The Istanbul Convention, Sofagate, and Turkey’s EU Candidacy: A Gender-Centric Convergence Analysis

Introduction

I am the first woman to be President of the European Commission. I am the President of the European Commission. And this is how I expected to be treated when visiting Turkey two weeks ago, like a Commission President, but I was not.

I cannot find any justification for the way I was treated in the European Treaties. So, I have to conclude, it happened because I am a woman. Would this have happened if I had worn a suit and a tie?

This is how Ursula von der Leyen chose to open her speech at the European Parliament Plenary meeting on 26 April 2021.^[1] She was referring to Sofagate, which is how the diplomatic protocol incident that occurred a few days earlier, on 7 April, during her official visit to Ankara with president of the European Council Charles Michel to meet Turkish president Recep Tayyip Erdoğan has come to be known.^[2] During the meeting, both the president of the Turkish Republic and of the European Council sat on chairs, while the president of the European Commission Ursula von der Leyen was offered to take a seat on a sofa in the same room. The move was construed to imply that von der Leyen was of lower rank and was attributed to her being a woman, as she herself stated during the European Parliament Plenary meeting mentioned above. The incident also sparked a discussion on the EU’s rules of diplomatic protocol (Nováky 2021). Irrespective of whether the incident was intentional or not, Sofagate is just the latest in a series of episodes which demonstrate the problems of EU–Turkish relations and Turkey’s protracted candidacy (see, for example, Dursun-Özkanca 2022; Müftüler-Baç 2017; Phinnemore and Içener 2016; Saatçioğlu et al. 2019). Indeed, the aim of the official visit by the two EU heads to Ankara was to try to place their bilateral relations on a new footing, with the accession negotiations having been described as being in a comatose state (Tocci 2016).

Around a month earlier, on 20 March 2021, Turkish president Erdoğan issued a presidential decree annulling Turkey’s ratification of the Istanbul Convention on preventing and combating violence against women, thus making it the first and only country to have withdrawn from the Convention.^[3] The decision was met with dismay and disappointment both by EU officials and the wider international community. For instance, in a statement issued on 21 March 2021, US president Joe Biden described the decision as deeply disappointing.^[4] Condemnation also came from the EU’s high representative and vice-president Josep Borrell who reacted by saying: “We therefore cannot but urge Turkey to reverse its decision. We hope that Turkey will soon join again the European Union in defending the rights of women and girls, a fundamental element of human rights, peace, security, and equality in the 21st century.”^[5] In a similar vein, in a resolution, the European Parliament

strongly condemns the Turkish Government’s decision to withdraw from the Istanbul Convention, bringing Turkey further away from EU and international standards […] putting into serious question its commitments to prevent violence against women and promote women’s rights, […] which is a clear sign of the serious deterioration of human rights in the country […].^[6]

Gender issues feature prominently on the EU’s agenda and in its policies (see, for example, Abels et al. 2021; Ahrens and Woodward 2021; Cullen 2021; Crowley and Sansonetti 2019). For instance, the Gender Equality Strategy 2020–2025 combating gender-based violence as well as protecting and supporting victims incorporates one of the five priority areas of the EU’s gender equality policy, which aims to achieve significant progress toward a gender-equal Europe by 2025.^[7] Moreover, one of the EU’s values is that “the principle of equality between women and men underpins all European policies and is the basis for European integration”.^[8] The Council of Europe Convention on preventing and combating violence against women and domestic violence was signed on 11 May 2011.^[9] However, the Council of Europe is not an EU institution. Thus, the Convention is not part of the acquis communautaire. Consequently, the Convention is not legally binding for EU member states or for candidate states such as Turkey. Currently, the Convention has been signed by 45 countries as well as the EU as a whole.^[10]

The two incidents described above are not necessarily causally related. Yet, even on a purely symbolic level, they highlight the significant and growing differences between the EU and Turkey. In a broader context, they attest to Turkey’s democratic backsliding and de-Europeanization (see, for example, Alpan 2021; Köker 2019; Cilliler 2021; Yilmaz 2016). This process involves a gradual move away from European norms, values, and policy demands in various areas, including gender issues and policies, as well as a broader distancing from “the West” (Bodur Ün and Arikan 2022; Bardakçı 2021, 2022). The process has been attributed to the rise of the Justice and Development Party (Adalet ve Kalkınma Partisi, AKP) in domestic politics and the increasingly authoritarian rule of president Erdoğan (see, for example, Akçay 2021; Arslanalp and Erkmen 2020; Waldman and Caliskan 2017; Yilmaz and Bashirov 2018).

Turkey is one of the eight countries that have been granted candidate status thus far.^[11] In fact, given that the relationship between the EU and Turkey dates to the 1963 Ankara Association Agreement, it is the country with the longest association and candidacy history (see, for example, Tocci 2016; Kollias 2023; Saatçioğlu et al. 2019; Bardakçı 2008). Turkey was granted candidate country status in 1999. Accession negotiations commenced in 2005. However, they have never been smooth and have progressed at a snail’s pace. Following the 26 June 2018 decision of the EU Council, accession talks came to a standstill and are now effectively frozen, but diplomatic efforts to restart them seem to be currently underway. As to be expected, Turkey’s protracted candidacy has attracted considerable attention in the relevant literature (see, for example, Phinnemore and Içener 2016; Müftüler-Baç and Cicek 2017; Dursun-Özkanca 2022; Müftüler-Baç 2019). The existing and growing literature on Turkey’s EU candidacy focuses on various aspects that seem to be impeding and hindering the accession negotiations and their progress. However, a comprehensive review of this expanding literature, is well beyond our scope here. Within the broader convergence literature examining whether candidate countries are converging with the EU, our study applies a gender-centric convergence analysis to the case of Turkey. Evidently, gender equality issues are not the epicenter of the accession process for any candidate country. However, in a broader sense, they are an integral part of the Copenhagen accession criteria and the EU’s norms and standards that candidate countries have to meet (Oğuzlu 2012; Dudley 2020; Kollias and Messis 2022; Lavenex and Schimmelfennig 2009).^[12] Indeed, as has been stressed “the state of women’s rights and gender equality in Turkey is a benchmark for the development of Turkey as a liberal democracy compatible with EU standards” (Glüpker-Kesebir 2016, 517). Moreover, as already noted, gender-related issues constitute an important policy area for the EU. This is in fact stressed by the president of the European Commission Ursula von der Leyen in the Gender Equality Strategy 2020–2025 factsheet, which states: “Gender equality is a core principle of the European Union […].”^[13] To the best of our knowledge, the existing literature on the convergence of the candidate countries has not addressed the gender equality issue in the context of Turkey’s EU candidacy. To explore the issue, we use five gender-focused indices compiled by the Varieties of Democracy (V-Dem) project.^[14]

In the following section, we offer a synopsis of the convergence literature and a comparative descriptive presentation of the five gender equality indices used in our study. We present and discuss the empirical methodology adopted to address the issue at hand along with our findings in section three, while section four concludes our study.

A Primer on Convergence and the Gender Indices

Convergence theory is closely related to theories of social progress and modernization. In broad terms, it may be defined as a gradual, long-term process during which differences between countries progressively become much less discernible (Ferdosi 2020). In the existing literature, a group of intertwined factors are identified as convergence drivers (see, for example, Dolowitz and Marsh 2000; Drezner 2005; Bennett 1991). Of these factors, the processes of regional integration and globalization are widely viewed as important promoters of convergence. In a nutshell, these processes increase the interaction between countries, societies, and cultures. This, in turn, facilitates and expedites the process of convergence in a wide variety of different spheres. Moreover, it is generally accepted that the process of integrating countries into supranational entities such as the EU acts as a facilitator and promoter of convergence (Kollias and Messis 2022; Surubaru 2021). Indeed, the EU constitutes an institutional framework within which the process of convergence is actively pursued.^[15] The broader European framework encourages countries to gradually move toward conditions of similarity, not only in terms of standards of living but also in terms of social and cultural norms, societal organization, supervision and regulation, governance, and institutional functioning. Thus, the gender-related issues and policies which we focus on in the present study should be viewed as part of the broader process of converging with EU norms, in this case by a candidate country during the candidacy phase.

One of the main attractions of EU membership has always been the opportunity it offers to candidate states to converge not only in terms of income levels and living standards but also in terms of social conditions, the functioning of democratic institutions, and polity. The appealing prospect of EU accession and membership is associated with the related accession conditionalities that candidate countries have to fulfill (Dudley 2020; Lavenex 2008). These conditionalities have generally proved to be a successful means of inducing countries aspiring to develop closer ties with the EU, and/or to apply for membership, to adapt and implement the acquis communautaire. That is, to converge. Indeed, convergence in various spheres has long been a pivotal policy that the EU has actively pursued – originally for its members. It subsequently also extended it to candidate and potential candidate countries, such as Turkey. The appealing prospect of future EU membership is the principal driver motivating states aspiring to join the EU to adopt and implement policies that meet the various accession criteria. Hence the EU’s transformative power (Lavenex 2008; Lavenex and Schimmelfennig 2009; Öner 2014). Convergence during the partnership or pre-accession phase facilitates and smoothens the process of integration once membership is eventually achieved.

As noted earlier, the various accession conditionalities include the Copenhagen criteria. Moreover, given the recent democratic backsliding in some Central and East European member states, the Commission has also proposed the imposition of conditionalities when it comes to flows of EU funds in cases where there are systematic rule of law infringements (Blauberger and van Hüllen 2021). Such conditionalities incentivize countries that are aspiring to membership to establish stable democratic institutions, promote effective democracy in their domestic political processes, encourage the rule of law, and make improvements in human rights practices, including respect for and protection of minorities or, for that matter, gender equality, which is our focus here. Nonetheless, although in theory such conditionalities are designed to serve as an incentive for EU aspirants to establish stable democratic institutions, Dudley (2020) finds that in some cases they fail to fulfill the conditionalities and that often the EU allows progress in the accession process without full compliance with the EU’s liberal democratic standards. For example, the promotion of gender equality and equal opportunities for women are specifically included in the EU’s European Neighborhood Policy (Debusscher 2012). Gender equality, or indeed the absence of it, can in fact be seen as a yardstick for adherence to the principles of liberal democracy and respect of human rights that candidate countries must satisfy before accession (Glüpker-Kesebir 2016). For instance, Chiva (2009) focuses on the impact of the EU’s gender equality-related conditionality during the pre-accession period in the cases of Bulgaria and Romania, both of which joined in 2007.

Gender issues in Turkey have also attracted attention with a particular focus on the era of the AKP’s rule (Inal 2020; Güneş-Ayata and Doğangün 2017; Mutluer 2019). For instance, Wuthrich (2021) attributes the increased number of women elected to the Turkish Grand National Assembly to a central AKP strategy with the aim of making electoral gains. That said, Wuthrich also notes that this increase has not changed the significant limitations on women’s access to political power in Turkey. A broadly similar contradictory picture is described by Güneş-Ayata and Doğangün (2017). They observe that despite significant legal advances in line with international conventions, gender equality in Turkey has not been fully attained. The same authors point out that Turkey not only hosted the signing ceremony of the Council of Europe Convention on preventing and combating violence against women and domestic violence in 2011, but it was also the first country to ratify the Convention without any reservations and indeed, in 2012, enacted a new law for the Protection of Family and Prevention of Violence against Women (Güneş-Ayata and Doğangün 2017, 614). Thus, Turkey’s withdrawal from the Istanbul Convention obviously represents a significant regression. This is stressed in the EU Commission’s Turkey 2021 Report.^[16] It notes that in “the area of gender equality there was important backsliding on women’s rights” due to Turkey’s withdrawal from the Council of Europe Convention, which, along with the aforementioned presidential decree, represents “a clear regression on the rights of women and girls. This decision compromises the women and girls’ rights and combatting gender-based violence in Turkey and sets a dangerous precedent.”^[17] The withdrawal from the Convention epitomizes the significant policy reversals in the area of gender equality, while Sofagate and the discourteous diplomatic treatment of Ursula von der Leyen are, on a symbolic level, testament to this backsliding during the AKP era. It has been shown that these reversals can be attributed to the discourses adopted by the ruling AKP party and the dominant elite, which have gradually degraded and delegitimized gender equality as an inferior Western import incompatible with domestic cultural and religious principles (Bodur Ün and Arikan 2022). As we will show, in terms of the five V-Dem gender indices, Turkey deviates significantly from the average EU scores as encapsulated and quantified by the indices used here.

As pointed out, even though Turkey’s candidate country status predates that of many more recent members, it was leapfrogged three times in the accession process. A total of 13 countries have joined the EU in the three successive enlargements – 2004, 2007, and 2013 – which have taken place since Turkey was officially declared a candidate country at the December 1999 European Council meeting in Helsinki. The Presidency Conclusions of the Helsinki Council emphasize the issue of human rights, since it is stated that this “will include enhanced political dialog, with emphasis on progressing towards fulfilling the political criteria for accession with particular reference to the issue of human rights […]”.^[18]

As can be seen from the relevant literature, the mere fact that Turkey’s association with the EU and the concomitant accession aspirations date back more than five decades strongly suggest that Turkey’s candidacy cannot be regarded as an ordinary one. In effect, its protracted candidacy is a manifestation of the multidimensional domestic as well as external factors that impede its convergence and eventual EU membership. The lengthy candidacy, along with the democratic backsliding witnessed in recent years, has resulted in a steadily growing skepticism regarding the country’s European prospects. Indeed, it has been argued that these developments indicate not only a lesser and more limited Europeanization, but also a process of de-Europeanization that seems to be gaining momentum (Aydın-Düzgit and Kaliber 2016). In fact, it has been suggested that Turkey’s membership prospects have always been vague and less credible than those of other accession candidates (Lippert 2021; Kollias 2023); hence the leapfrogging in the accession process. As already noted, following the EU Council decision of 26 June 2018, accession talks have come to a standstill. Among other issues, in its decision, the EU Council expresses concern regarding “the continuing and deeply worrying backsliding on the rule of law and on fundamental rights including the freedom of expression” and notes that “Turkey has been moving further away from the European Union.”^[19] In other words, it points to a process of divergence rather than convergence with the EU, its acquis, and, not least, policies such as the Gender Equality Strategy 2020–2025.

To explore the issue at hand in more depth, we use five indices constructed and published by the Varieties of Democracy (V-Dem) project (Coppedge et al. 2021, 2021a; Pemstein et al. 2020).^[20] The V-Dem project compiles over 350 indices with global coverage. The indices are designed to cover the many aspects of a country’s political system, quantifying the multidimensional nature of democracy.^[21] Some of the V-Dem indices assess gender-related dimensions of democracy, such as women’s political empowerment and participation (Arrington et al. 2021; Mechkova and Carlitz 2021). To the best of our knowledge, no other open and publicly accessible database offers such a range of indices quantifying the many different and important facets of gender equality over a long time period. The latter is an important characteristic that allows for the adoption of time series-based methodological tools that can provide better insights into the process of convergence – in this case convergence in terms of gender equality. Hence the decision to use the V-Dem indices in the analysis that follows. Nonetheless, a word of caution is in order here. Gender equality is a multidimensional qualitative concept. Reducing it to a unidimensional scale, as all indices invariably do, has significant limitations. That said, engaging with such issues comprehensively is well beyond the scope and the confines of the present article.

The five gender-related indices that we selected to examine the theme of convergence are as follows. The first is the Exclusion by Gender Index (henceforth exlgender). In a nutshell, this is defined by the V-Dem project as follows: “Exclusion is when individuals are denied access to services or participation in governed spaces based on their identity or belonging to a particular group.”^[22] For example, the five priority areas of the EU’s Gender Equality Strategy 2020–2025 include the promotion of equality between women and men at all levels and at all stages of policymaking. As has been noted, concerns regarding gender representation and women’s political presence need to be viewed in the context of the broader democratic backsliding and erosion that seem to be underway in a number of countries (Celis and Childs 2019). Just as the other four indices used here, exlgender is scaled from 0 to 1. However, it should be stressed that unlike the rest of the indices, in this case lower scores indicate a better situation (i.e., less exclusion), while higher scores mean the situation is worse (i.e., more exclusion).

The second index used is the Women Political Empowerment Index (gender): “Women’s political empowerment is defined as a process of increasing capacity for women, leading to greater choice, agency, and participation in societal decision-making,” incorporating “three equally-weighted dimensions: fundamental civil liberties, women’s open discussion of political issues and participation in civil society organizations, and the descriptive representation of women in formal political positions”.^[23] The scale range of gender is 0–1, with higher values reflecting a better situation in terms of women’s political empowerment. The Women Civil Liberties Index (gencl) quantifies whether (or not) “women have the ability to make meaningful decisions in key areas of their lives”.^[24] Just as before, higher values in the 0–1 range represent a better situation. The same applies to the other two indices. The Women Civil Society Participation Index (gencs) quantifies the ability of women “to express themselves and to form and participate in groups”. Lastly, the Women Political Participation Index (genpp) quantifies whether (or not) “women [are] descriptively represented in formal political positions”.^[25] All the V-Dem indices quantify aspects of gender equality that are among the central policy objectives of the EU’s Gender Equality Strategy 2020–2025 the aim of which is a gender-equal Europe, such that, irrespective of gender, its citizens “are free to pursue their chosen path in life, have equal opportunities to thrive, and can equally participate in and lead our European society”.^[26]

Figure 1 provides a comparative presentation of the five V-Dem indices used here: Exclusion by Gender Index (exlgender) in Figure 1A; Women Political Empowerment Index (gender) in Figure 1B; Women Civil Liberties Index (gencl) in Figure 1C; Women Civil Society Participation Index (gencs) in Figure 1D; and Women Political Participation Index (genpp) in Figure 1E. To offer a comparative perspective, the figures depict the difference between each country’s average score (y _ith ) and the EU28^[27] average ( y ‾ ) for the period 1999–2020. That is: y i t h − y ‾ . The start of the period is selected to coincide with Turkey being officially granted candidate country status at the 1999 Helsinki European Council meeting. The horizontal axis represents the EU28 threshold. Thus, in the case of the Exclusion by Gender Index (exlgender), countries with a positive deviation (positive values) – i.e., a higher average than the EU28 – perform worse vis-à-vis the EU28 average, showing more gender-based exclusion, while negative values indicate less gender-based exclusion and better performance vis-à-vis the EU28 average. The reverse is true for the other four indices. A positive deviation from the zero axis indicates a better score and performance in terms of the Women Political Empowerment, Women Civil Liberties, Women Civil Society Participation, and Women Political Participation Indices. The graphical representation of the five gender V-Dem indices (Figure 1A–E) reveals differences between the EU28 member states, while Turkey significantly deviates from all the EU member countries in all five indices.

Figure 1:

National averages versus EU28 average (zero axis), 1999–2020. Source: Authors’ calculations.

Starting with the Exclusion by Gender Index (exlgender), the average score for the entire EU28 group of countries is 0.07 (Figure 1A). As a reminder, like all the indices used here, exlgender is scaled from 0 to 1, but unlike the other four, higher scores indicate a worse situation, that is more exclusion, and lower scores mean less exclusion. Compared to the average EU28 exlgender score (0.07), Turkey’s average score in this index is 0.48, in other words appreciably higher even than the worst score among the EU28, which is 0.17 for Croatia, followed by Romania with 0.15, hence Turkey’s larger deviation from the zero axis which corresponds to the EU’s average. Romania is in fact one of the EU countries that systematically underperforms vis-à-vis the group’s average in all five indices. The same applies to Malta and Cyprus, as the figures reveal. All three, i.e., Romania, Cyprus, and Croatia are among the countries that leapfrogged Turkey in the accession process. Cyprus joined in 2004, Romania in 2007, and Croatia, as the newest member of the Union, in 2013. Their respective exlgender scores at their accession year were: Cyprus 0.14, Romania 0.15, Croatia 0.15, while the corresponding EU28 average was 0.06 at the time of these three enlargement rounds in 2004, 2007, and 2013. The corresponding scores for Turkey were 0.45 in 2004 and 2007, when Cyprus and Romania joined the EU, and 0.51 in 2013 when Croatia became a member.

Turning to the other four indices, that is gender, gencl, gencs, and genpp, where higher values in the 0–1 range reflect a better score and performance in terms of each criterion, Turkey once again significantly underperforms vis-à-vis the EU member states. The average score of the EU28 in the Women Political Empowerment Index (gender) displayed in Figure 1B is 0.92, while Turkey’s score is 0.57, which is lower than the lowest-scoring EU28 country, Malta, with 0.84, followed by Romania (0.85). Malta and Romania joined the EU in 2004 and 2007, respectively, leapfrogging Turkey. At the time of their EU accession, their scores on this specific index were 0.85 in both cases, Turkey’s was 0.61 (2004) and 0.62 (2007), while the EU’s average score was 0.92 in both years.

The same applies to the third index, the Women Civil Liberties Index (gencl). Turkey’s score is 0.66 compared to the EU28 average of 0.93 (Figure 1C). Romania has the lowest score among the EU28 (0.84). When Romania joined in 2007, its gencl score was 0.84 and Turkey’s was 0.76, while the EU’s average for that year was 0.94. Turkey’s deviation from the EU28 average score is even greater when it comes to the Women Civil Society Participation Index (gencs): 0.37 compared to 0.88 (Figure 1D). Hungary (0.86), Luxembourg (0.74), Malta (0.82), and Slovakia (0.85) are among the EU members that have the lowest average scores in this index, but all score significantly higher than Turkey with 0.37. Lastly, in the Women Political Participation Index (genpp, Figure 1E), Turkey’s score is 0.69, compared to an average of 0.95 for the EU28. Hungary (0.86), Malta (0.85), and Romania (0.85) are the EU member states with the lowest average scores and the larger deviation from the threshold of the horizontal axis which depicts the EU’s average of 0.95 for the period 1999–2020 presented in the figures. At the time of their accession, their genpp index scores were 0.86 for Hungary and Malta (2004) and 0.84 for Romania (2007). The corresponding year averages for the EU were 0.94 and 0.95 respectively, while Turkey’s scores in this index were 0.64 in 2004 and 0.71 in 2007.

As previously mentioned, Turkey is one of the eight countries that currently have candidate status. It is interesting to note that compared with the other candidate countries, Turkey is still the worst performer in terms of these five gender-centric equality indices. The average scores for the same period, i.e., 1999–2020 are presented in Table 1. As can be seen, all seven of the other candidates outperform Turkey in all five indices. For instance, in the Exclusion by Gender index (exlgender), where a higher value indicates more exclusion, Turkey’s average of 0.48 is by far the highest in the group. Montenegro (0.38), Bosnia and Herzegovina (0.25), and Albania (0.22) occupy the next three places. Ukraine (0.09) and Moldova (0.10) are the two candidate countries with the lowest gender-based exclusion levels. It is worth noting that all candidate countries’ average scores are higher than the corresponding EU28 average (0.07). A similar picture emerges for the other four indices, where a higher value suggests better performance. In all four, Turkey’s score is the lowest of the eight candidate countries, indicating a comparatively worse performance (Table 1). For example, Turkey’s score in the Women Civil Liberties Index (gencl) is 0.66 and the next lowest is 0.77 for Bosnia and Herzegovina, followed by North Macedonia with 0.79, and Albania (0.80). Ukraine and Serbia are the two candidates with the highest average score (0.87) in both of the aforementioned indices for the period examined.

Although the graphical representations and descriptive statistics for the five gender-centric V-Dem indices are clearly reveal significant differences between Turkey and the rest of the EU28, as well as compared to the other candidate countries, they are nevertheless a static depiction of the relative average performance during the period 1999–2020. Hence, they cannot offer any reliable insights into whether (or not) a process of convergence is taking place. For such inferences, we need to resort to other statistical tools that can capture such a process, if indeed it is taking place. We turn to this in the next section.

A Methodological Snapshot, Empirical Findings, and Discussion

The convergence hypothesis can be examined empirically using a variety of methodologies, ranging from σ and β-convergence to unit root tests. The latter are a tool that is increasingly used by many empirical studies examining convergence hypotheses (Chapsa and Katrakilidis 2014; Lau et al. 2016). For the purposes of the present paper, we opt to use four such tests to investigate the gender-centric convergence hypothesis in the case of Turkey. The unit root tests are deemed to be a more suitable tool in this case, since we are not examining the presence of a convergence process within a group of countries, as we would be if the issue was being investigated for all the EU28 countries. In our case, the question being addressed is whether Turkey, a candidate country, converges with the group of countries it aspires to join as a full member. As stressed by Glüpker-Kesebir (2016, 517), “the state of women’s rights and gender equality in Turkey is a benchmark for the development of Turkey as a liberal democracy compatible with EU standards”. We postulate that within the broader process of de-Europeanization, the democratic backsliding of Turkey, and failure to consolidate liberal democratic governance, as highlighted in a number of recent papers (see, for example, Turan 2023; Kollias 2023; Lippert 2021), Turkey also deviates and fails to converge in terms of gender equality vis-à-vis the EU28 and the average norms within this group of countries. In view of this, the unit root tests are a more suitable empirical tool compared to methodologies such as σ and β-convergence, which are used in cases where within-group convergence hypotheses are being tested.

Specifically, the unit root tests employed are the Dickey and Fuller (1979, 1981 ADF test, the Phillips and Perron (1988) PP test, the ADF-GLS test proposed by Elliot et al. (1996), and the NGP test by Ng and Perron (1995, 2001. To this effect, we adopt the methodological strategies employed by recent studies (Ceylan and Abiyev 2016; Kollias and Messis 2022; Tsanana and Katrakilidis 2014). These studies provide a detailed technical description of the four tests used. Consequently, for reasons of brevity, we refrain from providing a fundamentally similar detailed explanation here. In brief, in line with these studies, we use an average as the benchmark against which the convergence hypothesis is investigated. The benchmark is the EU average ( y ‾ ) of each of the five gender indices used here (exlgender, gender, gencl, gencs, genpp).

In line with the discussion in the previous section, we opt to focus on the entire period during which Turkey is officially recognized as a candidate country (1999–2020). As already noted, the prospect of impending EU membership has proved to be a successful means of encouraging candidate as well as potential candidate countries to adapt and implement the acquis communautaire in the various spheres. An important point of clarification is warranted here. Given that during the period 1999–2020, there were three EU enlargements – in 2004 with ten new members, in 2007 with the accession of two members, and finally in 2013 when Croatia joined –, in order to allow for the possible effect the accession of new members had on the average value ( y ‾ ) of the five indices, the average of the actual EU members is calculated for each specific year. We then calculate the difference y t u r − y ‾ between the time series of the indices, with y _tur being Turkey’s score in each of the five V-Dem indices. In other words, Turkey’s deviation, in the exlgender, gender, gencl, gencs, and genpp indices scores, from the corresponding EU average ( y ‾ ) for each year. Hence, y t u r − y ‾ demonstrates the difference between each series and the benchmark index. Testing for convergence involves identifying time series whose means and variances remain constant over time. To this end, following the aforementioned studies, the four unit root tests – ADF, PP, ADF-GLS, and NGP – are estimated using the time series which are generated by the methodological approach described above. Thus, in conjunction with the previous analysis, which relied on the graphical representations of the five indices, the unit root tests will provide additional information for the gender-centric convergence analysis.

The results from estimating the ADF, DF-GLS, PP, and NGP tests for the period 1999–2020 are shown in Table 2A. The findings do not offer any statistical evidence which supports the convergence hypothesis in terms of the five V-Dem indices that are used here. The only exception to this is the result of the DF-GLS test (with and without trend) in the case of the Women Political Participation Index (genpp). It is, however, the only one of the four tests that suggests convergence. The other three unit root tests – ADF, PP, NGP – reject the convergence hypothesis across all indices. Thus, given the strong consistency of the findings, it seems that in terms of these five gender-centric indices, Turkey does not appear to converge with the EU28 for the period 1999–2020.

Lastly, as a robustness test, the four unit roots for each of the five V-Dem indices were re-estimated for the entire Turkey–EU association period, which dates back to the 1963 Ankara Association Agreement. The results of the re-estimation for 1963–2020 are presented in Table 2B. Once again, the findings are very consistent across all tests except for DF-GLS (without trend at the 5 % significance level) in the case of the Women Political Participation Index (genpp). The 1963–2020 period results concur with the ones for the shorter period (1999–2020) presented in Table 1A. Overall, no process of convergence in terms of these V-Dem indices seems to be detected by the tests conducted.

As noted above, the EU actively pursues policies that facilitate and promote convergence, both among its current members and in countries, such as Turkey, that aspire to membership. The various accession conditionalities encapsulated by the Copenhagen criteria are designed to stimulate a convergence process in various spheres during the pre-accession phase, which will, in turn, facilitate the eventual integration of a country in the EU once membership is achieved. Essentially, in this context, convergence denotes the process generally referred to as “Europeanization”. As Alpan (2021) points out, the term Europeanization is used to describe how the EU affects domestic politics, policies, and polity in candidate countries, which leads to convergence via the adoption and implementation of the acquis communautaire. Gender issues are assigned a prominent position in the acquis, on the EU’s agenda, and in policies such as the Gender Equality Strategy 2020–2025, which is directly related to the theme of the present paper.^[28]

Turkey’s withdrawal from the Istanbul Convention and Sofagate are symbolic markers of backsliding on gender equality in Turkey, underscoring a much wider process of de-Europeanization and a significant democratic rollback over the past few years (see, for example, Yilmaz 2016; Bodur Ün and Arikan 2022). While Europeanization can be seen as a transformative process through which candidate countries, in this case Turkey, converge toward EU standards in many different spheres, de-Europeanization can be understood as the weakening or even failure of this process, signifying divergence (Aydın-Düzgit and Kaliber 2016). In other words, de-Europeanization entails the disappearance of EU norms as a key reference point stimulating progressive reform and change. The withdrawal from the Convention should be viewed as a further step in the broader context of Turkey’s ongoing process of de-Europeanization, in which the dominant official discourse portrays the EU as an intruder and imposer of norms that are both incompatible with and inferior to domestic cultural and religious values (Bodur Ün and Arikan 2022). In fact, with regard to gender-related issues, it has been argued that Turkey underwent a process of Europeanization without substance and that, despite vague commitments and policy initiatives, in practice gender equality was negated through policy reversals brought about by the adoption of pro-family and anti-gender religious conservative discourses by successive AKP governments (Aybars et al. 2019). Both the comparative descriptive presentation of the V-Dem indices and the empirical findings reported above concur with the analysis and conclusions of the existing literature on Turkey’s de-Europeanization and democratic regression.^[29]

Concluding Remarks

The integration of countries in supranational entities such as the EU acts as a promoter of convergence. Indeed, convergence among its members is actively sought by the EU in many different spheres. The promotion of convergence is not limited to EU member states but also includes candidate countries, since this facilitates the latter’s integration once membership is achieved. One of the main attractions of EU membership has, in fact, been the prospect of convergence, not limited to income levels and living standards, but also in terms of social conditions, the functioning of democratic institutions, and polity. The appealing prospect of EU accession that is tied to specific conditionalities has, in broad terms, proved to be a successful means of encouraging third countries to converge by adapting and implementing the acquis communautaire.

Turkey is one of the eight countries that are currently formally recognized as EU candidate countries. Within the broader discourse which examines whether candidate countries are converging with the EU, our study focused on a gender-centric convergence analysis in the case of Turkey. Turkey’s withdrawal from the Istanbul Convention on preventing and combating violence against women in March 2021 and the Sofagate diplomatic incident a month later when the president of the European Commission Ursula von der Leyen was subjected to discourteous diplomatic treatment which she attributed to the fact that she is a woman were the two events that served as impetus for this article.

Gender-related issues constitute an important policy area for the EU, as the Gender Equality Strategy 2020–2025 testifies, and is one of the themes usually examined in the Commission’s annual progress reports on Turkey’s candidacy. As mentioned above, significant regression on women’s rights were identified in the latest report, published in 2021. The report associates the backsliding on gender equality with Turkey’s withdrawal from the Istanbul Convention following president Erdoğan’s Decree on 20 March 2021.

The present study used five V-Dem indices to probe the gender convergence hypothesis: Exclusion by Gender, Women Political Empowerment, Women Civil Liberties, Women Civil Society Participation, and Women Political Participation. As revealed by the comparative descriptive analysis, Turkey lags behind in terms of all five indices, both when compared with the EU’s averages and with each of the current members. A battery of unit root tests, used extensively in the relevant convergence literature, was then applied to examine empirically whether convergence can be established. The findings reported in the present study do not point to any such process. Although gender issues are by no means the epicenter of the accession process, the preceding analysis points to yet another area where it can be argued that “Turkey has been moving further away from the European Union,” as noted in the EU Council’s Conclusions on the enlargement and association process of 26 June 2018.^[30] Our findings concur with the de-Europeanization and democratic backsliding described in the literature on Turkey and reveal another sphere where convergence is absent and a significant gap exists.