Labor_and_Employment_Relations_in_a_Globalized_World_New_Perspectives

144 T. Dereli this committee has proven to be an effective avenue for tripartite consultations, especially in preparing draft bills for the amelioration of the collective bargaining system of Turkey. 6.4 Flexibility 6.4.1 Types of Employment Contracts 6.4.1.1 Fixed-Term vs. Open-Ended Contracts; the Ongoing Controversy While open-ended employment contracts (contracts with an indefinite term) were generally the norm in Turkey in continual (permanent) employment relationships, pre-2003 labor legislation included provisions also on contracts with a definite (fixed) term, but only at a minimal level. The draft for the 2003 Act had to deal with this latter form of employment relationship as a flexible form of work generally more convenient for employers but at the same time with a motive to protect the worker and to prevent abuse. However both the definition and scope of fixed-term employment contract turned out to be somewhat different from the wording of the academic committee’s draft. Art. 11/I of Act no. 4857 provides that “an employ- ment contract is deemed to be open-ended (i.e. to have been made for an indefinite term) where the employment relationship is not based on a fixed term. A fixed-term contract is one that is concluded between the employer and worker, in written form, for work requiring a ‘specified term’ or based on objective conditions like the completion of certain work or the occurrence of a certain event” As seen from the way the Article is formulated, it seems that, in Turkey, there should be an objective reason necessary even in the first-time conclusion of a fixed-term contract or employment relationship. But in setting out the general principles and minimum requirements for fixed- term employment contracts, the main motive behind the EU framework agreement of the social partners on 18 March 1999 (which was given legislative effect through Council Directive 1999/70/EC) was the desire to improve the quality of fixed-term work by ensuring the observance of the principle of non-discrimination and to establish a framework to prevent abuse from the use of successive fixed-term employment relationships. In the exact text of the framework agreement there is no requirement for the establishment of objective reasons for the first use of fixed- term employment contracts, and most likely this was a deliberate choice as the framework agreement on fixed-term contracts obviously aimed at facilitating the use of such contracts (Hendricks 2004, pp. 6–7). The prevention of abuse is regulated only with regard to successive fixed-term contracts. Article 11 of Act no. 4857, by restricting the freedom of the parties to freely fix an expiration date in the first use of a fixed term contract, was apt to limit the scope of flexibility. While most EU jurisdictions have not required the presence of

6 Flexicurity and Turkey’s New Labor Act: Problems and Prospects 145 essential reasons for the first-time conclusion of fixed-term contracts, the Turkish Labor Act has gone further than the Directive in protecting the fixed-term worker by requiring an objective justification for making the fixed term contract. But this was not an option in the original draft of the academic committee—which simply mentioned for fixed- term contracts the phrase “whose duration is determined by the parties in terms of time and date”, rather than “for work of a specified term or duration.” The reasons for Art. 11 cited in the government’s text submitted to the Parliament also attests to the fact that the underlying rationale did not imply a requirement for the objective justification for the first conclusion of a fixed-term contract, although some academic circles had supported the opposite view. The prevention of abuse was regulated only with regard to successive “so-called chain” contracts. Art. 11/II provides that “an employment contract for a definite term must not be concluded more than once except when there is an essential reason which necessitates making repeated (chain) contracts. Otherwise the contract shall be deemed to have been made for an indefinite term from the very beginning.” This means that the first renewal will transform the contract into an open-ended one. “Chain contracts based on essential reasons shall maintain their status as fixed-term contracts.” In fact rulings of the Court of Appeals had confirmed this view even before the passage of Act no. 4857. It should nevertheless be noted that in the labor law of current Member States of the EU, a maximum total duration of successive fixed-term contracts or the number of renewals of such contracts which prevent those chains from being modified into open-ended contracts is often given. It appears that the Turkish Labor Act does not offer this flexibility in its present structure. The stipulation in Art. 11/I that all fixed-term contracts must be made in written form is clearly ill-conceived and contradictory to Art. 8/I which states that the written form is required only for fixed-term contracts of one year or more. According to one opinion, “combined with the presumption laid down in Art. 11 (first sentence), a fixed-term contract which is not in writing will be presumed to be an open-ended contract” (Hendricks 2004, p. 9). Yet this does not eliminate the said contradiction. The only solution to deal with the contradiction brought by Art. 11/I could be to treat the new “written form” notion of Art. 11 as a broad concept relating to the duty of the employer to provide the worker with a written document in cases where no written contract (fixed-term or open-ended) has been drawn up (Dereli 2012, p. 82). Parallel to the relevant provisions of Art. 5 on the principle of “equal treatment (non-discrimination)”, Art. 12 deals with the limitations on the distinction between fixed-term and open-ended employment contracts. An employee working under a fixed-term employment contract must not be subjected to different treatment in relation to a comparable employee working under an open-ended employment contract. “Divisible amounts for a certain time period relating to wages and monetary benefits to be given to a fixed-term employee must be in proportion to the length of time during which the employee has worked. In cases where seniority in the same establishment is treated as the criterion in order to have access to an employment benefit, the seniority criterion foreseen for a comparable employee

146 T. Dereli working under an open-ended contract shall apply to an employee with a fixed-term contract, unless there is a reason justifying the application of a different seniority criterion for an employee working under a fixed-term contract.” (Art. 12/I, II) These principles on equal treatment envisaged by Art. 12 are consistent with EU Council Directive 99/70. Since they were not included in previous legislation, the interpretation of the concept ‘comparable employee’ used to cause complications which were often resolved by expert opinion and court decisions. In the past the same principles were used in the computation of employees’ severance pay and entitlement to paid annual vacation. So practice and judiciary were familiar with applications. As for the restriction in the definition of “fixed term contract” which makes flexible practice somewhat difficult, it is advisable in a future amendment to the Act to carefully reconsider the definition of fixed-term contract to the extent that an objective reason need not be demonstrated in its first use, whereby the parties must be able to freely determine its duration and/or expiration date. 6.4.1.2 Part-Time Employment Contracts The previous legislation, Act no. 1475, did not carry any clarity on part-time work although this kind of employment was known in Turkish practice. The void was filled by court decisions. With a view to protect as well as to encourage part-time work for employment creation, Act no. 4857 brought provisions on this matter consistent with relevant EU norms. Domestic services, cleaning and preparatory work, work by company physician, lawyer, work by women and students in supermarkets were typical examples of part-time work in Turkey. Art. 13/I of Act no. 4857 has provided that “the employment contract is to be considered ‘part-time’ if the workers’ normal weekly working time has been fixed considerably shorter than a comparable worker working full-time.” In an attempt to clarify the meaning of “working time having been fixed considerably shorter than the normal weekly working time”, statement of reasons for Article 13 refers to “work which is less than at least two-thirds of the normal weekly working time.” Aiming to ensure non-discrimination, Art. 13/II stipulates that “an employee working under a part-time employment contract must not be treated differently in comparison to a comparable full-time employee solely because his contract is part- time, unless there is a justifiable reason for differential treatment. The divisible benefits to be appropriated to a part-time employee in relation to wages and other monetary benefits must be computed according to the length of his working time proportionate to a comparable employee working full-time. The comparable employee is the one who is employed full-time in the same or similar job in the establishment. If there is not such an employee in the establishment, an employee with a full-time contract performing the same or similar job in an establishment which falls into the same branch of activity shall be considered to be the compa- rable employee” (Art. 13/II, III) Thus a part-time employee will have access to all the fringe benefits (i.e. bonuses, premiums, allowances, holiday pay) granted to

6 Flexicurity and Turkey’s New Labor Act: Problems and Prospects 147 full-time employees but only in terms of divisible amounts proportionate to the length of his working time. Part-time employment may be based on an open-ended or fixed-term contract. In the computation of the worker’s length of service for various entitlements (for example to severance and notice pay) the total period between the beginning and ending dates of his employment contract shall be considered rather than the total working time actually worked. And termination of part-time employment contracts is subject to the same rules foreseen for full-time employees. Part-time workers have access to all the entitlements of freedom of association and collective bargaining. Adhering to the principle of equal treatment, their remuneration may be determined freely by the parties, provided that the wage to be paid for a given time period must not be less than the legal minimum wage corresponding to the same time slice. In line with the Council Directive 97/81, Art. 13/IV provides: “if there are vacant positions suited to their qualifications, employees’ requests to move into full-time from part-time jobs or vice versa must be given due consideration; and vacancies must be announced without delay.” The underlying motive is to further employ- ment expansion by encouraging the transfer of part-time workers to more stable employment. Yet the above-mentioned Matra Project did not find this article concrete or sufficient enough to meet the expectations of the said EU directive (Hendricks and Sengers 2006, pp. 92–93). 6.4.1.3 On-Call Work Contracts On-call work or call work is a special form of part-time employment. Differing from part-time work mainly with its irregular and casual character, it was not unknown in past Turkish practice either. It was referred to as a unique kind of part-time work in various decisions of the Court of Appeals. The 2003 Labor Act has chosen to regulate it as a flexible working arrangement to be accompanied by certain protective measures. As Art. 14 states, “employment relationship based on the performance of work by the worker upon the emergence of a need for his service, as agreed on by the parties in the written employment contract, is an on-call part-time contract. If the worker’s working time has not been determined by the parties in terms of time slices such as a week, month or year, the weekly working time shall be considered to have been fixed as 20 hours.” The worker must be paid his wages irrespective of whether or not he is engaged in work during the time announced for on-call work. Implicit here is the notion of flexibility in both “working time and wage”, construed as meaning the possibility of the parties’ agreeing to a period longer or shorter than 20 h. However some writers believe that the “20 hours” should be treated as a binding minimum rather than a fixed mandatory time slice, meaning that shorter hours may not be decided in individual contracts or collective agreements. When the employer needs the service of the worker, he must make the said call at least 4 days in advance unless the contrary has been decided. If the daily working time has not been decided in the contract, the employer must engage the worker in

148 T. Dereli work for a minimum of four consecutive hours at each call (Ar. 14/III, IV). Legislation has thus recognized freedom to the parties of on-call work while at the same time bringing protective measures to apply in the absence of explicit agreement. However labor was quite critical of on-call work during the legislative process on grounds of its alleged infringement on secure and stable employment. In practice due to the unions’ refusal to accept it as an acceptable employment form, on call work does not seem to have gained much acceptability in the formal and organized sectors of Turkey. Besides, where applied, the official inspection mech- anism falls short of implementing a sufficient monitoring of such work both in terms of the sufficient collection of social security contributions and implementa- tion of working time arrangements. 6.4.1.4 Temporary Employment Contract The original draft of the “academic committee” had foreseen in Article 93 the creation and licensing of temporary work agencies, but as a result of strong reactions by labor against this proposal, Art. 93 of the proposal was deleted from the text of Act no. 4857. In Art. 7 of the 2003 Act at present, therefore, there is a legal basis only for the staff leasing process between the employers within the structure of a holding company or the same group of companies. In Turkey this was a known practice anyhow whereby employers of related companies occasionally met their needs for skilled labor on joint projects or in their efforts to head off layoffs. Act no. 4857 has brought clarity to this somewhat controversial issue. From the wording of Art. 7, however, it is inferred that temporary work agencies which normally assign their workers to third parties (user enterprises) on a professional basis are not permitted to function. As Art. 7 provides, “a temporary employment relationship is established when the employer transfers his employee, after obtaining his written consent, to another establishment within the structure of the same holding company or the same group of companies. . .Temporary employment relationship which must be concluded in written form may be established for a period not to exceed six months, and if needed, it may be renewed twice, and the worker’s consent must be taken at each renewal.” Therefore the contract may be effective only up to 18 months. Art. 7 also provides other details relating to this relationship (i.e. certain limitations on implementation, duty of equal treatment, health and safety training, joint liability requirements for wages and social security contributions, union and collective bargaining rights). However, employers who seem happy for the legal clarification of this practice complain about the relatively short time span of using it envisaged by the Labor Act. On the other hand, the failure to legislate on temporary work agencies to operate with a view to create employment has been a matter of controversy, deplored both by the government and employer circles as a factor barring the further flexibilization of the Turkish labor market. However this type of triangular tempo- rary work relationship is still practiced as a commercial activity by various private

6 Flexicurity and Turkey’s New Labor Act: Problems and Prospects 149 employment agencies but informally and with no legal protection for the worker. In so doing these agencies rely for the most part on the general provisions of the Obligations (Contracts) Act, or the principal employer-subcontractor relationship, which by way of its quite different nature and the restrictive clauses brought by Article 2 of the 2003 Labor Act leads to controversial court cases. In an attempt to legalize temporary work agencies, a draft bill annexed Art. 7-A to Act no. 4857 in order to regulate temporary work agencies in accordance with the guidelines of EU Directive 2008/104/EC. But due to active lobbying by labor unions the President vetoed Art. 7-A on a number of grounds, among them basically for the void in respecting the principle of equal treatment. Following this, the draft for the Act of February 2011, no. 6111 revived Art. 7-A with more elaborate provisions on non-discrimination, but this proposal too was struck out during the final legislative process, again mainly for political reasons. Labor unions call temporary work “slave labor” and assert that temporary workers are doomed to employment mainly in less paying, substandard jobs with no employment protection. In addition to the difficulties encountered in organizing them, unions believe that this category of work clearly violates the Constitutional principle of the State’s duty to provide a “just wage” for the working people. These views notwithstanding, the submission of this issue to the agenda of the Parliament in the near future is quite likely. As a matter of fact, the Ministry of Labor and Social Security has prepared a new draft on temporary work agencies in 2011 and 2013, submitted to the assessment of social partners. However, texts of these drafts have also raised a few controversial points. It must be noted that in this effort matters like the optimum number of temporary workers to be employed in an establishment, their right to organize and bargain collectively, designation of the industry branch wherein they can organize, their job security as well as the “equal treatment” principle should be given due consideration. 6.4.2 Working Time Arrangements 6.4.2.1 Normal Working Time and Flexibility Measures The maximum weekly normal working time which was 48 h before was reduced to 45 h by Act no. 2869 in 1983. But the working time arrangements of the Labor Act, no. 1475, were quite rigid and the need for flexibilization was often voiced by employer circles during the 1990s. In responding to such demands as well as considering the relevant EU directives, the new Labor Act did bring various flexibilization measures and different types of more flexible employment contracts. In general terms, weekly working time is 45 h maximum which the parties may freely agree to reduce. Unless the contrary has been decided, 45 h shall be divided equally by the number of days worked at the establishment (Art. 63/I). So, as in the previous system, working time may be divided by the number of days on an equal

150 T. Dereli basis. If work is done 6 days of the week, daily working time is 7.5 h. If work is only 5 h on Saturday, daily working time shall be (45 À 5 ¼ 40:5 ¼ 8) 8 h. Provided that the parties have so agreed, however, working time may be distributed over the days of the week in different modalities, on condition that the daily working time must not exceed 11 h in any case. So within these limits, the practice of “compressed work week” is possible. But within a period of 2 months, the average weekly working time of the worker must not exceed the weekly normal working time (45 h); otherwise the employer must execute a transaction called the “balancing act”. Thus in so far as the weekly average does not exceed 45 h, there shall be no need to apply the so-called balancing even if the daily working times may have varied. This balancing period may be increased up to 4 months by collective agreements (Art. 63/II). Bringing flexibility to working time in this manner is consistent with EU directive 93/104. In this connection, the possibility exists to raise the weekly working time above 45 h in some weeks, provided that within a 2 month period the average weekly working time of the worker must be equaled to 45 h. Thus the implementation of “compressed workweek” is applicable. The workers’ consent may be obtained by way of new or changed individual employment contracts or by jointly inserting pertinent provisions into the estab- lishment’s personnel regulations or collective agreements. As for the workers concerned, balancing can be carried out on the basis of the whole establishment or in a section thereof or on the basis of individual employees. The balancing period is the time span (which may vary between 2 and 4 months) beginning from the first day of the application of the compressed week. The employer is held to have the imitative here through implicit or explicit agreement of the employees or through joint decision making by way of collective agreement. Because working time is a non-monetary issue, changes in the distribution of working time and adjustments to be made in respect to the balancing act are binding not only on the members of the signatory union but on all employees in the establishment in view of Art. 31 of Act no. 2821 on labor unions. Looking into major collective agreements, it is safe to say that, in contrast to many flexibility provisions of the 2003 Labor Act which require the workers’ consent, ‘compressed work week’ is, relatively speaking, more easily used, either by individual or by collective agreement. The collective agreement, if there is one, shall take prece- dence in any case, as it does have an automatic and binding effect on individual employment contracts. A collective agreement provision may foresee the distribu- tion of weekly working time to workdays –not to exceed 11 h daily – equally or on a differential basis. The ceiling on the length of the balancing act is of course mandatory and cannot be altered. In the balancing process the employee is not entitled to overtime pay even when his weekly working time exceeds 45 weeks in some weeks due to the application of “compressed workweek” (Art. 41/I). Because of the flexibility it carries, the “compressed workweek” is more beneficial to the employer than to the worker. However, the necessary adjustments to be made in the balancing process are complicated and require expertise in time management. Turkish firms were ill-equipped for such new functions. Employers would prefer a longer time span for the ceilings concerned. In fact the draft on Act no. 6111

6 Flexicurity and Turkey’s New Labor Act: Problems and Prospects 151 foresaw bringing an increase to the time span within which the balancing act is to be executed, but due to pro-labor reactions the proposed change was deleted from the draft text. Also, in view of Art. 67/II, “depending on the nature of the work or activity, the beginning and ending times of work may arrange differently for workers.” Thus, making flexi time arrangements is possible. 6.4.2.2 Overtime Work and Work at Extra Hours Overtime work is work which, under the conditions specified in Act no. 4857 (i.e. normal overtime work, compulsory overtime work, overtime work in emer- gency situations), exceeds 45 h a week. In cases where the balancing act mentioned above is applied, work which exceeds a total of 45 h a week shall not be considered overtime provided the average weekly working time of the worker does not exceed the normal weekly working time (Art. 41). It follows from this definition that the criterion for overtime work is work that exceeds 45-h weekly (but only up to 3 h) rather than work which lasts longer than the daily working time, which was the case under the previous Labor Act, no. 1475 when any work up to 3 h daily in excess of the daily working time, and not to be carried out more than 90 days in a year, used to be treated as overtime. The 2003 Labor Act also brought a new concept, namely “work at extra hours”. As Art. 41/III provides, “in cases where the weekly working time has been set by the employment contract at less than 45 h, work that exceeds that agreed upon average weekly working time . . . and which may last only up to 45 h weekly is deemed to be work at extra hours.” In normal overtime work, the worker’s consent must be obtained. The require- ment to receive the permission of the regional directorate of labor which existed under Act no. 1475 no longer exists. The worker’s consent must be obtained also in the case of work at extra hours. If the worker who has given his approval refuses to do overtime work, the employer may break the employment contract for just cause according to Art. 25/II (h) unless there is a collective agreement provision to the contrary. Art. 41/VIII states that the total overtime work shall not be more than 270 h in a year. As stated above, Act no. 4857 does not refer to any daily ceiling with respect to the maximum length of overtime The 3 h maximum that was foreseen in the draft bill was struck out during the legislative process, as this would conflict with the notion of “compressed week”. Recalling that even in the application of the “com- pressed work week” the worker’s maximum working time must not exceed 11 h daily (Art. 63/II), the 11-h daily maximum shall be included in the overtime hours that can be worked in a day, provided the total overtime worked in a year must not exceed 270 h. Unlike Act no. 1475, Act no. 4857 has not foreseen any penal sanctions for overtime work done in excess of the 270 h maximum. The only remedy is legal action where courts compel the employer to defray legal overtime wages in excess of this maximum. Thus, many stipulations on overtime work which

152 T. Dereli had to be respected for the past six decades or so were relaxed considerably by the 2003 Labor Act. Wages for each hour of overtime work must be remunerated at one and half times the normal hourly (or corresponding piece) rate. In work at extra hours, each extra hour shall be remunerated at one and a quarter times the normal hourly rate (Art. 41/II, III). If the employee who has worked overtime so wishes, instead of receiving overtime pay, he may use as free time one and half an hour for each hour worked overtime and 1 h and 15 min for each extra hour worked (Art. 41/IV). The employee shall use the free time within 6 months, within his working time and with no deductions from his wages. Though workers are inclined to accept doing overtime with a view to increase their earnings, the high degree of flexibility brought to overtime work is certainly more beneficial to the employer. Thus, the new Labor Act has brought a consider- able degree of flexibility to overtime work which is generally more beneficial to the employer. Having his workers do overtime in excessive amounts will increase the total production and profits of the employer under normal conditions rather than the labor productivity in the sense of output per man hour worked. In a quantitative field study conducted on the chemical and petroleum-plastics firms in Turkey, non-union firms were found to be more productive than unionized firms. In interviews, respondents referred to the strict application of working time regula- tions in unionized plants because of the union’s monitoring function while non-union firms were much more flexible in the application of overtime work.2 It seems that, despite all these detailed regulations, the new overtime arrangements are also quite difficult to control and monitor through the official inspection mechanisms. The maximum ceilings are usually violated partly due to the absence of penal sanctions for violators. 6.4.2.3 Short-Time Work In an attempt to head off mass dismissals in crisis situations, many employers even before the year 2003 allowed their workers to take paid or unpaid leaves or put them on a shortened workweek. To provide a legal framework for such initiatives, Act no. 4857 envisaged new rules in Article 65 under the title “short time work and its pay.” This provision was later transplanted, by way of a reference made, to Act no. 4447 on Unemployment Insurance. It provides that ‘in cases where work is 2 Responses to interview questions indicate the extent of violations in working time regulations, ranging from rest breaks to overtime rules. “In contrast to the previous Labor Act no. 1475, Act no. 4857 loosened the regulation of overtime work, thus paving the way for making workers do overtime in excessive hours more easily. In firms where there is a labor union which oversees the application of legal rules and collective bargaining terms, this is more difficult for the employer. Obviously this makes it possible for nonunion firms to reach higher production levels. In the so-called balancing period no overtime premium is required, provided that the daily working time does not exceed 11 hours.”

6 Flexicurity and Turkey’s New Labor Act: Problems and Prospects 153 suspended or short-time work is performed for at least 4 weeks due to a general economic crisis or force majeure, employees shall be paid short-time work benefits from the Unemployment Insurance Fund corresponding to the time not worked.’ To call on short-time work, however, the employer must fulfil certain procedures. He must communicate this matter, along with the reasons, immediately to the Employ- ment Organization, ISKUR, and to the union signatory to the collective agreement, if there is one. The acceptability of the request shall be decided by the Ministry. The relevant methods and procedures were indicated in the regulations of the Ministry of Labor and Social Security published in 2004 and 2009. Among the flexibility measures of the Labor Act, short-time work proved to be a widely used practice in times of economic crisis and especially after the 2008–2009 amendments (MESS I˙s¸veren Gazetesi 2010). The regulations of the Ministry brought clarity to implementation. For example, the criterion used in the definition of ‘short time work’ expressed as ‘time considerably shorter than the normal working time’ should be understood as “reduction of working time at a ratio of at least one-third of normal working time”. The notion ‘general economic crisis’ would refer to economic recession of a national or international scope as well as a sectoral crisis which impacts the national economy to a considerable extent. Force majeure refers to unanticipated events (e.g. fire, earthquake, flood or war) which are not attributable to the employer’s mismanagement and which partially or entirely disrupt the operation of the plant. The Ministry shall make a determination as to whether or not the case qualifies as an economic crisis or force majeure. Recourse to judicial process is possible against the determination of the Ministry. Because short-time work was supported by workers as well as employers, it was encouraged further by the publication of new regulations mentioned above. By annex articles to Act no. 4447, the maximum period of the availability of short-time benefits which was 3 months was elevated to 6 months and the rate of benefit was accorded a 50 % increase for the years 2008, 2009 and 2010. For force majeure short-time work must not exceed 3 months in any case. In order to be entitled to benefits, the worker must meet the conditions required for having access to unemployment benefits both in terms of his length of employment and the number of days for which unemploy- ment insurance contributions should have been paid. In the event of force majeure, payment of benefits shall commence after the lapse of the 1 week period envisaged in subsection III of Article 24 on termination of the contract by the worker for force majeure and Article 40 (that is, payment of half-wages up to 1 week). The employer may close the plant within the duration or at the end of the short-time work practice. Afterwards, an amendment to the Unemployment Insurance Act by Act no. 6111 of February 2011 bringing amnesty and restructuring to certain social insurance debts enlarged the scope of short-time work to include also cases of regional crisis situations and increased the rate of daily benefits to 60 % of the daily gross earnings of the worker computed as the daily average of the worker’s last 12 months’ earnings. Further, the government was empowered to extend the duration of short-time work benefits up to 6 months as well as to decide whether or not they will be deducted from unemployment benefits. Still further, a new Regulation was issued on 30 April 2011 replacing the regulation of 13 January 2009 and further

154 T. Dereli clarifying the relevant provisions on short-time work. The employer contemplating to implement short-time work for the above-stated reasons shall communicate his request to I˙S¸ KUR as well as to the labor union concerned, identifying also the workers who will do short-time work. The employer’s request shall be evaluated and then its feasibility decided upon by the management board of I˙S¸ KUR rather than by the inspectors of the Ministry of Labor -which was the case under the previous regulation-. To have access to benefits, the worker’s application is not required. The duration of the benefit is equal to the length of short-time work actually carried out, provided that it shall not exceed 3 months, to be computed monthly on the basis of the length of short-time work actually done. Its amount is the same as in the aforementioned Act no. 6111. 6.4.2.4 Compensatory Work The rigidities which existed in the Labor Act of 1971, no. 1475, did not permit employers to call on compensatory work in order to offset the time lost due to recession, force majeure, or similar reasons. The flexibility foreseen by Act no. 4857 makes this possible now. Article 64 of the new Labor Act provides that “in cases where time worked has been considerably lower than the normal working time, or where operations are stopped entirely due force majeure, or on the days before or after national and public holidays, or where the employee has been granted time off upon his/her request, the employer may call on compensatory work within 2 months in order cover the time lost due to unworked periods”. The Regulation on Working Time provides that the employer must specify according to which reason(s) indicated in Article 64 compensatory work shall be executed. Compensatory work shall not be considered overtime work; the wage the employee is entitled to receive is his normal wage. It may be performed before or after the emergence of the said conditions. Compensatory work shall not exceed 3 h daily, and in any case it must not be more than the maximum daily working time (11 h). It shall not be carried out on statutory or contractual holidays (Art. 64/II, III). Notwithstanding this restriction, it should be noted that, for employees who work 5 days of the week, thereby using the 6th day (say, the Saturday) as free time, compensatory work may be executed on Saturdays provided the daily working time on Saturday must not exceed 11 h. 6.5 Research on Flexibility Implementations Turkey seems to have been quite successful in eradicating the adverse effects of the global crisis. Among the G20 countries, the highest rate of employment creation was achieved by Turkey. The unemployment rate which had risen to 14 % in 2009 diminished by March 2011 to 10.8 %, which was the pre-crisis level. Of course it is

6 Flexicurity and Turkey’s New Labor Act: Problems and Prospects 155 difficult to measure how much of that employment creation was due to the impact of the new Labor Act. But the dominance of the informal labor market is a serious problem in Turkey, as emphasized in the progress reports on the projected EU accession of Turkey. Research studies show the share of informal employment as being 41.3 % of total employment. This is one of the reasons why overregulation of working conditions is often criticized for its adverse effects on the possible expansion of the formal market. As for flexibility, the limited research done so far shows that, perhaps after short- time work in its present form following the recent amendments, compensatory work is the most widely used flexibility measure in Turkish practice. Research conducted by MESS, the largest employers’ union in metal working industries, has yielded disappointing results concerning the effects of flexibility measures of the Labor Act no. 4857 (Metal Employer’s Union of Turkey 2010). MESS research has shown that in 97 % of the establishments the working time is still distributed equally by workdays, so the practice of compressed workweek is only minimal; the balancing act is executed in only 20.1 % of the workplaces. Short-time work is not practiced in 97 % of establishments. (It should be noted, however, that this research was done before the new flexibility rules were brought in 2009 and after.) On the other hand, the most common and popular flexible measure applied in MESS-affiliated estab- lishments proved to be compensatory work. Among the reasons cited by the MESS report which account for the insufficient level in the application of flexible work forms are the lack of awareness on the benefits of flexibility and the absence of applicability of short-time work in regional and sectoral crises—which, as mentioned above, were later provided for and led to more extensive use of short-time work. As cited in Su¨ral (2012), a more recent study made jointly by the Ministry of Labor and Social Security and the Turkish Personnel Management Association (PERYO¨ N) on 216 firms found that, of the 4,55,000 employees, only 5.9 % were working flexible (Turkish Personnel Management Association 2012). Of the female respondents, 3.4 % stated they were working according to a flexible arrangement; the ratio of male respondents was 2.5 %. 30.5 % of the respondents indicated that flexible work models help reduce their labor costs; 26.6 % mentioned increases achieved in their performance levels and competitiveness; 26.6 % referred to flexible work as being instrumental in adjusting working time to changes in their work load. Other reasons cited by the respondents are as follows: flexible working conditions are more compatible to working conditions desired by new generations; they are better able to solve work-family conflicts as employees are not restricted by official working hours; and they are helpful in planning the monthly working time of the enterprise. In response to the possible reasons accounting for the limited use of flexible arrangements, it was found that the 56 % of the establishments covered were not acquainted with the meaning as well as the pros and cons of flexible work models; 27 % believed that flexible work applications lead to losses in wages and employee benefits; 24.7 referred to loopholes in legislation; 26.3 said that flexible work was

156 T. Dereli not compatible with the nature of their operations. In terms of ranking the mostly used flexible work types, “compensatory work” topped the list with 24 % of the firms in the sample, followed by “short-time work” with 18 %. Services and office work is the branch of activity where flexible arrangements were most widely implemented. The research report concludes with a strong emphasis on the need to revise Act no. 4857 with a view to further flexibilize its relevant provisions on flexibility. 6.6 Concluding Remarks: Evaluation and Prospects The various problems mentioned above notwithstanding, one could safely say that the 2003 Labor Act of Turkey, a product of tripartite agreement based on social dialogue and mutual trust, is a piece of legislation more modern compared to the labor acts of the pre-2003 era. But has this Act been able to realize its avowed goal of achieving the right regulatory balance between flexibility and security? The answer can be only a qualified “yes”. In fact the 2003 initiative qualifies merely as a step towards approaching that balance. Comparing the rather limited scope of job security against the somewhat overregulated flexibility measures (except for the relaxed rules on overtime work), a reasonable degree of balance seems to have been struck. In passing, here it should be noted that, in order to approach that balance, the social partners must, to the extent possible, strive to implement flexicurity through the collective bargaining process between themselves. In so far as the successful transplantation of contemporary EU and ILO labor law into Turkish legislation is concerned, the draft of the “academic committee” was certainly successful (Hendricks 2004, p. 117). But one can hardly say the same thing for the final text. A number of intervening factors led to “unanticipated consequences” for the academic committees’ work and the stakeholders concerned, “unintended dysfunctions” so to speak. Due to intensive lobbying of the social partners before and during the legislative process, some articles were struck out or modified considerably; definition of the fixed-term employment contract, scope of job security: deletion of the proposed ‘workers representatives’ system from the final text altogether, obstruction of the collectively bargained arbitration in job security disputes by the Constitutional Court’s ruling, failure to legislate on the reform of the severance pay are only a few examples one could cite in this regard. The Act’s job security dimension was weakened by employers’ pressures while labor was responsible for the restrictions brought to flexibility measures (e.g. temporary work agencies, fixed-term work, and obstruction of the proposed reform in the severance pay system). How a relatively weakened labor movement, weaker in comparison to the pre-1980 times, could lobby so effectively is an interesting question. It should be recalled, however, that the preparation and enactment of that legislation and subsequent amendment initiatives coincided with the holding of national elections.

6 Flexicurity and Turkey’s New Labor Act: Problems and Prospects 157 Achieving a right balance in mathematical terms is certainly inconceivable in social relations. One can only think in terms of an ideal model for which the Dutch system is often cited as an example (Hendricks 2004, pp. 11–14). But one must recall that the Dutch model was started and developed in a milieu where employ- ment security was already solid and labor unions relatively stronger. Learning from the lessons of the last decade’s practices and considering the criticisms raised by the social partners, the Ministry of Labor and Social Security, following consultations with the social partners in the Tripartite Consultation Board in 2011 and 2013, has come up with new amendment proposals. At the time of this writing, two such proposals were already in place. Of these, in an effort to bring new atypical work categories, the draft amending Act no. 4857 has foreseen to include in its scope also “distant work” (or work from home, telework or telecommuting), as well as “on call work” and “work sharing”, both formulated now in more precise and concrete terms. But there is again a tendency to regulate these issues in terms stronger than the relevant EU norms. For further clarity in “flexitime”, the new drafts have defined the notion of “core time” which shall be determined by the employer, and stated that the worker may distribute the working hours which fall outside “core” time to the worked days of the week, provided that the daily working time shall not exceed 11 h; the worker can spend these times without working, but in this case a balancing act must be implemented within 1 month. However in this draft text the requirement for the presence of an essential or objective reason in the first- time conclusion of a fixed- term contract still remains. The second and third draft proposals aim to design temporary work to be carried out on a professional basis by licensed private employment agencies. The draft enumerates the cases for which such workers can be posted to user undertakings. These are 1. in order to temporarily replace the permanent workers of the estab- lishment who are unable to perform their duties for certain reasons (e.g. maternity, military duty, illness, etc.).; 2. in the case of an unforeseen increase in the workload of the undertaking; 3. for short term works of an intermittent nature; 4. in cases of an urgent need to take safety measures; 5. in jobs of an un-routine character; and 6 in seasonal jobs. Art. 5 of the draft proposal provides that in the first conclusion of the temporary employment contract, duration of the relationship shall not exceed 4 months. Where the need for the temporary worker continues, the contract can be renewed three times at the most, provided that the total duration shall not exceed 8 months. The number of temporary workers to be employed must not exceed one-fifth of the permanent workers of the user undertaking. Again, adhering to the in-built tradition of being protective towards labor, the draft bill seems to overregulate temporary work, a point which is at present the subject of ongoing controversy and criticism (Su¨ral 2012, p. 4; Aktekin 2012, p. 4). Also, the same draft on temporary work lacks an important dimension emphasized by the EU and CIETT (International Confederation of Private Employment Agen- cies), that is, regulations on temporary work must encourage the movement of temporary workers into permanent jobs of the user undertaking. Similarly, in Art. 7 (A) of the second draft, the provision that the temporary worker may benefit from

158 T. Dereli the terms of the collective agreement in force in the user undertaking by paying dues to the signatory labor union has also led to controversy, as it seems contra- dictory to the notion of temporary workers’ organizing and collective bargaining processes to be vis-a`-vis the temporary work agency itself, which is their main employer. This clause has been deleted from the text of the third draft. At the time of this writing, proposals for the revision of the severance pay system, amelioration of subcontracting arrangements and regulation of temporary work agencies were on the agenda of the government as three important issues. It is not clear at this point when and how the proposed changes will materialize. But one thing is certain. The search for a better balance between security and flexibility is very likely to continue for some time in Turkey’s labor relations discourse. References Aktekin S¸ (2012) Atipik ve Esnek C¸ alıs¸ma Bic¸imleri Bakanlıg˘ın Gu¨ndeminde. Metal Industry Employers’ Association (MESS), ˙Is¸veren Gazetesi Baydur R (2006) Zirvede 15 Yıl. Sinemis publications, Ankara. Court of appeals, 9th division, 26 January 2004, 23105/1204; general council decision, 11 November 2006, 9-613/644; 9th div., 28 January 2010, 14809/1480 Dereli T (2011) Partnership for developing human resources: Joint MESS-Turk Metal Training Project. In: International Labor Organization (ed) Freedom of association and development report. http://www.ilo.org/wcmsp5/groups/public/—ed_norm/–declaration/documents/publica tion/wcms_160208.pdf Dereli T (2012) Turkey. In: Blanpain R (ed) International encyclopedia for labour law and industrial relations. Kluwer Law International, The Hague Hendricks F (2004) Flexible work in EU labour law; challenges for Turkey. Unpublished paper submitted to the 2006 congress of the Turkish Industrial Relations Association Hendricks F, Sengers K (2006) The implementation of flexible work provisions in the labor act. In: Blanpain R (ed) Flexibilisation and modernisation of the Turkish labour market, report on the MATRA project. Kluwer Law International, The Hague Metal Employer’s Union of Turkey ˙Is¸veren Gazetesi (2010) The amendment to short-time work in 2009 made it possible for many firms to make use short-time work extensively and led to the protection of thousands of jobs. No. 844 Ruling of the Constitutional Court, 19 October 2005, 66/72, Official Gazette 24 November 2007, no. 26710 Su¨ral N (2012) Esneklik Du¨zenlenirken. Metal Industry Employers’ Association (MESS), Mercek Periodical (January) 18(69) Su¨zek S (2012) Is Hukuku, 8th edn. Beta Publication, Istanbul Turkish Personnel Management Association (2012) Esnek C¸ alıs¸ma Modelleri Bilinmiyor. http:// www.peryonkongre.com/kongre2013/index.html. Accessed January 2012 Uc¸um M, C¸ akmakc¸ı R (2003) Gerekc¸eli Atıflı-Kars¸ılas¸tırma Tablolu Eski ve Yeni Kanun Metinleri I˙le Birlikte ˙Is¸ Kanunu ve I˙lgili Temel Mevzuat. Legal Periodical, ˙Istanbul

Chapter 7 Prejudice Against Labor Unions and Effects on Membership Eul-Teo Lee and Tuncay Guloglu Abstract This chapter explores prejudice against labor unions by non-union mem- bers as one of reasons for the decline in union membership and workers’ avoidance to join a union. This is due to the fact that social changes with respect to the value of labor and movement have led workers to change their attitudes towards union participation and activities. Based on various theories of prejudice, this article suggests that prejudice against labor unions on the part of non-union members have been evoked through learning various negative effects of labor unions on a society. Particularly persistent global economies have been in recession, the unem- ployment rate has been increasing and a variety of mass media have reported social waste and damage brought forth by labor unions. As a result, non-union members have come to consider labor unions as socially irresponsible, and so they inevitably have prejudice against labor unions and avoid joining them. Also, this chapter explains cognitive process such as non-union member’s justifying their prejudice against labor unions based on social identity theory and the hypothesis of attitude similarity attraction. In this chapter, the practical implications of reducing prejudice against labor union are discussed. 7.1 The Aim of the Chapter The world has been experiencing crucial changes over the last three decades. The fact that the contemporary Information Age is presenting fundamental challenges to a broad spectrum of assumptions and theories upon which actions are based is becoming increasingly evident. There have been a series of global transformations. E.-T. Lee (*) Kunsan National University, Gunsan, South Korea e-mail: [email protected] T. Guloglu Cornell University, Ithaca, NY, USA e-mail: [email protected] T. Dereli et al. (eds.), Labor and Employment Relations in a Globalized World, 159 Contributions to Economics, DOI 10.1007/978-3-319-04349-4_7, © Springer International Publishing Switzerland 2014

160 E.-T. Lee and T. Guloglu Product markets have become more competitive. Labor markets have also become increasingly internationalized, as trade has been liberalized, immigration has increased, and capital markets have taken on a more global structure. A major slow-down in world economic growth and productivity and increased inflation following the oil shocks of the 1970s created adverse labor market situations in most countries. Unemployment rates soared and worker compensations decreased (Blanchflower 2007; Jones 2004). On the other hand, all of these crucial developments have influenced the composition of employment as it has shifted from unionized workers to sectors that have traditionally consisted of non-union workers. Due to these changes, labor union membership rates have gone down worldwide (Cregan 2005; Blanchflower 2007). The union membership rate in the U.S. dropped from 20.1 % in 1983 to 11.9 % in 2010 (U.S. Bureau of Labor Statistics). The United Kingdom’s rate fell from 38 % in 1990 to 26.6 % in 2010 (U.S. Department of Labor 2012). The rate in Turkey dropped from 9.5 % in 2002 to 5.9 % in 2011 (OECD Stat 2013), and the rate in Japan fell from 25.2 % in 1990 to 18.5 % in 2010 (U.S. Department of Labor 2012). It appears that there is no OECD country which experienced rising union density in recent years. The overall cause of decline was the growth in “never membership,” that is, persons who entered the labor market post-1980 and who increasingly never had a unionized job (Bryson et al. 2010). Nine primary factors have an impact on labor union membership: 1. Gender 2. Age of worker 3. Educational Attainment 4. Job Status 5. Ethnicity 6. Workplace 7. Industry 8. Region 9. Political Opinion The question of today is: Why are the labor union membership rates going down? Industrial labor relations scholars have sought to investigate why workers join unions or not in terms of the workers as individuals, the labor unions, the job, and the labor-management relationship (DeCotiis and LeLouarn 1981; Farber and Saks 1980; Maxey and Mohrman 1981; Cooke 1983; Angle and Perry 1986; Gallagher and Clark 1989; Kochan et al. 1986; Premack and Hunter 1988; Newton and Shore 1992; Cregan 2005). Also, wage increases for workers, changing of the industrial structure from manufacturing to service, the increase of temporary jobs, and the lack of participation by female workers in labor unions have been consid- ered some of the main reasons for the decline of labor union membership. There are good reasons for looking at gender differences. Female employment rates have risen rapidly over the period of decline (Machin 2004). The characteristics of union members show many similarities across countries: density rates are generally higher for men and public sector workers than for

7 Prejudice Against Labor Unions and Effects on Membership 161 women and private sector workers; union membership rates are higher in the manufacturing sector than services sector; membership rates are higher among the less educated than among the more educated; and the rates are higher among full-time workers than those who are considered part-time (Blanchflower 2007). In addition these features, a recent phenomenon shows that the density rate of union participation among young people is lower than others; in other words, the new situation has created a generation gap between younger and the older workers with respect to union membership (Blanden and Machin 2003). This chapter intends to shed light on why non-union members avoid joining a labor union in terms of their prejudice against labor unions. It can be reasonably inferred that both social changes on labor value and the phenomenon of labor movement have led workers to change their attitudes towards work ethic, union participation and union activities. Modern workers, unlike those in the past, no longer embrace the ideological values of labor unions (Cregan 2005). Although joining a labor union is a personal choice on the part of a worker, social influences play an important role in the decision whether or not to join a union. The decline in union membership is very likely linked to negative attitudes such as prejudice against labor unions at this point in time. For example, a non-union member can learn that a labor union’s ideology is good or bad through mass communication, education institutes or their family and friends. As a result, non-union members can have prejudice against labor unions. And prejudice shared by modern workers prevents non-union members from becoming attracted to the idea of labor unions, and it is very likely that they would hesitate to join them. On the other hand, sometimes career and professionalism are important factors for union membership. For example, professional worker’s attitudes towards unions are ambivalent. There is a degree of elitism among professionals who take great pride in their abilities, intelligence and accomplishments. There is a related ten- dency to look askance at unions as institutions better suited to the gritty world of blue-collar and low-wage service work. Meanwhile professional workers who feel that they are not given the respect that they deserve are self-confident enough to stand up and demand redress (Hurd 2001). 7.2 Previous Literature According to theories of union membership, the union status of workers is mainly determined as the result of separate decisions by workers and potential union employers. Generally, employees decide to join union or non-union jobs based on utilities. Workers become union members in two ways: by accepting existing union jobs or organizing a non-union workplace. On the other hand, unionized employers decide which of the workers who want union jobs to hire. In this way, the status of unions encompasses the job choices of individuals, collective choices of employees, and hiring decisions by employers (Kim and Kim 1997).

162 E.-T. Lee and T. Guloglu Previous studies about the decline in union membership and union membership reluctance can be divided into three areas. On the macro level, labor economists have tried to determine the reason for the decrease in labor union membership rates from economic indicators like wage increase rate, unemployment rate, and infla- tion. Industrial labor relation scholars have suggested that labor union membership is associated with changed production methods, decentralization of labor unions, and divergence of collective bargaining (Boyer 1988; Katz 1983). On the micro level, studies have shown that among the reasons for labor union membership reluctance is job satisfaction. Many non-union companies have attempted to develop human resource practices that could optimize organizational performance through the improvement of employee satisfaction and that, in many cases, this change preceded the experimentation with innovative work practices, setting the stage for their conceptual evaluation (Kochan 1979; Kochan et al. 1986), the characteristics of labor unions and union socialization (Youngblood et al. 1981; Cooke 1983; Fullagar and Barling 1989), the characteristics of the top management (Summers et al. 1986) and the workers’ attitudes towards labor unions (Newton and Shore 1992; Cregan 2005). For example, Youngblood et al. (1981) insisted that if non-union members sense that the labor union leadership is authoritative, destructive, and discriminatory, their decision as to whether to join the labor union will be extremely negative. Newton and Shore (1992) classified workers’ perceptions of their labor union into four types, focusing on the instrumental and ideological axes. They called the case that is ideologically and instrumentally positive based on the two axes “union attachment,” and the reverse case “union opposition.” Unlike Kochan et al. (1986), who asserted that labor unions should stress instrumental appeals rather than ideological appeals to maintain the unionists’ support, they asserted that labor unions should pursue their ideologies and instrumentalities simultaneously. Cregan (2005) directly asked 1,216 Australians what attitudes led non-union members to refuse to join a labor union. The results showed that the non-union members’ attitudes that made them refuse to join a labor union were cost benefit ineffectiveness, ideological anti-union stance, job norms, instrumental and ideo- logical disillusionment, instrumental individualism, and pro-business ideology. From the sociological perspective, some studies have focused the social relation- ships of workers. According to Gordon et al. (1980), the attitudes of the workers’ family members and friends strongly influence the workers’ conviction as far as the ideology of labor unions is concerned. In other words, if the workers’ family members or friends positively view the ideology of labor unions, the workers develop a positive conviction with regard to labor unions, promoting participation and commitment to them, and that the reverse is also true. This research showed that family members and friends have a huge effect on non-union members’ entry into a labor union. Fukami and Larson (1984) insisted that co-workers strongly influence the union members’ conviction with regard to labor unions as well as their commitment to their company’s labor union, and that the reverse is also true. Fiorito and Maranto (1987) suggested that those who are not part of a labor union are

7 Prejudice Against Labor Unions and Effects on Membership 163 opposed to its collective activity because they believe that it does not make any significant contribution to the development of their society. 7.3 Prejudice Against the Labor Union and Membership 7.3.1 Prejudice Against the Labor Union Prejudice means “an unfavorable opinion or feeling formed beforehand or without knowledge or reason” or “any preconceived opinion or feeling, either favorable or unfavorable” (Random House 1997). Prejudice is also a negative or exclusive feeling towards the members of a group or towards the group itself (Brigham 1971). According to Fiske and Taylor (1984), prejudice is a negative feeling towards an exterior group or its members, defined by the statement: “All the exterior groups other than the group to which an individual belongs have bad attributes, all of them are identical, and they are different from us.” Prejudice matters in social activities because people give less compensation to those belong- ing to exterior groups than to those belonging to interior groups, help them less, and show them less friendly non-verbal cues (Brewer 1979; Tajfel 1982; Crosby et al 1980). The fact that prejudice exists can be explained through the conception of the categorization of the social identity theory presented by Tajfel (1982). Categoriza- tion is simply classifying people as “us” and “them” without any realistic conflict, which can in turn act as a starting point of antagonism among categories. When an outstanding group category exists, people try to find their identity through mem- bership in a group. Therefore, they show the conviction, attitude, and action of preferring the interior group to which they belong while looking down on exterior groups. Another reason that prejudice exists is associated with information processing. Since people cannot deal with every piece of complex information due to a limited cognition-accepting ability, they process it more quickly and easily by means of the expectations or frames which already exist (Ashmore and DelBoca 1981; Bodenhausen and Lichtenstein 1987; Bodenhausen and Wyer 1985). Social- cultural learning is another reason for prejudice. People learn prejudice against exterior groups as a kind of cultural norm through the socialization process. People learn prejudice against exterior groups from family, friends, education, and cultural norms. Furthermore, mass media accelerates prejudice and spreads it to others. The first antecedents of prejudice against unions is major company’s unions going on strike. We found that nowadays ordinary people do not regard strikes by labor unions, especially those of major companies, aimed at raising the workers’ pay or improving working conditions as necessary. The reason for this is due to the fact that the unions have already placed excessive pressure on companies to provide better wages and benefit packages. In Korea, companies with over 1,000 employees had more strikes compared to smaller companies. In the case of the U.S. auto

164 E.-T. Lee and T. Guloglu industry, many car makers were in the midst of a management crisis due to uncontrollable welfare benefits for their employees and the burden of personnel expenses, which once led the U.S. Congress to pour public funds into the car industry. Thus, it can be said that the attitude of ordinary people towards strikes by labor unions is not as positive as it used to be. Another reason is that current global economies have been in recession, the unemployment rate has been increasing, and in particular, the number of temporary workers is increasing. The percentage of temporary workers in the workplace in Korea is almost 40 %. Thus, the impact of strikes by major companies is perceived as having caused relative deprivation by temporary workers, unemployed people, low salaried workers and so on. The more important thing is that mass media outlets, such as TV networks, newspapers, and the internet, are social phenomena in the modern world. They have reported strikes as causing social losses and damage to society. In fact, strikes by major manufacturing unions result in huge losses, not only for the numerous sub- contractors supplying parts to the company, but also for the local economy. Mass media tends to focus on reporting the social losses caused by strikes rather than the necessity and reasons for strikes. As a result, the members of the society have come to consider labor unions as socially irresponsible, and so they inevitably have prejudice against unions. In particular, prospective employees developed their prejudice against unions even before getting employed, and have a negative attitude towards joining a labor union. According to Fiorito and Maranto (1987), those who are ideologically exclusive towards labor unions do not think that these are beneficial to the society and are opposed to collective action. The following statements can thus be made based on the aforementioned dis- cussion: Despite the increased unemployment rate and the number of temporary workers due to the economic recession and the changing of the industrial structure, strikes organized by unions, especially those of major companies, accelerate the prejudice against unions. Mass media such as TV networks, newspapers and the internet have reported such strikes as causing social loss. Thus, even before working at a company, prospective employees believe that unions do not help social growth. Consequently, they already have formed a prejudice against labor unions before being employed, and, therefore, they will not join them. In addition, the unions generally are not able to be successful in marketing campaigns designed to deliver their messages to non-union workers. For this reason, unions must learn how to market themselves to non-union workers by chaptering and approaching techniques from contemporary membership-based institutions such as Facebook and other successful online networking communities (Bryson et al. 2010).

7 Prejudice Against Labor Unions and Effects on Membership 165 7.3.2 Justifying the Process of Prejudice and Its Results The important thing is that if a newly employed worker who is prejudiced against unions encounters a union member who is highly oppositional toward the company, and behaves in a manner that can be considered aggressive and belligerent, their prejudice towards unions might be justified, and so they will not join. This element in the decision whether or not to join labor unions can be explained by the social identity theory. This theory states that a person’s judgment of a member of an exterior group is less classified, and therefore simpler than that same person’s judgment of a member of an interior group. People generally view the members of exterior groups as less homogeneous and more radical, and have a large tendency to generalize the characteristics of the whole group based only on the characteristics and behaviors of a single member that they observe (Park and Rothbart 1982; Linville 1982; Quattrone and Jones 1980; Wilder 1984). Thus, when a non-union member who is prejudiced against unions meets a union member, he/she will not judge him or her favorably and will likely regard the union member as heterogeneous, ultimately justifying his/her negative attitude towards labor unions. This is an error in simple information processing as a result of a kind of categorization. The hypothesis of attitude similarity attraction can also be applied. This hypoth- esis states that if a person feels attracted to someone who is similar to him/her in terms of attitudes, such as personality, values, and faith, the person would willingly participate in their group and activities (Byrne 1971; Byrne et al. 1986). The reverse is also true. Non-union members who are prejudiced against labor unions are highly unlikely to have a similar attitude to unionists whom they encounter because non-union members are likely to feel unfriendly towards the union members in terms of opinions, values, and personalities. Thus, because a prejudiced non-union employees has a different attitude toward the union, he/she is highly unlikely to feel attracted to union members when he/she meets them. Newton and Shore (1992) reported that the previous studies were not able to prove anything with regards to the moderating effects of ideological similarity by non-union member with the labor union. They concluded that it is quite difficult for those who are not friendly to the labor union to join it and to be committed to it even if the labor union mobilizes unionists and promotes its instrumentality among non-union members through personal contacts with them. While Newton and Shore (1992) did not include any direct discussion on attitude similarity, their position is that labor unions cannot lead non-union members, who have no strong value-related or ideological attitude in common with members of labor unions, to decide to join them and to commit themselves to them even when unionists appeal for the union’s instrumentality from non-union members. The following statements can thus be made based on the aforementioned dis- cussion: non-union members who are prejudiced against unions will likely regard unions as heterogeneous and will unlikely to perceive attitude similarity with

166 E.-T. Lee and T. Guloglu unionists when he/she meets them. As a result, non-union members tend to justify their prejudice against unions and they will not join them. 7.4 Practical Implications How can employees’ prejudice against unions be altered? This chapter proposes unions should contribute to the growth of the local community. It also proposes unionists should manage their first impression when they meet a non-union member for the first time. 7.4.1 Unions Should Contribute to the Growth of the Local Community For labor unions to remove the social prejudice against them and to encourage non-union members to join unions, they should contribute to the growth of the local community. First of all, they should pursue the agenda of social responsibility towards the local society. As labor strikes may negatively impact the economy of a local community, the local residents do not support them. If a labor union should go on a strike, they should consider the negative impact that their strike may have on the local community, and the reasons for going on a strike should be such that the workers and the local community can support them. Labor unions should refrain from holding a strike that is only essentially a power struggle with the management. Unions also should make efforts to break away from the somewhat belligerent or militant attitude of threatening to go on a strike if the collective bargaining with the company does not go well. They should make an effort to solve their labor-related issues through persistent dialogues with the company’s management. Strikes should be the final recourse to the resolution of issues. Also, unions could hire local residents as staff members, initiate various local events, support the alienated classes in the local community and request a social contribution funds from the company as part of collective bargaining. To illustrate this, it was recently reported in a TV news program that the labor union of Hyundai Motor Company was exerting various efforts to stimulate the traditional market in the local community where the company’s factory is located. Such events will help promote a positive attitude towards the union among local residents and among viewers who saw the report. The labor unions’ contributions to the local commu- nities will be then reported by the media and spread among residents through word of mouth. Public promotion will thus help positively impact the image of the union. As mentioned by Gordon et al. (1980), once the members of the society develop a positive attitude towards labor unions, non-union members will follow suit and will thus consider joining their company’s labor union.

7 Prejudice Against Labor Unions and Effects on Membership 167 On the other hand, new immigrants constitute the most problematic group within the local communities. The unions should support new immigrants by the language education and the training programs. The unions can encourage participation from the immigrants by putting flyers into many different languages, and respecting specific cultural traditions of each group. In addition, in contemporary diverse societies the unions must develop training programs for the highly diverse work force in local level. 7.4.2 First Impressions of Union Members on Non-union Members When a non-union member who is prejudiced against unions meets a union member for the first time, it is highly likely for him or her to justify their prejudice if the union member demonstrates an aggressive or belligerent attitude towards the company. Thus, this chapter proposes that a positive first impression management of union members on non-union members is necessary for labor unions to eliminate non-union members’ prejudices against unions. Union members should make an effort to refrain from behaving in an aggressive or belligerent manner towards the company when they meet a non-union member. When a labor union member meets a non-member for the first time, it is important for the former to show the latter that the labor union is a partner of the company and will always support the company, rather than aggressively introducing the union’s ideals based on class conflict. Labor unions also should take note of the characteristics of those members tasked to meet non-union members. Similarities such as ages, genders, educational background between these employees need to be addressed. For example, while the chapter conducted by Kim and Kim (2006) is not directly related to the demo- graphic similarity between union and non-union members, the two researchers empirically showed that when a new employee perceives that the age, gender, and educational background of union members are not similar to his/hers, he/she is also likely to perceive the group’s value as low. Also, personal values, tendencies, and even the appearance between a non-union member and a union member need to be addressed. Byrne et al. (1986) argued that when one is attracted more to the physical charm of the other, the former will be attracted more to the latter, which will eventually lead to the desire to become a part of the group to which the latter belongs. If the non-union member is attracted to the appearance of the union member, the former will evaluate the union more positively and consider joining the union.

168 E.-T. Lee and T. Guloglu 7.5 Concluding Remarks This chapter explored precursors of prejudice against unions and the relationship between prejudice against unions and union membership. Based on discussions, it also suggested how to overcome prejudice against unions on non-union members. This chapter suggested that unions make contributions to the local community and that unions should manage the first impressions of unionists on non-union members to mitigate prejudice against unions. This chapter has implications for new strategies for labor unions to revitalize their membership rates. A drop in union membership rates is an indication that the traditional model of labor unions is no longer attractive to workers. Labor unions should establish new strategies for recruiting and retaining members in terms of socio-psychological perspectives, such as proactively working against prejudices. From the findings of this chapter, it is clear that labor unions should make an effort to increase their positive image and attempt to alter the culture of prejudice against unions. Also, labor unions have to play a role as a seller from the viewpoint of marketing. Labor unions, like other sales organizations, should investigate what the non-union members want, and, after finding out what these wants are, must promote them. Further research should identify the more concrete precursors of prejudice against the labor unions, such as interest and past career of family and friends on labor movements, education about unions at school or company, individual dispo- sitions on collective movements, and so on. The limitation of this chapter is that we could not demonstrate empirical results, and it is our hope that further chapter can help to empirically identify aspects of the relationship between prejudice against labor unions and the decision to join a union by non-union members. References Angle HL, Perry JL (1986) Organizational commitment and labor-management relationship climates. Acad Manag J 29(1):31–50 Ashmore RD, Del Boca FK (1981) Conceptual approaches to stereotypes and stereotyping. In: Hamilton D (ed) Cognitive process in stereotyping and intergroup behavior. LEA, Hillsdale, pp 1–36 Bodenhausen GV, Wyer RS (1985) Effects of stereotypes on decision making and information- processing strategies. J Pers Soc Psychol 48:267–282 Blanchflower DG (2007) International pattern of union membership. Br J Ind Relat 45:1–27 Blanden J, Machin S (2003) Cross-generation correlations of union status for young people in Britain. Br J Ind Relat 41(3):391–415 Bodenhausen GV, Lichetenstein M (1987) Social stereotypes and information-processing strategy: the impact of task complexity. J Pers Soc Psychol 52:871–880 Boyer R (1988) Wage/labour relations, growth, and crisis: a hidden dialectic. In: Boyer R (ed) The search for labour market flexibility. Clarendon, Oxford, pp 3–25

7 Prejudice Against Labor Unions and Effects on Membership 169 Brewer MB (1979) In group bias in the minimal intergroup situation: a cognitive-motivational analysis. Psychol Bull 86:307–324 Brigham JC (1971) Ethnic stereotypes. Psychol Bull 76:15–33 Bryson A, Gomez R, Willman P (2010) Online social networking and trade union membership: what the Facebook phenomenon truly means for labor organizers. Labor Hist 51(1):41–43 Byrne D (1971) The attraction paradigm. Academic, New York Byrne D, Clore GL, Smeaton G (1986) The attraction hypothesis: do similar affect anything? J Pers Soc Psychol 51(6):1167–1170 Cooke WN (1983) Determinants of the outcomes of union certification elections. Ind Labor Relat Rev 36:402–414 Cregan C (2005) Can organizing work: an inductive analysis of individual attitudes toward union membership. Ind Labor Relat Rev 58(2):282–304 Crosby F, Bromley S, Saxe L (1980) Recent unobtrusive studies of black and white discrimination and prejudice: a literature review. Psychol Bull 87:543–563 DeCotiis TA, LeLouarn J (1981) A predictive chapter of voting behavior in a representation election using union instrumentality and work perceptions. Organ Behav Hum Resource Perform 27:103–118 Farber H, Saks D (1980) Why workers want unions: the role of relative wages and job character- istics. J Polit Econ 38:349–369 Fiorito J, Maranto CL (1987) The contemporary decline of union strength. Contemp Policy Issues 5:12–27 Fiske ST, Taylor SE (1984) Social cognition. Addison-Wesley, Reading Fukami CV, Larson EW (1984) Commitment to company and union: parallel models. J Appl Psychol 69:367–371 Fullagar C, Barling J (1989) A longitudinal test of a model of the antecedents and consequences of union loyalty. J Appl Psychol 74:213–227 Gallagher DG, Clark PF (1989) Research on union commitment: implications for labor. Labor Stud J 14:52–71 Gordon ME, Philpot JW, Burt RE, Thompson CA, Spiller WE (1980) Commitment to the union: development of a measure and an examination of its correlates. J Appl Psychol 65:479–499 Hurd RW (2001) Professional employees and union democracy: from control to chaos. In: Estreick S, Katz HC, Kaufman BE (eds) The international governance and organizational effectiveness of labor unions. Kluwer Law International, New York, pp 159–172 Jones S (2004) Towards a theory of industrial relations for a knowledge economy. Ger J Ind Relat 11:15–26 Katz HC (1983) The decentralization of collective bargaining: a literature review and comparative analysis. Ind Labor Relat Rev 47(1):3–22 Kim DO, Kim S (1997) The effect on union membership of race and immigration status focusing on Asian Americans. J Appl Behav Sci 33(3):378–396 Kim MS, Kim SY (2006) The effects of demographic dissimilarity and person-group value fit on group members’ attitudes. Korean J Ind Organ Psychol 19(3):395–420 Kochan TA (1979) How American workers view labor unions. Mon Labor Rev 102(4):23–31 Kochan TA, Katz HC, McKersie RB (1986) The transformation of American industrial relations. Basic Books, New York Linville PW (1982) The complexity-extremity effect and age-based stereotyping. J Pers Soc Psychol 38:689–703 Machin S (2004) Factors of convergence and divergence in union membership. Br J Ind Relat 42 (3):423–438 Maxey C, Mohnman SA (1981) Worker attitudes toward unions: a chapter integrating industrial relations and organizational behavior perspectives. Proc Ind Relat Res Assoc 326–337 Newton LA, Shore LM (1992) A model of union membership: instrumentality, commitment, and opposition. Acad Manag Rev 17(2):275–298 OECD (2013) OECD.Stats extracts. http://stats.oecd.org/. Accessed 15 Dec 2012

170 E.-T. Lee and T. Guloglu Park B, Rothbart M (1982) Perception of out-group homogeneity and levels of social categoriza- tion: memory for the subordinate attributes of in-group and out-group members. J Pers Soc Psychol 42:1051–1068 Premack SL, Hunter JE (1988) Individual unionization decisions. Psychol Bull 103:223–234 Random House (1997) Webster’s universal college dictionary. Random House, New York Quattrone GA, Jones EE (1980) The perception of variability within ingroups and outgroups: implications for the law of small numbers. J Pers Soc Psychol 38:141–152 Summers TP, Betton JH, DeCotiis TA (1986) Voting for and against unions: a decision model. Acad Manag Rev 11:643–655 Tajfel H (1982) Social psychology of intergroup relations. Annu Rev Psychol 33:1–39 United States Department of Labor (2012) Bureau of labor statistics. http://www.bls.gov/. Accessed 20 Dec 2012 Wilder DA (1984) Intergroup contact: the typical member and the exception to the rule. J Exp Soc Psychol 20:177–194 Youngblood SA, Mobley WH, DeNisi AS (1981) Attitudes, perceptions, and intentions to vote in a union certification election: an empirical investigation. Proc Ind Relat Res Assoc 244–253

Chapter 8 Unionization Status and Productivity Nurgun Komsuoglu Yilmaz Abstract The purpose of this study is to review the effects of unionization on productivity in the Turkish Chemical sector. The data proceed from the firms listed in the first and second 500 firms in the review conducted by ICI (Istanbul Chamber of Industry) between 1998 and 2006. In this study one-way ANOVA analysis was employed in two different ways. First, by including the totality of the chemical sector firms and second it is applied sub-sectors. In conclusion, nonunion firms were found more productive compared to the union and “nonbinding agreement firms”. 8.1 Introduction The impact of unionization on productivity has been the subject of an ample number of studies, both theoretical and empirical in the fields of industrial relations and labor economics. This study is mainly based on author’s Ph.D. dissertation (Yilmaz 2012). These have often resulted in contradictory findings. The diversity of the findings, which are known to cause disputes, has often been explained by the variation in data, definitions, techniques and methodology. In this study, by using analysis of variance analysis, current position in the chemistry sector according to its unionization status is determined. Apart from the theoretical analyses, empirical studies on the subject matter can also be found in abundance. These have also suggested contradictory findings, where certain studies suggesting a positive impact and others show negative effects. Furthermore, some research has demonstrated a statistically insignificant relation- ship between unionization and productivity. These studies have been carried out in a variety of sample spaces, covering industry, inter-industry, and firm levels. The distinct industries that have been subject to research were cement, coal, N.K. Yilmaz (*) 171 Istanbul Aydin University, Istanbul, Turkey e-mail: [email protected] T. Dereli et al. (eds.), Labor and Employment Relations in a Globalized World, Contributions to Economics, DOI 10.1007/978-3-319-04349-4_8, © Springer International Publishing Switzerland 2014

172 N.K. Yilmaz construction, hospital, banking, furniture and education, among others, as exempli- fied in the literature review. In this study ANOVA analysis was employed in two different ways. First, by including the totality of the chemical sector firms considered, the degrees of productivity of the firms and factors that may affect the productivity were investi- gated in terms of unionization status (union, non-union and “nonbinding collective agreement” groups). Then, after the firms were arranged within subgroups (general chemicals, pharmaceuticals, oil and plastics), ANOVA analysis was applied again to investigate the differences between the effects of unionization on corporate productivity with regard to these subgroups. 8.2 Literature Review A pioneering study by Brown and Medoff (1978) employed the Cobb–Douglas productivity function modified with a unionization variable, using US manufactur- ing industry data from 1972. Holding the capital/labor ratio and the worker quality as constants and using gross added value as the dependent variable, they have found unionization to have a positive impact on productivity. It was concluded that the companies with higher unionization density show higher productivity levels by 20– 25 %. The findings of the Brown–Medoff study are supportive of another study by Frantz (1976) 2 years earlier. In his study, Frantz found that unions had caused a 15 % increase in productivity levels in the wooden household furniture industry. Clark (1984), in that he has also kept value added as a dependent variable in the production function. The study includes inter-industry research that uses sample data from 250 large companies within 1970–1980 timeframe. The research dem- onstrates that the union effect on productivity is 0.01 where control variables are used regardless of the dependent variable. Clark concludes that unions have barely any effect on productivity. In his study, Clark (1980a) has used data from 119 union, and 9 non-union plants for 4 years during the early 1970s. He has applied several controls, including ratios of capital to labor and supervisors to labor, size of the facility, average utilization, vintage, and plant-specific effects. The empirical evidence from the study suggests a positive productivity effect of union- ization within a range of 6–8 %, in both cross-section and time-series data. Another study on the same industry was performed by Clark (1980b), where he similarly used the tons of cement production per worker as the indicator of productivity. This study empirically demonstrates that unionization has led to increased productivity, primarily due to substantial changes in procedures and management staff. Allen (1984) has utilized the Brown–Medoff methodology to estimate the union effect on productivity in the American construction industry, using value added as an indi- cator of productivity. The estimated effect decreases to 17–22 % when value added is deflated by the construction price difference between areas. Allen (1986) esti- mates the effect of public and private ownership on productivity in the construction

8 Unionization Status and Productivity 173 of hospitals and nursing homes. The union/non-union productivity differential was shown to be generally positive, but quite imprecise. 8.3 Data Companies active in the chemical sector between the years 1998–2006 and regis- tered in ICI first and second 500 have been included in the ANOVA analysis. Data including prices have been deflated using the 1994 consumer price index figures. The companies included in the study have been specified as giving data continu- ously as well as with missing data only for a maximum 2 years. The missing data have been filled in by using the SPSS 15 program. Data has been standardized based on the q/l value which shows the gross added value per capita. Based on a distribution with zero mean and 1 standard deviation, records with +3, À3 averages and with deviation below or above the standard deviation were detected as outliers and extremes. The data set has been set in line with the normal distribution as much as possible, and the two companies creating outlier and extreme values, which are Tu¨rk Shell and Tu¨pras¸ A.S¸, have been excluded from the data set. Six hundred and three data of 67 companies have been included in the study. The analysis made has been carried out with the computer program of PASW 19. The data set frequency chart to be examined according to union status is presented in Table 8.1. 8.4 Implementation As it is known, in the ANOVA test groups more than two can be examined from the perspective of some criteria. The per capita gross added value mentioned as productivity indicator in the literature, per capita capital and the number of workers for the differences in the unionization status have been examined. The union status has been specified, in three groups as unionized companies, non-union companies and companies with “nonbinding union agreements”. The one way ANOVA test has been implemented in two different ways. Firstly, all chemical companies have been included in the research and subjected to the data set analyses where the ANOVA test is applied. It has been specified whether there is a significant difference from the statistical point of view among the three groups, unionized companies, non-union companies and companies with “nonbinding union agreements”. Secondly, taking the same criteria into consideration, the ANOVA was applied again but, this time in more detail for the subsectors of the chemical sector in Turkey. With a view to answer the question if the results obtained for each subgroup would be the same as the ones for the entire sector. The ANOVA test aim to find out whether there was a statistically significant difference among the

174 N.K. Yilmaz Table 8.1 Descriptive statistics of union status Frequency Percent Valid percent Cumulative percent Valid Nonbinding union agreement 36 6.0 6.0 6.0 Non-union firm 396 65.7 65.7 71.6 Union firm 171 28.4 28.4 100.0 Total 603 100.0 100.0 three groups (unionized, non-union companies and “companies with nonbinding union agreement”). 8.4.1 ANOVA Test for the Entire Chemical Sector Through the ANOVA test, unionized companies, non-union companies and com- panies with nonbinding union agreements are compared in dealing with the pro- ductivity outcomes. For the three groups specified according to their status, the differences between q/l (per capita productivity), c/l (per capita capital), l (number of laborers) were examined. For ANOVA test, firstly, variance homogeneity Levene’s test (test on the absolute or squared deviation of scores from their own group means) was applied. Its significant results were found smaller than 0.05 for q/l, c/l, and l, Welch and Brown–Forsythe tests were applied afterwards. Multiple comparison tests were applied at 95 % interval to determine which group is different among the mentioned three groups. Scheffe and Tamhane tests were applied by taking into consideration whether the variances were homogenous or not, and results of the two tests were provided with explanations. 8.4.1.1 Productivity In the measurement of productivity where the per capita gross added value was taken as the criterion, there was statistically a significant difference between the unionized and non-union groups according to the “Tamhane” test results. The difference between the unionized and non-union is specified as À298.879 (unionized–non-union). Between the unionized and the “nonbinding union agree- ment group” a significant result of À222.272 has been obtained. When the Scheffe test is considered, the difference between the unionized and non-union group has been found as À298.879, thereby supporting Tamhane. No significant difference has been obtained between the “nonbinding agreement group” and the unionized and non-union groups (see Table 8.2).

8 Unionization Status and Productivity 175 Table 8.2 Multiple comparisons for q/l Dependent (I) Union (J) Union status Mean Std. error Sig. variable status difference 189,492.558 0.922 Nonbinding (I–J) q/l Scheffe Non-union union 99,608.181 0.011 firm agreement 76,607.011 189,492.558 0.922 199,611.241 0.538 Nonbinding Union firm 298,879.542* union Non-union firm À76,607.011 99,608.181 0.011 agreement Union firm 222,272.530 199,611.241 0.538 Union firm Non-union firm À298,879.542* 103,490.385 0.843 Nonbinding À222,272.530 Tamhane Non-union 74,644.048 0.000 firm union 76,607.011 103,490.385 0.843 agreement 88,249.822 0.044 Nonbinding Nonbinding 298,879.542* union union À76,607.011 74,644.048 0.000 agreement agreement 222,272.530* 88,249.822 0.044 Union firm Union firm Non-union firm À298,879.542* Union firm À222,272.530* Non-union firm Nonbinding union agreement * Statisticaly significant 8.4.1.2 Capital When the capital per labor is examined through ANOVA analysis; according to the Tamhane test results, no significant difference has been obtained between groups. In the same way, no significant difference has been noticed between the groups with the Scheffe test. Multiple comparison tests results are presented in Table 8.3. 8.4.1.3 Labour When the number of wage workers in each of the three groups is compared, some significant differences between the three groups have been found. According to the Tamhane test results, the difference between the means of unionized and non-union groups is 802.463 and the difference between unionized and “nonbinding union agreement group” has been found to be 638.643. When we look at the Scheffe results for the same factor, the same results are obtained. The difference between the unionized and non-union groups is 802.463 and the difference between the unionized and nonbinding union agreement group is 638.643 (results are presented in Table 8.4).

176 N.K. Yilmaz Table 8.3 Multiple comparisons for c/l Dependent (I) Union status (J) Union status Mean Std. Sig. variable Non-union firm difference error 0.708 Nonbinding (I–J) c/l Scheffe union 302.655 agreement À251.591 Nonbinding Union firm À16.169 159.093 0.995 union Non-union firm 251.590 302.655 0.708 agreement Union firm 235.421 318.817 0.761 Union firm Non-union firm 16.169 159.093 0.995 Nonbinding À235.422 318.817 0.761 Tamhane Non-union firm union À251.591 390.495 0.892 agreement Nonbinding Nonbinding À16.169 139.642 0.999 union union 251.590 390.495 0.892 agreement agreement 235.421 394.027 0.911 Union firm Union firm Non-union firm 16.169 139.642 0.999 Union firm À235.422 394.027 0.911 Non-union firm Nonbinding union agreement Table 8.4 Multiple comparisons for labor Dependent (I) Union status (J) Union status Mean Std. Sig. variable Non-union firm difference error 0.487 Nonbinding union (I–J) 136.519 Lbr Scheffe agreement À163.819 0.000 71.762 0.487 Nonbinding union Union firm À802.463* 136.519 0.000 agreement Non-union firm 163.819 143.809 0.000 Union firm 71.762 0.000 Union firm Non-union firm À638.643* 143.809 Nonbinding union 802.463* 0.080 Tamhane Non-union firm 638.643* 71.586 agreement 0.000 Nonbinding union Nonbinding union À163.819 105.506 0.080 agreement 71.586 0.000 agreement À802.463* 125.053 0.000 Union firm Union firm 163.819 105.506 0.000 Non-union firm 125.053 Union firm À638.643* Non-union firm 802.463* Nonbinding union 638.643* agreement * Statisticaly significant

8 Unionization Status and Productivity 177 8.4.2 ANOVA Test for Each Sub Sector The ANOVA test has been applied for the subgroups of the chemical field where there are companies in the same study data set. The aim here is to specify whether there is a difference in the chemical sector in general according to the criteria on the number of workers, the gross added value, capital, and according to the unioniza- tion status in each subsector. The research question required a more detailed investigation on this subject. The data set comprising companies active in subfields such as power supply, dye, gas, general chemical, fertilizer, pharmaceutical, plastic, petroleum has been put into four main groups. Since all the companies in the field of tyres are unionized, analysis under a separate sub-group was not deemed necessary. The groups specified as subsector have been set as general chemical, pharmaceu- tical, oil and plastic. The ANOVA test has been applied for four groups. 8.4.2.1 General Chemicals Criteria such as number of workers, the gross added value and capital in the general chemical field were evaluated, with an attempt to find out whether there was a difference between the groups. The Levene test has been applied in the data set and the result was obtained as sig. < 0.05 for each criterion. For q/l, c/l, l, Welch and Brown–Forsythe tests have been applied. The Welch and Brown–Forsythe tests have given significant results for each criterion as sig. < 0.05. There are multiple comparisons in the post hoc test and also data frequency is presented in Table 8.5. Productivity There is a statistically significant difference between the unionized and non-union groups according to the Tamhane test results, at which measurement for per capita gross value added was taken as the productivity criterion and the unionized and non-union difference was specified as À338.397 (union–non-union). Between the unionized and “nonbinding union agreement” group, a significant difference of À723.580 was obtained. When the Scheffe test considered, the difference between the unionized and non-union groups was found as À338.397 (union–non-union), supporting Tamhane. No significant difference has been found between the nonbinding union agreement and the other two, unionized and non-union groups (see Table 8.6).

178 N.K. Yilmaz Table 8.5 Descriptive statistics of union status (general chemical) Frequency Percent Valid percent Cumulative percent Valid Nonbinding union agreement 9 3.7 3.7 3.7 Non-union firm 162 66.7 66.7 70.4 Union firm 29.6 29.6 100.0 Total 72 100.0 100.0 243 Table 8.6 Multiple comparisons for q/l (general chemical) Dependent (I) Union status (J) Union status Mean Std. Sig. variable Non-union firm difference error 0.476 Nonbinding union (I–J) 315.573 q/l Scheffe agreement À385.183 0.036 130.516 0.476 Nonbinding union Union firm 338.397* 315.573 0.087 agreement Non-union firm 385.183 325.788 0.036 Union firm 723.580 130.516 0.087 Union firm Non-union firm À338.397* 325.788 Nonbinding union À723.580 0.093 Tamhane Non-union firm 163.175 agreement À385.183 0.002 Nonbinding union Nonbinding union 98.064 0.093 agreement 338.397* 163.175 0.002 agreement 385.183 147.065 0.002 Union firm Union firm 723.580* 0.002 Non-union firm À338.397* 98.064 Union firm À723.580* 147.065 Non-union firm Nonbinding union agreement * Statisticaly significant Capital The Scheffe and Tamhane tests have been applied for the per capita capital criterion in the field of general chemical. Both tests (Tamhane and Scheffe) have obtained the same results. The statistically significant difference between the nonbinding union agreement group and the non-union group has been specified as 3,322.357, and the statistically significant difference between the nonbinding union agreement group and the unionized group has been specified as 3,716.570. Multiple compar- ison results for general chemicals are presented in Table 8.7. Labour When the average of the workers in the field of general chemical is examined, significant differences have been found between the three groups. According to the Tamhane test results, the difference between the unionized and non-union group

8 Unionization Status and Productivity 179 Table 8.7 Multiple comparisons for c/l (general chemical) Dependent (I) Union status (J) Union status Mean Std. Sig. variable Non-union firm difference error 0.000 Nonbinding union (I–J) 700.751 c/l Scheffe agreement À3,322.357* 0.398 289.820 0.000 Nonbinding union Union firm 394.213 700.751 0.000 agreement Non-union firm 3,322.357* 723.435 0.398 Union firm 3,716.570* 289.820 0.000 Union firm Non-union firm À394.213 723.435 Nonbinding union À3,716.570* 0.010 Tamhane Non-union firm 839.179 agreement À3,322.357* 0.257 Nonbinding union Nonbinding union 234.466 0.010 agreement 394.213 839.179 0.005 agreement 3,322.357* 833.533 0.257 Union firm Union firm 3716.570* 234.466 0.005 Non-union firm À394.213 833.533 Union firm À3,716.570 Non-union firm Nonbinding union agreement * Statisticaly significant has been obtained as 493.4310, and unionized and “non-binding agreement” group difference as 602.418. The difference between the non-unionized and “non-binding group” is 108.987. When we look at the Scheffe results for the same factor the results are the similar. The difference between the unionized and non-union groups is 495.255, and the unionized nonbinding union agreement group difference has been found to be as 602.418 (see in Table 8.8). 8.4.2.2 Pharmaceuticals In the pharmaceuticals subgroup, which is a subsector of the field of Chemical, the ANOVA test and the statistical analysis were applied. For the unionized, non-union and “nonbinding union agreement” group, analysis focused on whether there was a statistically significant difference from the q/l, c/l, l criteria. Also with multiple comparisons, effort was made to specify between which groups this difference existed and what its total was. In the data set, first the Levene’s test was applied and the results for q/l, were found as sig. > 0.05. For q/l, it shows the homogeneity in the variances as well as the ANOVA results. For c/l, l Welch and Brown–Forsythe tests have been applied (descriptive statistics for pharmaceuticals are presented in Table 8.9).

180 N.K. Yilmaz Table 8.8 Multiple comparisons for labor Dependent (I) Union status (J) Union status Mean difference Std. variable Non-union firm (I–J) error Sig. Nonbinding union Lbr Scheffe agreement 108.987 152.342 0.774 Nonbinding union Union firm À493.431* 63.006 0.000 agreement Non-union firm À108.987 152.342 0.774 Union firm À602.418* 157.273 0.001 Union firm Non-union firm Nonbinding union 493.431* 63.006 0.000 602.418* 157.273 0.001 agreement Tamhane Non-union firm Nonbinding union 108.987* 35.921 0.025 Nonbinding union agreement À493.431* 88.393 0.000 agreement Union firm À108.987* 35.921 0.025 Non-union firm À602.418* 91.598 0.000 Union firm Union firm 88.393 0.000 Non-union firm 493.431* 91.598 0.000 Nonbinding union 602.418* agreement * Statisticaly significant Productivity Concerning productivity in the pharmaceuticals sector, for q/l groups significant difference could not be specified in ANOVA test. In the post hoc tests table multiple comparisons have been carried out. As a result of the q/l, Tamhane and Schefe tests, no significant difference was obtained between the groups (see in Table 8.10). Capital In the pharmaceuticals, significant difference was obtained at the Tamhane and Scheffe tests. Both tests give a significant difference of 804.453 between the unionized and non-union groups. Multiple comparison results for Pharmaceuticals’ c/l are presented in Table 8.11. Labour In the pharmaceuticals sector, examining the average number of workers Scheffe and Tamhane tests were applied. In the Tamhane test, a significant difference was between “the nonbinding union agreement group” and the other two groups, that is, the unionized and non-union groups. The difference between the nonbinding union agreement group and the unionized group has been found as 334.157. The differ- ence between the “nonbinding union agreement” group and the non-union group is 413.180. Multiple comparison results for Pharmaceuticals’ number of labor are presented in Table 8.12.

8 Unionization Status and Productivity 181 Table 8.9 Descriptive statistics of union status (pharmaceuticals) Frequency Percent Valid percent Cumulative percent Valid Nonbinding union agreement 18 15.4 15.4 69.2 Non-union firm 63 53.8 53.8 53.8 Union firm 36 30.8 30.8 100.0 Total 117 100.0 100.0 Table 8.10 Multiple comparisons for q/l (pharmaceuticals) Dependent variable (I) Union status (J) Union status Mean Std. Sig. q/l Scheffe Non-union firm difference error 0.511 Nonbinding union (I–J) 122.306 Tamhane Nonbinding union agreement À142.174 1.000 agreement 95.611 0.511 Union firm À0.491 122.306 0.564 Union firm Non-union firm 142.174 132.106 1.000 Union firm 141.683 95.611 0.564 Non-union firm Non-union firm 132.106 Nonbinding union 0.491 0.349 Nonbinding union À141.683 92.658 agreement agreement 1.000 Nonbinding union À142.174 106.373 0.349 Union firm 92.658 0.553 agreement À0.491 118.010 1.000 Union firm 142.174 106.373 0.553 Non-union firm 141.683 118.010 Union firm Non-union firm 0.491 Nonbinding union À141.683 agreement 8.4.2.3 Oil For the subgroup oil, which is a subsector of the field of Chemical, the ANOVA analysis was made. In the oil field, there are no nonbinding union agreement group firms in our data set. For the two groups i.e. unionized and non-union firms, research was done to see whether or not there is statistically a significant difference from the q/l, c/l, l; in the data set first the Levene’s test has been carried out (see Tables 8.13 and 8.14), the results were found to be sig. > 0.05 for c/l. For q/l, and l, sig. < 0.05 is was found. Productivity In the measurement, the per capita gross added value (q/l) was taken as the productivity criterion; according to Brown–Forsythe and Welch tests, there was a statistically significant difference between the unionized and non-union groups (see

182 N.K. Yilmaz Table 8.11 Multiple comparisons for c/l (pharmaceuticals) Dependent variable (I) Union status (J) Union status Mean Std. Sig. difference error 0.863 c/l Scheffe Non-union firm Nonbinding union (I–J) 299.016 agreement À162.674 0.004 Nonbinding union 233.751 0.863 agreement Union firm À804.453* 299.016 0.144 Non-union firm 162.674 322.974 0.004 Union firm Union firm 233.751 0.144 Non-union firm À641.779 322.974 Tamhane Non-union firm Nonbinding union 804.453* 0.583 641.779 140.455 Nonbinding union agreement 0.035 agreement Nonbinding union À162.674 305.558 0.583 140.455 0.129 Union firm agreement À804.453* 311.183 0.035 Union firm 162.674 305.558 0.129 Non-union firm 311.183 Union firm À641.779 Non-union firm 804.453* Nonbinding union 641.779 agreement * Statisticaly significant Table 8.12 Multiple comparisons labor (pharmaceuticals) Dependent (I) Union status (J) Union status Mean difference Std. variable Non-union firm (I–J) error Sig. Nonbinding union Lbr Scheffe agreement À413.180* 77.058 0.000 Nonbinding union Union firm À79.023 60.239 0.426 agreement Non-union firm 413.180* 77.058 0.000 Union firm 334.156* 83.232 0.001 Union firm Non-union firm 79.023 60.239 0.426 Nonbinding union À334.156* 83.232 0.001 Tamhane Non-union firm agreement À413.180* 58.338 0.000 Nonbinding union Nonbinding union À79.023 58.107 0.443 agreement agreement 413.180* 58.338 0.000 Union firm 334.156* 56.623 0.000 Union firm Non-union firm 79.023 58.107 0.443 Union firm À334.156* 56.623 0.000 Non-union firm Nonbinding union agreement * Statisticaly significant

8 Unionization Status and Productivity 183 Table 8.13 Descriptive statistics of union status (oil) Valid Non-union firm Frequency Percent Valid percent Cumulative percent Union firm Total 63 77.8 77.8 77.8 18 22.2 22.2 100.0 81 100.0 100.0 Table 8.14 Comparison of q/l, c/l, labor (oil) q/l Non-union firm N Mean Std. deviation Std. error c/l Union firm Labor (l) Total 63 2412.168 2080.452 262.112 Non-union firm 18 863.459 339.783 80.087 Union firm 81 2068.011 1949.022 216.558 Total 63 2409.627 1508.860 190.098 Non-union firm 18 2823.929 942.138 222.064 Union firm 81 2501.694 1408.214 156.468 Total 63 629.777 489.337 61.650 18 4688.944 1019.082 240.200 81 1531.814 1813.755 201.528 Table 8.15). Where equal variances are not assumed, the non-union and unionized difference has been specified as 1,548.708 (non-union–union). Capital In the chemical sector for oil the per capita capital (c/l); has been examined; according to the ANOVA results between the unionized and non-union groups, a statistically significant difference has not been found (see Table 8.15). Labour The difference between the groups in the number of workers has been examined through the Brown–Forsythe and Welch tests; according to the results of the test, a statistically significant difference has been found between the unionized and non-union groups (see Table 8.15). The unionized non-union difference is 4,059.162 (unionized–non-union). 8.4.2.4 Plastic For the subgroup plastic, which is a subsector of the field of Chemical, the ANOVA test and the analysis have been applied. For the unionized, non-union and “nonbinding union agreement” groups, whether or not there is a statistically significant difference from the q/l, c/l, l criteria was examined. Also with multiple

184 N.K. Yilmaz Table 8.15 ANOVA test (oil) Sum of squares df Mean square F Sig. q/l Between groups 33,578,975.939 1 33,578,975.939 9.813 0.002 Within groups 270,316,174.418 79 3,421,723.727 Total 303,895,150.357 80 c/l Between groups 2,403,041.485 1 2,403,041.485 1.215 0.274 Within groups 156,242,455.877 79 1,977,752.606 Total 158,645,497.361 80 Labor (l) Between groups 230,675,676.389 1 230,675,676.389 560.702 0.000 Within groups 32,500,989.833 79 411,404.935 Total 263,176,666.222 80 comparisons tests, the attempt was made to specify between which groups this difference exists and what its total is. In the data set first the Levene’s test was applied. According to which, the significance value for q/l was obtained above 0.05. The significance value for c/l and l was below 0.05, and by assuming that the variances were not homogenous, Brown–Forsythe and Welch tests were employed. Descriptive statistics of union status (plastic) are presented in Table 8.16. Productivity Productivity was examined in the plastic sector, the significant values of the criteria were studied according to the ANOVA test results, for q/l no significant difference was found between the groups (sig. ¼ 0.175). Multiple comparisons for q/l in plastic sector’ productivity is represented in Table 8.17. Capital For the c/l, the per capita capital changes according to unionization status in the plastic sector were examined by using the Welch and the Brown–Forsythe tests. When the Tamhane test result is considered, the difference between the non-union and the “nonbinding union agreement” groups has been found as 838.545. Besides, the difference between the union firms and non-union firms’ group averages was obtained as À770.722. Scheffe test could not obtain a statistical significant differ- ence between the groups (see Table 8.18). Labour When the number of employees in the plastic subgroup is considered; according to Tamhane test, differences were obtained between all group averages. The differ- ence between the non-union group and the unionized group has been found to be 300.334. The difference between the non-union group and the “nonbinding union

8 Unionization Status and Productivity 185 Table 8.16 Descriptive statistics of union status (plastic) Frequency Percent Valid percent Cumulative percent Valid Nonbinding union agreement 9 7.1 7.1 92.9 Non-union firm 108 85.7 85.7 85.7 Union firm 7.1 100.0 Total 9 7.1 100.0 126 100.0 Table 8.17 Multiple comparisons for q/l (plastic) Dependent variable (I) Union status (J) Union status Mean Std. Sig. q/l Scheffe Non-union firm difference error 0.244 Nonbinding union (I–J) 224.022 Tamhane Nonbinding union agreement À378.179 0.632 agreement 224.022 0.244 Union firm À214.844 224.022 0.866 Union firm Non-union firm 378.179 304.386 0.632 Union firm 163.335 224.022 0.866 Non-union firm Non-union firm 214.844 304.386 Nonbinding union 0.173 Nonbinding union À163.335 180.677 agreement agreement 0.245 Nonbinding union À378.179 118.847 0.173 Union firm 180.677 0.804 agreement À214.844 195.887 0.245 Union firm 378.179 118.847 0.804 Non-union firm 163.335 195.887 Union firm 214.844 Non-union firm Nonbinding union À163.335 agreement agreement” group was 204.556. Additionally, the difference between the nonbinding union agreement group and unionized group was obtained as 95.778. According to the results of the Scheffe test, the difference between the non-union and unionized groups has been found to be 300.334, giving the same results as Tamhane. Scheffe and Tamhane’s results are presented in Table 8.19. 8.5 Interpretations and Conclusion Companies active in the Chemical Sector in Turkey between the years 1998–2006 and registered in the ICI first and second 500 groups have been subject to the ANOVA analysis. The ANOVA test was applied in two different ways. First, all the chemical companies were included in the scope of the research and all the data set was subjected to analysis, applying the ANOVA test. Tamhane’s test results revealed statistically significant differences between the mean values relating to unionized firms and the other two groups, that is, non-union firms and firms with

186 N.K. Yilmaz Table 8.18 Multiple comparisons for c/l (plastic) Dependent variable (I) Union status (J) Union status Mean Std. Sig. difference error 0.181 c/l Scheffe Non-union firm Nonbinding union (I–J) 450.052 agreement 0.235 Nonbinding union 838.545 450.052 0.181 agreement Union firm 450.052 0.994 Non-union firm 770.722 611.501 0.235 Union firm Union firm À838.545 450.052 0.994 Non-union firm À67.823 611.501 Tamhane Non-union firm Nonbinding union À770.722 0.002 212.277 Nonbinding union agreement 67.823 0.000 agreement Nonbinding union 136.624 0.002 838.545* 212.277 Union firm agreement 0.972 Union firm 770.722* 168.104 0.000 Non-union firm À838.545* 136.624 0.972 168.104 Union firm À67.823 Non-union firm À770.722* Nonbinding union 67.823 agreement * Statisticaly significant non-binding agreements. Considering the groups’ mean values, there is a tendency towards productivity increase successively from the unionized to non-unionized firms, meaning non-union firms are more productive (see Tables 8.20 and 8.21). The reason for this tendency was probed by in-depth interviews conducted with informants. An informant representing one of the largest non-union firms (AYGAZ) in the petroleum industry attributed this difference to low productivity possibly resulting from lack of discipline in unionized firms due to the fact that workers see their union as a powerful force shielding them against disciplinary actions by the employer. It is in fact conceivable that those who violate work discipline in non-union settings can be terminated more readily than similar workers in unionized setting. Union members have a stronger feeling of job security as they view their union as a protective force against dismissals and other disci- plinary actions by the employers. Similar views were also supported by respondents representing union as well as unionized firms and firms with non-binding agree- ments. Likewise, in an interview with a representative from a firm in the “non- binding agreements” group (ALK˙IM), the higher productivity compared to that of the unionized group was attributed to the absence of union power behind workers, resulting in the workers increased commitment to work since they lack the support and protection of the union. Union pressures forcing employers to accept job descriptions, thereby limiting the scope of worker’s activities and effort, may also be the cause of lower productivity. Strict adherence by the union to working time rule and avoiding undertaking longer hours are also seen as factors resulting in lower productivity in union

8 Unionization Status and Productivity 187 Table 8.19 Multiple comparisons for labor (plastic) Dependent (I) Union status (J) Union status Mean Std. error Sig. variable Non-union firm difference 101.447883 0.135 Nonbinding union (I–J) Lbr(l) Scheffe agreement 101.447 0.015 204.556 101.447 0.135 Nonbinding Union firm 137.840 0.786 agreement Non-union firm 300.334* 101.447 0.015 Union firm À204.556 137.840 0.786 Union firm Non-union firm Nonbinding union 95.777 39.640 0.000 Tamhane Non-union firm À300.334* agreement 31.205 0.000 Nonbinding Nonbinding union À95.777 39.640 0.000 agreement 27.096 0.017 agreement 204.556* 31.205 0.000 Union firm Union firm 27.096 0.017 Non-union firm 300.334* Union firm À204.556* Non-union firm Nonbinding union 95.777* À300.334* agreement À95.777* * Statisticaly significant Table 8.20 ANOVA findings (1) ANOVA criteria ANOVA test for all Turkish ANOVA test for general Productivity per labor chemical sectors chemical sectors Capital per labor Union < non-union Union < non-union Union < nonbinding Union < nonbinding No significant relationship Nonbinding > union Nonbinding > non-union Labor Union > non-union Union > non-union > nonbinding Union > nonbinding settings. Non-union firms tend to employ workers at lower labor costs and/or keep them working longer hours, refraining from paying the legally mandated overtime premiums. Certainly work along these lines is apt to increase production. It was common practice for some firms, while progressive employers kept paying the time and a half or even the union-negotiated “higher” rates for all overtime hours worked. The flexibility brought in the overtime work rules in the Labor Act of 2003 paved the way for abusing these rules in non-union work places. Differently from the previous regulations in Act no 1475, the obligation of dividing weekly work hours equally into work days in a week was abolished, thereby, distributing normal weekly working time days in a week in different proportions through agreements was made possible. In the case of balancing (63/II), on the other hand, working over 45 h in some weeks is not considered as overtime work as long as weekly average 45 h work is not exceeded within the

188 N.K. Yilmaz Table 8.21 ANOVA findings (2) ANOVA ANOVA test for ANOVA test for criteria Productivity pharmaceuticals sectors oil sectors ANOVA test for plastic sectors No significant relationship per labor No significant relationship Union < non- union Non-union > union Capital per labor Union > non-union No significant Non-union > nonbinding relationship Non- Labor Nonbinding > non-union Union > non- union > nonbinding > union union Nonbinding > union equalization period (2 or 4 months) (Labor Act Article 41/I) (Topc¸uog˘lu 2009, p. 8). A simple example can be given to better understand. Accordingly, in case 55 h work is exceeded in five work days in the first 2 weeks and 30 h work is exceeded in the third and fourth weeks of the equalization period, total work hours will be 110 + 60 ¼ 170 in the 4 week equalization period and as weekly average work period is 170 Ä 4 ¼ 42.5 h, no overtime payment will be made to the worker as the weekly average work does not exceed 45 h in average. If they worked for 45 h in a week in the third and the fourth weeks, the average work period would be 50 h and they would earn overtime payment for 5 h for each week (Topc¸uog˘lu 2009, p. 76). But this flexibility measure has not been used properly by non-unionized firms; that is, while no overtime pay was granted daily the weekly working time was not reduced adequately in the following weeks so as to balance out the average weekly working time at 45 h within the given period of adjustment. Thus the non-union firms took advantage of the new flexibility measures illegally. Furthermore, some firms did not respect the legally mandated rest breaks. As it is possible for union firms to implement the labor law on a wider scale when compared to non-union firms, one is likely to find their productivity at a lower level. Collective agreements regulate working time in a more orderly fashion. They foresee the application of rest periods, tea or meal breaks along legal lines. It is difficult to estimate how these adjustments are implemented even at legal mini- mums at non-union firms. The hypothesis that unions tend to be organized in firms with high per capita capital was tested and confirmed for this sector; when the three groups were examined in terms of per capita capital, no statistically significant difference was found between them. According to the results of the Tamhane’s test, it was found that compared to the other two groups, more workers were employed on average in the “unionized firms” group. When the groups’ mean values are considered, the number of workers employed in respective groups tends to increase as the rate of unionization rises, whereas the trend of the gross value added per capita tends to decrease. While the number of workers employed is the highest in unionized firms, it tends to fall in firms with “nonbinding union agreements” and reaches the lowest level in

8 Unionization Status and Productivity 189 non-union firms, in that order. Common views of respondents indicate that the reason accounting for this difference could be attributed to the larger volume of employment (overmanning) prevailing in the unionized firms. Furthermore, the respondents also made reference to the difficulties involved in terminating workers in unionized firms, due to the additional protection provided by the labor union. This leads to continued employment of unproductive workers. To offset lower productivity, unionized firms have to hire additional workers. On the other hand, this may be given as a reason for the high number of employees in union firms. Besides, unions are also opposed to the employer’s use of subcontractors in establishments where they are active and try to restrict subcontracting of employ- ment. As the subcontracting applications are welcomed by the non-union firms, the required labor force is met by contract labor. The number of employees in the non-union firms is higher than unionized firms because contracted employees are not included in the workforce of the firm. Concerning the results obtained for the unionized group (i.e. the large volume of employment and lower added value per capita), the same interviewee noted the probability of the unionized group supply- ing correct information reflecting the real figures. Likewise; the KI˙PLAS represen- tative from the employers’ union active in the chemicals, petroleum and tire sector made the following remarks on the aforementioned topics; data on workers are recorded and documented more correctly in unionized firms. This may be one reason why the number of workers appears to be higher in such firms. This may also explain the lower productivity computed as a ratio of production per worker. Numerical data reflect the real situation more correctly in unionized firms. Further, employment of subcontractors and their workers is hardly acceptable in unionized firms; and if there is such employment, it is easily observable. Non-union firms, on the other hand, rarely state the number of contractor’s employees correctly. As a further step in the analysis, the ANOVA test was applied again, taking the same criteria into account but this time for the subsectors of Turkey’s chemical sector and in a more detailed fashion, in order to find out whether the results obtained for the whole sector also applied in the case of subsector comparisons. The purpose is to test if there is a statistically significant difference between the three groups (unionized, non-unionized and nonbinding union agreements group) in terms of each subsector of the chemicals sector. According to the ANOVA test results, in which the general chemical field groups are taken into consideration in terms of per capita productivity, the union- ized firms are less productive than the other two groups. The difference (in terms of the q/l mean) between unionized and non-union, as well as between the unionized and nonbinding union agreement group, were found to be statistically significant. Unionized firms give lower rates of productivity when compared to the other two groups. As for the q/l between the non-union and the nonbinding union agreement groups, no statistically significant difference was found. When statistical relevance is ignored and only the group averages are considered, the nonbinding group can be suggested as the most productive. Considering the results obtained for the general chemicals sector, the findings concerning the comparison between the unionized and non-unionized firms reflect the real situation. But the nonbinding union

190 N.K. Yilmaz agreement group’s appearance of having the highest productivity may be mislead- ing since the sample includes only one firm representing this category. Therefore the comparison involving the non-union group may not be considered reliable. Concerning the interview question as to why the non-unionized firms’ per capita productivity (in the general chemical sub group) has been found to be higher than unionized firms, a human resources manager representing the “nonbinding union agreements” group in the general chemicals sector (ALKI˙M) responded by saying that this finding was quite natural. According to him workers’ stronger commitment led to this finding. Thus, its reflection to the worker’s total productivity in the firm is only natural. Similarly, a respondent from the non-union general chemicals sector representing KAYALAR repeated his earlier statements with a further remark pointing also to the possible effects of lower unionization in this sector. In terms of per capita capital comparisons, no significant difference was found between the unionized and non-union groups, but the “nonbinding union agree- ments” group is characterized by statistically significant higher rates of per capita capital in comparison to the other two groups. In a group comparison of the number of laborers, the Tamhane test results showed clearly that the unionized group had more workers than the non-union group and nonbinding union agreement group. The companies in the “nonbinding union agreement group” are the companies that provide the least amount of employment among the three groups. On account of the insufficiency of data on the “nonbinding agreements firms”; it may be deceptive to conclude that the nonbinding agreement firms employ the lowest number of workers possible. Including the companies in the area of pharmaceuticals, when the per capita productivity is examined, the ANOVA test results showed no significant difference between the three groups, i.e. unionized, non-union and “nonbinding union agree- ment” as a result of the Scheffe’s test. The Tamhane’s test was applied in examining per capita capital for these groups in order to disclose the differences between averages of unionized and non-union firms, referring to the situation when variances are not homogenous. Per capita capital was found higher in unionized firms, meaning they are mostly capital intensive organizations. The same result was obtained by applying the Scheffe’s test (with homogeneity of variances under the same assumption). As a result of the Tamhane’s test result, in the comparison of the average of number of workers, the “nonbinding union agreement” group companies were found having more workers than the other two groups. No significant difference was found between the unionized and non-union groups. Concerning the particular findings for the oil-field which is a subsector of the chemicals industry, non-unionized companies were found to be more productive than unionized. When analyzed in terms of per capita capital, no difference was found between union and non-union group averages. The number of employees was again analyzed by using the ANOVA analysis. It was discovered that union firms had more workers than non-union firms. Findings in the petroleum field support the general chemical field. Moreover, certain possible reasons for the higher number of employees in the union firms than the non-union

8 Unionization Status and Productivity 191 ones have been detected as a result of the interview conducted with the employers’ union. Concerning the higher number of workers employed in oil sector firms, the representative of K˙IPLAS refers to the uniqueness of the petroleum sector where the number of workers employed is overstated since some of them, though seem- ingly employed and paid wages, work only for temporary periods as they are also land owners. The interviewee from AYGAZ, representing the non-union group in the petroleum sector, referred to the higher number of workers employed in this sector as a factor arising from the predominance of public ownership and a result of politically motivated recruitments in this sector. Unions are organized more easily in the public sector and the performance of Petrol-I˙s¸ is noteworthy in this respect. In the plastic industry, the application of the Tamhane’s and Scheffe’s tests showed that there was no difference in productivity between the “unionized” and “non-union” groups. Here, in terms of per capita capital; the non-union group represents more per capita capital; the “unionized” and “nonbinding union agree- ments” groups seem to have lower per capita capital than the “non-union” group. On the other hand, the situation is somewhat different in the plastic industry which is also subsector of the chemical industry. Here, in terms of the number of workers employed, statistically significant differences were found. Non-union firms tend to employ the highest number of workers, followed by companies in the “nonbinding union agreements” group, while the unionized firms employ the least number of workers. (However, one must interpret these findings for the plastic sector with some caution as the “unionized” and “nonbinding union agreement” firms are under-represented in the sample with the inclusion of only one firm from each.) References Allen SG (1984) Unionized construction workers are more productive. Q J Econ 99:251–274 Allen SG (1986) The effect of unionism on productivity in privately and publicly owned hospitals and nursing homes. J Labor Res 7:59–68 Brown C, Medoff JL (1978) Trade unions in the production process. J Polit Econ 93:355–378 Clark K (1980a) The impact of unionization on productivity: a case study. Ind Labor Relat Rev 33:451–469 Clark K (1980b) Unionization and productivity: microeconomic evidence. Q J Econ 45:613–640 Clark K (1984) Unionization and firm performance: the impact on profits, growth and productivity. Am Econ Rev 74:893–919 Frantz JR (1976) The impact of trade unions on production in the wooden household furniture industry. Senior Honors Thesis, Harvard University Topc¸uog˘lu I˙ (2009) 4857 Sayılı I˙s¸ kanunu Is¸ıg˘ında Tu¨rk Hukukunda Fazla Saatlerde C¸ alıs¸ma. TU¨ H˙IS Yayın No. 58, Ankara Yilmaz NK (2012) The relationship between unionization, productivity and firm efficiency: evidence on the chemical industry in Turkey. Ph.D. Thesis, Isik University

Part IV Labor Market Challenges, the New Workplace, Informal Employment and Skills Training

Chapter 9 The New Labor Skills in the New Economy Pınar Falcıog˘lu Abstract In the last 30 years a dramatic change has been observed in the supply and price of skilled labor relative to unskilled labor. The aim of this chapter is to review the recent empirical literature on the changing demand for labor skills, to define the scope and direction of change in labor skills and to explore the reasons of the observed change, especially on the labor skills demanded in knowledge based global economy. Findings on the reasons of skill change are important for the study aims of this book because they lead to results and consequently to implications to be used in the distribution of earnings and employment of the workforce. The empir- ical research identifies the main reasons of the observed change in skills as the interactions of technological change, organizational change and globalization. 9.1 Introduction The increasing emphasis on science, technology and knowledge based economic development creates higher complexity for studies examining the pattern and reasons of change in labor skills demanded. In the last 30 years a dramatic change has been observed in the supply and price of skilled labor relative to unskilled labor. In line with the observed changes in skill demand, research has been conducted to understand whether the simultaneous change in technology or organizational prac- tices or changes related to globalization have effect on the skill change process or each have interactive effects on each other has also been investigated. In the empirical literature, technological change has been identified as the “first suspect” of the increase in the demand for skilled labor as well as the degree of inequality in the distribution of earnings. The “new suspects” are globalization and organiza- tional change. Findings on the reasons of skill change are important for the study P. Falcıog˘lu (*) 195 Faculty of Economics and Administrative Sciences, Department of Management, Is¸ık University, S¸ ile, Istanbul, Turkey e-mail: [email protected] T. Dereli et al. (eds.), Labor and Employment Relations in a Globalized World, Contributions to Economics, DOI 10.1007/978-3-319-04349-4_9, © Springer International Publishing Switzerland 2014