. The yes-vote won only with a small margin: 51.4% voted in favour of the new constitution and 48.6% voted against it (Zirh et al. 2020; see Table 1). In Turkey itself, there was a split between urban and rural areas. In particular, the east of the country (with a predominantly Kurdish population) along with the bigger cities and richer tourist areas on the Mediterranean voted against the constitutional amendments. The rural population in central and northern Turkey, which is a stronghold of Erdogan’s ruling party AKP voted in favour of the constitutional amendments.
  As a result of the binding constitutional referendum, the 2017 constitutional reform introduced a strong presidential system (see Makovski 2017. The approved amendments to the constitution abolished the office of Prime Minister and strengthened the executive role of the President within the Turkish political system (Öney 2018). This change from a parliamentary to a presidential system included considerable expansion of the presidential rights. For example, the Turkish president can now appoint the Supreme Board of Judges and Prosecutors (Hâkimler ve Savcılar Yüksek Kurulu, HSYK). While a small number of provisions in the referendum package strengthened the Parliament (the number of seats in Parliament was increased from 550 to 600), in general, power was centralised within the person of the President and checks and balances were weakened (Yilmaz 2020).
  Turkey became an electoral democracy with many characteristics of modern authoritarian regimes (Baghdady 2020). As in the early 2010s, the 2017 referendum did not follow the constitutional review process (see Petersen and Yanasmayan 2020). However, a referendum did not eventuate in the 2010s, although a relatively broad constitutional debate did. In contrast, the 2017 referendum was characterised by a lack of freedom of information and repression. Moreover, the referendum was implemented during a state of emergency with the aim of strengthening presidential powers.

  Table 1 Results of the 2017 Turkish Constitutional Referendum and the Russian Nationwide Vote 2020

  Turkey, 2017				Russia, 2020
  Response	Votes	%		Response	Votes	%
  Yes	25.2 Mio.	51.4		Yes	57.7 Mio	78.6
  No	23.8 Mio.	48.2		No	15.7 Mio	21.4
  Registered voters/turnout	58.2 Mio.	85.4		Registered voters/turnout	109.2 Mio	67.9

  Source: Central Election Commission Turkey; Levada Center 2020.

  Since, the collapse of the USSR, the former communist states have carried out numerous plebiscites, most of them taking place in modern authoritarian states such as Azerbaijan, Belarus, Tajikistan, Turkmenistan, Kazakhstan and Uzbekistan. In most cases, the main purpose of the vote was to obtain popular endorsement for constitutional changes designed to extend the political power and/or abolish the constitutional terms that limit the power of the relevant political incumbent (Hill and White 2018). After the national referendum that approved the new constitution in 1993, Russia did not hold another nationwide referendum or plebiscite until the top-down initiative of March 2020. This Russian vote represents an example within a weak party system, with consolidated electoral authoritarianism and a high level of political personalism (Golosov 2011; Hale 2005). Thus, on 1 July 2020, the Russian people voted to approve the constitutional reform proposed by President Putin in January 2020. According to figures provided by the Central Electoral Commission of Russia, 78.6% of voters voted in favour, with a turnout of 67.9% of the eligible voting population (see Table 1). This constitutional revision opened the way for Vladimir Putin to accept another term in office in 2024. The 2020 reform was presented by the President as intending to ensure stability, security and prosperity for the country. However, in actuality, the State Duma approved the changes weeks before the national vote and there was no legal requirement for a referendum (Hutcheson and McAllister 2021: 357).
  Indeed, the 1993 Russian Constitution distinguished between two mechanisms. The mechanism provided for in Article 136 concerns constitutional revision, i.e. amendments that do not affect the constitutional foundations but require Parliament to adopt a constitutional law. The second mechanism, provided for in Article 135, allows a constituent assembly to amend the core of the constitution, with the new constitution to be adopted either by the constituent assembly or by popular vote. The constitutional law that revised the constitution on 14 March 2020 was expressly presented as falling under Article 136. Therefore, by proposing this so-called ‘nationwide vote’, national vote or ‘all-Russian vote’, the Kremlin was purely seeking greater popular legitimacy through a non-binding referendum.
  The constitutional vote had an ambiguous legal status and legal experts claimed that it was unacceptable to pass constitutional amendments in a “package” not united by a common subject.²x www.bbc.com/russian/features-51402865. Regarding the broader political agenda, the referendum came when Putin’s popularity rating reached a historically low record of 59% in March, as the country struggled amidst the Covid-19 crisis.³x Indeed, during the Covid-19 pandemic, Putin preferred to leave unpopular decisions to others and make regional governors managers, a tactic that was also criticised. See “Putin’s Approval Rating Drops to Historic Low: Poll”, in the Moscow Times. www.themoscowtimes.com/2020/05/06/putins-approval-rating-drops-to-historic-low-poll-a70199 (accessed 10 March 2022). However, criticism of the voting process was widespread. In a statement issued as the polls closed, the independent election watchdog, Golos, highlighted serious shortcomings in the way the campaign and referendum were conducted including accusations of ballot stuffing, voter coercion and double voting.⁴x www.golosinfo.org/articles/144477?fbclid=IwAR3cF_Yeej5IuA4pkT7vT6siDnTlRunaW55lJESd9HXouoqp0r1-yzsyQJA. Immediately thereafter, the European Union also called for an investigation into the alleged irregularities (Radio Free Europe, 2020).
  The amendments focused on creating new powers enabling the State Duma to approve the appointment of a Prime Minister, even against the will of the President, although the President retained the power to remove the Prime Minister from office. Furthermore, numerous social welfare regulations were included in the constitution including the regular indexation of pensions, a minimum wage above subsistence level. Besides this new social welfare programmes, nationalistic sentiments as well as homophobic attitudes were obvious. Amendments were restricting different homosexual rights and they were restricting marriage to being between one man and one woman. The reforms also included other more symbolic nationalistic elements. Finally, the power of the incumbent President was strengthened. With the new constitution and its amendments, any presidential terms held before the new constitution came into effect are not counted. Thus, the new constitution allows Putin to stay in power (so long as he is successfully re-elected) until 2035. Moreover, it makes it more complicated for challengers to run against him because it restricts which candidates are eligible to run for President, Prime Minister and judgeships by disallowing candidates with formerly non-Russian foreign citizenship and requiring a minimum period of residency of 25 years. This last point is specifically designed to exclude key figures in the Russian opposition who have lived abroad and prevent them from running for the presidency.
  State authorities unfolded a large-scale media campaign involving celebrities which was designed to promote participation. Achieving high turnout rates (more than 50%) and high numbers of votes in favour of the amendments (at least 70%) were the priorities of the plebiscite organisers. According to the independent pollsters, voter turnout was high – 68%. About 74 million votes were cast by the 109 million registered Russian voters. Approximately 79% of the votes cast were yes-votes and 21% were no-votes. Invalid and blank votes constituted less than 1% of all votes (data from Levada Center 2020). Despite the relatively high turnout and triumphant results for the President and his supporters, how the vote was conducted prompted numerous allegations of violations and fraud.⁵x See “EU Calls for Investigation into Irregularities in ‘Triumphant’ Vote for Putin”, Radio Free Europe, www.rferl.org/a/eu-calls-for-investigation-into-irregularities-in-triumphant-vote-for-pu-tin/30702503.html (accessed 10 March 2022).

• 5 Referendum Integrity: Turkey and Russia

  In the following section, we present some of the key results of the expert survey on integrity. The analysis focuses on the items where expert opinions showed clear agreement as well as on any particularly controversial scores provided by our experts. The media coverage and campaign finance of the two case studies are analysed and compared in more detail in the following text.
  Please note, we do not focus on the political implications of the results as the concept of referendum integrity is process focused and, therefore, excludes the outcomes of any given referendum per se. It is logically possible that both an integrous referendum and a referendum without integrity could produce any given result (see Kersting and Grömping 2021).

  5.1 Turkey 2017

  For the Turkish referendum of 2017, the expert survey suggests that the overall integrity of the referendum was relatively low. Our analysis of the eleven referendum stages compared the pre-referendum phase to the campaign phase to polling day and the post-referendum phase (see Figure 2). The scores for the eleven stages in the spiderweb graph are averages based on the percentage of scores (from 1 to 5) given to over thirty statements in the surveys. The spiderweb graph shows the overall score calculated based on the variables measured for each dimension. A high level of integrity would be indicated by a score of 100% and the lower the score, the lower the integrity.

  Integrity Scores for the 11 Stages of the Referendum Cycle (Turkish Constitutional Referendum 2017)
  Note: DDI-Index Turkey 2017, a higher percentage denotes higher referendum integrity (‘strongly agree’ or ‘agree’ for positively worded statements). N = 45 in Turkey.

  All phases show a low level of integrity (below 40%). However, the scores are slightly higher in the voter registration phase and for the procedures, the voting process on polling day and the vote count. However, the adherence to referendum law, the campaign, the media coverage and campaign financing were all highly problematic in Turkey.

  Referendum Integrity Turkey 2017: Campaign Finance
  Note: Scored on a 5-point Likert scale ((strongly) disagree – (strongly) agree). The middle category (‘neither agree nor disagree’) is not shown but can be deduced.

  In the campaign financing stage, the indicators show that nearly all experts agreed that state resources were inappropriately used for campaigning (see Figure 3). In this regard, the campaign financing did not provide an even playing field and give the opposition the opportunity to secure a comparable level of financial support. This became even more evident when the experts made it clear that during the referendum, both sides did not publish transparent financial records. Furthermore, both sides did not have the same access to private sponsors and public finance. Therefore, more than 90% of the experts identify a lack of integrity in all these indicators.
  In Turkey, in 2017, the media coverage was affected by the tense political situation (see Figure 4). The public media were essentially muzzled by the government, which controlled important critical TV media outlets and is known to have threatened independent journalists. Approximately 90% of the experts claimed that newspapers did not provide balanced and fair information about the referendum.

  Media Coverage
  Note: Direct Democracy Integrity Index Turkey 2017: 5-point Likert scale ((strongly) disagree – (strongly) agree). The middle category (‘neither agree nor disagree’) is not shown but can be deduced.

  The reason for this lack of unbiased, informative coverage was that neither side had fair access to broadcasting and advertisement avenues. A clear majority of the experts (over 80%) stated that the media reporting about the campaign was often based on fake news and actually inaccurate. Nearly two-thirds indicated that fake news caused a great deal of confusion and this notion was only dismissed by ten per cent of the experts. The experts revealed that the TV news, in particular, strongly supported the status quo and the ruling government.

  5.2 Russia 2020

  In Russia, the overall evaluation again showed a low level of integrity (see Figure 5). While around 40% of the experts were rather satisfied with the voting procedures, the DDI expert survey presented a relatively critical picture of the ‘referendum legislation’ dimension. Our experts criticised the legal framework of the top-down plebiscite, which included the restriction of citizens’ rights (e.g. same-sex marriage) while favouring the status quo and the strategies of the incumbent government.
  In relation to the initiation of the referendum, experts were much more critical. The Kremlin used the civil society outreach groups and the parliamentary process to strengthen its powers. It was not regarded as entirely legitimate or fair because the government strongly dominated the agenda-setting. The majority of the experts also claimed that representative democracy was weakened by the reforms. Although citizens were partly included in some regions in symbolic outreach programmes, the overall lack of citizen engagement was criticised.

  Integrity Scores for the 11 Stages of the Referendum Cycle (Russian Constitutional Plebiscite 2020)
  Note: DDI-Index Russia 2021 Higher percentage denotes higher referendum integrity (‘strongly agree’ or ‘agree’ for positively worded statements). N = 50 in Russia).

  The government was also criticised because the Kremlin strategically used social welfare programmes, nationalistic sentiments as well as homophobic attitudes in the citizenry to get support for political reforms and the constitutional review. Less than half of the experts criticised the integrity of the Russian referendum 2020.
  In regard to voter registration, these figures are even lower. On this point, only around 20% of the experts claimed that the referendum was fair. The voter roll was not accurate and there were problems with voter registration in the pre-referendum phase that had more to do with the exclusion of important social subgroups and less with minimising the registration of ineligible voters (ghost voters).
  In the campaign phase, the campaign itself focused on the package referendum. This means that citizens had to decide on too many complicated issues with different facets in one vote. On the ballot for the final referendum ‘Do you agree on the new constitution’, there was only a ‘yes’ and ‘no’ option and no option to provide a more detailed decision. The regional outreach programmes were supposed to bring higher legitimacy to the process; however, the Kremlin left itself open to criticism here. Thus, only a small number of experts evaluated this dimension positively. Most criticised the lack of information and the ambiguous wording of the question.
  In the third phase, the referendum procedures on polling day received relatively higher scores from experts, although still more than half of them criticised the Direct Democracy Integrity of this dimension. According to them, the number of polling stations and voter information provided were adequate and it was easy to cast a vote.
  The pre-referendum phase was the most significant and a number of malpractices occurred in Russia during this phase, especially in the media coverage and campaign finance stages. The campaign finance was highly criticised by a clear two-thirds majority (see Figure 6). Only a very small number of experts did not criticise this dimension for a lack of integrity. The transfer and publishing of financial accounts, as well as the equitable access to public political subsidies and political donations were highly biased.
  Campaign finance was supported by a predominantly pro-Kremlin media. State resources were improperly used for extensive campaigning. Meanwhile, the Kremlin had more and better access to free radio and television broadcast time. Access to political communication was also structurally unequal as almost all Russian media are directly or indirectly controlled by the state (through companies like Gazprom). Indeed, it has been alleged that “the first clear indicator of Putin’s authoritarian bent was his aggressive move to eliminate independent ownership of Russia’s major television stations” (Puddington 2017: 11). Thus, it is no surprise that the campaign media coverage in Russia was heavily criticised as having low integrity (see Figure 7).

  Russia 2020 Referendum Integrity: Campaign Finance
  Direct Democracy Integrity Index Russia 2021: 5 Likert scale ((strongly) disagree – (strongly) agree). The middle category (‘neither agree nor disagree’) is not shown but can be deduced.

  The lack of media coverage can also explain the low evaluation the experts gave for referendum integrity. Not only did the Russian media focus strongly on the Kremlin’s position, it was readily observed that the campaigns of particular groups were hindered and that politicians were offering political patronage to others. It seemed that even independent journalists were not particularly focused on the fair coverage of the pros and cons of the referendum. The pro-government media skewed the subject of the referendum by presenting amendments that were more likely to win support, for example, relating to culture and history, while the independent media tended to highlight the reforms that sought to strengthen the government’s powers.
  After the opinion poll in early 2020 indicated that the referendum may be defeated, the government launched a broad campaign incorporating different forms of deliberative participatory instruments for discussing the referendum issues (Hutcheson and McAllister 2022: 358). This propaganda strongly focused on the benefits of the new constitution and sought to mobilise the broader population. In contrast, the opposition did not have the opportunity to access the public media and deliberative instruments that were used to disseminate the government’s propaganda as opposition leaders were side-lined and repressed. The relatively clear result reflected the fact that the opposition struggled to show their positions in this political environment and the unequal means different parties had available to reach the electorate, notably, due to the government-defined agendas and official government control of the media. Thus, the media conveyed an open letter signed by people from civil society who deplored the referendum as an “anti-constitutional coup”⁶x Dozhd, 16 March 2020.
  The expert survey also indicated that 70% of respondents considered the questions listed on the ballot ambiguous. As previously indicated, citizens had to vote on a very broad constitutional amendment package that included presidential competencies in addition to homophobic regulations and social welfare policies. The questions relating to whether there was sufficient time for deliberation and sufficient information about the topics were evaluated negatively. The majority of respondents would have liked better information and more time to debate the issues. Despite the various outreach programmes, this process was criticised for not successfully providing broad and open debates. Instead, this dimension was strongly criticised by 72% of the experts who indicated that they believed some groups were excluded from campaigning and restricted holding campaign rallies. About 43% claim that politicians and the ruling party offered patronage to their supporters and 19% disagree.

  Russia 2020 Referendum Integrity: Campaign Media Coverage
  Direct Democracy Integrity Index Russia 2021: 5 Likert scale ((strongly) disagree – (strongly) agree). The middle category (‘neither agree nor disagree’) is not shown but can be deduced.

• 6 Conclusions

  Our main research question focused on the integrity of constitutional referendums in modern authoritarian systems. Integrity is considered a normative standard that incorporates free and fair institutions, including elections, and focuses on the protection of minority rights. In recent years, deliberative instruments have been included in referendum processes and direct and deliberative instruments have been combined. Our research also focused on how this deliberation is implemented in constitutional referendums in modern authoritarian systems. To analyse these two questions, we applied the new DDI-Index through expert surveys.
  Russia and Turkey, our two case studies, are modern authoritarian regimes that are thought to have used contentious Napoleonic plebiscitary referendums to strengthen their presidential systems. Their use of symbolic direct democracy is typical of ‘sultanistic’ regimes with strong autocratic presidential and weak party systems. The DDI expert surveys for Russia and Turkey showed a variation in experts’ responses regarding the different stages of the referendum cycle. These pilot case studies give us valuable insight into where the experts interviewed in the two countries differ significantly in their assessment and where their assessments converge. Thus, the study has highlighted the importance of the pre-referendum stage in two modern autocracies. Results clearly show that there is an asymmetry of means to convince the electorate, notably due to the context of government-defined agendas and official government control of the media. A number of malpractices seem to have occurred during the pre-referendum stage in our case studies, in particular, in the form of non-compliance with the legal framework of the respective country. As such, the Russian and Turkish referendums are fertile pilot cases for our research instrument, especially in relation to the campaign and media coverage stages.
  The Turkish referendum in 2017 was deeply criticised by electoral observer groups for having an uneven playing field in the pre-referendum phase. In Turkey, after the attempted coup, the country was put into a state of emergency and a phase of repression and violation of civil rights and liberties ensued. During this time, the extreme media dominance of Erdogan’s ruling party AKP became evident and the constitutional debate did not trigger a broad constitutional deliberation (as it did in Turkey from 2011 to 2013). Although this constitutional review process focused on an important topic, repression led to greater disengagement and apathy. In Turkey, the initiation phase and the legal requirements were not highly onerous or problematic (only a mandatory referendum after the vote). Yet, in the context of such political repression, the unfair electoral laws, biased campaign media and unequal access to financing were highly problematic. The post-referendum environment was also still characterised by an atmosphere of fear that led to further disengagement and apathy in parts of the population.
  In Russia, in 2020, being required to vote on a package deal that presented measures strengthening presidential powers alongside nationalist, right-wing, anti-minority, homophobic, populist reforms and social welfare measures considerably diminished the integrity of the issue being voted on. However, other areas were also severely lacking in integrity.
  The Russian initiation phase (top-down by the president) did not comply with Russian legislation for direct democracy and, because of this, the electoral laws and the registration were regarded as problematic by most experts and the plebiscite was referred to as a Russian national vote. In the Russian referendum of 2020, which was affected by the Covid-19 pandemic in the month following its announcement, it was not clear whether the plebiscite would pass. In fact, opinion polls predicted a loss. From this point on, an enormous propaganda machine sought to mobilise the population in favour of the incumbent president: this included the launch of new forms of participation and outreach. As a result of this mobilisation, the experts identified the campaign phase as the most problematic and of the lowest integrity. The campaign media and public broadcasting were viewed as highly biased and not neutral mobilisation strategy as accompanied by repression and a selective demobilisation of the opposition. This created a lack of integrity and prompted some oppositional and minority groups to boycott the vote.
  Both the Russian and Turkish plebiscites were supposed to demonstrate symbolic legitimacy and public endorsement of their president’s power, that is, of Vladimir Putin and Recep Tayyip Erdogan, respectively. In both countries, shortcomings on the polling days and during the post-referendum phases have been identified by independent watchdogs. These include voter coercion, violence against journalists, ballot stuffing and multiple voting.
  The political contexts of both countries are largely biased because of the dominance of state-controlled media and the strong coercion of voters. Both countries combine mobilisation by propaganda and demobilisation by repression. These regimes restrict human rights as well as political rights and civil liberties. In 2017, Turkey was more oriented towards a mode of repressive communication leading to demobilisation, apathy and fear. In contrast, in 2020, Russia differed from traditional authoritarianism in that it behaved, at least symbolically, as a deliberative authoritarian state, which is defined as a type of rule in which decision-making powers are concentrated but which permits communicative contexts that generate limited influence. However, the state’s strong media campaign and participatory outreach programmes were characterised by manipulation and propaganda, not by open and fair deliberation. Thus, Russia’s deliberative authoritarianism was only symbolic, and the processes used were strongly founded on propaganda and mobilisation. The lack of democratic integrity was strongly associated with the repression of oppositional minority rights.
  Authoritarian systems often misuse referendums to strengthen their legitimacy and discipline opposition parties and civil society groups. Historically, referendums and elections authoritarian systems have lacked integrity as they were manipulated by repression and vote rigging (stuffed ballot boxes, fraudulent vote counting). However, in Russia in 2020, the integrity of the initiation and campaigning phases had become more important than the issues on the polling day.
  The new research instrument, the DDI-Index, has provided initial insight into the integrity of and how referendums and deliberative instruments are being used in modern authoritarian political systems. Further qualitative and quantitative research is necessary to evaluate the deliberative instruments used in referendums in more authoritarian systems. However, it will be difficult to secure research support and free and fair access for the necessary investigations. Nevertheless, future studies and expert surveys on direct democratic integrity should include case studies of additional authoritarian, hybrid and democratic systems.

References

• Altman, David (2010). Direct democracy worldwide. Cambridge; Cambridge University Press.

• Baghdady, Gabriela (2020). Turkey’s electoral authoritarianism. In: Journal of International Relations, 4 (2020).

• Bowler, Shaun and Donovan, Todd (2002). Democracy, institutions and attitudes about citizen influence on government. In: British Journal of Political Science, 32, 371-339.

• Collin, Katherine (2014). Do referendums resolve or perpetuate contention? In: P. Norris, R. W. Frank and F. Martinez i Coma (Eds.), Contentious elections (pp. 111-130). New York: Routledge.

• Coppedge, Michael, Gerring, John, Knutsen, Carl Henrik, Lindberg, Staffan I., Skaaning, Svend-Erik, Teorell, Jan and Ziblatt, Daniel (2018). V-Dem, Dataset v8.

• Golosov, Grigorii V. (2011). The regional roots of electoral authoritarianism in Russia. In: Europe-Asia Studies, 63(4), 623-639.

• Hale, Hale E. (2005). Why not parties in Russia? Democracy, federalism, and the state. Cambridge: Cambridge University Press.

• He, Baogang and Warren, Mark E. (2011). Authoritarian deliberation: The deliberative turn in Chinese political development. In: Perspectives on Politics, 2(9) 269-289.

• Hill, R. J. and White, S. (2018). Russia, the Former Soviet Union and Eastern Europe. In: Qvortrup, Matt Referendums Around the World (pp. 113-145). Cham: Palgrave Macmillan.

• Hutcheson, Derek S. and McAllister, Ian (2021). Consolidating the Putin regime: The 2020 referendum on Russia’s constitutional amendments. In: Russian Politics, 6(3), 355-376.

• Kaufmann, Bruno, Büchi, Rolf, & Braun, Nina (2010). Guidebook to direct democracy in Switzerland and beyond. Marburg: Initiative & Referendum Institute Europe.

• Kersting, Norbert (2023). Constitutional referendums and deliberation: Direct democratic integrity in Russia, Italy, and Turkey. In M. Reuchamps and Y. Welp (Eds.), Deliberative Constitution-making: Opportunities and Challenges (pp. 180-198). Abingdon & New York: Routledge.

• Kersting, Norbert and Grömping, Max (2021). Direct democracy integrity and the 2017 constitutional referendum in Turkey – A new research instrument. In: European Political Science, 21, 216-236.

• Kersting, Norbert and Regalia, Marta (2023). Direct democracy and its integrity. The Italian 2020 and the Turkey 2017 Constitutional Referendums. In: Zeitschrift für Vergleichende Politikwissenschaft, 16, 527-544.

• Levitsky, Steven and Way, Lucan (2002). The rise of competitive authoritarianism. In: Journal of Democracy, 13(2), 51-65.

• Mendelsohn, Matthew and Cutler, Fred (2000). The effect of referendums on democratic citizens: Information, politicization, efficacy and tolerance. In: British Journal of Political Science, 30, 685-698.

• Mommsen, Margareta (2019). Dossier Russland. Bundeszentrale für Politische Bildung, pp. 30-67.

• Monje, Sergio Velasco and de la Cruz, Alberto Penadés (2019). The role of direct democracy under autocratic rule. Paper presented at the XIV Congreso AECPA.

• Norris, Pippa (2014). Why electoral integrity matters. New York: Cambridge University Press.

• Norris, Pippa, Frank, Richard W. and Martínez i Coma, Ferran (2014) Measuring electoral integrity around the world: a new dataset. In: PS: Political Science and Politics, 47(4), 789-798.

• Norris, Pippa and Max Grömping (2019). Perceptions of electoral integrity (PEI-7.0). V2. Harvard dataverse.

• Öney, Berna (2018). Constitutional debate in Turkey: From dominant party hierarchical strategy to single-party hegemony, quantitative text analyses. In: Mediterranean Quarterly, 29(3), 33-54.

• Petersen, Feix and Yanasmayan, Zeyneo (Eds.) (2020). The failure of popular constitution making in Turkey. Cambridge: Cambridge University Press.

• Puddington, Arch (2017). Breaking down democracy: Goals, strategies, and methods of modern authoritarians. Washington, DC: Freedom House.

• Qvortrup, Matt (Ed.) (2018). Referendums around the world. London: Palgrave.

• Setälä, Maja & Schiller, Theo (2009). Referendums and representative democracy: responsiveness, accountability and deliberation. London: Routledge.

• Schedler, Andreas (Ed.) 2006: Electoral authoritarianism. Boulder: Lynne Rienner

• Welp, Yanina and Soto, Francisco (2020). Beyond fashion and smokescreens: Citizens’ deliberation of constitutional amendments. Cost Constdelib Working Paper. https://constdelib.com/wp-content/uploads/2020/06/WP7-2020-CA17135.pdf.

• Yilmaz, Zafer (2020). Erdoğan’s presidential regime and strategic legalism: Turkish democracy in the twilight zone. In: Southeast European and Black Sea Studies, 20(2), 265-287.

• Zirh, Besim and Gül, Volkan (2020). Country report Turkey. COST Action “Constitution-making and deliberative democracy”. Retrieved from https://constdelib.com/countries/reports (last accessed 10 July 2023).

Noten

• 1 In the following text, we use the terms ‘modern authoritarian regimes’ and ‘modern autocracies’ interchangeably.

• 2 www.bbc.com/russian/features-51402865.

• 3 Indeed, during the Covid-19 pandemic, Putin preferred to leave unpopular decisions to others and make regional governors managers, a tactic that was also criticised. See “Putin’s Approval Rating Drops to Historic Low: Poll”, in the Moscow Times. www.themoscowtimes.com/2020/05/06/putins-approval-rating-drops-to-historic-low-poll-a70199 (accessed 10 March 2022).

• 4 www.golosinfo.org/articles/144477?fbclid=IwAR3cF_Yeej5IuA4pkT7vT6siDnTlRunaW55lJESd9HXouoqp0r1-yzsyQJA.

• 5 See “EU Calls for Investigation into Irregularities in ‘Triumphant’ Vote for Putin”, Radio Free Europe, www.rferl.org/a/eu-calls-for-investigation-into-irregularities-in-triumphant-vote-for-pu-tin/30702503.html (accessed 10 March 2022).

• 6 Dozhd, 16 March 2020.