STUDY Trade Unions in Turkey 2018 ALPKAN B İ RELMA November 2018 n One out of every four workers in Turkey was protected by a collective bargaining agreement at the end of 1980s when the almost continuous decline of the coverage began. The lowest membership rate was 6 per cent in 2013; since then it has risen slightly to 7.6 per cent in 2017. n The number of union workers rose from 1 million to 1.8 million between January 2013 and July 2018. Approximately 500,000 of these new members are public-sector subcontracted workers. Union density in the private sector(excluding public subcontractors) stagnates. n Changes in much-criticized union laws of both workers and civil servants in 2012 are far from aligned with ILO conventions. The authorization process for collective bargaining of newly unionizing workers is still extremely difficult and the protection against unfair dismissal on the grounds of trade union membership has been weakened. Additionally, civil servants still do not enjoy the right to collective bargaining because the system introduced in 2012 does not include negotiation but merely consultation. Among other problems, civil servants are still not allowed to strike. n The right to strike has also been further restricted. Since early 2017, the government has banned seven strikes by more than 150,000 workers. Due to this rising number of strike bans and arrests of union leaders, the International Trade Union Confederation declared that Turkey was among the ten worst countries for workers in 2018.

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Contents Introduction 3 I. Worker Unionism 3 1. An Overview of Workers’ Unionism in Turkey 2. Legal framework: Unions and Collective Agreement Law of 3. Trends in Workers’ Union Membership and 10 4. A Closer Look at the Three 16 II. Civil Servant Unionism 21 1. Amendments in the Union Law of Civil Servants in 21 2. Changing Landscape of Civil Servants’ Unionism.............................. Conclusion 23 References 25

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Introduction This report is an extension and update of Dinler’s Trade Unions in Turkey report published in 2012 by Friedrich Ebert Stiftung. 1 Dinler provides an effective summary of the historical background and the basic characteristics of Turkish unionism. This report focuses on the changes between 2012 and 2018 and rests more on statistical data, some of which were published after 2012 by several state agencies. There have been some important changes in the union landscape of Turkey since 2012. Firstly, a new Unions and Collective Agreement Law was ratified in 2012. Secondly, since 2013, the Ministry of Labor and Social Security(MoLSS) has declared the numbers of members of all trade unions biannually. This is considered a reliable source by scholars and unions. Before that, membership statistics were entirely fictitious and exaggerated. New data allows observers to capture trends in union membership twice a year. Last but not least, the Turkish political scene has changed tremendously since 2012, and this transformation has implications for unions. There is a clear legal distinction between workers and civil servants in Turkey. All employees except civil servants are defined as»workers« by Turkish Labor Law 4857. Workers and civil servants work under different labor laws(4857 and 657) and organize according to different union laws(6356 and 4688). The main difference in the labor laws of workers and civil servants is the much greater job security of civil servants compared to workers; nevertheless, civil servants are not allowed to strike. Public workers, who are mostly manual workers 2 employed by State Economic Enterprises or various public institutions such as ministries, municipalities or banks, should not be confused with civil servants. Like private sector workers, public workers operate under labor law of 4857 and unionize according to union law of 6356. The number of public workers was nearly 500,000 at the beginning of 2018. 3 This number almost tripled in April 2018 for reasons explained below. 1. Dinler, 2012. 2. Manual worker refers to workers whose jobs involve mostly physical tasks. The term includes blue-collar workers but also service workers like janitors, security guards, municipal workers, postal workers etc. 3. See the website of Bütçe ve Mali Kontrol Genel Müdürlü ğ ü<http:// www.bumko.gov.tr/TR,908/kadro-istatistikleri.html> There are approximately 3 million civil servants, and around 17 million workers(including 3 million Turkish-­ citizen 4 informal workers) in Turkey as of 2018. This report focuses primarily on worker unionism, while the second section briefly summarizes recent developments in civil servant unionism. Table 1 outlines different groups of employees in Turkey, all of which have significant differences in terms of their access to rights to organize and collective bargaining. Türkİş , Hakİş and D İ SK are the confederations of the trade unions representing workers in private and public sectors, whereas Memur-Sen, Türkiye Kamu-Sen and KESK represent civil servants as confederations of trade unions. Table 2 illustrates the membership and emblems of the six strongest confederations functioning under two different legal frameworks. I. Worker Unionism 1. An Overview of Workers’ Unionism in Turkey Workers’ unions in Turkey have not been always weak. They grew in the 1960s and 1970s and survived an anti-labor military rule between 1980 and 1983. They organized the strongest strike waves in the nation’s history in the first half of 1990s and raised real wages, even exceeding pre-coup levels. 5 However, since then they have been losing power almost unceasingly. Turkey did not publish reliable data on the number of union members before 2013. The membership data before 2013 was exaggerated for several historical reasons. However, MoLSS provides relatively credible data based on the number of workers who have worked under a collective bargaining agreement since the introduction of that right in 1963. The industrial relations system of Turkey is reminiscent of the United Kingdom(UK) and the United States(US), in the sense that there are no particularly supportive mechanisms for collective bargaining, such as exten4. TURKSTAT provides an estimation about the number of informal workers, but its data only includes Turkish citizens. The rising number of informal migrant workers are excluded. The arrival of 3.6 million Syrian refugees(because of Syrian Civil War) since 2011 has significantly increased the number of informal migrant workers in Turkey. 5. See Birelma, 2017: 273. 3

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 1. Different groups of employees in Turkey Level 1 Level 2 Level 3 Definition Approx. numbers Union density Civil servants Civil servants Civil servants Mostly white-collar public employees who work and organize according to different laws than workers(all the above). 3 million Very high Public workers Public workers Mostly manual workers working in the public enterprises and institutions. 0.5 million (until April 2018)* Very high Public-sector subcontracted workers Workers of subcontractors working for the public enterprises and institutions. 1 million (until April 2018)* High(After an amendment in 2014) Workers Private sector workers Formal private sector workers Informal workers I Formal private sector workers (other than the public-sector subcontracted workers) Turkish citizens who work in an unregistered way 12.5 million 3 million Low Zero** Informal workers II Non-citizens who work in an unregistered way ?*** Zero** TOTAL 20 million * In April 2018, the government transformed nearly 900,000 public-sector subcontracted workers into public workers. ** Informal workers cannot become a union member. *** Only Syrian informal workers are estimated to be around 400,000. Source: http://www.hurriyet.com.tr/ekonomi/kayitli-3-686-kayit-disi-400-000-40024074, accessed July 1, 2018. sion procedures one sees in coordinated market economies like Germany. 6 Just like in the US and UK, Turkish unions must organize workplace by workplace to enlarge collective bargaining coverage. Therefore, collective bargaining coverage is a good proxy for union density, which is moderately higher than the bargaining coverage. Graph 1 displays the trajectory of workers’ collective bargaining coverage in Turkey since 1988. The ratio is calculated by dividing the number of workers covered by a collective bargaining agreement by the total number of workers. The total number of workers stands for all employees(including informal workers) except civil servants. One out of every four workers was protected by a collective bargaining agreement at the end of 1980s, when the almost continuous decline of union membership be6. Frege and Kelly, 2004: 38. gan. The lowest ratio was 6 per cent in 2013, and has risen slightly to 7.6 per cent in 2017. Organizing in the private sector is extremely difficult in Turkey because the authorization process for collective bargaining is complicated as explained below. In the most comprehensive quantitative research on this issue, Özveri investigated the cases of unionization in four industries between 1983 and 2009 that resulted in lawsuits over authorization. 7 He found that in 73 per cent of the cases in which the court authorized the union as a result of the lawsuit, the employer busted the union during the lawsuit, which lasted an average of 424 days. In these 73 per cent of the cases, unions could not sign a collective contract despite they once organized the majority of the workers in the workplace as confirmed by the court. In a qualitative approach, Birelma scrutinized three cas7. Özveri, 2013: 379. 4

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 2. Membership of main trade union confederations in Turkey Confederations representing workers(July 2018) Türkİş 958,618 Hakİş 654,722 Confederations representing civil servants(July 2018) D İ SK 160,568 Memur-Sen 1,010,298 Türkiye Kamu-Sen 394,423 KESK 146,287 Graph 1. Collective bargaining coverage among workers% 30,0 25,0 20,0 15,0 10,0 5,0 0,0 Source: Ministry of Labor and Social Security(MoLSS) provides the number of workers for whom a union signs a collective contract within a year. The annual number of workers working under a collective contract is calculated as suggested by Çelik and Lordo ğ lu(2006: 19). Turkish Statistical Institute (TURKSTAT) provides the numbers of all employees and civil servants. 5 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY es of unionization in the private sector ethnographically, exposing immense difficulties in the unionization and authorization process. 8 The research uncovers years of secret organizing, firings of nearly 80 workers due to union activities, and picket lines and protests of workers. In one of the three cases the union was busted even though it won the lawsuit, which took three and a half years. Building on the data from newspaper reports of labor unrest, Emek Çalı ş maları Toplulu ğ u(Labor Studies Group) found that at least 2,258 workers among 81 cases were reported to be fired because they participated in union organizing in 2015.9 This number does not include the cases in which the unionizing workers did not organize any protest. The next year, that number was 1,359 in 42 cases. Less labor unrest is related to the political turmoil caused by the coup attempt in July. 10 In 2017 the number was 857. Again, even fewer protests can be attributed to the state of emergency declared after the coup attempt and continued until July 2018. 11 The increase in coverage since 2014, on the other hand, is due mostly to the unionization of workers employed by subcontractors working for public institutions. The number of these public-sector subcontracted workers is more than 1 million, which corresponds to a potential of a 6-point increase in the coverage. Thanks to the mobilization of these workers and some unions since the mid-2000s, the governing Justice and Development Party(AKP) enacted an amendment to Turkish Labor Law in 2014 that facilitates unionization of public-sector subcontracted workers. The amendment states that when a group of public-sector subcontracted workers successfully unionizes, they bargain and sign a collective agreement with the principal employer, namely the concerned public institution. Instead of the subcontractor, public institutions pay the additional wages and rights gained with collective bargaining. As expected, this amendment encouraged subcontractors to be less resistant to unionization because it would cause no financial burden for them. Since its enactment, many unions, even those that so far had ignored subcontracted workers, began campaigns to unionize them and in many cases have signed collective agreements. 8. Birelma, 2014. 9. Emek Çalı ş maları Toplulu ğ u, 2016: 26. 10. Emek Çalı ş maları Toplulu ğ u, 2017: 22, 28. 11. Emek Çalı ş maları Toplulu ğ u, 2018: 29. But how could workers and unions manage to win such a victory? Efforts of unions, other worker associations and ordinary subcontracted workers in covert or transparent ways, through formal or informal channels, accumulated and built a significant line of resistance. The bulk of public-sector subcontracted workers makes up the social base of the governing party. They are employed mostly through party channels and this has strengthened their hand. This fact rendered AKP more fragile to the demands and protests of this social base. Moreover, the forthcoming elections offered a political opportunity for the movement at a time when the vast AKP support of the last twelve years seemed to be eroding, as became obvious in the June 2015 elections when the party lost the majority in the parliament. In addition, an increasing number of industrial homicides in the last decade, the devastating massacre of 301 miners in May 2014, and their relationship with subcontracting mobilized a public outcry supporting workers’ causes and demands. In this political environment, before the elections in June 2015 all three main opposition parties declared they would put an end to public subcontracting and transfer all of the existing subcontracted workers to regular positions in public institutions. 12 In the face of ongoing demands of unions and subcontracted workers and leading up to elections in June 2018, AKP finally transitioned most the public-sector subcontracted workers(around 900,000) into regular public workers in April 2018. Because it is much easier for public workers to unionize, we can foresee that the increase in the collective bargaining coverage since 2013 will continue for the next couple of years and the coverage will probably reach around 10 per cent. The number of striking workers is another important indicator to estimate the power of workers’ unions. MoLSS provides the number of workers who participate in the legal strikes. According to workers’ union law the only legal way to stage a strike is if the collective bargaining process does not result in an agreement. Therefore, MoLSS’ striking worker data does not include wildcat strikes or other spontaneous strikes of non-union workers; however, it still provides an important indicator of union power and activity. Graph 2 reveals the number of participants of legal strikes organized by unions 12. For a detailed analysis of subcontracted workers’ struggle and victory see Birelma, 2017. 6

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Graph 2. Striking workers and wage share 250.000 100 90 200.000 80 70 150.000 60 50 100.000 40 30 50.000 20 10 0 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Participants of legal strikes Adjusted wage share% Source: The number of participants of legal strikes is provided by MoLSS. Adjusted wage share is calculated by AMECO, the annual macro-economic database of the European Commission's Directorate General for Economic and Financial Affairs. on the left y-axis and wage share on the right y-axis. Wage share is the part of the national income allocated to wages and is probably the best indicator to estimate the total value of wages in a country in general from a distribution of income perspective. 13 As Graph 2 illustrates, the strike waves in 1990 and 1991 achieved a great increase in wages and wage share. The economic crisis of 1994 led employers to reduce wages which resulted in another strike wave in which nearly 200,000 workers participated in 1995. However, this time their strikes were largely unsuccessful and the year 1995 marked the last massive outcry of the labor movement. In the next twenty-two years, legal strikes almost vanished. Inevitably, wage shares reached their lowest levels. The reason for declining strikes is not only unions’ growing hesitation due to their shrinking power resources. Table 3 presents the strikes banned by the AKP governments since 2002. Birle ş ik Metalİş Union of D İ SK Confederation and Kristalİş Union of Türkİş Confederation are leading by each having four banned strikes. As men13. Bengtsson and Ryner, 2015; ILO and OECD, 2015. tioned above, after the coup attempt in July 2016 the government declared a state of emergency, which continued until July 2018. The table reveals that the government imposed strike bans much more often under the state of emergency. While there were eight strike ban cases involving nearly 40,000 workers in fourteen years between 2003 and 2016, the recent period of one-anda-half years since the beginning of 2017 witnessed seven strike ban cases involving more than 150,000 workers. Due to this rising number of strike bans and cases of arrests of union leaders, the International Trade Union Confederation(ITUC) declared that Turkey was among the ten worst countries for workers in 2018. 14 Dinler notes that numerous new mechanisms of tripartite social dialogue have been introduced in Turkey»especially as part of the European Union accession process.« 15 Despite the proliferation of new tripartite bodies,»uneven government representation« and»subsequent state control,« among other factors, inhibit»the development of a bottom-up practice that would empower partners 14. ITUC, 2018: 27. 15. Dinler, 2012: 11. 7

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 3. Strikes banned by AKP governments(2003–2018) Year Workplaces Number of workers approx. 1 2003 Petlas 350 2 2003 Ş i ş ecam 5,000 3 2004 Ş i ş ecam 5,000 4 2004 Pirelli, Goodyear, Bridgestone 5,000 5 2005 Erdemir Mining 400 6 2014 Ş i ş ecam 5,800 7 2014 Çayırhan and Çöllolar Coal Mines 1,500 8 2015 Turk. Employers Association of Metal Industries 15,000 9 2017 Asil Steel 600 10 2017 Employers Union of Electromechanical Metal Ind. 2,200 11 2017 Akbank 14,000 12 2017 Ş i ş ecam 6,500 13 2017 Mefar Pharmaceuticals 500 2018 Turkish Employers Association of Metal Industries 14 (MESS) 130,000 15 2018 Soda Sanayii A. Ş .( Ş i ş ecam) 540 Total 192,390 Unions Petrolİş Kristalİş Kristalİş Lastikİş T. Madenİş Kristalİş T. Madenİş Bir. Metalİş Bir. Metalİş Bir. Metalİş Banksis Kristalİş Petrolİş Türk Metal, Birle ş ik Metalİş , Çelik İş Petrolİş Source: Çelik(2018), Hürriyet Newspaper, http://www.hurriyet.com.tr/petrol-is-soda-sanayiinde-isciler-isyerini-te-40845838. equally and shape policy outcomes.« 16 The quality of tripartite social dialogue has worsened recently as criticized by various unions, especially after the declaration of the state of emergency in July 2016. As of July 2018, it had been a year since the Tripartite Consultation Board was convened by MoLSS. The government did not convene the board even to consult about the most dramatic development of recent years, namely, the transformation of nearly 900,000 public-sector subcontracted workers into regular public workers in April 2018. 2. Legal framework: Unions and Collective Agreement Law of 2012 Military rule between 1980 and 1983 enacted two very restrictive union and collective agreements laws that went into effect in 1983. One of the main reasons why 16. Ibid. union density has fallen since the end of the 1980s was that the 1983 union laws made the unionization of new workplaces much more difficult. The unions, European Union and the ILO raised the need for a new union law in the mid-2000s. However, the government, influenced by business interests, decided to keep the main restrictions of the former union laws intact. The Unions and Collective Agreement Law of 2012 number 6356 is not based on consensus between all social partners. One of the three main workers’ union confederations, D İ SK and several member unions of Türkİş Confederation opposed various articles of the new law. 17 Even though the new law introduces limited improvements regarding to the founding and internal functioning of unions and union membership, it maintains many restrictions of the previous law, especially related to collective bargaining and striking. 17. Çelik, 2013. 8

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY The Founding and Internal Functioning of Unions and Union Membership The 2012 law simplified the founding and internal functioning of unions. The obligation of notarization to join and leave a union has been abolished. This obligation was one of the most pronounced complaints of unions about the 1983 law because it caused a significant financial burden and unwieldy bureaucracy. The new law replaced notarization with an online system in which the worker becomes a union member through an online portal of the Turkish state. Thus, once a worker applies for union membership online, MoLSS is notified of the membership, which is then submitted to the union for approval. This however results in infringements of the privacy of personal information. As one of many problematic outcomes of this procedure, unions claim that some employers are asking for their workers’ portal passwords to check if they are union members, especially at the point of hiring. By keeping the former limitation, the new 2012 law only permits unions to be established at the level of industry and does not allow unions at the levels of workplace or profession. A union organized only at a single workplace cannot sign a collective agreement, because it will not exceed the industrial threshold. A union of a particular profession is not allowed either because what determines which union to join is not the individual’s profession but the industry of that workplace. A union can organize only workers of a specific industry and the composition of any industry branch is defined by the government in an arbitrary way. The new law limits the right to sue for union compensation in case of dismissal due to trade union activity. Union compensation is a special compensation(different than and in addition to the severance payment). If a worker is dismissed due to union membership or activity and can prove this at the court, he or she will get union compensation, which cannot be less than the worker’s annual wage. With the new law, the workers of workplaces employing fewer than 30 workers lose the right to union compensation. The Joint Report of EU-Turkey Joint Consultative Committee points to this change as one of the three areas of particular concern about the new law, because an estimated 95 per cent of companies are small-scale enterprises and 50 per cent of the registered workers work for those companies. 18 This limitation functions as an increased restriction of collective bargaining. Authorization of Collective Bargaining The new union law kept two thresholds of the abrogated union law for the authorization of collective bargaining with only minor changes. The first threshold concerns workplace membership. While the law maintains the workplace membership threshold at 50 per cent, the threshold for establishments with more than one branch has been reduced to 40 per cent. The second threshold is industry wide. Before the 2012 law, a union had to represent at least 10 per cent of the total number of registered workers in the relevant industry to be authorized for collective bargaining. The new law reduced this threshold to 1 per cent for unions that are members of one of the three main confederations named above. However, the threshold for other unions(which were not members of these three confederations) was 3 per cent. Upon the application of the main opposition party, in 2015 the Constitutional Court lowered the threshold to 1 per cent for all unions. However, this reduction from 10 to 1 per cent is not as dramatic as it sounds, because the new law also changed the statistical system of the membership count. Before the enactment of new union law, the most recently announced official rate of union density in 2009 was 60 per cent. This rate was based on unreliable and deliberately exaggerated membership records. The very same rate first declared by the new system in 2013 was 9 per cent. So, mathematically the actual decrease in the industry threshold was not 90 but 33 per cent. Furthermore, the number of industries was reduced from 28 to 20, which caused an automatic increase in some industries’ thresholds in terms of number of workers. Although it is still an improvement, the existing 1 per cent threshold practically precludes new, independent unions from obtaining authorization. The problems related to the authorization of collective bargaining are not limited to the thresholds. The mechanism itself is problematic and the new law kept this mechanism intact. Even if a union exceeds two thresholds and applies to the ministry for authorization, and the minis18. EU-Turkey Joint Consultative Committee, 2013, pp. 5. 9

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY try authorizes the union after checking the membership numbers, the employers have the right to reject the authorization by filing a lawsuit. The employers can file a lawsuit, for example, claiming that they employ more workers than it appears in the records of MoLSS, or claiming they have another branch. Such allegations would raise the workplace threshold, which would mean that the union had not met the representation requirement. Claiming that the workplace belongs to another industry than the union’s industry is also another common legal tactic. These lawsuits are aimed less at winning than at buying time as employers rely on the excessive length of legal proceedings in Turkey to wear down their workers. Unions do not have authority to start collective bargaining until the court declares it, which takes nearly two years. In this way employers buy great amount of time for union-busting. Employers usually fire leaders of unionization and unions mostly respond by building picket lines in front of the workplace. Meanwhile, management tends to intimidate, suppress and/or bribe other union workers. In many cases, after nearly two years of proceedings and union-busting campaigns, many workers get exhausted and resign from the union. During this process, union workers cannot legally strike, because according to the union law workers can strike only if their union has agreement authority in their workplace. In numerous cases, by the time the court authorizes the union, most of the frustrated and demoralized members have already resigned the union. 19 Özveri’s research has uncovered how in most cases unions cannot sign a collective agreement due to union busting even though they won the lawsuits at the end. 20 In only 27 per cent of the cases in which the court authorizes the union as a result of the lawsuit the union could sign a collective contract. In the rest of the cases the union was busted. Restrictions on the Right to Strike Just like the law it replaced, the new law bans all strikes other than those organized in case of disagreement during collective bargaining. In other words, slowdowns, 19. For a detailed ethnographic account of such a case see Birelma, 2014, pp 153–206. 20. Özveri, 2013: 379. solidarity strikes, general strikes, strikes to enforce collective agreement, etc. continue to be illegal. Moreover, the new law maintains the government’s power to ban any kind of strike for the sake of national security or public health. 21 Therefore, one of the three main areas of concern of the Joint Report about the new law was the »continued limitations on the right to strike.« 22 The new law also keeps the general ban on strikes in banking and urban transportation; however, with legal pressure from the main opposition party, the Constitutional Court lifted the ban in these industries in 2014. In November 2016, the government made a move to bypass that ruling by exploiting the state of emergency. Using an executive order, the government ratified an amendment to the Unions and Collective Agreement Law 6356 that extended the conditions under which the government can ban a strike. Hereafter governments can ban strikes in banking and urban transportation even only if the strikes»pose a threat to the economic and financial stability.« Before this amendment, governments could only ban a strike if it posed a threat to»national security or public health,« which they can interpret very broadly. 3. Trends in Workers’ Union Membership and Landscape In January 2013, the Ministry began to announce biannually the new union membership data based on an online system that is generally acknowledged as reliable. The number of union workers rose from 1 million to 1.8 million in five-and-a-half years between January 2013 and July 2018. The share of union workers among registered workers rose from 9.2 to 12.8 per cent in that period. The real union density of workers, which is the share of union workers among all workers including informally employed Turkish citizens, rose from 7.8 to 10.9 per cent during the same period. Graph 3 illustrates the union density of workers together with collective bargaining coverage between 2012 and 2017. The gap between the two has risen from 1.7 in 2012 to 3.2 points in 2017, which reveals that a growing number of union workers cannot use the right of collective bargaining. This is due largely to the problematic mechanism of authorization for collective bargaining, which kept intact by the new union 2012 law as noted above. 21. Caniklio ğ lu, 2013. 22. EU-Turkey Joint Consultative Committee, 2013, pp. 4. 10

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Graph 3. Union density of workers 12,0 10,0 8,0 6,0 4,0 2,0 0,0 2012 2013 2014 Union density of workers 2015 Bargaining coverage 2016 2017 Source: Ministry of Labor and Social Security(MoLSS) provides the number of union workers. Turkish Statistical Institute(TURKSTAT) provides the numbers of all employees and civil servants. The rise of union workers within five-and-a-half-years is significant. As mentioned above, public-sector subcontracted workers are a large share of this newly-added 800,000 union workers. A calculation based on the membership increase of the unions, which mainly recruited public-sector subcontracted workers in the last five-and-a-half years, yields an estimation that more than 500,000 of the new members are public-sector subcontracted workers. According to this calculation, the union density in the private sector(excluding public subcontractors) clearly stagnates. The newly-added 800,000 union workers since 2013 led to a significant transformation of the union landscape in the last years. If we assume that no union worker has changed her union, the share of confederations among new members can be illustrated as in Graph 4. As a result, shares of the three main confederations among union workers have changed significantly since 2013. While the share of Hakİş grew from 16.6 to 36.3 per cent, the shares of Türkİş and D İ SK decreased from 70.8 to 53.2 per cent and from 10 to 8.9 per cent respectively. The share of other minor confederations and independent unions fell from 2.6 to 1.6 per cent in that period. In July 2018 Türkİş , Hakİş and D İ SK had 958,618, 654,722 and 160,568 members respectively. As mentioned before, according to Turkish union law, unions can be established only at the level of industry, and the industries are defined by the government. The 2012 union law defined twenty different industries and these show very different union density rates, as Table 4 summarizes. MoLSS provides the number of formal workers and union workers in each industry. Table 4 presents approximate union density figures among each industry’s workers including informal workers. TURKSTAT produces the data of informal workers in eighteen industries defined according to international standards. MoLSS, however, defines twenty industries, some of which are defined in a rather arbitrary way as explained below. This information is calculated by adjusting the data of these different sets of industries. First, note the great difference in the number of workers employed in different industries as defined by MoLSS. While the communication industry employs nearly 62,000 formal workers, the curiously combined indus11

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Graph 4. Shares of Confederations Among New Members Btw. 2013–2018 D İ SK 8% Hakİş 61% Other 0% Türkİş 31% Graph 5. Shares of confederations among members% 80,0 70,0 Türkİş Hakİş D İ SK Others 60,0 50,0 40,0 30,0 20,0 10,0 0,0 2013-I 2013-II 2014-I 2014-II 2015-I Note: The suffix»-I« stands for January, while»-II« stands for July. 2015-II 2016-I 2016-II 2017-I 2017-II 2018-I 2018-II 12

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 4. Approx. union density including informal workers among 20 industries(2018, July) Union density including informal workers (approx.%) 1 Security serv. and arms manufacturing 37.4 2 Municipal work and general services 34.7 3 Banking& finance 33.5 4 Energy 26.5 5 Communication 19.3 6 Cement& glass manufacturing 17.2 7 Mining 16.8 8 Metal industry 15.8 9 Food manufacturing 10.7 10 Petro-chemistry 9.7 11 Transportation 8.7 12 Textile& garment manufacturing 7.8 13 Tree& paper manufacturing 7.8 14 Shipbuild., sea transp.& warehousing 7.6 15 Media 6.6 16 Health and social services 6.5 17 Agriculture, fishing, forestry 5.8 18 Trade, retail, office& education 4.4 19 Hospitality and entertainment 2.5 20 Construction 2.3 Total Union density among formal workers(%) 41.1 43.5 34.4 27.4 21.0 19.8 17.2 18.2 12.3 11.1 10.4 9.0 9.0 9.1 7.2 12.0 22.4 5.2 3.4 3.1 Number of formal workers of the industry 295,265 1,012,090 290,563 248,051 68,725 180,393 206,273 1,582,714 610,841 491,679 732,647 1,055,766 240,117 177,756 91,443 406,501 156,968 3,473,776 994,168 1,805,928 14,121,664 Number of union workers 121,397 439,903 99,877 68,047 14,417 35,755 35,480 287,428 75,162 54,696 76,465 95,278 21,569 16,219 6,596 48,686 35,108 180,787 33,873 55,412 1,802,155 try of»trade, retail, office and education« employs nearly 3.5 million formal workers. Just like the»trade, retail, office and education« industry, some other industries above need clarification, since their composition is not coherent with any international standard.»Security services and arms manufacturing« combines security services provided by private sector with arms manufacturing, where public employment has a significant share. Because public institutions have subcontracted their security services since the 2000s, there is a significant share of public-sector subcontracted workers in this industry.»Municipal work and general services« combines most of the services provided by municipalities with so-called»general services« which mostly implies janitorial and cleaning services. Historically, public workers employed by municipalities lead the unionization in this sector. Public-sector subcontracted workers working in municipalities or as cleaners in other public institutions have joined them since 2014.»Shipbuilding, sea transportation and warehousing« is another arbitrarily defined industry. While all other forms of transportation were united in 2012 in the transportation industry(eleventh in the above table), sea transportation was excluded. According to international standards, warehousing is also mostly considered to be a part 13

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 5. Top 15 worker unions with highest number of members(2018, July) 1 Hizmetİş 2 Türk Metal 3 Genelİş 4 Belediyeİş 5 Tez-Koopİş 6 Tesİş 7 Koopİş 8 Yolİş 9 Teksif 10 Öz Büroİş 11 Çelik İş 12 BAS İ SEN 13 Petrolİş 14 Öz Finansİş 15 Güvenlikİş Confed. Hakİş Türkİş D İ SK Türkİş Türkİş Türkİş Türkİş Türkİş Türkİş Hakİş Hakİş Türkİş Türkİş Hakİş Türkİş Date of est. 1979 1963 1962 1974 1962 1963 1964 1963 1951 2011 1965 1964 1950 2013 2011 Industry Municipal w. Metal ind. Municipal w. Municipal w. Trade, office Energy Trade, office Construction Textile Trade, office Metal ind. Banking Petro-chem. Banking Security serv. Number of members 286,356 209,429 76,925 74,904 63,969 60,627 59,021 53,070 48,196 46,291 45,121 41,242 37,796 37,380 34,205 Share among formal workers of the industry(%) 28.3 13.2 7.6 7.4 1.8 24.4 1.7 2.9 4.6 1.3 2.9 14.2 7.7 12.9 11.6 of transportation industry, while shipbuilding companies are usually included in the metal industry. 23 The top three industries with highest union density are »security services and arms manufacturing,«»municipal work and general services« and»banking and finance.« They all have a greater share of public workers and/or public-sector subcontracted workers. The next three industries(energy, communication and cement& glass manufacture) seem to share the experience of relatively recent privatization that did not result in the total erosion of the membership base under the rule of new private employers. The industry with greatest union density and highest private sector membership base is the metal industry, which includes the manufacturing of consumer durables and automobiles. There are 169 active worker unions as of July 2018 according to MoLSS. Table 5 illustrates the top fifteen unions with the highest number of members. Although 23. Akdemir and Odman, 2008, 77. the Hakİş union Hizmetİş has taken first place, Türkİş confederation maintains its supremacy with ten unions among the top fifteen. The only D İ SK union among the top fifteen is Genelİş . Unions of municipal work and general services dominate the top of the list. When the change in the number of members between January 2013 and July 2018 is scrutinized, the ranking alters(see Table 6). Although Türkİş dominates all other confederations regarding the number of member unions with the highest number of members, Hakİş unions had the highest increment of new members between 2013 and 2018 as demonstrated in Table 6. Hizmetİş welcomed more than 235,000 members in five years, which marks a nearly six-fold increase in its membership. The bottom of this list is also interesting as illustrated in Table 7. There we find those unions that lost the most members in the last five years among the 164 active workers’ unions. With 6,649 members, Teksif(a Türkİş union of textile workers) has lost the most members. Two independent unions, BANKS İ S and Turkonİş , which 14

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 6. Top 15 worker unions with highest number of additional members(2013–2018) 1 Hizmetİş 2 Türk Metal 3 Öz Büroİş 4 Genelİş 5 Öz Finansİş 6 Güvenlikİş 7 Belediyeİş 8 Koopİş 9 Özİş 10 Öz Sa ğ lıkİş 11 Öz Güven-Sen 12 Yolİş 13 Öz Ta ş ımaİş 14 Çelik İş 15 Tesİş Confed. Hakİş Türkİş Hakİş D İ SK Hakİş Türkİş Türkİş Türkİş Hakİş Hakİş Hakİş Türkİş Hakİş Hakİş Türkİş Date of est. 1979 1963 2011 1962 2013 2011 1974 1964 2011 2014 2015 1963 2013 1965 1963 Industry Municipal w. Metal ind. Trade, office Municipal w. Banking Security serv. Municipal w. Trade, office Security Health Security serv. Construction Transport. Metal ind. Energy Additional members 235,277 57,695 40,303 35,459 34,437 33,966 33,590 30,932 27,415 25,741 25,478 20,685 20,525 17,628 14,745 Members in July 2018 286,356 209,429 46,291 76,925 37,380 34,205 74,904 59,021 29,351 28,548 28,491 53,070 21,407 45,121 60,627 Table 7. Bottom of the list of unions according to difference in membership(2013–2018) 160 Tek Gıdaİş 161 G. Madenİş 162 Turkonİş 163 BANKS İ S 164 Teksif Confed. Türkİş Türkİş Indepen. Indepen. Türkİş Date of est. 1952 1946 1992 1983 1951 Industry Food manu. Mining Hospitality Banking Textile Additional members – 3,269 – 3,478 – 3,972 – 5,266 – 6,649 Members in July 2018 27,910 7,940 3,222 6,318 48,196 lost nearly 5,000 and 4,000 members respectively, and another Türkİş union, Genel Madenİş , which lost 3,478 members, follow Teksif in reverse order at the bottom of the list. One week after the coup attempt in July 2016, the government closed down the small workers’ union confederation Aksiyonİş with an executive order. The reason was the alleged allegiance of Aksiyonİş to a particular religious community that the government alleged was the power behind the coup. Aksiyonİş was founded in 2014 and its twenty unions had nearly 30,000members in January 2016. Dozens of executives of the confederation were detained in March 2018. 15

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY 4. A Closer Look at the Three Confederations Türkİş Formed in 1952, Türkİş (The Turkish Confederation of Workers’ Unions) is the oldest workers’ union confederation and historically it was mainly based on unions of public workers. Türkİş has adopted a position of»above party politics« and maintains a conciliatory attitude visà-vis governments as noted by Dinler. 24 Although its supremacy has been weakened, most union workers in Turkey are still Türkİş members. As noted above, the share of Türkİş member workers among all union workers have decreased from 71 to 53 per cent in the last five years. As Graph 6 reveals, Türk Metal is by far the largest union of Türkİş , comprising almost the quarter of the confederation’s membership. As Table 8 illustrates, Türk Metal has added more than 50,000 new members in the last five years. Nichols and Sugur rightfully state that»any 24. Dinler, 2012. ranking of the trade unions that are important to the modern sector would have to put Türk Metal at the top of the list.« 25 On the other hand, based on their fieldwork, they conclude that Türk Metal»embodies the worst aspects of the corporatist ideology and practice that characterised the early years of the Republic.« 26 The leadership of Turk Metal changed in 2009, which might have led to some moderate change in Turk Metal’s practices compared to the period of Nichols and Sugur’s research. However, in 2015, ten thousands of Türk Metal members in more than ten factories staged wildcat strikes to protest the union and the collective agreement it signed. 27 This massive and unexpected outburst of workers’ protest seemed to ignite a new wave of change in Turk Metal, which has become relatively more militant as indicated by the increasing number of protests it staged while organizing new workplaces in 2017. 28 25. Nichols and Sugur, 2004: 165. Similarly, Wannoffel(2011: 556) underlines that in contrast with other unions of the industry Türk Metal enjoys the support of MESS, the employer federation of metal industry. 26. I bid. 27. For details see Emek Çalı ş maları Toplulu ğ u, 2016; Korkmaz, 2015. 28. Emek Çalı ş maları Toplulu ğ u, 2018: 35. Graph 6. Türk-İş unions’ share among Türk-İş membership Others Demiryolİş T. Sa ğ l ı kİş T. Harbİş T. Çimseİş T. Madenİş Tek G ı daİş Güvenlikİş Petrolİş BAS İ SEN Teksif Yolİş Türk Metal Belediyeİş Tez-Koopİş Koopİş Tesİş 16

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 8. Türk-İş unions in detail, ranked according to the number of members Industry Date of est. 1 Türk Metal Metal ind. 2 Belediyeİş Municipal w. 3 Tez-Koopİş Trade, office 4 Tesİş Energy 5 Koopİş Trade, office 6 Yolİş Construction 7 Teksif Textile 8 BAS İ SEN Banking 9 Petrolİş Petro-chem. 10 Güvenlikİş Security 11 Tek Gıdaİş Food manu. 12 T. Madenİş Mining 13 T. Çimseİş Cement& glass 14 Türk Harbİş Security 15 T. Sa ğ lıkİş Health 16 Demiryolİş Transport. 17 Havaİş Transport. 18 Toleyis Hospitality 19 BASS Banking 20 T. Haberİş Communication 21 Ş ekerİş Food manu. 22 Tarımİş Agriculture 23 TÜMT İ S Transport. 24 G. Madenİş Mining 25 Türk Denizİş Shipb., sea tran. 26 Kristalİş Cement& glass 27 A ğ açİş Paper manu. 28 Selülozİş Paper manu. 29 T. Dok Gemiİş Shipb., sea tran. 30 Deriteks Textile 31 Basınİş Media 32 TGS Media 33 T. Ormanİş Agriculture Total 1963 1974 1962 1963 1964 1963 1951 1964 1950 2011 1952 1958 1963 1963 1961 1952 1962 1977 1972 1962 1963 1961 1949 1946 1983 1965 1949 1952 1947 1948 1963 1952 1975 Rank among unions of industry 1/12 3/12 1/15 1/5 2/15 1/10 1/15 1/6 1/6 1/10 2/9 1/7 1/7 4/10 2/7 2/9 3/9 1/11 3/6 1/3 3/9 2/7 4/9 2/7 1/6 2/7 2/8 3/8 3/6 4/15 2/4 3/4 3/7 Number of members 209,429 74,904 63,969 60,627 59,021 53,070 48,196 41,242 37,796 34,205 27,910 26,718 26,028 23,238 19,095 18,309 18,097 15,765 14,710 13,490 12,986 9,838 8,601 7,940 6,864 6,825 4,900 4,036 3,865 3,380 2,245 1,256 63 958,618 Formal share in the indus.% 13.2 7.4 1.8 24.4 1.7 2.9 4.6 14.2 7.7 11.6 4.6 13.0 14.4 7.9 4.7 2.5 2.5 1.6 5.1 19.6 2.1 6.3 1.2 3.8 3.9 3.8 2.0 1.7 2.2 0.3 2.5 1.4 0.0 Share among Türkİş members% 21.8 7.8 6.7 6.3 6.2 5.5 5.0 4.3 3.9 3.6 2.9 2.8 2.7 2.4 2.0 1.9 1.9 1.6 1.5 1.4 1.4 1.0 0.9 0.8 0.7 0.7 0.5 0.4 0.4 0.4 0.2 0.1 0.0 100 Difference in the# of members since 2013 57,695 33,590 13,650 14,745 30,932 20,685 – 6,649 3,111 10,404 33,966 – 3,269 2,517 5,886 2,104 13,831 3,746 4,600 1,753 4,264 – 2,713 – 2,681 – 115 1,826 – 3,478 2,328 78 2,454 768 1,620 1,576 454 439 -661 17

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Dinler underlines a development in Türkİş as a possible sign of change in the confederation. 29 Ten unions within Türkİş , called the Platform of Unification of Union Power (Sendikal Güç Birli ğ i Platformu), challenged the confederation leadership at the 2011 congress. Their candidates could not be elected to the executive board of the confederation, but they created a power bloc within the confederation and decided to act together on issues regarding labor legislation, international relations and women’s rights. They aimed to build a radical alternative to the dominant bureaucratic structure and conciliatory stance of the confederation. This initiative seemed to dissolve by 2015, likely due to changes in leadership of several unions of the Platform, such as of Havaİş in 2013. As one of the members of this Platform, TÜMT İ S(transportation industry), with the support of Global Unions, won an impressive series of organizing victories in largescale, international firms in the 2010s. 30 TÜMT İ S has attracted international attention not only because of its 29. Dinler, 2012. 30. See Birelma, 2018. inspiring victories but also because of the heavy prison sentences received by its fourteen Ankara branch officials. In April 2017, the Turkish Court of Appeal approved the decision of the local court and TÜMT İ S officials received prison sentences ranging from 1.5 years to 6.5 years»for recruiting new members and obstructing the freedom of conducting business.« 31 This irrational reasoning was shocking even in terms of the declining legal standards in Turkey. Hakİş Turkish labor unions are divided among ideological lines similar to those in France and Italy. 32 Embracing an Islamist ideology Hakİş (The Right Confederation of Workers’ Unions), established in 1976, is close to the unionism of social Catholicism in Europe as scrutinized by Hyman. 33 It embraced a non-confrontational, integrationist 31. ITUC, 2018: 27; Birelma, 2018: 222. 32. For France and Italy see Frege and Kelly, 2004: 38. 33. Hyman, 2000: 38–65. Graph 7. Hak-iş unions' share among Hak-iş membership Öz Taşıma- İş Öz Orman- İş Others Öz Güven-Sen Öz Sağlık- İş Öz- İş Hizmet- İş Öz İplik- İş Öz Gıda İş Öz Finans- İş Çelik İş Öz Büro- İş 18

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 9. Hak-İş unions in detail, ranked according to the number of members 1 Hizmetİş 2 Öz Büroİş 3 Çelik İş 4 Öz Finansİş 5 Öz Gıdaİş 6 Öz İ plikİş 7 Özİş 8 Öz Sa ğ lıkİş 9 Öz Güven-S. 10 Öz Ormanİş 11 Öz Ta ş ımaİş 12 Oleyis 13 Öz A ğ açİş 14 Enerjiİş 15 Limanİş 16 Medyaİş 17 Öz Toprakİş 18 Öz İ leti ş imİş 19 Öz Petrolİş 20 Öz İ n ş aatİş 21 Öz Maden-İş Total Industry Date of est. Municipal w. Trade, office Metal ind. Banking Food manu. Textile Security Health Security Agriculture Transport. Hospitality Paper manu. Energy Shipb., sea tran. Media Cement& glass Communication Petro-chem. Construction Mining 1979 2011 1965 2013 1976 1978 2011 2014 2015 2003 2013 1947 1980 2013 1963 2012 2014 2014 2003 2015 2011 Rank among unions of industry 1/12 3/15 2/12 2/6 1/9 2/15 2/10 1/7 3/10 1/7 1/9 2/11 1/8 2/5 2/6 1/4 3/7 2/3 4/6 2/10 3/7 Number of members 286,356 46,291 45,121 37,380 32,416 29,827 29,351 28,548 28,491 25,130 21,407 13,085 11,624 6,717 5,119 2,697 2,139 912 894 625 592 654,722 Formal share in the indus.% 28.3 1.3 2.9 12.9 5.3 2.8 9.9 7.0 9.6 16.0 2.9 1.3 4.8 2.7 2.9 2.9 1.2 1.3 0.2 0.0 0.3 Share among Hakİş members% 43.7 7.1 6.9 5.7 5.0 4.6 4.5 4.4 4.4 3.8 3.3 2.0 1.8 1.0 0.8 0.4 0.3 0.1 0.1 0.1 0.1 100 Difference in the# of members since 2013 235,277 40,303 17,628 34,437 11,445 12,821 27,415 25,741 25,478 1,350 20,525 6,728 4,244 6,661 1,979 2,137 1,380 730 433 567 495 approach based on harmony between employers and employees. 34 Hakİş leadership has championed a unionism based on the concept of social dialogue, and they are proud that the concept has become widely accepted among those in the labor movement lately. Many scholars claim that AKP, the governing party since 2002, supports Hakİş due to ideological affinities and the same scholars explain the rapid growth of Hakİş especially in 2010s with this support. 35 As noted above, 34. Dinler, 2012. 35. Çelik, 2015; Do ğ an, 2013; Erdinç, 2014; Gürcan and Mete, 2017. Hakİş ’ share among all union workers grew from 17 to 36 per cent since 2013, which corresponds to nearly half million new members. As calculated by Hakİş unions’ reports, more than 400,000 public-sector subcontracted workers seem to have become Hakİş members since 2013. Indeed, more than 200,000 of them have become member of Hizmetİş as early as 2016. 36 Hizmetİş , the biggest workers’ union in Turkey, which grew nearly six-fold in the last five years, comprises 44 per cent of Hakİş membership. 36. Hizmetİş , 2016: 4. 19

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Twelve new Hakİş unions were established since 2011 in eleven industries, none of which Hakİş represented before. Those which grew rapidly organized mainly public-sector subcontracted workers. The only exception is Öz Finansİş in the banking sector, which organized public workers employed by two public banks, Ziraat Bankası and Halk Bank. 37 On the other hand, Çelikİş , Öz Gıdaİş and Öz İ plikİş are the three biggest Hakİş unions with members predominantly employed by the private sector. D İ SK D İ SK(The Confederation of Progressive Trade Unions), founded in 1967, is close to the class unionism of three ideal types of trade unionism as defined by Hyman. 38 As Dinler underlined it remains»the most radical confed37. Nearly 45 thousand employees of public banks(mostly white-collar) are not civil servants but public workers. 38. Hyman, 2000. eration in its critical attitude towards government policy and labor rights.« 39 D İ SK lost a large portion of its membership to Türkİş between 1980 and 1991, when it was suspended by the 1980 military coup. D İ SK never recovered from this loss, and with the rise of Hakİş in the 2000s, it became the third confederation in terms of membership. Its share among union workers decreased from 10 to 8.9 per cent since 2013. D İ SK has the most unbalanced distribution in terms of the shares of unions within the confederation membership. Two unions, Genelİş and Birle ş ik Metalİş comprise 67 per cent of the confederal membership. Ten of twenty-one D İ SK unions have less than five hundred members. Only seven of twenty-one D İ SK unions have the authorization to sign collective agreements, while the rest do not exceed the industrial threshold. 40 39. Dinler, 2012. 40. Two of the seven authorized D İ SK unions do not actually exceed the industrial threshold, but they are deemed exempt from the threshold, because they used to exceed the threshold before the introduction of the new membership registration system in 2013. Graph 8. DİSK unions’ share among DİSK membership 2018 Others Nakliyat- İş Güvenlik-Sen Sosyal- İş Tekstil Genel- İş Lastik- İş Bir. Metal- İş 20

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Table 10. DİSK unions in detail, ranked according to the number of members 1 Genelİş 2 Bir. Metalİş 3 Lastikİş 4 Tekstil 5 Sosyalİş 6 Güvenlik-Sen 7 Nakliyatİş 8 Gıdaİş 9 D. Turizm İş 10 Enerji-Sen 11 Dev Sa ğ lıkİş 12 Dev. Yapıİş 13 Basınİş 14 C. Keramikİş 15 Limterİş 16 Tümkaİş 17 D. Maden-S. 18 Bank-Sen 19 Sine-Sen 20 Bir.Tar.Or. İş . 21 D. İletişim-İş Total Industry Date of est. Municipal w. Metal ind. Petro-chem. Textile Trade, office Security Transport. Food manu. Hospitality Energy Health Construction Media Cement& glass Shipb., sea tran. Paper manu. Mining Banking Trade, office Agriculture Communication 1962 1949 1949 1965 1966 2013 1975 1947 2011 2007 1974 1970 1947 1968 1976 1971 1959 1972 1978 2014 2013 Rank among unions of industry 2/12 3/12 2/6 3/15 4/15 5/10 5/9 4/9 4/11 3/5 3/7 3/10 4/4 4/7 4/6 5/8 4/7 6/6 9/15 4/7 3/3 Number of members 76,925 31,058 14,050 11,666 10,623 5,621 4,019 1,730 1,515 676 538 445 398 272 271 267 205 109 103 62 15 160.568 Formal share in the indus.% 7.6 2.0 2.9 1.1 0.3 1.9 0.5 0.3 0.2 0.3 0.1 0.0 0.4 0.2 0.2 0.1 0.1 0.0 0.0 0.0 0.0 Share among Hakİş members% 47.9 19.3 8.8 7.3 6.6 3.5 2.5 1.1 0.9 0.4 0.3 0.3 0.2 0.2 0.2 0.2 0.1 0.1 0.1 0.0 0.0 100 Difference in the# of members since 2013 35,459 4,997 6,882 1,463 3,377 5,593 1,230 – 58 1,508 397 -696 408 -148 263 137 – 326 27 – 384 88 58 7 Genelİş is a D İ SK union that has organized the most new members since 2013 by far. Similar to other unions in municipal work and general services, almost all of these new members seem to be public-sector subcontracted workers. While Lastikİş and Birle ş ik Metalİş ’ growth is based on private sector organizing, Güvenlik Sen’s new members are also mostly public-sector subcontracted workers. Internal debates have always been present in D İ SK. The general assembly of 2016 revealed that there are some serious tensions within the confederation, which seem to have further reduced the confederation’s already limited capacity. However, after the recent change of D İ SK leadership in May 2018, the disagreements seem to be in the process of resolution. II. Civil Servant Unionism 1. Amendments in the Union Law of Civil Servants in 2012 In Turkish law, civil servants’ individual and collective rights are regulated by entirely different laws than those for workers in the private and public sectors. As mentioned earlier, there are 3.1 million civil servants in Turkey as of 2018. After an amendment to the constitution in 2010 and an additional amendment to the Public Employee Unions and Collective Agreement Law numbered 4688 in 2012, civil servants acquired the right to collective bargaining 21

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY for the first time. Ever since, every two years a collective agreement has been signed. However, negotiations are limited to less than one month and the law maintained the ban on strike for civil servants. If the social partners cannot reach an agreement during the bargaining, a special arbitration board makes the final decision. The government determines most of the members of this board. Therefore, the Joint Report of EU-Turkey Joint Consultative Committee states that the salary fixing system as foreseen by Law No. 4688 for civil servants cannot be called collective bargaining. According to the report, this system does not involve»negotiation« but merely»consultation.« In some respects, the new law makes collective bargaining even more restrictive. For example, in the former law a majority of votes in each of the groups within the Collective Agreement Committee(30 members: 15 employers, 15 employees) was required to sign the agreement. However, according to the 2012 law, only the heads of both groups, representing the largest organizations, sign on behalf of 3 million civil servants. 41 In the same vein, the law maintained the ban on unionization of several civil servants such as judges, public prosecutors, policemen or military personnel. 42 Nearly 600,000 civil servants in 2018 were not legally eligible to join a union. As the Joint Report of EU-Turkey Joint Consultative Committee reveals, the three main civil servants’ union confederations did not consider the new law an improvement. 2. Changing Landscape of Civil Servants’ Unionism There are three main confederations of unions representing civil servants: Türkiye Kamu-Sen(est. 1992), KESK(est. 1995) and Memur-Sen(est. 1995). Civil servant unions are also divided along ideological lines. Türkiye Kamu-Sen is close to the nationalism of the National Action Party, KESK is left-leaning, while Memur-Sen has ties to Islamism. 43 In terms of politics and ideology KESK and D İ SK on the one hand, Memur-Sen and Hakİş on the other hand broadly correspond to each oth41. EU-Turkey Joint Consultative Committee, 2013, pp. 7. 42. Çelik, 2014: 297–298; EU-Turkey Joint Consultative Committee, 2013, pp. 6. 43. See Koç and Koç, 2009: 99; Koç and Koç 2014: 19. Specifically for Türkiye Kamu-Sen, see<http://www.ortadogugazetesi.net/haber. php?id=57807>. er 44 ; however, the same cannot be said for the pairing of Türkiye Kamu-Sen and Türkİş . While Türkİş embraces a position of»above party politics« and always try to build close relationships with the governing parties, Türkiye Kamu-Sen has a stronger ideological tendency. 67 per cent of nearly 2.5 million civil servants eligible to join a union were union members in 2018. This ratio has slightly increased from 62 per cent in 2003. The landscape of civil servants’ unionism, on the other hand, has changed even more tremendously than the worker unions. Memur-Sen had 42,000 members in 2002, which grew to more than one million in 2018. This corresponds to an increase from 6 to 60 per cent of union member civil servants as revealed by Graph 9. As of July 2018, Memur-Sen, Türkiye Kamu-Sen and KESK have 1,010,289; 394,423 and 146,287 members respectively. Despite its skyrocketing membership, ITUC refused Memur-Sen’s application for membership twice, in 2006 and 2011, claiming that Memur-Sen’s independence as a union is questionable. In 2018, MoLSS nominated Memur-Sen as the workers’ delegate to ILO Conference, because its membership exceeded that of Türkİş in 2017. Until 2018, Türkİş was the workers’ delegate of Turkey to ILO Conferences since it was established in 1952. Memur-Sen’s nomination provoked a reaction from Türkİş , D İ SK, KESK and Türkiye Kamu-Sen, which led to ITUC’s objection to ILO’s Credentials Committee. The Committee approved the objection and criticized the unilateral determination of the delegation by the government without the agreement of the most representative organizations. 45 After the coup attempt in July 2016 and the ensuing state of emergency, the government closed down Cihan-Sen, a small civil servants’ union confederation, with an executive order. Similar to the case of Aksiyonİş , the government charged allegiance of Cihan-Sen to a particular religious community it claimed was the power behind the coup. Cihan-Sen was founded in 2014 and its ten unions had 22,000 members in July 2016. KESK has suffered since 2012 from arrests and trials. In 2012, 72 KESK members and executives were arrested based on alleged relations with the PKK. In 2013, all of them were released, but the trial continued until 2017. 44. See Erdinç, 2014: 166. 45. ILO, 2018: 28–30. 22

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Graph 9. Shares of confederations among union member civil servants(%) 70,0 Memur-Sen Türkiye Kamu-Sen KESK Other 60,0 50,0 40,0 30,0 20,0 10,0 0,0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 During the state of emergency between July 2016 and July 2018 nearly 126,000 civil servants were fired. As of August 2018, almost 7,000 of them were reinstated. A KESK report in 2017 notes that 67 per cent of the fired civil servants were union members, while KESK’s share was 3 per cent(3,249 people). 46 The number of fired KESK members increased to 4,218 by January 2018. 47 Besides dismissals, other repressive measures towards KESK members continued during the state of emergency. In August 2017, 682 teachers in the southeastern provinces, all members of KESK, were subject to compulsory reassignment to other provinces for»participating in marches contrary to the Turkish Republic’s fight against terrorism.« 48 Furthermore, several KESK leaders were arrested during the state of emergency. According to a KESK report, 66 KESK members and executives were prisoners as of January 2018. 49 46. KESK, 2017. 47. KESK, 2018: 3. 48. ITUC, 2018: 27. 49. KESK, 2018: 8. Conclusion Unions in Turkey are neither helpless nor hopeless. There are 1.7 million unionized civil servants and 1.8 million union workers, which makes a total of 3.5 million union members in the country. Among nearly 20 million total employees(including informal workers), 3.5 million union members marks a significant social base and power resource. Despite its drawbacks, the mobilization and ensuing victory of public-sector subcontracted workers exemplifies the potential of the unions, especially if they apply cumulative pressure. Workers and unions forced the government to implement this huge de-privatization of nearly one million subcontracted workers. To assess the significance of this, one should recall that AKP was the power behind the skyrocketing subcontracting in the public sector. Due to ideological and competitive differences, the three confederations of workers’ unions did not coordinate or collaborate in this struggle. Nevertheless, they acted for the same cause and the cumulative pressure had positive effects. 23

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY Furthermore, wildcat strikes on an unprecedented scale by metal workers in May 2015 and by shoe-maker artisans and workers in September 2017 50 may be the symptom of a more common subjectivity among workers that unions can tap into and mobilize. In the former case, the massiveness and militancy at the very heart of the Turkish economy, in the latter case, the cooperation of Turkish, Kurdish and Syrian workers of subcontracting micro-enterprises against big brands were more than inspiring for the future of the labor movement. For workers, the authorization process for collective bargaining as the law prescribes is the major obstacle to unionization in the private sector. Without the right to strike, civil servants are still far from enjoying the right to collective bargaining. Demanding relevant and specific amendments to the union laws, or mobilizing for the annual tripartite negotiations for the workers’ minimum wage(which also affects civil servants’ wages) can de facto unite unions to put cumulative pressure on the government. Although the rise of authoritarianism and arbitrary rule in Turkey renders not only unions but all egalitarian and democratic forces of the country increasingly vulnerable, AKP is highly dependent on and sensitive to the support of the working class, and the party leadership is well aware of that. 50. Emek Çalı ş maları Toplulu ğ u, 2018: 52–54; Hak İ nisiyatifi, 2017. 24

ALPKAN B İ RELMA| TRADE UNIONS IN TURKEY References Akdemir, N./ Odman, A.(2008):»Tuzla Tersaneler Bölgesi’nde Örülen ve Üstü Örtülen Sınıfsallıklar,« Toplum ve Bilim, 113, pp. 49–89. Bakır, O./ Akdo ğ an, D.(2009):»Türkiye’de Sendikala ş ma ve Özel Sektörde Sendikal Örgütlenme,« Türkİş Dergisi, 383, pp. 88–95. Bengtsson, E./ Ryner, M.(2015):»The(International) Political Economy of Falling Wage Shares: Situating WorkingClass Agency,« New Political Economy, 20(3), pp. 406–430. Birelma, A.(2014): Ekmek ve Haysiyet Mücadelesi: Günümüz Türkiyesi’nde Üç İş çi Hareketinin Etnografisi, İ stanbul: İ leti ş im Yayınları. Birelma, A.(2017):»Subcontracted Employment and Labor’s Response in Turkey,« in Uncertain Times: Anthropological Approaches to Labor in a Neoliberal World, ed. P. Durrenberger, Boulder: University Press of Colorado. Birelma, A.(2018):»When Local Class Unionism Meets International Solidarity: A Case of Union Revitalisation in Turkey,« Global Labour Journal, 9(2), pp. 215–230. Caniklio ğ lu, N.(2013)»6356 Sayılı Kanuna Göre Grev Yasakları ve Grevin Ertelenmesi,« Çalı ş ma ve Toplum, 4, pp. 289–316. Çelik, A.(2013): A General Evaluation of Turkey’s New Act on Trade Unions and Collective Agreements, İ stanbul: Friedrich-Ebert-Stiftung. Çelik, A.(2014): Avrupa Birli ğ i, Sosyal Politika ve Türkiye, İ stanbul: Kitap Yayınevi. Çelik, A.(2015):»Turkey’s New Labour Regime Under the Justice and Development Party in the First Decade of the Twenty-First Century: Authoritarian Flexibilization,« Middle Eastern Studies 51(4), pp. 618–635. Çelik, A.(2018):»Bu ülkede grev hakkı ayaklar altındadır!« Birgün Newspaper; https://www.birgun.net/haber-detay/ bu-ulkede-grev-hakki-ayaklar-altindadir-202053.html. Çelik, Aziz and K. Lordo ğ lu(2006):»Türkiye’de Resmi Sendikala ş ma İ statistiklerinin Sorunlari Üstüne,« Çalı ş ma ve Toplum, 2, pp. 11–30. Dinler, Demet(2012): Trade Unions in Turkey, İ stanbul: Friedrich Ebert Stiftung. Do ğ an, G.(2013)»1980 Sonrası Sendikal Hareket: Türkiye’de Sendikacılı ğ ın Ku ğ u Ş arkısı«, in Tanzimattan Günümüze Türkiye İş çi Sınıfı Tarihi 1839–2014, ed. D. Çetinkaya and M. Alkan, İ stanbul: Tarih Vakfı Yurt Yayınları. Emek Çalı ş maları Toplulu ğ u(2015): İ sçi Sınıfı Eylemleri Raporu 2015, İ stanbul: Emek Çalı ş maları Toplulu ğ u. Emek Çalı ş maları Toplulu ğ u(2017): İ sçi Sınıfı Eylemleri Raporu 2016, İ stanbul: Emek Çalı ş maları Toplulu ğ u. Emek Çalı ş maları Toplulu ğ u(2018): İ sçi Sınıfı Eylemleri Raporu 2017, İ stanbul: Emek Çalı ş maları Toplulu ğ u. Erdinç, I.(2014):»AKP Döneminde Sendikal Alanın Yeniden Yapılanması ve Kutupla ş ma: Hakİş ve Ötekiler,« Çalı ş ma ve Toplum, 2, pp. 155–174. EU-Turkey Joint Consultative Committee(2013): Joint Report: Trade Union Rights Situation in Turkey, Brussels: EUTurkey Joint Consultative Committee; https://www.eesc.europa.eu/resources/docs/joint-report-on-trade-union-rights_final.doc. Frege, C./ Kelly, J.(2004):»Union Strategies in Comparative Context« in Varieties of Unionism: Strategies for Union Revitalization in a Globalizing Economy, ed. Frege, C. and J. Kelly, Oxford: Oxford University Press, pp. 31–44. Gürcan, E./ Mete, B.(2017): Neoliberalism and the Changing Face of Unionism, Cham: Palgrave Macmillan. Hak İ nisiyatifi(2017): Sayacılık Sektöründe Ya ş anan Hak İ hlallerine Dair Gözlem Raporu, İ stanbul: Hak İ nisiyatifi. Hizmetİş (2016): Hizmetİş Yayın Organı, 148, İ stanbul: Hizmetİş . Hyman, R.(2001): Understanding European Trade Unionism: Between Market, Class and Society, London: Sage Publications. ILO/ OECD(2015): The Labour Share in G20 Economies, Antalya: ILO and OECD. ILO(2018): Reports on credentials: Second report of the Credentials Committee, ILO: Geneva. ITUC(2018): ITUC Global Rights Index, International Trade Union Confederation; https://www.ituc-csi.org/IMG/pdf/itucglobal-rights-index-2018-en-final-2.pdf. KESK(2017): KESK’in Sesi İ hraç Özel Sayısı, Ankara: KESK; http://www.kesk.org.tr/wp-content/uploads/2017/06/I%CC %87hrac%CC%A7O%CC%88zelSay%C4%B1.pdf. KESK(2018): 15 Temmuz Darbe Giri ş imi Sonrasi Sivil Darbe Sürecinde Ya ş anan Hak İ halleri, Ankara: KESK; http://www. kesk.org.tr/wp-content/uploads/2018/01/19_01_2018-rapor.pdf. Koç, C./ Koç, Y.(2009): KESK Tarihi – I: Risk Alanlar Yolu Açanlar(1985–1995), Ankara: Epos. Koç, C./ Koç, Y.(2014): Memur-Sen’in Sicili(1995–2014), Ankara: Epos. Korkmaz, E.(2015): Unexpected Wave of Strikes in Turkish Automotive Industry, İ stanbul: Friedrich Ebert Stiftung. Özveri, M.(2013): Türkiye’de Toplu İş Sözle ş mesi Yetki Sistemi ve Sendikasızla ş tırma, Ankara: Üniversitesi Siyasal Bilgiler Fakültesi. Wannoffel, M.(2011):»Trade Unions in Turkey: past, present and future developments,« SEER Journal for Labour and Social Affairs in Eastern Europe, 4, pp. 545–569. 25

About the author Alpkan Birelma is an assistant professor in the Department of Humanities and Social Sciences at Özye ğ in University. As a sociologist, his research focuses on labor movements, workingclass subjectivity, social policy, and sociology of work. As an activist, he has volunteered in various labor unions and associations in Turkey since his undergraduate years. He is a member of Labor Studies Collective, which has published annual reports on working-class protests in Turkey since 2015. His last two publications in English were on the successful mobilization of subcontracted workers and on a revitalizing logistics workers’ union. Imprint Friedrich-Ebert-Stiftung| International Policy Analysis Hiroshimastr. 28| 10785 Berlin| Germany Responsible: Dr Marc Meinardus| Coordinator Trade Union Programmes Europe and North America Phone:+49-30-269-35-7744 www.fes.de/ipa To order publications: info.ipa@fes.de Commercial use of all media published by the Friedrich-EbertStiftung(FES) is not permitted without the written consent of the FES. The views expressed in this publication are not necessarily those of the Friedrich-Ebert-Stiftung. This publication is printed on paper from sustainable forestry. ISBN 978-3-96250-229-4