Collective bargaining, wage formation and unemployment in ...131260/FULLTEXT01.pdf · Collective...
Transcript of Collective bargaining, wage formation and unemployment in ...131260/FULLTEXT01.pdf · Collective...
Uppsala University Department of Economics
Collective bargaining, wage formation and unemployment in Russia
- effects of the degree of centralisation in wage bargaining among trade unions in 10 sectors
Bachelor thesis: January 2006 Thesis advisor: Andreas Westermark
Author: Kajsa Borgnäs
Abstract
Calmfors and Driffill in 1988 argued that there is a humpshaped relation between the degree of centralisation in wage bargaining structures within an economy and unemployment. They collected aggregate economic data from 17 different OECD economies and ranked them according to their relative degree of centralisation to prove their model. The model was further developed by Rowthorn who in 1992, using individual data from the same countries, concluded that there is a negative linear relationship between the degree of centralisation in wage bargaining structures and wage dispersion. During the past two decades the Russian economy, as well as the Russian trade union movement, has transformed greatly. Membership rates in trade unions have fallen and bargaining leverage of trade unions vis-á-vis employers has decreased. Using data from ten sectors within the Russian economy (collected in interviews with trade union representatives in Moscow, June 2006) this essay questions whether the theoretical assumptions above hold in the Russian context. By ranking the sectors according to their relative degree of centralisation in wage bargaining structures and using these rankings as explanatory variables in econometric analyses with unemployment rates and wage dispersion rates as dependent variables, this essay finds little proof that the theoretical framework of Calmfors and Driffill holds within the Russian economy. However, Rowthorn’s model of centralisation and wage dispersion seems to be more valid. Keywords: unemployment, wage dispersion, centralisation, wage bargaining, trade unions.
2
1. INTRODUCTION 1.1 Introduction..............................................................................................................4 1.2 Structure of the paper...............................................................................................5 1.3 Background – a brief history of the trade union movement in Russia.....................6 1.4 Method.....................................................................................................................7
1.5 Definitions...............................................................................................................8
2. THEORIES ON CENTRALISATION, UNEMPLOYMENT AND WAGE DISPERSION
2.1 The Calmfors/Driffill theory on centralisation and unemployment.............................9 2.1.1 Centralised wage bargaining.............................................................................10 2.1.2 Branch level wage bargaining...........................................................................11 2.1.3 Decentralised wage bargaining………………………………….……………11
2.2 Rowthorn´s theory on centralisation and wage dispersion.........................................12 2.3 Data.............................................................................................................................13
3. A DESCRIPTION OF THE STRUCTURE OF TRADE UNIONS IN RUSSIA
3.1 The financial structure of trade unions…………………………................………...14 3.2 The formal structure of wage bargaining among trade unions...................................15
3.2.1 Tripartite agreements........................................................................................15 3.2.2. Central level bargaining...................................................................................15 3.2.3. Branch level bargaining…………………………………….……………..…16 3.2.3. Regional bargaining.........................................................................................16 3.2.4. Local bargaining..............................................................................................16
3.3 The level of centralisation among employers.............................................................17
4. CENTRALISATION OF COLLECTIVE BARGAINING
4.1 Comments on the method of ranking the ten trade unions.........................................17 4.2 Short description of the trade unions..........................................................................19 4.3 Unemployment, wage dispersion and centralisation; data analysis……..............…..23
4.3.1. A Visual Interpretation of Data……………………….……………….........….23 4.3.2. Regression Analyses of Data…………………………………………..……….25
4.4 Summary of results…………………..…………………................………...………27
5. CONCLUDING REMARKS.............................................................................................28
REFERENCES.....................................................................................................................29
Tables and Figures: Table 1 Sectors, Rank Orderings, Dispersion, Unemployment……………………......................................21 Table 2 Unemployment, Dispersion and Bargaining Structure with observation no.3..................................25 Table 3 Unemployment and Bargaining Structure without observation no 3................................................26 Table 4 Unemployment and Bargaining Structure without the squared variable, with observation no 3.....26 Table 5 Unemployment and Bargaining Structure without the squared variable, without observation no. 3..................................................................................................................................................27 Figure1 Calmfors and Driffill hump-hypothesis…………........………………….……................................10 Figure2 Scatterplots of Unemployment vs Centralisation...……………………………...............................23 Figure3 Scatterplots of Wage Dispersion vs Centralisation...........................................................................24 Appendix A.....................................................................................................................................................31
3
1.1 Introduction
After the break up of the Soviet Union in 1991, a fast and thorough remodelling of the
Russian economy has taken place in which both domestic as well as international private
companies have come to play increasingly important roles, alongside the public sector.
The planned economy operating during Soviet times was soon replaced by a market-
economy and many old state-owned companies were swiftly privatized or simply closed
down. For many workers in Russia the swift transformation of the economy has led to a
lowering of wages and a worsening of conditions at the work place which neither public
authorities nor newly born trade unions have been able to put a stop to.1
Alongside the larger trade unions that survived the Soviet break-down, many smaller
trade unions in various parts of the country and different sectors have started up.
However, despite fast rearrangement and reforming of the old structure of the trade union
movement, there is a widespread public suspiciousness against most trade unions in
Russia. Trade unions, operating on an increasingly internationalised, privatised and
deregulated market, are generally associated with the old Soviet system and nation-wide
membership in trade unions has fallen dramatically during the past decade.2
Since the early 1980´s economists have been aware of the fact that the degree of
centralisation in wage bargaining structures within an economy has a great impact on the
economic performance of the economy as a whole, measured by the level of
unemployment, inflation rates and wage-dispersion. More specifically, Calmfors and
Driffill (C/D) in 1988 argued that there is a hump-shaped relation between the degree of
centralisation in wage negotiations and unemployment (and consequently: a U-shaped
relation between centralisation and employment).3 C/D also claimed that there is
empirical evidence that economies in which wages are mainly set at a central level the
lowest levels of unemployment can be found. Economies in which wage setting occur at
a semi-centralised level (branch-level) are expected to show the highest levels of
unemployment and in economies where wages are set at the local level, unemployment- 1 LO-TCO Secretariat of International Trade Union Development Co-operation 2 Surin, 2006 3 Calmfors, Driffill, 1988
4
rates should be low. The rationale behind this assumption is that in economies where
wage-bargaining is conducted at the central level by a few encompassing trade unions
representing most workers, all parties will take the inflationary impact of higher wages
into consideration and thus moderate wage-demands. In decentralised economies trade
unions are either very weak or non-existent and thus unable to raise wages at all. In semi-
centralised economies, on the other hand, no party takes inflationary tendencies into
account but bargain to and are often successful in raising wages above what companies
are actually able to pay their work-force. Unemployment thus rises accordingly.4
The C/D theory of centralisation has been challenged by some economists, but the main
rationale behind the theory still seems to hold. Rowthorn thoroughly estimated the model
and in 1992 analysed the relationship between the degree of centralisation and wage
dispersion, finding a linear negative relationship between centralisation and dispersion.
He also pointed to a negative relation between dispersion and unemployment, arguing
that greater wage dispersion would lead to more people being employed at the bottom
end of the wage-line, thus lowering unemployment rates.5
This paper aims at analysing the degree of centralisation in wage bargaining structures
within trade unions in ten different sectors in Russia. Second, I examine whether and how
the degree of centralisation in wage bargaining affects the economic performance of the
sector, indicated by unemployment rates and wage dispersion. My study thus combines a
qualitative understanding of the Russian trade union movement with quantitative analyses
of data.
1.2 Structure of the paper
This paper begins with a brief history of the trade union movement in Russia, followed
by an explanation of the method used in this paper as well as definitions. Then there is a
description of and discussion about the C/D and Rowthorn analyses on centralisation in
collective bargaining. In addition I also modify the theories so as to fit my research. In
4 Calmfors, Driffill, 1988, p. 15 ff 5 Rowthorn, 1992, p. 510 ff
5
the second part of the paper I examine the structure of the trade unions and try to rank
them. This ranking will be done both on the basis of the structure of the trade union itself
as well as on the basis of the level of centralisation among employers in the sector. In the
third part of the paper I will run regressions on the effect of centralisation in wage
bargaining on the level of unemployment and wage dispersion in the different sectors.
The interpretation of the regression analyses will be discussed in greater detail in this
chapter. The paper finishes with a concluding discussion on how centralisation in wage
bargaining seems to affect unemployment and wage dispersion.
1.3 Background - a brief history of the trade union movement in Russia
The history of the Russian trade union movement goes back to the late nineteenth
century. In 1895 the first trade union was set up in St.Petersburg and in the following
decades the trade union got engaged particularly in the struggle for shorter working
hours. After the 1917 revolution the trade union was abolished as Lenin saw no need for
a trade union in the ´dictatorship of the proletariat´. During the 1940´s the trade union
was started up again and took the name All-Union Central Council of Trade Unions
(AUCCTU). The AUCCTU was mainly responsible for the members’ physical well-
being and the organisation owned about 16.000 sanatoriums and rest homes for the use of
members. The trade union was also the administrative body of several social insurances.
Despite great membership levels (membership was often obligatory) and an advanced
formal internal structure, the trade union had no influence on wage levels or wage setting
procedures. In all sectors of the economy, wages were set by state officials.6
After the fall of the Soviet Union in 1991, the structure of the AUCCTU was immediately
broken up. In several sectors and many parts of the country new trade unions were started
up both at local and at branch levels. However, in parts of the country and some sectors,
no trade union managed to organise the number of workers to survive. Today, the only
nation-wide trade union of any vigour is the Federation of Independent Russian Trade
Unions (FNPR). The FNPR is the largest trade union association in Russia, organising
about 30 million workers in 300.000 primary trade unions in 48 sectors. About 90% of all
6 LO-TCO Secretariat of International Trade Union Development Co-operation
6
trade union members are organised by a trade union belonging to the FNPR.7 The FNPR
is also the organisation which is actively participating in international bodies on behalf of
the Russian trade union movement, such as the International Labour Organisation (ILO),
the International Confederation of Free Trade Unions (ICFTU) as well as in other
regional and international organisations. The FNPR sometimes organises protest actions,
rallies, marches and other forms of collective actions, but the main objective of the FNPR
is to negotiate for national minimum wages and laws on labour rights within the tripartite
commissions with representatives of the government and employers.8
Given that the Russian trade union movement has been operating more than a century one
might expect that the representatives of trade unions should have acquired great skills in
negotiating over wages and working conditions. Unfortunately, as is being discussed
further below, the main objective of the ACCFTU has historically been to look to the
possibilities for the recreation of members rather than taking an active part in bargaining
over working conditions or wages. In this respect, the trade union has historically had too
little independence from the employer, i.e. the state, to be able to bargain fully on behalf
of its members. Thus, although it might seem like the trade union movement in Russia is
old, the structure it has today (and which resembles the structure of trade unions in other
parts of the world) is, in fact, very new.
1.4 Method
The empirics and data in this essay were collected during a weeks´ trip to Moscow, where
I met and conducted interviews with 12 representatives both of the different trade unions
as well as of the FNPR and the ICFTU. In cooperation with LO-TCO Secretariat of
International Trade Union Development Co-operation (LO-TCO Biståndsnämnd) in
Stockholm, I contacted the Russian trade union representatives and they also helped with
logistical matters before, during and after the trip. The interviews were all performed in a
similar way. Most of them were oral interviews taking place in Moscow, but two of the
trade unions are not located in Moscow and they therefore responded to my questions via
7 Ibid. 8 Surin, 2006
7
e-mail instead. One interview was conducted in Stockholm. These interviews serve as the
main empirical material in the paper when examining and interpreting the level of
centralisation in wage negotiations among Russian trade unions and how it affects
unemployment and wage dispersion in the different sectors.
The purpose of my paper is twofold, one part building directly upon the other. The first
focus is to examine and describe the structure of the trade union movement in the ten
Russian sectors. This is to give the reader a more qualitative understanding of the Russian
trade union movement today. On the basis of the answers and the data I will thereafter try
to rank the different sectors according to how centralised they are in wage bargaining. I
will do this following the C/D method as closely as possible. The ranking will then serve
as a tool when analysing, in an econometric analysis using Minitab, how the degree of
centralisation affects the economic performance in different sectors, measured by the
levels of unemployment and wage dispersion.
1.5 Definitions
Whereas both C/D and Rowthorn conducted their analyses on seventeen economies
(countries), I will instead treat the ten different Russian sectors in a corresponding way.
Within each of the ten sectors there is one main central trade union which organises
regional and local trade unions. Thus, I will treat these sectors as economic entities,
corresponding to the economies that C/D and Rowthorn use in their analyses. I am aware
that one cannot compare sectors and entire economies without problems, as countries are
very different from sectors within an economy. This is particularly true as the most
central degree of bargaining within an economy, the state level, is somewhat “lost” in an
analysis that takes sectors within an economy as analytical objects. Speaking in the terms
of C/D, the right end of the scale on the X-axis might not be reached, and the ´hump´
might thus be cut off somewhere just after the top of the hump.9 However, for this study
to be manageable and the results somewhat comparable to those of C/D’s and
Rowthorn’s, I will treat the different sectors as separate economies.
9 See discussion on page 9 ff
8
The concept of sector is also something that needs to be briefly discussed, as well as the
reason for me choosing to study ten sectors specifically. There is no clear definition of
what constitutes a certain sector or how it should be circumscribed. As neither C/D nor
Rowthorn deals with sectors as analytical objects, little help can be found there on how to
categorize a sector. For simplicity, I have decided to take the Swedish trade unions that
cooperate with the Russian trade unions as givens for my definition of sectors within the
Russian economy. The trade unions in Russia, with which Swedish trade unions
cooperate through the LO-TCO Secretariat of International Trade Union Development
Co-operation, thus constitute the trade unions of separate sectors, but do so by my
definition only, rather than through some universal criteria. There are eleven Swedish
trade unions that have a cooperative project with Russian trade unions, but two of them
are cooperating with the same Russian trade union, and the number of
sectors/economies/analytical objects in my study is thus ten.
2. THEORIES ON CENTRALISATION, UNEMPLOYMENT AND WAGE
DISPERSION
2.1 The Calmfors/Driffill theory on centralisation and unemployment
Lars Calmfors and John Driffill10 (C/D) were among the first to develop a theoretical
framework of how to explain how macroeconomic performance of an economy is
correlated to the level of centralisation in wage bargaining structures. They argued that
the level of centralisation in wage bargaining has a great impact on the ability of an
economy to provide employment for its workers. C/D proved that the relative levels of
centralisation correlates in a U-shape with employment rates, or correspondingly, that
there is a hump-shaped relation between the relative levels of centralisation and wages.11
C/D presented a theory about how an economy works at different levels of aggregation:
from fully decentralised, where single companies and local trade unions are the primary
agents in wage-setting negotiations, via more aggregated sectors where bargaining is 10 Calmfors, Driffill, 1988 11 Calmfors, Driffill, 1988 p. 15
9
conducted at branch-level institutions to a fully centralized economy where wage
bargaining is conducted by a few large trade unions and a few centralised employers´
associations.
FIGURE 1 Calmfors and Driffill hump-hypothesis
Above is the hump-shaped and U-shaped relation between centralisation and real wages,
unemployment and employment respectively. The three points that are of certain interest
are points A, B and C.
The interpretation is straightforward; according to the C/D analysis countries where wage
bargaining is centralized (point C) would be expected to show low levels of real wages
and unemployment. In countries where wage bargaining is mainly conducted at a semi-
centralized level (point B), real wages and unemployment should be expected to be high,
and in countries where employers and workers negotiate at the local level (point A),
unemployment is supposed to be low.12 Below I will discuss the rationale behind the
theory in more detail.
2.1.1 Centralised wage bargaining
In economies where centralisation is high and wage bargaining is conducted by a few
large all encompassing organisations, trade unions will take the inflationary impact of
12 Calmfors, Driffill 1988 p. 15. Rowthorn, 1992 p. 506.
1
higher wages on the economy as a whole into account, and thus lower their demands on
wage-increases. The net effect for trade unions will not be positive if the rise in wage
levels is out-weighed by the price-rise so that real-income among members is in fact not
affected at all, but the companies are forced to fire workers to stay afloat. Therefore, the
ideal wage level of centralized trade unions is rather low since they have to take the
consequences of their actions on the economy as a whole into account. Lower wage
demands allow for firms and enterprises to hire a larger number of workers, thus
lowering unemployment.
2.1.2 Branch-level wage bargaining
Economies in which wage bargaining is conducted at the semi centralized or at branch
level each sector only constitutes a small part of the economy, and trade unions can not
be expected to take into account how wage levels in one specific sector affects the
general prise level of the economy. Demands for wages can thus be expected to be higher
than in fully centralised economies. Moreover, compared to fully decentralized systems
trade unions in semi centralized economies often attain a stronger position vis-à-vis
employers through reducing job competition among workers in the sector and can be
expected to be more successful in pushing through wage demands (increasing
unemployment). However, the net-effect of the rise in nominal wages might be nullified
by the general price-rise, decreasing real wages.
2.1.3 Decentralised wage bargaining
If wage bargaining is being conducted mostly at company level by many weak unions (or
by the employees themselves), wage levels are expected to remain low as the trade
unions generally have a weaker position vis-à-vis the employers and the market forces
are therefore allowed to play a larger role, thus keeping wages down. Also, the local trade
unions have to take into account the fact that if the company faces increased costs due to
higher wages, its competitiveness decreases and it will have to fire workers (members).
In their 1988 paper, C/D ranked 17 OECD countries on the degree of centralisation in
wage bargaining between the years 1974 and 1985. They specifically used the ranking of
the countries in their analyses instead of any nominal value, as they had more confidence
1
in the relative levels of centralisation among countries than of the magnitude of
differences. Schmitters (1981), Cameron (1984) and Blyth (1979) have also ranked the
same countries according to similar criteria, and gotten a similar ranking. Schmitters,
however, ranked the countries only on the basis of the trade union side, not the
aggregated level of centralisation, but did not get very different results from the others.
One of the modifications that C/D themselves pointed to will also prove important in my
analysis. They admit that within economies open to international trade, prices of products
are more or less taken as givens. A semi centralized system of wage bargaining in such
an economy will thus be more similar to a fully decentralized system in the case sectors
are chosen as analytical objects, as is the situation in the case of this paper. When treating
the sectors of the Russian economy as independent economies, the in-and-out-flow of
goods and services between sectors might be expected to have the same effect on prices
and wages as open international trade has on an economy.
Another point worth mentioning, which was briefly discussed earlier, is the fact that it
might not be possible to attain the perfect C/D hump shaped relation when analysing
sectors within an economy as the most central degree of bargaining, state level, is by
definition excluded from the analysis, indicating that the hump might be ´cut off´ half
way. However, limited mobility of the work force between sectors indicates that there
might be room for wage increases in semi centralised sectors resulting at least in a
positive linear relationship between centralisation and unemployment.
2.2 Rowthorn´s theory on centralisation and wage dispersion
Rowthorn´s (1992) aim was not to challenge the fundamental ideas of C/D but to extend
their analysis in another direction. Rowthorn criticized C/D on the basis of their concern
mainly with aggregate variables such as average wages and total employment. C/D did
not consider asymmetrical bargaining structures but assumed that all unions had the same
degree of organized power in the economy. However, Rowthorn pointed out that
empirical evidence indicates that wage differentials are much larger in some economies
than in others and that bargaining structures have an impact on the level of wage
1
dispersion. He concluded in his paper that economies where wage bargaining is
centralized wage dispersion is relatively low, whereas economies where bargaining is
heavily decentralized wage dispersion is much higher. Moreover, Rowthorn in his
analyses also found a negative relationship between wage dispersion and unemployment
as high wage differentials would mean low salaries and unskilled labour being absorbed
at the bottom end of the wage-scale.13
The C/D and Rowthorn theories above can be summarized in the following equations:
U = B(0) + B(1)C + B(2)C² (1)
WD = B(0) + B(1)C (2)
Equation (1) above is the equation used in the econometric analyses below of
centralisation and unemployment. The second equation above is the equation used in the
econometric analysis of centralisation and wage dispersion. B(0) is the constant variable,
B(1)C the linear variable and B(2)C² the quadratic variable. C = centralisation. WD =
wage dispersion.
2.3 About the data
In all sectors examined I have managed to get figures of nominal wages (minimum and
average) as well as unemployment figures. These data, however, are not by any common
standard among the trade unions and the way of measuring them may differ between
sectors and unions. Moreover, in no case have I been able to get a specific figure for
wage dispersion and I will thus use the somewhat more imprecise way of measuring
dispersion by dividing the average wage by the minimum wage in each sector. As briefly
discussed earlier, herein lies one obvious weakness of my research, as the data is different
and also varies in reliability among the sectors and unions. Some unions have very
reliable data, others have less reliable sources.
Moreover, both the level of unemployment in an economy (or sector) as well as wage
13 Rowthorn, 1992 p. 510 ff
1
dispersion can be expected to be affected by many other circumstances but centralisation
in wage bargaining structures only. However, in this study, I will take centralisation as
the single explanatory variable, as this was also the single most important explanatory
variable in both the C/D and Rowthorn analyses.
3. A DESCRIPTION OF THE STRUCTURE OF TRADE UNIONS IN RUSSIA
Most trade unions in Russia, as well as the ten trade unions I have examined, show a
similar internal structure. In fact, there is a formalized structure of central, regional and
local levels of organization and negotiations in all sectors with some degree of
unionisation. The financial structure is also similar among the trade unions as there are
national laws and regulations for trade union membership.
3.1 The financial structure of trade unions
Today it is no longer obligatory for workers to be members of a trade union, as it was
during Soviet times. In all trade unions examined, membership dues are the most
important and only really reliable source of income. All members of trade unions pay 1%
of their reported salary to the trade union in which they are members. From this 1%
roughly 65% stays with the local office, 25% goes to the regional office, 6% goes to the
central, or national, office to be redistributed between trade unions or used in nation-wide
campaigns and negotiations, and 4% goes to the FNPR, the central union that bargains for
minimum standards in legislation within the tripartite negotiations.14
According to four of the interviewees this financial structure is a great weakness since
different trade unions in the same sector differ greatly in strength at local level and there
are little resources being redistributed among different local unions. The strength of the
local union thus depends heavily on how many members it organises locally rather than
how many workers are organised in the entire sector. Moreover, according to many of the
interviewees, another weakness with the financial structure is that local trade union
representatives sometimes are more likely to spend money on meeting short sighted
popular demands by local members, such as providing dachas and banjas (Russian
14 Surin, 2006
1
recreation houses and saunas) for the use of members. These policies contributed to the
popularity of trade unions during Soviet times and today, for local representatives
depending on local popular support for their continued work, spending money on social
activities is often an easy way to keep members happy. However, because of this
structure, the leverage of trade unions in central bargaining procedures is not as large as it
could potentially be.
3.2 The Formal Structure of wage bargaining of trade unions
3.2.1 Tripartite agreements
In the national tripartite commission 30 representatives for Russian workers, employers
and the government respectively meet twice a year (the FNPR appoints 25 of the 30
representatives of workers. The other five are appointed by the much smaller national
trade unions VKT and the KTR). According to Russian general law, all legislation that is
to be put forward to the parliament that concerns labour rights or the situation of workers
must be discussed by the tripartite commission before-hand. However, according to the
international secretary of the FNPR; Anatolij Surin, this might give the trade union
movement valuable insight into the legislation process, but the commission rarely adopts
any proposals that are being put forward by the representatives of the workers. Instead,
the employers and the government representatives are often successful in agreeing on
laws that effectively reduce workers´ rights.
3.2.2 Central level bargaining
The ten trade unions examined in this essay are all members of the FNPR. There are also
two other central trade unions; the VKT and the KTR, which organize around 1,2 million
workers each. However, the FNPR is the only union that is recognized at state, as well as
at the international, level.
The FNPR organizes about 30 million workers at the moment (down from 42 million a
decade ago). The workers are members of 42 different primary sector trade unions or 79
different regional trade unions. The FNPR negotiates with the government (as well as in
the tripartite commission) over general agreements where minimum wages and standards
1
are indicated. These general agreements, however, are recommendations and the unions
may, or may not, be able to agree on the same or higher minimum standards when
bargaining at branch level. The general agreement minimum standards, moreover, are
often set extremely low and are thus of little or no practical use.
3.2.3 Branch level bargaining
All local unions are members of a certain branch trade union. At this branch level of the
specific sector, the representatives of trade unions meet with representatives from
relevant ministries or with employers´ associations´ representatives (if there are any).
Here, representatives try to reach branch agreements on minimum standards for the
specific sector. These might thereafter be implemented or renegotiated to the better at the
regional and local levels. However, if the employers´ association or individual employers
think they are not able to meet up to the demands or minimum standards of the
agreement, they can apply to the relevant ministry not to adhere to the agreements or
laws. The ministry might also, without consulting the trade unions, agree to accept lower
standards within a certain sector or for certain employers.
3.2.4 Regional bargaining
At the regional level, trade unions try to bargain with the regional employers´
representatives or regional public authorities to improve the general standards put
forward in the national general agreement and the branch level agreements. The success
of the trade unions at this level depends heavily on the strength of the trade unions
regionally and how important the workers of the sectors are considered to be for the
regional economy. However, according to most of the interviewees, it is at the regional
level that most trade unions manage to get some minimum standard improvements
through. These agreements formally cover all workers in the sector in the region, and
standards will differ less between individual enterprises in the same sector if the trade
union manages to get its demands through at this level.
3.2.5 Local bargaining
The ability to implement general standards at the local level depends on the strength of
the local trade union. Sometimes, general agreements, branch level agreements or
1
regional agreements are not being obtained, as the local union is not strong enough to
force the local employers to implement common policies. At other times, the local trade
union is strong enough to greatly improve the standards and have the leverage to force
the employers to agree upon and implement higher standards. Since the financial
structure of trade unions all give considerable decision power over economic resources to
local representatives, ambitions and skills of local representatives affect the effectiveness
of trade unions implementation of policies and standards may thus differ widely for
workers in the same sector at different enterprises.
3.3 The level of centralisation of employers
A feature that all interviewees mentioned as an important obstacle in the wage bargaining
process was the lack of centralisation among employers in the sectors. There is no
formalised structure of employers´ associations and most employers are not part of any
association at all. This makes it very difficult for the trade unions to secure that common
policies are being implemented, as many employers are not part of the agreements and
are thus not bound by them. It also seems like the sectors in which the trade unions
themselves think they are effective there is at least some degree of centralisation among
employers. The MMWU, a trade union in a sector with an effective employers’
organisation even helped start it up a decade ago and today the representative of the
MMWU is very satisfied with having a well-functioning employers’ association to work
with. In the transport sector, an employers association has also been started up (simply
through the trade union, TU, making employers in the sector meet at a conference) and
the trade union representatives are very confident about more effective negotiations in the
future. In sum, when asked what aspect needs to be improved first in order to get more
effective negotiations, almost all interviewees stated that centralisation and organisation
of employers in the sectors is the most important means.
4. CENTRALISATION OF COLLECTIVE BARGAINING
4.1 Comments on the method of ranking the ten trade unions
1
As the trade unions examined all show a similar structure of local, regional, central and
national representation and negotiations, as described above, I will try to measure the
level of centralisation by two standards that do not take the formal structure of the trade
unions into account. I have not been able to get any figure of some objective
measurement of the ranking of centralisation among the trade unions. Instead, I will rank
the unions on basis of what percentage of the workers in the specific sectors they cover.
Thus, a branch trade union that covers many of the workers in a particular sector will be
ranked ‘high’ on the centralisation ranking, and a trade union that covers few of the
workers in a sector will be given a lower position in the ranking. The second measure by
which I have chosen to rank the sectors is a weighted average of the level of
centralisation among the employees in the sector as well as among the employers. This
will be done in a more qualitative way completing the first ranking of employees with an
analysis of the arguments given by the interviewees about how centralised they consider
the employers in the sectors to be.
Another important note is that figures on average wages and wage dispersion may differ
widely between sectors not only because there is a great difference of productivity
between workers in different sectors but also because the trade unions might organise
workers at enterprises located in different parts of the country. For example, this seems to
be particularly important in the case of the telecom sector, where the minimum wage is
very low but the average wage is very high. One of the reasons for this is that most
companies in the telecom sector are located in Moscow, St. Petersburg and other major
Russian cities where wages are generally much higher. Thus, sectors mostly located in
the cities tend to show higher average wages than those located all over the country. A
table of average wages, minimum wages (measured in Rubles) and unemployment rates
for the different sectors can be found in the Appendix A below.
As for unemployment rates, some of the interviewees had figures for unemployment in
the sector, others estimated the unemployment rates during or after the interviews.
However, I have decided to make use of the figures given to me by the trade union
representatives without validating them further. This might also be a weakness in my
1
analysis, but is necessary in order to be able to conduct the research at all.
4.2 Short description of the trade unions
1. International Federation of Builders and Wood-workers (BWI) is a trade
union that today organizes 700.000 construction site workers (down from 3 million
during Soviet times) all around Russia and some former Soviet states (i.e. Ukraine). It
is a rather fragmented trade union, where companies are restructured all the time and
only 60-70% of the workers in the sector are actually covered by the collective
agreements agreed upon centrally. On the employers´ side there is some parallel
structure of employers organized into what resembles an employers´ association.
However, it does not represent all employers and the agreements are thus not always
implemented at company-level.
2. Health Workers´ Union of the Russian Federation organises about 3 million
state hospital workers, which is about 90% of all workers in state hospitals. The union
is the only union that organises publicly employed workers (except medical doctors)
in this sector. According to the interviewee it is a rather well-organised and not very
fragmented union. However, no trade union has access to the private hospitals where
about 10% of the total health work-force is employed. The unemployment rate is low
as there is a lack of nurses that are willing to work in the state sector. There is an
employers´ organisation to bargain with – representatives for the state or regional
authorities.
3. The All Russia Life Support Workers Federation organises about 1,5 million
workers which is about 70% of the total work-force in the sector. There is a high level
of local implementation of agreements that are being agreed upon at the central level,
as the union is rather well-organised and there is quite a level of coordination among
the local trade unions. On the employers´ side there are a few different employers´
associations that intervene at different levels in the negotiations.
1
The Russian Communication Workers´ Trade Union (RCWTU) organises
about 750 000 workers in the telecom and postal sectors. I will treat these sectors
separately as they show very different features.
4. In the Telecom Sector, which today is mainly privatised, the workers are
hardly
unionized at all as there are many private enterprises where unionization is
prohibited. There is no employers´ association. Since much of the telecom sector
is located in Moscow and the larger cities, average wage is very high although
minimum wage is still very low.
5. In the Postal sector, on the other hand, about 80% of all workers in the sector are
unionised. However, most of these are employed in the state owned companies. In
private companies (such as DHL, FedEx) there are no trade unions at all. In the
state run postal service trade union representatives bargain with representatives
for the public authorities at different levels. Among private employers, there is no
coordination.
6. The Russian Mining Workers Trade Union (MMWU) is one of the oldest and
most well organised trade unions in Russia. Roughly 95% of all workers in the
sector are members of the union, although it today is a rather privatized sector.
According to the interviewee many of the local unions manage to get through
agreements that give the workers higher standards than are being stipulated in the
central agreements. There is a well-functioning association of employers which
cover most employers in the sector.
7. The Education and Science Employees´ Union of Russia (ESEUR) is a well
organised trade union for education sector workers. It organises 82% of all
workers in the sector and there are no other unions in the sector. The ESEUR has
about 5,5 million members, of which 1,5 are still students. On the employers’
side, the ESEUR relies on the state and regional authorities for implementation.
2
8. The Agro-Industrial Trade Union Federation (IUF) covers 65% of all
workers in the agro-industrial and food-processing sector. Membership number is
about 2,7 million. There is no employers´ association but the union can have a
constructive dialogue with state owned enterprises. According to the interviewee
himself, it is a very decentralised and fragmented sector, where the local unions
try to bargain with little or no coordination among themselves.
9. The Russian Transport General Machinery Workers’ Trade Union (TU) is
a union covering 82% of all workers in the sector. Although the workers are rather
well organised, the employers in the sector are not very coordinated or organised.
The trade union has about 1 million members and according to the interviewee the
wages differ widely among workers in the sector in different enterprises and
different parts of the country.
10. The Russia Medical Doctors’ National Association (RNA) is the only nation-
wide trade union for medical doctors. However, membership levels are low and many
medical doctors prefer to work in private hospitals instead where trade union
membership is prohibited. In fact, the trade union only covers about 55% of all
medical doctors. There is hardly any coordination among the private employers, and
the state officials are rather weak, so minimum standards and wages are low.
Below is a table showing the data of the different trade unions as well as their ranking
and the calculated wage gap within the sectors.
2
TABLE 1 Sectors, Rank Orderings, Dispersion, Unemployment
Sector Covering Rank -Sq Rank(weighted) -Sq Dispersion Unemployment
RCWTU (tel) ”low” 1 1 1 1 17,60 4%
RNA 55% 2 4 2 4 5,33 5%
IUF 65% 3 9 3 9 8,75 20%
BWI 65% 4 16 5 25 2,67 5%
Life Support.. 70% 5 25 4 16 3,23 7%
RCWTU (post) 80% 6 36 7 49 1,60 8%
TU 82% 7 49 6 36 5,07 14%
ESEUR 85% 8 64 9 81 1,36 10%
Health workers.. 90% 9 81 8 64 4,00 ”low”
MMWU 95% 10 100 10 100 4,09 6% Notes: Rank = ranking on basis of trade union centralisation. Rank (weighted) = calculated as (level of centralisation of trade unions*level of centralisation of employers)/2. –Sq = squared. Dispersion = average wage/minimum wage. Some of the above figures need commenting. First of all, wage dispersion in Rowthorns’
essay is measured by using weighted observations in which earnings in each category are
weighted by the number of people employed and coefficients of variation are computed
for men and women separately. However, these kinds of individual data is very difficult
to obtain and I have thus used a different (and perhaps less precise) way of computing
intra-sector wage dispersion. Wage dispersion in this essay is computed by dividing the
average wage by the minimum wage in each sector. Second, in some of the following
regression analyses, I have chosen to exclude the figure of unemployment of the IUF
(observation no 3). This is because the figure is so extreme as to be expected to affect the
overall results significantly. I present the results of the regressions both with and without
2
this observation.
4.3 Unemployment, wage dispersion and centralisation; data analyses
4.3.1 A visual interpretation of data
Below, in figure 1 and 2, are the scatter plots of the observations of the ten sectors. In
figure 1 unemployment is measured on the Y-axis and the ranking of centralisation in
measured on the X-axis. In the two lower scatter plots the extreme observation no 3 has
been excluded. The interpretation of the dotted line (the observations of the
unemployment rates in the ten sectors) is not fully straight-forward, but it is possible to
interpret a hump-shaped relation between centralisation and the degree of unemployment.
This is also confirmed by the line illustrating the OLS-line15 as the level of
unemployment is lowest at the lowest and highest levels of centralisation and the level of
unemployment is higher in sectors which have semi-centralised wage bargaining
structures.
FIGURE 2 Scatterplots of Unemployment vs Centralisation
Ranking
Unem
ploy
men
t (%
)
1086420
20,0
17,5
15,0
12,5
10,0
7,5
5,0
Unemployment vs Centralisation
With observation no 3
Ranking (weighted)
Unem
ploy
men
t (%
)
1086420
20,0
17,5
15,0
12,5
10,0
7,5
5,0
Unemployment vs Centralisation (weighted ranking)
With observation no 3
15 OLS = Ordinary Least Squares
2
Ranking
Unem
ploy
men
t (%
)
1086420
14
12
10
8
6
4
2
Unemployment vs Centralisation
Without observation no 3
Ranking (weighted)
Unem
ploy
men
t (%
)
1086420
15,0
12,5
10,0
7,5
5,0
Unemployment vs Centralisation (weighted ranking)
Without observation no 3
Notes: The square dots and the dotted lines in the above scatter plots show the rate of unemployment (%) in the different sectors. The thicker lines show the corresponding quadratic regression lines. The ranking of centralisation ranges from 0-10 where 0 indicates the sector with least centralised wage bargaining and 10 indicates the sector where wage bargaining is most centralised.
In the following scatter plots, figure 2, the Y-axes show wage dispersion rates in the
different sectors and the X-axes still show the relative levels of centralisation. There
seems to be a linear negative relationship between the two variables, as wage dispersion
is lowest in the sectors that have the highest levels of centralisation. These results, both in
figure 1 and 2, would, if shown to be significant in a regression analysis, comply with the
C/D and Rowthorn analyses described at the beginning of the paper.16
FIGURE 3 Scatterplots of Wage Dispersion vs Centralisation
Ranking
Dis
pers
ion
1086420
20
15
10
5
0
Dispersion vs Centralisation
Ranking (weighted)
Dis
pers
ion
1086420
20
15
10
5
0
Dispersion vs Centralisation (weighted ranking)
16 Rowthorn, 1992 p. 510
2
Notes: The square dots and the dotted lines in the above scatter plots show the rates of wage dispersion in the different sectors. The thicker lines show the corresponding linear regression lines. 4.3.2 Regression analyses of data
The results presented in Table 2 and 3 help cast some light on how to understand the data
in Table 1 as well as in the scatter plots in figure 1 and 2. As for the data on
unemployment, the linear rank variable is positive and the squared variable is negative,
which indicate a hump-shaped relationship between the relative levels of centralisation
and unemployment. This is true both when the ranking of trade unions only is taken as
the explanatory variable as well as when the weighted measure is taken to be the
explanatory variable. However, the results become significant only when running a
regression analysis without the extreme observation no 3, and even in that case the results
are only significant when taking the level of centralisation among workers only as the
explanatory variable. Between wage dispersion and the relative levels of centralisation
there is an obvious negative linear relationship significant at the 5%-level.
TABLE 2 Regression Analysis: Unemployment, Dispersion and Bargaining Structure with observation no 3
Explanatory Variables
Centralisation of trade unions
Dependent variable Rank (Rank)-Sq R-Sq t-value(Rank) t-value((Rank)-Sq)
Unemployment 2,64 -0,230 13,1% 0,95 -0,92
Wage Dispersion -0,984 - 38,6% -2,24** -
Centralisation of trade unions and employers, weighted average
Rank(w) (Rank(w))-Sq R-Sq t-value(Rank) t-value((Rank)-Sq)
Unemployment 2,60 -0,229 12,7% 0,93 -0,93
Wage Dispersion -1,06 - 45,2% -2,57** -
Notes: The constant term is not shown in the equations in Table 2 and 3 but included in the analyses. Rank (w)= weighted ranking as shown in Table 1. * (**) denotes significance at the 10% (5%) level using a one-tailed t-test.
2
TABLE 3 Regression Analysis: Unemployment and Bargaining Structure without observation no 3
Explanatory Variables
Centralisation of trade unions
Dependent variable Rank (Rank)-Sq R-Sq t-value(Rank) t-value((Rank)-Sq)
Unemployment 2,47 -0,175 49,9% 1,81* -1,42
Centralisation of trade unions and employers, weighted average
Rank(w) (Rank(w))-Sq R-Sq t-value(Rank) t-value((Rank)-Sq)
Unemployment 2,41 -0,177 41,2% 1,62 -1,35
As can be seen in the tables hardly any of the results on the relation between
centralisation and unemployment is significant either at the 10% or the 5% level.
However, the linear negative relation between centralisation and wage dispersion seems
to hold at the 5% level.
TABLE 4 Regression Analysis: Unemployment and Bargaining Structure without the squared variable, with
observation no 3
Explanatory Variables
Centralisation of trade unions
Dependent Variable Const Rank R-Sq t-value(Const) t-value(Rank)
Unemployment 7,945 0,163 0,8% 2,04** 0,24
2
Centralisation of trade unions and employers, weighted average
Const Rank R-Sq t-value(Const) t-value(Rank)
Unemployment 8,47 0,059 0,1% 2,21** 0,09
Notes: the constant term is shown in these equations. Rank (w)= weighted ranking as shown in Table 1. * (**) denotes significance at the 10% (5%) level using a one-tailed t-test.
TABLE 5 Regression Analysis: Unemployment and Bargaining Structure without the squared variable,
without observation no 3
Explanatory Variables
Centralisation of trade unions
Dependent Variable Const Rank R-Sq t-value(Const) t-value(Rank)
Unemployment 4,19 0,593 29,6% 1,85* 1,59
Centralisation of trade unions and employers, weighted average
Const Rank R-Sq t-value(Const) t-value(Rank)
Unemployment 4,82 0,464 19,9% 2,03** 1,22
Notes: the constant term is shown in these equations. Rank (w)= weighted ranking as shown in Table 1. * (**) denotes significance at the 10% (5%) level using a one-tailed t-test.
Tables 4 and 5 show the results from estimations that are computed without the squared
variable. The constant term is significant at the 5% level in all estimations, although the
linear term is not. However, when calculated without the extreme observation no. 3, the
results of the estimations are close to being significant at the 5% level (Table 3 and 5).
4.4 Summary of results
After having examined the relation of the relative levels of centralisation among workers,
2
as well as a weighted measure of the relative levels of centralisation among workers and
employers, to the levels of unemployment in the sectors, it seems difficult to interpret any
hump-shaped relation between centralisation and unemployment. When including the
squared term in the analysis (Tables 2 and 3) the linear term is positive and the squared
term negative, which indicates a hump shaped relation between centralisation and
unemployment (as can also be seen in Figure 1). However, only one of the terms is
significant. When excluding the squared variable (Tables 4 and 5) the constant term is
significant in all analyses, and when excluding the extreme observation no 3 the linear
variable is close to being significant at the 5% level (Table 5). This might give some
support to the theory above that the hump might be ‘cut off’ half-way in an analysis of
sectors within an economy and instead prove a positive linear relation between
centralisation and unemployment.
The interpretation of the relation between wage dispersion and centralisation is perhaps
more straightforward. As the theory of Rowthorn’s suggests, there seems to be a linear
negative relationship between the relative levels of centralisation in wage bargaining and
wage dispersion, meaning that the sectors in which wage bargaining is most centralised,
there is less wage dispersion than in sectors where wage bargaining is decentralised. This
can be seen both in figure 2 as well as in Table 2, where both simulations show
significant results at the 5% level.
After having examined trade union structure in ten Russian sectors, one can conclude that
the Rowthorn theory on centralisation and wage bargaining largely seems to hold
whereas the C/D theory on how centralisation in wage bargaining structures affects
unemployment rates seems less valid. However, the reason for some of the results not
being significant is impossible to know, whether it be the data not being valid, because
there are too few observations or whether the relation actually does not exist in the
Russian context.
5. CONCLUDING REMARKS
Trade unions in Russia today face severe challenges in how to expand their activities and
2
improve effectiveness. Membership levels have dropped dramatically in all trade unions
and the internal financial structure makes is difficult to improve the work done at the
central level.
After having examined the ten sectors in focus of this essay there seems to be some
evidence that parts of the theoretical framework developed by Calmfors, Driffill and
Rowthorn, discussed at the beginning of this paper, holds also in the Russian context. A
linear negative relation between the relative levels of centralisation and wage dispersion
seems to exist, meaning that the higher the level of centralisation in wage bargaining
among workers (as well as among employers) the lower the level of wage dispersion.
However, the relation between the relative levels of centralisation and unemployment is
more difficult to interpret. There seems to be either a hump shaped or a positive relation
between the two variables when looking at the scatterplots and interpreting the signs of
the terms in the simulations although few of these results are significant. When excluding
the extreme variable no 3, a larger number of the results are either significant or close to
being significant at the 5% level, computed both with and without the squared variable.
One of the possible weaknesses with my analysis is that I have tried to apply the C/D and
Rowthorn analyses, which take countries as analytical objects, on sectors within the
Russian economy instead. This might be a difficulty both because one can expect the
dynamics of sectors within an economy to be very different from that of between
economies, but also because it is not clear what indicators C/D have used in order to rank
the countries. The ranking in this essay thus becomes somewhat arbitrary and a different
ranking would probably have affected the overall results significantly. Moreover, as
discussed earlier, the data is not fool-proof and in many cases rather subjective, which
also has a great impact on the results, as does the fact that I have only used ten
observations in my analysis.
Thus, taking these weaknesses into consideration while studying the results of the
estimations in this essay, the conclusion of my study has to be that the Rowthorn model
seems to hold in the Russian economy whereas the C/D model, although it at first glance
2
looks as if the hump shaped relation exists, is more difficult to interpret.
REFERENCES: Litterature: Rowthorn, R. E. “Centralisation, Employment and Wage Dispersion”. The Economic Journal, 102 (May 1992), pp. 506-523. Great Britain. Calmfors Lars; Driffill, John. “Bargaining Structure, Corporatism and Macroeconomic Performance”. Economic Policy, Vol. 3, No. 6 (Apr., 1988), pp. 13-61. LO-TCO Secretariat of International Trade Union Development Co-operation. http://www.lotcobistand.org/country/129, April-June 2006. Contact Person: Anders Åhlin. Interviews: Andreevitj, Misha, International Secretary and Program Coordinator. International Federation of Builders and Wood-workers. 10 May 2006, Stockholm. Lavrova, Olga, international Secretary. Health Workers’ Union of the Russian Federation. 20 June 2006, Moscow. Klimova, Olga, International Secretary. The All Russia Life Support Workers Federation. 20 June 2006, Moscow. Vinogradova, Olga, Main Program Coordinator. The Russian Communication Workers´ Trade Union. 20 June 2006, Moscow. Kirillova, Irina, Central Committee International Secretary. Russian Mining Workers Trade Union. 21 June 2006, Moscow.
3
Borisov, Vadim, Internatinonal Secretary. International Confederation of Free Trade Unions. 21 June 2006. Kolbashkin, Nikolai, Coordinator. Education and Science Employees´ Union of Russia. 22 June 2006, Moscow. Shapoval, Bogdan, International Secretary. Agro-Industrial Trade Union Federation. 23 June 2006. Surin, Anatolij, Main International Coordinator. National Federation of Russian Trade Unions (FNPR). 23 June 2006. Sarkisova, Valentina, General Secretary. Russia Medical Doctors’ National Association. Via e-mail, June 2006. Egorov, Alexey, International Secretary. Russian Transport General Machinery Workers’ Trade Union. Via e-mail, August 2006. Appendix A Trade union Minimum wage (R) Average wage (R) Unemployment rate
(%) BWI Subsistence level 4500 5 Health Workers'.. 750 3000 “low” Life Support.. 1860 6000 7 Telecom Subsistence level 13200 4 Postal.. 3500 5600 8 MMWU 1100 4500 6 ESEUR 1100 1500 10 IUF 800 7000 20 TU 750 3800 14 RNA 900 4800 5
3