Collective bargaining systems, which are not fully decentralized, are also correlated with lower wage inequality for full-time workers (Figure 3.3) compared to the D9/D1 rate, i.e.dem ratio of wages in the 9th decile of the wage distribution to wages in the 1st decile. This association is present in the lower and upper half of the wage distribution17 Similar results are obtained when taxonomy indicators are replaced by centralization and coordination indicators – see Annex 3.B.  Cardoso, A. and P. Portugal (2005), “Contractual wages and the wage cushion under different bargaining settings”, Journal of Labor Economics, Vol. 23/4, pp. 875-902. Thematic Forum “Freedom of association and effective recognition of the right to collective bargaining: a basis for decent work”. 29. ←.
For example, in the round of negotiations in 2016, the “cost bar” was set at about 2.5%, but licensed practical nurses received an agreed salary increase of about 3.5%. All the social partners agreed on this exception because of a relatively small increase in the salaries of women workers, despite the shortage of manpower in their profession. 26. ←. In systems where negotiations take place mainly at company level, representation in the workplace and coverage of collective agreements go hand in hand. At the same time, Chapter 2 shows that, in these countries, the proportion of workers represented in any form in workers` representation in the workplace is lower (whereas it tends to be high in multilevel systems characterised by complementarity between sectoral agreements and company agreements). Therefore, the transfer to company-level negotiations could result in less coverage.  Doucouliagos, H., R.
Freeman and P. Laroche (2017), The economics of trade unions: A study of a research field and its findings, Routledge, London. The treaty is considered a constitution drafted by the trade union and management representative`s points conference in the form of a compromise or trade agreement. The agreement sets out the enforcement and interpretation mechanism for the industry. The right of initiative is limited within the framework of the legislation.  Ibsen, C. (2016), “The role of mediations institutions in Sweden and Denmark after centralized bargaining,” British Journal of Industrial Relations, b. 54/2, pp. 285-310.  Schulten, T.
and R. Bispinck (2017), “Varieties of decentralisation in German collective bargaining: Experiences from metal industry and retail trade”, CSDLE “Massimo D`Antona” Working Paper, no. 137/2017. In countries where coordination works well, it tends to be strongly supported by employers` organisations, as it mitigates wage growth and trade unions, as it guarantees a high level of employment. To be effective, coordination requires strong and self-regulating social partners, as well as effective mediation bodies. Worker i`s wage is measured per hour and survey weights are used to better fit the sample to the actual labour force. Control variables, x i b, include models for age, gender, education, company size, type of contract (indefinite or temporary), duration of employment, sector and occupation. Some control variables are not available for some countries. The comparison of the estimated coefficients, β ^ for the same variables, makes it possible, for example, to study the differences in the gender gap or the training premium between workers in collective bargaining and those not covered by collective agreements.
A prosperous industrial unit in the same region can accept with the unions a substantial increase in wages and other social benefits, while a losing industry cannot. There is still pressure on losing industries to provide wages and benefits similar to those of other (relatively prosperous) units in the same region. . . .