You are on page 1of 180

Striking Numbers

Van der Velden

ISSN 0927-4618
IISH Research Paper 50

Striking Numbers.

New approaches to strike research.

2
3
01 4
56789

Editedvan
by der
Sjaak
van der
Sjaak
Velden
(ed)Velden

Striking Numbers
New approaches to quantitative strike research
Edited by Sjaak van der Velden

IISH-Research Paper 50
2012

ISSN 0927-4618
IISH Research Paper 50
A complete list of IISH-Research Papers can be found after page 176

Copyright 2012, the contributors and the International Institute of Social History
All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system, or transmitted, in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without the prior permission of the publisher.
IISH-Research Papers is an online publication series, originally inaugurated in 1989, by
the International Institute of Social History (IISH) to highlight and promote sociohistorical research and scholarship. Through distribution of these works the IISH hopes
to encourage international discussion and exchange. This vehicle of publicizing works
in progress or in a prepublication stage is open to all labour and social historians. In this
context, research by scholars from outside the IISH can also be disseminated as a
Research Paper.

Contents
Editors foreword

Introduction

The influence of business cycles on strike activity in Austria, Germany and


Switzerland
Julia Casutt

13

Strike volume in Ghana: Trajectory of labour strikes in Ghana


Edward Fokuoh Ampratwum

59

The structure and dynamics of the workers protest movement at the beginning
of the 20th century in Russia: Database analysis
Leonid Borodkin, Irina M. Pushkareva and Irina V. Shilnikova

71

To participate or not to participate? Demographic aspects of the first mass strike


in Sundsvall, Sweden in 1879
Maria Bergman

99

Why do workers strike? Looking for an answer using micro data on Leiden strikers
in 1914
Sjaak van der Velden
119
Beyond the average and the aggregate: Researching strikes in Canada
Linda Briskin

137

Building a repository for strike data: The search for micro data
Sjaak van der Velden

165

Contributors

175

Editors Foreword
In the early nineteen eighties, interest in the history of strikes and industrial conflicts
dwindled. This waning interest was probably related to the fact that the number of strikes
and industrial conflicts in the real world was also visibly diminishing. Just as the growing
interest among historians in the nineteen seventies had been a reflection of the upturn in the
workers struggle in the years around the protests of 1968, so did the decline in the number
of strikes from the mid-seventies result in a lessening of interest among historians.
Since this relationship between the social reality and the degree of interest
among historians has already manifested itself in the past, it is legitimate to question
whether that relationship will continue to exist. Will the unprecedented number of
industrial conflicts in booming China and the social unrest associated with the economic
crisis that started in 2008 lead to an increase in the number of historians investigating
this aspect of labour history? Some signs are already indicative of such growth, but it
may still be too early to draw firm conclusions.
If the growth continues, the profession will nevertheless look different from the
way it appeared in the nineteen seventies. In those days, many researchers - including
those studying labour disputes - focused on the lives of ordinary men and women,
whereas colleagues in a previous era tended to concentrate foremost on the history of
organisations. We have now moved on another 30 years and new research will probably
adopt a different perspective from research carried out in a different period.
The digitisation of research is a strong influence on current researchers. Data can
be stored on the Internet, where it is universally accessible, and can also be exchanged
over the Internet. Since the seventies, society has become increasingly individualistic,
which has also left its mark on the profession.
In 2008, digitisation and the resulting options for global research prompted the
International Institute of Social History (IISH) to launch the Collaboratories in Social
and Economic History. One of these was the Labour Conflicts Collaboratory. As part
of the initiative, I was given the opportunity to invite a number of colleagues to come
and discuss potential new perspectives for future research on strikes, based on their own
papers and how they themselves shape their research. Present in that workshop were
Paul Jonker (Finland), Carlindo Rodrigues de Oliveira (Brazil), Stefan Dormans (the
Netherlands), Edward Ampratwum (Ghana), and Leonid Borodkin and Irina Shilnikova
(Russia). Unfortunately, Oscar Edoror Ubhenin (Nigeria) and Julia Casutt-Schneeberger
(Switzerland) were prevented from attending due to circumstances, but we discussed
their papers nevertheless. Dr. Dave Lyddon from Keele University was invited to act as
external reviewer.
After the meeting, I made a selection from the papers that were discussed and I
approached a few more researchers who are also engaged in new ways of researching
strikes. The result is the book in front of you. Before the publication, Dr. Kurt Vandaele
of the European Trade Union Institute (ETUI) was asked to appraise the chapters, and
Prof. Lex Heerma van Voss submitted the selected texts to a final evaluation.
Hopefully this selection will provide an insight into the ways in which labour
historians are currently carrying out, and are able to carry out, their research.
Sjaak van der Velden, March 2012

Introduction
Sjaak van der Velden
Strikes in modern capitalist societies have been the subject of study since the first one
occurred. Because striking was considered a breach of law, the police were the first to
collect data on strikes. Strikes also being a clear expression of the social question,
economists, historians and national statistical bureaus soon followed suit. Labour unions
and socialist parties regarded strikes as a means to achieve power for the working
classes and their organisations. They therefore also published data and the stories of
strikes. A good overview of the different attitudes towards strikes is presented by
Roberto Franzosi in his illuminating study (Franzosi, 1995).
From the first issue of the Yearbook of the International Labour Organization
(ILO) in 1935, international-level data on strikes has been published. This data consists
of three indicators: the number of strikes (or labour conflicts), the number of strikers
and the number of working days lost due to strikes. Data from different countries is not
always comparable, because the national bureaus that provide information to the ILO
often differ in the ways they collect their data. Some countries include political strikes,
some set a minimum for the number of workers involved, others separate strikes from
lockouts. There are many such differences that the ILO can only mention (Chernyseev,
2003). Data published annually by the ILO therefore forms only a small proportion of
the data collected and published by national bureaus.
Using data published by the ILO and national bureaus, researchers have tried to
compare the strike patterns of different countries to obtain answers to questions
concerning the propensity to strike of workers from certain professions or regions. The
most well-known, much debated, but still quoted example is the publication written by
Clark Kerr and Abraham Siegel in 1954 (Kerr and Siegel, 1954). Their main theme was
the question of which professions are more prone to strike than others. Another
important theme was the general question of why workers strike. A few scientists tried
to answer this question in the early days of relevant research by comparing economic
cycles and strike development. They discovered that in general, workers are more likely
to strike during economic boom periods than during periods of recession, as the risk of
getting sacked is low and the chances of winning are higher. The easiest way to answer
the general question is to investigate strikers demands, which show that most workers
strike for higher wages or a reduction in working hours. The problem is that all workers
probably want such improvements to their working conditions, but not all workers will
engage in a strike to accomplish these aims. To answer the question of why workers act
differently requires more data than the ILO publishes. Their data has a strictly economic
basis and is insufficient to answer questions from a sociological or psychological angle.
Strike data for a number of levels can be published. The first is the national
level, for which the ILO and other supranational bureaus provide data. This data can be
used to study between-country differences, in other words, which countries are more
strike prone than others? The second level is the playing field of national bureaus and
also includes data at professional and/or regional levels. This data can be used to study
within-country differences between regions or professions. The third level comprises
microdata on individual strikes. Data at this level forms the basis of the first two levels,
but using it enables more questions to be answered using methods such as cross-

references between regions and professions. This cannot be carried out using only the
data published by the ILO and national bureaus. This third-level data can also be used to
supplement official data, which is sometimes incomplete. On the fourth level, we find
data about the individual strikers who took part in a labour conflict. This data is very
rare, but in combination with other information it may possibly be used to answer the
question of why individual workers do or do not participate in strikes.
In this volume, a number of studies are published that try to find answers based
on one of these four levels, while one study goes even further than these.
Schneebergers The influence of business cycles on strike activity in Austria, Germany
and Switzerland is a good example of using the data from the first two levels for
research into comparative history. With official data from three countries, sometimes
supplemented with additional data, she explores a problem that has been paramount
from the beginning of strike research: Is there a relationship between business cycles
and strike activity? Using highly sophisticated statistical techniques she has confirmed
previous findings, which suggest that short business cycles in particular structure the
occurrence of strikes. It is shown that short business cycles played a role during the
entire period investigated, at least in the case of Austria and Germany, although the
intensity of the influence and the strength of the correlation varied over time. Business
cycles were able to explain the timing of strikes in specific periods characterised by a
particular combination of economic development and institutional setting. However,
economic fluctuations alone cannot determine the extent of the increase or decrease of
these periodic movements. For a complete picture of strike activity, it is important to
consider short-term as well as long-term fluctuations.
Like many other researchers, Schneeberger encountered the problem of missing
data. This brings us to the question of how data is collected. Further to the commonly
agreed problems resulting from different methods of collecting data, there is an
additional problem in developing countries due to political issues. In Strike volume in
Ghana. Trajectory of Labour strikes in Ghana, Edward Fokuoh Ampratwum presents a
short history of the Ghanaian strike movement. He also mentions existing problems
with collecting strike data and provides new, micro-level data for the recent period. This
data is not available in international publications by the ILO and therefore unique for
strike researchers unfamiliar with African labour history.
The collection of strike data described by Edward Fokuoh Ampratwum is at the
third level. Another example of using such data is presented by Borodkin, Pushkareva
and Shilnikova in The structure and dynamics of the workers protest movement at the
beginning of the 20th century in Russia. Database analysis. Soviet historiography deals
with strikes as an expression of the working class. However, it does not deal with the
working class as such, but with a class acting under the leadership of the communist
party. Therefore strikes are mostly viewed from a political angle. A unique meta source
made it possible to build a database and add information on strikes to the often quoted
official statistics. The new insight resulting from Borodkin, Pushkareva and
Shilnikovas research is that most strikes were about labour conditions such as wages,
rather than about political issues. However, they also discovered a decreasing
proportion of strikes at individual enterprises, whereas general strikes, demonstrations
and meetings increased in number. This proves a growth of organisation and an increase
in the political component of strike dynamics. Nevertheless, at the same time strikes did
not lose their importance as a means of solving labour conflicts.

By correlating variables which characterise the different aspects of labour


conflicts, the database analysis offers an opportunity to verify hypotheses formulated
earlier on the basis of case studies or incomplete aggregated data. The Russian work is
only the beginning of such analyses and in the future, more sophisticated statistical
techniques could be applied. The publication of the database and the first efforts to
analyse the microdata are only the start of a new era in Russian strike research.
In addition to data concerning the first three levels, it is also possible to collect
data at the level of individual strikers. Is it possible to detect the reasons why certain
individuals join or even start strikes while others do not? Modern psychologists and
sociologists are sometimes in a position to carry out survey research and ask people
during or (shortly) after strikes about their reasons for joining or abstaining. In historical
research this is not possible, so we have to look for answers in historical sources. Such
research has been carried out by Maria Bergman from Sweden and Sjaak van der
Velden from the Netherlands.
Bergman looks for demographical differences between strikers and non-strikers
during the 1897 Sundsvall sawmill strike. She concludes that older married workers with
families were the most cautious and hesitant in making the decision of whether to join the
strike. Van der Velden, in his efforts to discover the demographical characteristics of the
1914 cotton-printing strikers in Leiden, came to similar findings. The profile of the 1914
striker in Leiden is one of a semi-skilled male who earned slightly more than non-strikers
prior to the strike. He was married and young, while he and his wife had few children. The
striker acted regardless of his religious beliefs and was not very loyal to the company. The
conclusions by Bergman and Van der Velden are based on two small examples, but invite
more historical research at the individual level.
All the research mentioned above primarily uses the three strike indicators that
are found in official data: the number of conflicts, the number of participants and the
number of working days lost. Linda Briskin in her contribution Beyond the average
and the aggregate: Researching strikes in Canada shows that there is more to strikes
than just these numbers. She advocates a labour militancy framework, which presents
an alternative to the employer focus on time lost, the governmental concern of
measuring the economic impact of stoppages, and the scholarly emphasis on strike
determinants. She wants to go beyond the aggregates and illuminate workers
experiences of strikes. Her approach makes visible the local and the personal aspects,
and supports new ways of looking at overall patterns of labour militancy, highlighting
very long and very short strikes, strikes by a few workers or those involving many
thousands, and strikes for first agreements. By going further than the bare numbers she
enriches quantitative data by not only offering a window into the lived experience and
texture of strikes, but also giving access to the voices of strikers.
All the researchers put forward a strong argument for not only using microdata
covering the usual indicators, but using a multitude of data. In the final chapter, Van der
Velden once again strongly stresses the need for the collection of such microdata. He
also offers a format that can be used to build standardised databases. With as much data
as possible it will be feasible for researchers to construct their own sets of variables
instead of being dependent solely on official data. When the focus had not shifted to the
economic consequences of strikes, statistical bureaus often published microdata. A first
step could be to make this data available electronically. For later periods, additional data
must be collected.

10

The International Institute of Social History has opened a repository where all
kinds of strike data can be stored. Hopefully this will be the beginning of a large
international database and the start of digging deeper into the subject of strikes and the
motives for striking.

11

References
Chernyshev, Igor (2003), Decent work statistical indicators: strikes and lockouts
statistics in the international context, Bulletin of labour statistics no. 3, XIIIXIV.
Franzosi, Roberto (1995), The puzzle of strikes. Class and state strategies in postwar
Italy, Cambridge.
Kerr, Clarke and Abraham Siegel (1954), The interindustry propensity to strike an
international comparison, in: Arthur Kornhauser et al., Industrial Conflict, New York,
p. 190-212.

12

13

The influence of business cycles on strike activity in Austria,


Germany and Switzerland
Julia Casutt
1. Introduction
Labour disputes, such as strikes or lock-outs, are an integral part of industrial relations.
The 20th century appears especially conflictual, considering the high level of strike
activity in the 1920s and the period after World War II, and the international wave of
strikes in the late 1960s and early 1970s. After this peak the following twenty-five years
showed a considerable decline in labour unrest, particularly in the industrialised
European and North American countries. However, increasing strike activity has been
reported for developing, and less developed, countries such as South Korea, Nigeria,
Brazil, China and Russia.1 Academic research in this field is primarily interested in
analysing and explaining the long-term patterns and trends of industrial conflict.2 The
shape and determinants of short-term fluctuations have been addressed to a much lesser
extent. This is especially the case for the countries under analysis.
This first section of the paper analyses the short-term fluctuations of strike
activity in the mainly German-speaking countries of Austria, Germany and Switzerland
in the period from 1901 to 2004. The focus lies on the relationship between strike
patterns and business cycles, the latter being generally assumed to exert influence on the
cyclical structure of industrial conflicts. Is there an impact of the business cycle on the
occurrence of strikes, and what form does the potential relationship between the two
take?
Clearly, this kind of research question might not be able to explain why the
intensity of strike activity varies over time, and in what way. Moreover it may not
contribute to the more fundamental question of why strikes occur at all. It is, however,
able to clarify whether the timing of strike activity exhibits a cyclical pattern and what
role might be attributed to business fluctuations with regard to these periodicities.
Section 2 reviews existing literature and findings concerning the cyclical pattern
of strike activity and its relationship with business fluctuations. After that, the applied
method and utilised data are introduced in sections 3 and 4. Section 5 presents and
discusses the empirical results for all three countries under analysis. The last section
presents conclusions.
2. The link between business cycles and strike activity
Although strike activity did not follow a predictable pattern during the 20th century, the
occurrence and rise of labour disputes in this period cannot be viewed as random. As a
result of the industrial revolution in the 19th century, the structure of European and North
1

For a discussion of recent trends in global strike activity, see Perry and Wilson (2004).
A study of labour conflicts from a global-historical perspective is by Silver (2005). For the period from
1968 on, see Van der Velden et al. (2007).
2

14

American labour markets started to change. Whereas in the centuries before, the majority of
population worked dependently or independently in agriculture, industrialisation gave
rise to mass employment in this emerging sector. This shift in the structure of the working
population enhanced the proportion of dependent employment, which in turn had
implications for the occurrence of industrial conflicts. Considerable parts of the labour force
were now increasingly dependent on salaries and wages. Thus, the outcome of negotiations
and disputes affected not only the level of payments or working conditions, but in a broader
sense general living conditions. In this regard, strike activity could be viewed as one
manifestation of an ongoing struggle for power between social classes over the distribution
of resources, principally (though not exclusively) national income.
However, to date there is no generally accepted theory which is able to explain
the occurrence, frequency and incidence of strikes. There is still no consensus as to
which conditions trigger the turning of structural opposition between employers and
employees into open conflict and what role may be attributed to business fluctuations
within this framework.
One of the earliest observations concerning the cyclical pattern of strike activity
is that it increases during economic expansions and declines in periods of contraction.
Karl Marx had already described this phenomenon in 1853: (...) the great alternative
phases of dullness, prosperity, over-excitement, crisis and distress, which modern
industry traverses in periodically recurring cycles, with the up and down of wages
resulting from them, as with the constant warfare between masters and men closely
corresponding with those variations in wages and profits. (Marx, 1853).
The quantitative analysis of strike activity has a long tradition3 but the existence
of a co-movement between strikes and business cycles is still contested: () there is
persuasive although not conclusive evidence that the frequency (and more importantly
and more doubtfully) the incidences of strikes are positively related to general cyclical
movements in the economy. (Kennan, 1986: 1133).
Explanations for such a (pro-cyclical) pattern of business and strike fluctuations
have mainly focused on the behaviour of workers and trade unions. Economic upswings or
boom phases are considered generally to improve the bargaining position of labour, because
unemployment rates tend to be low and claims for proper, or higher, proportions of national
income arise in a context of higher economic output. This framework is regarded as
conducive to the more militant behaviour of labour and its institutions when advancing
claims. Therefore, the probability of the occurrence of strikes increases during these phases
of the economic cycle. By contrast, in phases of economic contraction the position of labour
is weakened as unemployment rises and the overall distribution leeway decreases. This in
turn reduces the probability of strike action.
As a consequence, the strike pattern reflects the structure of the cyclical business
pattern. However, this correlation does not imply that the business cycle is the sole
determinant of strike activity. The causes of strikes are manifold and each represents a
singular event in its own context with its own history. Clearly, economic causes such as
wage claims are often an important factor in calling a strike. With regard to the cyclical
structure, the timing of strikes seems to be particularly influenced by economic
fluctuations. Albert Rees was one of the first scholars to take up this idea (Rees, 1952).
3

The majority of related studies originate in Anglo-American countries and also deal with this region. A
potential reason for this fact might be the early availability of ample data on strikes and business
indicators, particularly in North America. For an overview on quantitative strike research see the
contributions of Shalev (1980), Kennan (1986) and Franzosi (1989; 1995).

15

According to him, a positive correlation between strikes and business cycles results
from the view that strikes are typically conducted by business unions which are
always seeking gains for their members, and will adjust their strategy to business
fluctuations so as to maximise these gains (Rees, 1952: 371).
Rees was analysing the experience in the United States, which differs in many
ways from the European countries that are the subject of this paper. In particular,
differences concern the institutional setting, such as the level of centralisation of
collective bargaining or the general organisation of trade unions.4 The countries under
analysis, especially Austria, are characterised by comparatively well-regulated industrial
relations, with the strong involvement of unions in political processes. Although the
notion of business union might be misleading for these countries, the analysis will
reveal whether these unions adopted strike strategies adjusted to the flow of business
fluctuations.
The specific question concerning institutional development, and its relation to
short-term strike fluctuations viewed from a long-term perspective, is also addressed in
a study by David Snyder (Snyder, 1975). The author analyses annual strike activity in
Italy and the United States between 1900 and 1970, as well as in France from the 1870s
until 1966. He addresses the question of whether changing patterns of institutional
settings show an influence on the economically determined, short-term fluctuations in
strike activity. Snyder argues that previous research found a relationship between strikes
and economic fluctuations only in specific periods, characterised by the distinctive
institutional settings found in pronounced and stable labour institutions. Splitting the
investigation period into pre-war and post-war phases, regression results indicate,
especially in the case of the United States, a strong correlation between the institutional
setting and the underlying fluctuations in labour conflicts. For the United States,
economic determinants appear to be more important to strike fluctuations in the postwar period than in the pre-war period (Snyder, 1975: 26970).
The author explains this result by the weak and less marked industrial relations
dominant in the United States in the pre-war period. This kind of institutional setting is
usually characterised by the absence of stable, functional unions as well as by
institutionalised collective bargaining. As a result, industrial conflict is likely to be
determined to a great extent by political factors. However, once stable industrial relations
are established, economic determinants tend to become more important for the structuring
of strike activity: To the extent that national labor relations systems move towards
increased size and stability of union membership, institutionalization of collective
bargaining and political integration of labor, the determination of strike fluctuations shifts
from a primarily political to an economic process (Snyder, 1975: 275).
Although Snyders results apply mainly to the United Stated, he presents a
comprehensive approach which is able to take institutional, as well as long-term
economic, developments into account with regard to the explanation for short-term
fluctuations in strike activity. This is particularly the case when analysing longer
periods of strike behaviour in countries that have experienced several political and
economic interruptions. Snyders analysis moreover reveals the restricted nature of
studies which assume that the collective bargaining context is the natural framework of
strike activity.5
4

With regard to European trades unions, particularly the German example, see Hyman (2001).
Starting with the bargaining theory of Ashenfelter and Johnson (1969), subsequent literature,
particularly in North America, tended to restrict the analysis of strikes to the context of bargaining

16

A further aspect in the quantitative analysis of strike activity concerns waves of


strikes and the so-called Kondratieff swings. Kondratieff tried to identify economic
cycles of about fifty years length: the so-called long waves that he attributed to
technical innovations (Kondratieff, 1926). The long wave theory has always been
controversially discussed among economists, but gained popularity in other social
sciences.6 Ernesto Screpanti investigated a hypothesis of Kondratieff which links strike
waves to long economic cycles (Screpanti, 1987). According to Kondratieff, class
struggle tends to increase during economic upswings as a result of increasing social
tensions caused by fast economic growth. Employing principal component analysis for
France, Germany, Italy, United Kingdom and the United States for the period from
1864 to 1982, this hypothesis seems to be confirmed. The big strike waves, however, do
not occur during upswings. The major outbreaks of labour unrest tend to be centred on
the upper turning points of the long cycles (Screpanti, 1987: 111). During periods of
prosperity there is an intensification of conflicts but these never reach the levels of a big
upheaval. Screpanti explains this finding by accumulated tension and frustration on the
part of the working class, which tend to explode at the end of the period forming the
peak of the prosperity phase (Screpanti, 1987: 111). For the upswing periods, strike
activity is described as follows: () it seems to follow a regular course, as if the script
of the industrial relations comedy would have required workers, in order to gain the
welfare improvements to which they feel prosperity entitles them, to play a moderate
periodic intensification of strike activity through the alternative vicissitudes of the
shorter cycle (Screpanti, 1987: 111). Taking this consideration into account, the
analysis of the short-term regularities of social conflicts may also contribute to a deeper
historical understanding, and possibly shed light on less visible phenomena. Moreover,
an analysis of this kind is able to integrate business fluctuations in a meaningful way,
because economic activity is primarily dominated by shorter cycles.
For the countries under analysis, the cyclical structure of strikes and their
relationship with the business cycle has not been investigated extensively.7 Figures 1
through 3 and Table 1 depict days not worked8 and strike frequencies for all three
countries during the period under investigation. Particularly after World War II, strike
activity in Austria and Germany declined substantially compared to the inter-war years
and before. This pattern is also apparent in Switzerland. This decline of strike activity
was also accompanied by economic and political changes. Whereas the pre-war period
was characterised by economic and political instability, after 1945 these countries
experienced a stable period of growth with a considerable integration of the working
class and their representatives into the post-war societies and states. Figures 4 and 5
situations. This trend was also reflected in the change to using contract data instead of aggregated strike
activity. Pioneer studies in this field are Tracy (1986) or Vroman (1989). A study which analyses both
aggregated and contract data for Canada is by Harrison and Stewart (1994).
6
see Hobsbawm (1964: 148).
7
Studies dealing with short-term fluctuations of strike activity are Kaelble and Volkmann (1972) and
Hagelstange (1979) for Germany, Hirter (1988) for Switzerland, and Botz (1978) and Otruba (1981) for
Austria.
8
Days not worked (DNW) is the most comprehensive strike measure. It captures the annual number of
workdays cancelled due to strikes and lock-outs. Thus, this indicator implicitly measures the duration, as
well as the involvement, in conflicts. Alternative labels for this measure are work days lost, man days
lost, man days idle, strike days and strike volume. The present study prefers to use days not worked or
strike days, because these terms tend to be the most neutral. A detailed discussion on the applied strike
measures, the time periods covered by the indicators and their sources follows in the data section.

17

present actual GDP per capita growth rates for all three countries. The first graph
illustrates the economic struggle in the first half of the last century, with negative
growth rates for long periods. This stands in contrast to the unprecedented growth in the
period following World War II which, except for Switzerland, is curtailed only from the
1970s and 1980s onwards. In this period, social democratic parties gained increasing
political power and started to participate in governments, and trade unions started to be
sustainably respected as bargaining partners in collective agreements.
The main goal of this analysis is to identify and explore the common pattern of
strike activity and business fluctuations in the 20th century, in order to determine
whether short-term fluctuations in strike activity conform with business activity.
Table 1. Days not worked and strike frequency in Austria, Germany and
Switzerland, 1900-2004

1900-1909
1910-1919
1920-1929
1930-1939
1940-1949
1950-1959
1960-1969
1970-1979
1980-1989
1990-1999
2000-2004

Austria
Germany
Average per decade
Average per decade
Frequency DNW
Frequency DNW
592
197
1.814
337
98
1.833
322
189
2.272
n.d.
n.d.
n.d.
n.d.
14
n.d.
n.d.
10
56
n.d.
16
55
13
3
155
6
1
n.d.
2
1
n.d.
5
40
n.d.

n.d. = incomplete data


Source: see section data

111
149
277
n.d.
n.d.
16
5
19
8
4
1

Switzerland
Average per decade
Frequency DNW
134
116
n.d.
n.d.
n.d.
n.d.
24
13
24
11
5
2
2
1
8
1
2
0
4
1
6
3

18

Figure 1: Days not worked and strike frequency in Austria, 1900-2004


Days not Worked

Frequency

4000000

1200

3500000
1000
3000000
800
2500000

2000000

600

1500000
400
1000000
200
500000

19
00
19
04
19
08
19
12
19
16
19
20
19
24
19
28
19
32
19
36
19
40
19
44
19
48
19
52
19
56
19
60
19
64
19
68
19
72
19
76
19
80
19
84
19
88
19
92
19
96
20
00
20
04

Year

Days not worked

Frequency

Figure 2: Days not worked and strike frequency in Germany, 1900-2004


Days not worked

Frequency

40000000

6000

35000000
5000
30000000
4000
25000000

20000000

3000

15000000
2000
10000000
1000
5000000

Year

Days not worked

Frequency

2002

1999

1996

1993

1990

1987

1984

1981

1978

1975

1972

1969

1966

1963

1960

1957

1954

1951

1948

1945

1942

1939

1936

1933

1930

1927

1924

1921

1918

1915

1912

1909

1906

1903

0
1900

19

Figure 3: Days not worked and strike frequency in Switzerland, 1900-2004


Days not worked

Frequency

900000

300

800000
250
700000
600000

200

500000
150
400000
300000

100

200000
50
100000
0

19
00
19
03
19
06
19
09
19
12
19
15
19
18
19
21
19
24
19
27
19
30
19
33
19
36
19
39
19
42
19
45
19
48
19
51
19
54
19
57
19
60
19
63
19
66
19
69
19
72
19
75
19
78
19
81
19
84
19
87
19
90
19
93
19
96
19
99
20
02

Year

Days not worked

Frequency

Figure 4: Real GDP per capita growth rates, 1900-1949


Percent
40%
30%
20%
10%
0%
-10%
-20%
-30%
-40%
-50%
-60%

Year
Austria

Source: see section data

Germany

Switzerland

19
48

19
46

19
44

19
42

19
40

19
38

19
36

19
34

19
32

19
30

19
28

19
26

19
24

19
22

19
20

19
18

19
16

19
14

19
12

19
10

19
08

19
06

19
04

19
02

19
00

-70%

20

Figure 5: Real GDP per capita growth rates, 1950-2004


Percent
20%

15%

10%

5%

0%

-5%

20
04

20
02

20
00

19
98

19
96

19
94

19
92

19
90

19
88

19
86

19
84

19
82

19
80

19
78

19
76

19
74

19
72

19
70

19
68

19
66

19
64

19
62

19
60

19
58

19
56

19
54

19
52

19
50

-10%

Year

Austria

Germany

Switzerland

Source: see section data

3. The use of spectral analysis in the investigation of strike activity


For the identification of cyclical patterns of variables and their interrelation, spectral
analysis techniques are applied. Basically it is assumed that time series exhibit a
diverse, cyclical structure that is difficult to identify based on purely descriptive
methods. This concerns in particular simple calculations and visual analysis, as were
applied in the beginnings of quantitative research into strikes.9 Further, regression
analysis cannot identify the cyclical structure, because the estimated parameters do not
provide information about potential cycles and their intervals.
To identify the complex cyclical structure of a time series, spectral analysis
techniques are useful, because they are able to determine the dominant cycles in univariate
series as well as the dominating common cyclical patterns of two or more variables. With a
technical extension it is moreover possible to evaluate changes in the cyclical structures and
their co-movement over time. The applied method thus reveals the cyclical structure of the
strike series in the analysed countries during the investigation period, and further explores
this structure in interaction with the business cycle.
To become more familiar with the applied method, the present section discusses
the contribution of Gerald Scully, who investigated the cyclical structure of strikes and
business cycles in the United States for the period from 1919 to 1969, applying spectral
analysis techniques (Scully, 1971). According to Scully, the feasibility of decomposing
the cyclical structure of a given series is one of the advantages of the applied technique:
9

See Hansen (1921), Yoder (1940) or Jurkat and Jurkat (1949).

21

In regression analysis all the frequencies are lumped together. Spectral analysis
examines all possible frequencies and produces a catalogue of all the frequencies in the
series scaled in terms of their respective contributions to the entire variance in the
series. As a result, all important cycles can be determined (Scully, 1971: 364). This
kind of decomposition is also possible for the multivariate case, which is important with
regard to the question of a potential correlation between strikes and business cycles.
Scully fails to produce evidence for this thesis and concludes that strike and business
fluctuations have been generated by fundamentally different factors (Scully, 1971: 370).
According to his findings, strike fluctuations are dominated by cycles of twelve months,
whereas the series of business fluctuations is characterised by cycles at lower
frequencies, corresponding with intervals of 64 to 96 months. Co-movement between
strikes and business cycles could not be found in the classic business cycle frequency
ranges, although there was a correlation for very short cycles of two years.
Scully infers that the business cycle has no impact on strike occurrence and
fluctuations, thus: government, labor, and business will gain little in improving the
predictability of the outcome of a labor-management contract dispute by focusing
attention on the point on the business cycle () (Scully 1971: 374).
Scullys study has to be regarded as a pioneering work in this field, although he
disagrees with the prevalent opinion on the pro-cyclicality of strikes. One weakness in the
study is the fact that the author presents average spectral estimates for the entire period
under investigation. Thus, in the univariate case for the single series, as well as for the
comparison of two series, the results shrink to one single number representing the entire
investigation period. Forcing the relationship between macroeconomic fluctuations and
strikes to be constant over time seems to be overly restrictive. As suggested by previous
findings, such as by Snyder (1975), business cycles might help to explain strike fluctuations
only during specific periods, i.e. periods of growth with stable political and industrial
relations. Only in phases of growth do bargaining parties have leeway for concessions,
which in fact is a precondition for a bargaining situation to emerge. In periods characterised
by economic decline and political instability, other determinants might have a stronger
impact with regard to the cyclical pattern of strike activity. To overcome this restriction, the
method applied herein follows Scully and utilises frequency-domain techniques, but the
approach is extended to allow for variation over time, generating spectral estimates for each
year of the investigation period.
4. Method
Spectral analysis converts a time series into a set of superimposed harmonic waves. This
conversion is called a Fourier transform. The spectrum measures the importance of each
frequency ( ) in terms of the contribution of a cycle with a particular frequency to the total
variance of the series. The frequency is measured in cycles per unit of time, dependent on the
underlying series. Basing the analysis on vector autoregressive models (VARs) it is possible
to generate time-varying results, because the parameter matrices of the VAR model can be
treated as time-dependent. Thus, instead of generating an average result for the entire period
under investigation, spectral results can be obtained for each year of the period. This
extension allows eventual changes in the cyclical structure of the analysed series, and their comovement over time, to be followed. The appendix provides a formal description of the
applied method. This section will omit technical details, but instead present the three

22

important measurements that can be derived from spectral decomposition in order to examine
the cyclical structure of one or more time series.
The first measurement to be discussed in the results section is the proportion of
total variance of the strike series. This reports the proportion of a particular cycle length
in respect of the total variance of the series for each year of the investigation period. In
the results section, this proportion is calculated for the ranges three to five, five to seven
and seven to ten years, which represent traditional business cycle durations. Spectral
analysis does not identify causalities but correlations. However, it is safe to assume that
it is the economic environment which influences strike activity, and thus restrict the
analysis to the investigation of these classic business cycle lengths. Accordingly, no
other potential cycles of strike activity will be investigated because, even though such
cycles may exist, they are irrelevant with regard to the research question here.
Nevertheless, there may be genuine strike cycles of these respective durations which
exist independent of potential economic impacts. Such cycles can be identified by the
proportion of total variance.
For the analysis, the classic ranges of the business cycle have been chosen. The
seven to ten year cycle is known as the Juglar cycle and is generally regarded as the
classic business cycle. This cycle is attributable to equipment investment. The Juglar
cycle is superimposed by the shorter three to five year Kitchin cycle, which is rooted in
inventory investment (Zarnowitz 1992: 239). According to literature, it is this cycle in
particular which is supposed to structure rises and falls in strike activity.10 The presence
of these cycles tends to be a well established empirical fact in related literature.11
The dominating frequencies as reported by the proportion of total variance
contain important information concerning the structure of a series, such as days not
worked. For the research question, it is of particular interest to examine the interrelationship between two different series: namely strike days and business fluctuations.
For this purpose, the cross spectra (off-diagonal elements of the spectral density matrix)
are used to calculate the squared coherency sc( ) , a measurement that is similar to the
correlation coefficient.
This can be used to analyse the relationship between a series Yt (e.g. strike days)
and a series X t (e.g. business activity) at a frequency by measuring the extent to which
the variance of cyclical components of the series Yt in a particular interval, such as three to
five years, can be attributed to corresponding cyclical components in the series X t . This
measurement is called explained variance and is also reported in the results section.
Explained variance does not provide information about the co-movement of two series. For
this purpose, it is possible to decompose explained variance into in-phase and out-of-phase
components. This measures the extent to which the two series reach upper and lower
turning points at the same time in a particular frequency interval. Figure 6 gives a graphical
illustration of the spectral decomposition used in this chapter.

10

See the discussion of Levitt (1953) and Goldner (1953) with regard to this question.
Evidence for the appearance of these cycles in industrial output for Germany in the late 19th century up
to 1913 has been found in AHearn and Woitek (2001). Bergman et al. provide evidence for the presence
of a short cycle of about five years in Germany from 1948-1995 (Bergman, Bordo et al. 1998: 74). Other
related studies are Hillinger and Sebold-Bender (1992), and Woitek (1996). Investigations on business
cycles in Austria can be found in Brandner and Neusser (1992) or Scheiblecker (2002). A general
discussion of historical business cycles is by Solomou (1998).
11

23

Figure 6: Spectral decomposition12

5. Data
The present analysis is based mainly on data derived from work stoppage statistics
compiled by the International Labour Office (ILO).13 The ILO data has been amended
where possible with additional information taken from various sources, such as
historical statistics and archives. Since 1927, the ILO has published annual data
concerning dispute frequencies, workers involved and days not worked due to strikes
and lock-outs. The conflicts are disaggregated at branch level whenever possible. The
ILO receives its data from various institutions, such as national employment agencies,
employers associations and trade unions. From the very beginning, differences in
collection practices, definitions and procedures caused problems. There are no general
standards on which strikes to include. This especially concerns the question of the type
of strike, such as economic versus political conflicts. In addition, there are differences
with respect to the so-called minimum criteria. These criteria determine on a national
level the requirements for a conflict to be included in the statistics. They vary not only
between countries, but also within countries over time. Moreover, varying legal
frameworks influence the quality of the data. In some countries, for example Germany,
public institutions are compelled by law to report strike activity. In contrast, no such
legal requirements exist in Austria and Switzerland.14
12

All computer programs which are necessary to apply spectral analysis techniques as well as Figure 6
have been provided by Ulrich Woitek. The programs are written in MATLAB.
13
This data is compiled in the Yearbook of Labour Statistics which is published annually. For the present
study, all volumes from 1936/1937 to 2004 have been used (International Labour Office ILO, vol.
1936/1937-2004).
14
For a discussion on collecting problems see Perry and Wilson (2004), Lyddon (2007) and the
International Labour Office (1993).

24

To overcome these complications, the common practice for comparative analysis


is to focus on strike days adjusted by labour force or population data, as is the case in
this present study. The days not worked indicator indirectly measures workers involved,
as well as the duration of strikes. Obviously, this series tends to be dominated by larger
stoppages that result in a considerable number of strike days. Higher minimum criteria
for a strike to count, or the failure accurately to record small scale stoppages, will
therefore have only a minor impact on the total number of stoppages (Perry and Wilson,
2004: 24). However, there will be differences with regard to the strike frequency
indicator. The present study also concentrates primarily on days not worked, mainly
because Austria and Germany did not report strike frequencies for several decades
during the post-war period.
Apart from differences in compilation methods, an equally serious problem
appears to be that of missing data. Social and political upheaval or war can have an
adverse effect on the recording of stoppages, as well as affecting their actual number in
the first place (Perry and Wilson, 2004: 1). This is particularly the case for data on
strikes from Austria and Germany during the war years from 1914 to 1918 and the
periods of fascist dictatorships. Although strike activity continued - though at a lower
level - there is either no relevant information, or the reliability of available data from
these periods has to be called into question, as workers institutions such as unions (as
well as the labour parties) were banned. In the case of Germany there is no annual strike
data available for the period of fascist rule. This period has therefore been excluded
from the analysis.
However, it was possible to amend the German series for the time before the
ILO statistics had been introduced. This concerns the strike frequency from 1890 to
1932 and the strike days series from 1899 to 1932.15 Therefore, spectral results could be
obtained for the strike frequency series from 1892 to 1932 as well as for the strike days
series from 1901 to 1932 and from 1950 to 2004. Further, it was possible to obtain
strike frequencies and a more accurate days not worked series from historical strike
statistics for the period from 1949 to 1980 (Spode et al., 1992). The ILO data for days
not worked in the respective period has accordingly been replaced by this dataset.
In the case of Austria, additional data for the period of the Habsburg Empire
before 1918, as well as for the actual territory after 1918, is available dating back to the
year 1896 (Tomandl, 1965: 30-41).16 Strike records for the years 1915 to 1917 are
provided by Boleslawski (1919), although this data does not cover all parts of the
Habsburg Empire (Boleslawski, 1919: 146). Moreover, data for nearly the entire period
of Austro-fascist rule, up to the year 1937, has been recorded. Due to the inconstancy
prior to 1918, doubts as to reliability in the 1930s,17 and also because the gap for the
period of World War II is too great for the method to handle, spectral analysis is
restricted to the period after the end of World War II. For the post-war period, it is
possible to exploit sources from the Austrian Trade Union Congress (sterreichischer
Gewerkschaftsbund, GB). The GB is an umbrella organisation and compiles
information on strikes received from its various member unions. In publications such as

15

For data and sources, see Screpanti (1987).


The data before 1918 is derived from the Office of Labour Statistics (Arbeitsstatistisches Amt) and
covers the Austrian territory within its boundaries at that time.
17
This data was provided by the fascist Austrian administration.
16

25

Arbeit und Wirtschaft,18 strike activity (usually measured in annual strike hours as well
as involved workers) is publicly reported. Although information on strike frequencies is
available, labour institutions restrict themselves to publishing these particular indicators.
For its internal documentation, the GB archives more material with regard to this
topic. This source includes press releases, internal statistics and public information
derived from its publications. This documentation on strike activity can be obtained
from the GB archives. A comparison of the ILO data with the union data reveals
similar numbers of strike days using a working day of eight hours as the basis.19 One
exception is the year 1965, which according to the union sources, is the year with the
fourth highest number of strike hours since 1950 (some 3,387,787). Converted into
strike days, this would equate to approximately 423,473 days. However, the ILO reports
only 151, 261 strike days for this year. This deviation might originate from the fact that
in this year, a huge 24 hour warning strike by postal and railway workers took place
(Kropf, 2004: 157). This warning strike was probably not captured by the ILO series.
Therefore, the ILO record for this year has been replaced by the newly calculated value.
Strike frequencies are regularly reported only for the period prior to World War
II. It was possible to obtain information on the number of strikes from additional
sources, such as the study by Ferdinand Karlhofer (1983). Karlhofer investigated
wildcat strikes in the 1970s and was able to make use of the internal strike records of
trade unions from 1970 to 1981. From the 1980s onwards it is possible to reconstruct
strike frequencies using the previously mentioned sources provided by the GB.20
However, the obtained series is still too short to be included in the spectral analysis for
the post-war period. Thus, results could only be achieved for the days not worked series
from 1947 to 2004.
For Switzerland, the ILO data provides complete series for both indicators for the
period from 1927 to 2004. Although the country did not experience a period of fascist rule,
and labour organisations and workers were not exposed to the risk of prosecution as was the
case in Germany and Austria, information on strike activity during the war period has to be
assessed with caution. This particular time saw restrictions, such as the prohibition of the
communist and other left-wing parties such as the Fdration Socialiste Suisse. It also
involved the imprisonment of known and alleged communists or left-wing activists. Due to
this political climate, and also because the Social Democrats put into practice a party truce,
strike activity tended to be oppressed and underreported. One example is the year 1940,
which according to the ILO statistics had six strikes with 578 workers involved and 1480
strike days. The number of strikes cannot be verified, but in that year a major strike took
place in the arms industry, at the Bhrle & Co. factory in Zurich-Oerlikon. The strike was
led by a workers commission headed by a communist worker and lasted about two weeks.
The main demands concerned security improvements and wage increases. Roughly 1700
out of 3000 workers followed the strike call (Koller, 2008: 220). Converted into strike days
and assuming a working week of six days, this strike alone would account for about 20,000
strike days, which is about thirteen times the total number of strike days reported for the
18

The Arbeit und Wirtschaft is a monthly magazine released by the GB and the Federal Chamber of
Labour (Arbeiterkammer). It was initially published only by the Chamber of Labour, but in the year 1962
merged with the Gewerkschaftliche Rundschau, a magazine for union activists in the plants (Klenner and
Pellar 1999: 587). Strike records before 1962 are published in the Gewerkschaftliche Rundschau.
19
Series of strike hours and workers involved until 1999, which are based on union sources, can be found
in Klenner and Pellar (1999: 584).
20
Although not explicitly recorded, it was possible to reconstruct the annual strike numbers based on
these sources from 1981 up to 2004.

26

year. Strike data in Switzerland is available for the period from 1880 to 2004. The data
before 1914 is derived from Hirter (1988). Spectral results for Switzerland will be discussed
for the period from 1929 to 2004 for both strike frequency and days not worked.
Before the results are presented, some additional constraints regarding the data need
to be discussed. First, it is not possible to separate strikes from lock-outs for the entire
investigation period for all countries. Both are clearly labour disputes, but obviously they
are of completely different origin, which in turn has an impact on the interpretation of the
results. Moreover, it is not possible to distinguish between authorised, wildcat or political
strikes for large parts of the period under investigation. Sometimes it is even not clear
whether unauthorised and political strikes are included in the data. Where information can
be obtained about the relationship between these three kinds of disputes, authorised strikes
usually dominate. An exception might be the case of Germany in the 1960s and early 1970s
(Table 2) although, according to Kablitz (1972: 189192), due to shorter duration and a
smaller spread of wildcat strikes, strike days series in Germany tend to be dominated by
authorised strikes. It is also known that lock-outs in Germany are particularly frequent in
proportion to other forms of disputes, making up 44 per cent of all disputes in the decade up
to 1981 (Perry and Wilson, 2004: 5).

Table 2: Strikes in Germany 1949-1976 (percentage of row), 4 year intervals,


disaggregated by recognition status and strike cause

1949-1952
1953-1956
1957-1960
1961-1964
1965-1968
1969-1972
1973-1976

Union strikes
Economic
Political
49.0
58.9
35.8
38.1
11.3
5.2
4.3

Source: Spode et al. (1992: 391, 395).

4.3
0.6
2.2
0.0
0.6
0.2
0.0

Wildcat strikes
Economic
Political
45.4
39.3
61.9
61.9
87.0
93.3
95.4

1.3
1.2
0.0
0.0
1.1
1.4
0.3

For Austria, Karlhofer ascertained that although the overall strike level decreased in the
1970s, the proportion of wildcat strikes increased considerably (Table 3). Wildcat
strikes, at least during the 1970s, seem to be included in the aggregate series provided
by the ILO, as a comparison with the data from Karlhofer reveals.

27

Table 3: Wildcat strikes in Austria 1970-1981 (percentage of annual total)


1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981

Union Strikes

Source: Karlhofer (1983).

53.8
47.8
53.6
55.0
76.9
90.0
42.9
0.0
60.0
62.5
75.0
100.0

Wildcat strikes

38.5
52.2
46.4
45.0
23.1
10.0
57.1
100.0
40.0
37.5
0.0
0.0

Not declared

7.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
25.0
0.0

Switzerland also experienced a series of wildcat strikes in the 1970s. These conflicts
started as unofficial disputes which were later accepted by unions. As a result, these
strikes tend to be included in the aggregate strike series provided by the ILO. Taking
these shortcomings of aggregate data into account, the interpretation of the results
becomes somewhat more difficult. With regard to the research question, another
important remark concerns the choice for the measurement of strikes. The strike days
series is certainly not the best indicator, because a high number of strike days might
simply reflect a small number of long strikes or conflicts involving many workers.
However, a comparison of the series for strike days and strike frequency (Figures 1
through 3) reveals that years with high numbers for either of the indicators are
accompanied by high numbers for the respective other measurement. Thus the cyclical
pattern is not disturbed. What changes over time is, however, the relative position of the
strike measurements. In the Swiss case, the 1970s seems to be characterised by higher
strike numbers compared to strike days. In the 1950s and 1960s this relation is reversed
for some years. These changes in the relative position of the indicators may affect the
interpretation, because high numbers of strike days as well as of strikes sometimes
reflect different origins, particularly with regard to union behaviour as will be revealed
in the results section. For Austria and Germany, however, it seems as if the cyclical
pattern for the strike measurement deviates in some periods. To obtain a complete
picture, spectral results of strike frequencies will be additionally examined where it is
possible to derive estimates based on this series.
For the analysis, the strike series have been adjusted using population data as
approximations for the development of the labour force. As a measurement of economic
activity, annual data on gross domestic product (GDP) per capita at 1990 prices are
analysed. The sources for actual GDP per capita and population are the data from
Maddison up to 2001 (Maddison, 2004) and from the Total Economic Database of the
Conference Board for the years up to 2004.21 The data takes geographic changes into
consideration. From 1950 onwards, it represents the German territory as of 1991. A
detailed description of the sources of the processed data is given in Maddison (1991:
203204). For Austria, GDP estimates start in the year 1870, but have been corrected
21

The Conference Board (2009).

28

for territorial changes and refer to the present boundaries (Maddison, 1991: 202). This
restriction also hampers the inclusion of Austrian strike data prior to 1918 in the
analysis, because these series capture, with some territorial inconsistencies, the
Habsburg Empire. The data and estimates from Maddison are, like most historical data,
vulnerable for many reasons. The present study, however, refrained from using national
sources for GDP. Apart from the fact that Maddison relies heavily on existing national
sources for his estimates, this study prefers to utilise his data in order to ensure
comparability between the three countries.
6. Results
The main empirical results reveal that strike activity is characterised by cycles of three
to five years for all three countries over a large part of the period under investigation. In
Austria and Germany, there is a temporary correlation between strike days and business
fluctuations in the three to five year range, particularly during the period of economic
prosperity and low unemployment up to the 1970s. Strike activity during this period
exhibits a pro-cyclical pattern that becomes weaker in the course of the economic
slowdown in recent decades. In Switzerland, the correlation between strike activity and
business fluctuations is weak and does not exhibit a strong pro-cyclical pattern.
The following paragraph presents the results for Germany. The spectral results
are displayed in figures in order to facilitate the understanding of the applied method, as
well as the related findings. For the other two countries, a table which captures all
important spectral results, along with economic indicators such as growth and
unemployment rates, is utilised.
Germany
The first step in the analysis is to investigate the cyclical structure of the stoppage
indicators for the three countries through examination of the proportion of total
variance. As discussed in the method section, this spectral measurement reveals the
dominating cycles of a given time series.
Examining the cycle ranges of three to five, five to seven and seven to ten years,
Figure 7 shows that the strike days series for Germany is clearly dominated by the three
to five year range. About 40 per cent of the variation of the series can be attributed to
this short cycle in the period from 1901 to 1970. After this, the power at the
corresponding frequency drops to about 34 per cent.

29

Figure 7: Proportion of total variance, Days not worked in Germany 1901-2004

Percent
50%
45%
40%
35%
30%
25%
20%
15%
10%
5%

19
01
19
04
19
07
19
10
19
13
19
16
19
19
19
22
19
25
19
28
19
31
19
34
19
37
19
40
19
43
19
46
19
49
19
52
19
55
19
58
19
61
19
64
19
67
19
70
19
73
19
76
19
79
19
82
19
85
19
88
19
91
19
94
19
97
20
00
20
03

0%

Year
7-10 years

5-7 years

3-5 years

Source: see section data

This result indicates that strike activity measured in days not worked occurs in
intervals of three to five years. It takes three to five years to get from one upper turning
point to the next. Within this interval, it takes about 18 to 30 months to reach a lower
turning point (see Figure 8).
Figure 8: 3 to 5 year cycle

30

It should be repeated at this point that the cyclical pattern only provides information about
the structure of the occurrence and increases and decreases of strike activity over certain
intervals, such as three to five years. The magnitude and intensity of these movements does
not become apparent. Thus, it is not possible to make a statement with regard to the level of
strike activity and its long-term development based on spectral analysis.
Due to the fact that Germany does not report the annual number of strikes on a
regular basis, strike frequency estimates could only be obtained for the period from
1892 to 1932. Figure 9 reveals that this series is also dominated by the three to five year
cycle. About 35 per cent of the strike frequency variance until 1910 can be attributed to
this cycle. Up to 1916, however, the five to seven year cycle seems to be nearly as
important for this series as does the dominating short cycle.
Figure 9: Proportion of total variance, strike frequency in Germany 1892-1932

Percent
45%
40%
35%
30%
25%
20%
15%
10%
5%

19
32

19
30

19
28

19
26

19
24

19
22

19
20

19
18

19
16

19
14

19
12

19
10

19
08

19
06

19
04

19
02

19
00

18
98

18
96

18
94

18
92

0%

Year
7-10 years

5-7 years

3-5 years

Source: see section data

Despite a loss in explanatory power over the course of the subsequent decades, the short
cycle remains important compared to the other two ranges.
The German data allows for comparison of both applied strike measurements for
the period up to 1932. The shortfall in the days not worked measurement, which has to
be used for the major parts of the analysis, has been discussed in the previous section.
The analysis reveals that both strike days and strike frequency are dominated by the
same cycle. Considering explanatory power, the days not worked measurement seems to
exhibit a more pronounced cyclical structure with regard to the selected cycle lengths.
Because the cycle lengths have been chosen in accordance with the generally accepted
business cycle lengths, this result might reflect that in particular, short-term business
fluctuations structure strike activity. This is also a finding of previous research
(Goldner, 1953: 580). A confirmation of this hypothesis will be obtained through
examining the squared coherency, which can reveal to what extent a correlation of two
series in a particular interval exists.

31

Figure 10: Explained variance, strike frequency in Germany 1892-1932

Percent
14%

12%

10%

8%

6%

4%

2%

19
32

19
30

19
28

19
26

19
24

19
22

19
20

19
18

19
16

19
14

19
12

19
10

19
08

19
06

19
04

19
02

19
00

18
98

18
96

18
94

18
92

0%

Year
7-10 years

5-7 years

3-5 years

Source: see section data

Figures 10 through 11 illustrate explained variance, the measurement derived from the
squared coherency, which gives information about the extent to which the variance of
the strike series in each of the selected intervals can be attributed to corresponding
cyclical components in the business activity series. Explained variance of the strike
frequencies reveals that most of the correlation between the number of strikes and
business fluctuations is found in the five to seven year interval up to the year 1916
(Figure 10). Explanatory proportions of about 11 to 13 per cent leave a considerable
part of the variation unexplained. With regard to the potential influence of business
fluctuations on strike activity, however, it is this cycle of medium length which seems
to be temporarily the most important. Although the univariate analysis indicated that the
dominant cycle of strike frequencies for the entire period is the three to five year one, a
considerable part of the series variance could also be attributed to the five to seven year
cycle (Figure 9). The deviating dominant cycles in the proportion of total variance, as
well as in explained variance, might also point to the existence of a genuine, unique
strike cycle that is independent of business fluctuations but exhibits the same length as
the short business cycle. Such a short strike cycle might be the result of grouped
contract expirations, which seem to be plausible triggers for a potential cyclical pattern
of strikes.

32

Figure 11: Explained variance, days not worked in Germany 1901-2004

Percent
18%
16%
14%
12%
10%
8%
6%
4%
2%

19
01
19
04
19
07
19
10
19
13
19
16
19
19
19
22
19
25
19
28
19
31
19
34
19
37
19
40
19
43
19
46
19
49
19
52
19
55
19
58
19
61
19
64
19
67
19
70
19
73
19
76
19
79
19
82
19
85
19
88
19
91
19
94
19
97
20
00
20
03

0%

Year
7-10 years

5-7 years

3-5 years

Source: see section data.

Because strike frequency series end in 1932, the analysis will focus on the strike days
series that capture the entire investigation period. Considering this indicator, it can be
seen that the three to five year business cycle accounts for only about 4 per cent and less
of explained variance in the period from 1901 to 1932 (Figure 11).
The low correlation between the business cycles of various lengths and strike
activity in the first half of the 20th century might be explained by the deep social,
economic and political crises during this period. Particularly in the years following
World War I, Germany experienced monetary instability and rising unemployment,
both of which were associated with increasing social tensions.22 A short recovery period
started in 1924, but unemployment was still increasing. The German economy was
severely affected by the Great Depression at the end of the 1920s. In the course of this
crisis (1929-1932), industrial output dropped by almost 40 per cent (Feinstein et al.,
1997: 107) and unemployment reached its highest levels, of about 30 per cent, in the
early 1930s (Mitchell, 2003). According to Borchardt, the period between the wars was
characterised by turbulent economic cycles of large amplitudes. It can be regarded as an
exceptional period in terms of cyclical fluctuations (Borchardt, 1982: 79).
During World War I, the German trade unions behaved loyally to the war regime
and rarely engaged in industrial disputes. With increasingly deteriorating working and
living conditions during wartime, this policy lost the acceptance of the rank and file and
ended with the German revolution of 1918. This was one of the significant strike waves
of the 20th century, starting with the January strikes (Januarstreik) of 1918 (Schneider,
22

An overview on the interwar economy can be found in Feinstein et al. (1997). For the inflationary
spiral see particularly pp. 38-42.

33

1987: 284). The early 1920s saw further political strikes, such as the first German
general strike against the Kapp Putsch in 1920 with about twelve million workers
involved (Schneider, 1987: 329). As described before, the German economy underwent
a deep crisis in the aftermath of World War I, with its effects exceeding the usual ups
and downs of business activity and also exacerbating existing social tensions. These
developments found their expression in several years of mass political strikes, which did
not follow any clear cyclical pattern. The subsequent years saw a slight recovery of
cyclical strike activity. Its low correlation with business cycles might be attributed to
the fact that the economy did not recover in terms of a pronounced cyclical pattern.
Furthermore, German trade unions were organised as partisan units. A partisan structure
usually promotes competition between the different unions in industrial sectors, which
might have also contributed to less co-ordinated proceedings with regard to the business
environment. Nevertheless, the enforcement of widespread collective agreements was
an achievement of the German revolution and collective bargaining already became a
feature during the turbulent inter-war years. The establishment of normal conflict
behaviour in the framework of collective bargaining was particularly the case in the
short period between 1924 and 1928, which was characterised by a slight economic
stabilisation (Kittner, 2005: 500). However, this period was too short for a pronounced
cyclical pattern of strike activity to develop, as well as any distinct co-movement with
business fluctuations.
A very different picture emerges for the post-war period. After 1951, the short
economic cycle becomes increasingly important, with an explanatory power up to 15
per cent in 1970 (Figure 11). Following this, its impact declines rapidly (0.06 per cent in
2003), except for a short period from 1978 to 1984. A potential explanation for this
development might be the extraordinary growth period from 1950 to 1970, which was
the foundation for the moderate behaviour of social conflict parties with regard to strike
activity. In this period, unions increasingly gained acceptance as bargaining partners; a
development which went hand in hand with the emergence of national labour
agreements. Social peace and favourable economic conditions gave rise to the
appearance of wide-spread labour agreements (Flchen- bzw. Verbandstarifvertrag).
These agreements dominate German industrial relations, although contracts also exist at
company level. The most prominent of these is the contract of the car manufacturer
Volkswagen (VW). However, in the overall landscape of employment contracts, these
kinds of agreements play only a minor role. The main feature of the prevailing
Tarifvertrag is negotiation at branch level with the respective employers associations.
These settlements usually apply to affected employees and employers in an area or the
entire country (Bispinck, 1993: 57).
The dual structure of workers representation is a central characteristic of the
German model of industrial relations: collective bargaining between unions and
employers associations at the macro-level, as well as between works councils and
management at the company level (Jacobi et al., 1998: 190). Although union density in
Germany is traditionally low, with rates declining from about 30 per cent to 20 per cent
over the last fifty years (Visser, 2009), the major associations such as the German
Federation of Trade Unions (Deutscher Gewerkschaftsbund, DGB) reveal a
comparatively high degree of centralisation and are able to exert considerable influence
on their affiliates and membership (Jacobi et al., 1998: 191). This is reflected to some
extent in the high coverage rate of collective agreements, which ranges around

34

comparatively high levels of about 80 per cent in the 1960s. In recent decades, this rate
has declined to about 60 per cent (Visser, 2009).
The enforcement of comprehensive labour agreements brought both advantages
and disadvantages for the unions and their membership. The widespread tariff areas had
already been established during the fascist period and adopted after the war. Mainly due
to pressure from employers, initially regional negotiations were extended to centralised,
nationwide ones. Clearly, employers were mindful of the high levels of industrial action
in the previous decades. On the other hand, local as well as nationwide agreements
provided benefits for employers in the specific economic framework of the period.
Centralised collective bargaining attributes higher importance to macroeconomic
variables for orientation. It cannot consider the singular interests of either employers or
employees, but instead follows general trends in price increases and economic growth
with regard to wage expectations and other claims. One effect which becomes
particularly important within this framework during phases of economic prosperity is
that the concession leeway of comparatively profitable enterprises, areas and branches is
not exhausted. Therefore the leeway for economic concessions for employers, as well as
for works councils in certain areas and companies, is increased. They are able to
negotiate out-of-tariff payments and in turn foster positions of power in companies.
Moreover, out-of-tariff payments are obviously easier to handle compared to
stipulations in contracts lasting over longer periods. Another effect of centralised
collective bargaining is the increasing importance of macroeconomic advisors in the
headquarters of both negotiating partners (Mller-Jentsch, 1973: 155).
Barry Eichengreen describes this particular form of German industrial relations
and the leading role of the metal workers union (Industriegewerkschaft Metall, IG
Metall) within the bargaining process as follows: Although the DGB was not directly
involved in wage negotiations, it provided a forum for discussions among union leaders,
and its researchers provided economic analysis to help frame the annual wage round.
This enabled the metal workers to pick a level of wage increases appropriate for the
economy as a whole and encouraged other unions to follow. (Eichengreen, 2007: 96).
It was mainly the union representatives of the metal workers, and not the rank
and file, who picked the appropriate wage level. This temporarily led to considerable
discontent among rank and file union members, especially at the end of the 1960s and
early 1970s when a wave of unofficial strikes swamped the country. Overall, collective
bargaining in the post-war period was characterised by the centralisation of
negotiations, which went hand in hand with an increasing orientation towards general
economic developments with regard to union demands. According to Mller-Jentsch,
the hierarchic structure of this framework was conducive to the incorporation of unions
in similar forms of corporate relations, such as collaboration within the so-called
Concerted Action (Konzertierte Aktion). The Concerted Action was an attempt to
overcome the economic crisis of 1966/67 by the joint efforts of all social forces.23

23

In der tarifpolitischen Praxis hatte sich seit Mitte der fnfziger Jahre die Orientierung an
gesamtwirtschaftlichen Daten durchgesetzt. Durch die Zentralisierung der lohnpolitischen
Entscheidungen und Tarifverhandlungen konnten die Gewerkschaftsspitzen zentral getroffene
Entscheidungen relativ problemlos innerhalb der Organisation nach unten durchsetzen. Damit waren die
westdeutschen Gewerkschaften aufs beste gerstet, jene Rolle zu bernehmen die ihnen das staatliche
Krisenmanagement zugedacht hatte: im Interesse der Krisenberwindung eine Lohnpause einzulegen und
diese ihren Mitgliedern schmackhaft zu machen (Mller-Jentsch 1973: 157).

35

With regard to strike action, the post-war period was accompanied by a decline
in overall strike activity, which was doubtlessly a result of co-operative industrial
relations and the increasing leeway for concessions due to favourable economic
conditions. Moreover, strike activity was restricted by law during this period. Since the
1950s, the so-called ultima ratio principle dominated union strike activity: before
calling a strike, the responsible union representatives are supposed to explore all
bargaining and mediation possibilities. Moreover, they are constrained by having to
consider the overall economic conditions in terms of the potential success or failure of a
strike and its possible effects on enterprise and the economy.24 Offences under this law
tend to be accompanied by organisational and financial risks to unions and this has
consequently reduced overall strike levels. Moreover, the widespread tariff areas
include a considerable volume of workers, and strike payments in Germany have been
comparatively high. As a result, strike actions tend to be very expensive for large
bargaining units, which affected the strike tactics of German unions in later decades
(Dribbusch, 2007: 277). IG Metall, for example, implemented the so called warning
strikes (Warnstreik) which lasted no longer than a couple of hours. Warning strikes can
include a large number of workers, but remain comparatively cost-saving due to their
brevity. Usually these strikes are also included in strike records. Although overall strike
levels have declined, the cyclical activity of strike action has tended to exhibit a
pronounced pattern, especially in terms of its correlation with business fluctuations.
Explained variance indicates an increasing importance of the short business cycle for
the strike days series during the boom years, but also reveals that not the entire strike
activity has been affected by these fluctuations (Figure 11).
With the noted historical developments in mind, it would not be surprising that
during a period in which wage policy and other claims were adjusted according to
economic development, strike activity might also follow this logic. In addition,
examination of the in-phase proportions of explained variance supports this argument.
The in-phase proportion measures the extent to which the two series reach upper and
lower turning points at the same time.
Figure 12 reveals that the short cycle in the German strike days series is in phase
with the corresponding business cycle, especially during the post-war period. As
recently as during the period of prosperity from 1951 to 1970, explained variance is
dominated by in-phase movements, which account for up to 14 per cent in 1970.
Considering that the short business cycle accounts for 16 per cent of the explained strike
variation in 1970, an in-phase movement of 14 per cent implies that up to 87.5 per cent
of explained variance is in phase. This result suggests that strike peaks tend to coincide
with economic peaks in this particular interval.

24

Strike guidelines of the DGB, cited in Abendroth (1989: 69-70).

36

Figure 12: Explained variance in phase, days not worked in Germany 1901-2004
Percent
16%

14%

12%

10%

8%

6%

4%

2%

19
01
19
04
19
07
19
10
19
13
19
16
19
19
19
22
19
25
19
28
19
31
19
34
19
37
19
40
19
43
19
46
19
49
19
52
19
55
19
58
19
61
19
64
19
67
19
70
19
73
19
76
19
79
19
82
19
85
19
88
19
91
19
94
19
97
20
00
20
03

0%

Year
7-10 years

5-7 years

3-5 years

Source: see section data.

The analysis reveals that business cycles have an impact on strike activity in Germany.
Most of the correlation between these variables has been found in the three to five year
cycle over the entire investigation period, but particularly from 1950 to 1970. A potential
explanation for this result might be the moderate behaviour of unions and employers during
this period. Their co-operation was not free from conflicts, because that would imply no
strikes at all. However, the disputes that actually took place seemed to be more influenced
by the possibilities and limits dictated by the phase of the business cycle, as was the case at
other times during the investigation period. The moderate behaviour of both parties was
possibly determined by the extraordinary economic growth during this period, in which it
was much easier to bring wage expectations and other claims into line with profits. This
interpretation would apply especially to the behaviour of unions in collective bargaining
processes, who could adjust the regularly occurring bargaining process according to the
possibilities of the business cycle. It is a generally accepted fact that strike activity,
particularly in the 1960s and 1970s, was dominated by struggles for higher wages, and
towards the end of this period by struggles against impending wage reductions (Bergmann
et al., 1975: 209). As the 1950s still experienced some political strikes, such as the struggles
with regard to co-determination in the beginning of the decade, it seems plausible that the
dominance of wage strikes was conducive to the increasing impact of business cycles on
strike activity. The related claims were obviously adjusted to overall economic conditions,
which in turn generated pro-cyclical behaviour as revealed by spectral analysis of this
period. However, the fact that strike activity (particularly in the late 1960s and early 1970s)
was dominated by unofficial disputes, hampers drawing a conclusion based on economic

37

conditions and the changing cyclical pattern, that exclusively explains union behaviour.25
Moreover, it is known that the larger strikes by the influential metal workers union tended
to follow an anti-cyclical pattern (Bergmann et al., 1975: 209). This does not seem to be the
case for aggregate strike activity as revealed by spectral results, although a considerable part
of the strike activity obviously does not follow a pro-cyclical pattern. In the case of
Germany, business fluctuation seemed to especially structure the willingness of the rank
and file to strike in the 1960 and 1970s.
An explanation for this phenomenon might be the incorporation of the unions in the
state-driven income policies after the economic crisis of 1966/67. As a result of this crisis, a
new law was introduced which included an extension of the fiscal authority of the federal
government, the states and the municipalities in order to improve the co-ordination of
expenditure in line with economic targets (Altmann, 2004: 49). The unions collaborated and
guaranteed wage moderation. In return, they received assurances with regard to the
preservation of jobs and their demand for more distributive justice.
Wage moderation was accepted by the rank and file during crisis years, but not
in the subsequent period that brought a perceptible economic upturn. Thus the procyclical co-movement between strike activity and business fluctuations endured and
even increased. The actors, however, seemed to change and the rank and file took the
lead in strike activity. Although these struggles were not initiated by unions in the first
place, it was likely that the favourable economic environment determined the timing of
strike action. Business cycles, therefore, tend to be able to explain the timing of the
actual ups and downs of strike activity. However, they cannot explain the extent of
these fluctuations: the increasing discontent of workers with regard to union politics,
and frustration due to inadequate participation in the economic upturn, led to
exceptionally high strike activity considering the overall level of industrial conflict in
post-war Germany.26
One could argue that the rank and file is usually not as well informed about
economic conditions as are their leaders and thus is not sufficiently equipped with the
necessary knowledge to adjust strike action in line with business fluctuations. Franzosi
remarks upon this concern in the following: () workers do not have to look at
production figures to know that things are going better for the employer and that times
are favourable to demand higher wages. One breathes a special atmosphere in a plant
where a slump is giving way to a full productive swing: trucks are coming and going,
the pace of production picks up, and working hours get extended into overtime.
Workers have only to look at their employers face to know whether or not their
demands are likely to be met. (Franzosi, 1995: 59).
After 1973, the economic prerequisite for easy co-operation, as well as ongoing
improvements in respect of living and working conditions, was challenged. Germany
had to face two major recessions in the years 1974/75 and 1981/82. Slower economic
growth was accompanied by rising unemployment. Contrary to the boom years, where
the loss of jobs tended to be cyclical in character, it turned out to become a persistent
phenomenon and reached the levels of the immediate post-war period again in the
25

For an analysis of wildcat strikes during the period of economic prosperity in Germany, see Birke
(2007). Another contribution which discusses the occurrence of wildcat strikes and their relationship to
business cycles in Germany in this period is by Hagelstange (1979).
26
For an explanation of the German strike wave, which takes into account international economic
development as well as the fact that this wave was part of an international movement of increasing
militancy in workers, see Soskice (1978).

38

1990s. In economic terms, one of the notable problems of this period was a decline in
profitability, which was accompanied by a decrease in investment (Carlin, 1996: 474).
Deteriorating economic conditions increased the opposition between capital and labour,
because possible solutions were sought which required freeing up market forces. These
solutions included rationalisation, re-structuring and the reduction of the workforce at
the plant level. On the macroeconomic level, it led to deregulation of the labour market
by cutting unemployment benefits and attempting to decrease legal working hours.
Thus, the hard-won gains of the 1960s and 1970s (Glyn, 2006: 49) became
increasingly challenged. Moreover, the previously welcomed centralisation of collective
bargaining came under pressure in favour of more flexible and decentralised structures.
After 1974, labour was forced to struggle primarily for the preservation of the status quo
with regard to wages and working conditions, even though unions reacted in the first
place with an offensive strategy. They intensified the fight for the reduction of working
hours (to a 35 hour working week) in order to preserve jobs. This in turn led to an
increase in the level of strike activity at the end of the 1970s and the early 1980s (see
Figure 2). The metal workers union in particular pressed ahead with this claim and
initiated major conflicts such as the strike in the steel industry in 1979 (Martens, 1979)
and the strike in the car industry in 1984. The 35 hour working week was not achieved,
but the dispute led to a reduction of weekly working time to 38.5 hours. The 35 hour
working week was eventually introduced in the 1990s. The reduction in working time,
however, was generally accompanied by concessions concerning flexibility among
workers and overtime deals so that the overall operating time of machines was not
affected (Armstrong et al.: 266). This conflict lasted for seven weeks and was the largest
post-war dispute led by the metal workers union. It marked a turning point in German
strike history. From 1985 onwards, overall strike activity declined and was mainly
characterised by the dominance of defensive disputes. The toughening of the situation
between social forces was reflected in an increasing use of lock-outs on the part of
employers, which posed a major threat to union strike tactics. Particularly during the
metal workers strike in 1984, lock-outs were systematically used to boost the strike
costs to IG Metall. Moreover, workers affected by so-called cold lock-outs (kalte
Aussperrung) outside their bargaining area lost their entitlement to social benefits. Cold
lock-outs are the closure of establishments due to supply shortages in the course of
industrial conflicts. Up to 1984, workers affected by cold lock-outs outside of their
bargaining area received a degree of unemployment benefits provided by the Federal
Employment Service. The abolishment of these payments during the 1984 conflict and
the adoption of this into law two years later, led to significant constraint of the use of
the strike instrument (Dribbusch, 2007: 281).
Overall, the period from 1974 onwards led to a retreat by labour and an overall
weakening of its position in the social fabric. Considering the cyclical pattern, the
dominant three to five year cycle of the strike days series somewhat loses its
explanatory proportion, from 40 per cent in 1970 to about 34 per cent in 2004, which
might be explained by troubled industrial relations and collective bargaining. Contrary
to the previous period, decreasing proportions of explained variance and in-phase
proportions indicate a weakening correlation with business fluctuations. Explanatory
proportions decrease from about 16 per cent in 1970 to about 5 per cent in 1978. 1984
follows a fresh increase to about 9 per cent (Figure 12). Although these proportions
cannot compete with the values of the boom period, it seems as if the unions offensive
struggles with regard to working time reduction were adjusted to business fluctuations,

39

unlike the defensive struggles in the subsequent decade. Collective bargaining following
the mid-1990s mainly involved negotiated concessions concerning working time and
wages, in return for job guarantees. Disputes occurred as a result of structural changes,
such as plant closures, relocation and outsourcing. These developments went hand in
hand with a decreasing significance for industrial relations. This change is reflected in
the general weakening of the two constituent institutions of German industrial relations:
collective bargaining agreements tend to cover the decreasing part played by employers,
which is the result of general pressure of centralised bargaining. Moreover, recent
decades have seen a decreasing coverage by works councils, the crucial institutions for
bargaining at the plant level: A parallel decline in works council and collective
bargaining coverage, however, leads to a shrinking regulative capacity of the German
industrial relations institutions (Hassel, 1999: 57).
In consequence, conflicts exhibited a weaker correspondence with the rhythm of
the bargaining agenda, as well as with the fluctuation of the business cycle. Contrary to
the boom period, traditional roles in the collective bargaining process collapsed and the
employees side was no longer the initial part to which the other side reacted
(Dribbusch, 2007: 283). In times of slow economic growth the range of concessions for
both sides, unions/workers and employers, decreases. Existing opposed interests resurface and due to limited distribution leeway it becomes increasingly difficult to reach
compromises. This is expressed not only in an overall decline in strike activity but also
in the changing cyclical pattern of strike activity.
Austria
Due to reasons of space, Table 3 summarises the main empirical results for Austria and
Switzerland. These results are related to a compressed description of the economic
development during the period under investigation in order to highlight the main findings of
the present study. Columns 1 and 2 show the number of years with negative and positive
deviations from the average growth rate of each country, in order to determine the general
condition of the economies for each decade. Columns 3 and 4 calculate the averages of
negative and positive deviations in each decade to specify the previous results. The average
unemployment rate, as an additional important socio-economic indicator, is displayed in
Column 5. Columns 6 and 7 show the average number of strikes and strike days per head, to
capture long-term trends in strike activity. Columns 8 and 10 present the average explained
variance (the spectral measurement that highlights the correlation between the two series of
interest) of strike frequencies and strike days in the three to five year range. These cycles
turned out to be the most important in explaining the relationship between strikes and
business cycles in the period after World War II, as well as for the strike series in the
univariate case. Columns 9 and 12 calculate the average in-phase proportions as a
percentage of explained variance per decade.
For Austria, a similar pattern to the German case becomes clear, with an
increasing correlation between strike activity and business fluctuations from the 1950s
onwards. However, the strongest correlations between these variables were found in the
1970s and particularly the 1980s, with about 66 per cent of explained variance on
average in these decades. The strong correspondence is also exhibited by high
proportions of in-phase components with about 91 per cent of the explained variance of
strike activity running in phase with the corresponding business cycle in the three to
five year range. The 1990s also reveal high average numbers of explained variance and
in-phase proportions. This period can be already regarded as the beginning of a

40

downward trend in respect of the relationship of interest. Considering the economic


indicators, it can be observed that during the 1950s and 1970s there are only a few years
with negative deviations from average growth, whereas the 1980s and 1990s are
characterised by a considerable number of years with negative deviations (15 years in
total).27 With regard to the average unemployment rates, Table 4 shows that
comparatively low levels dominated up until the 1990s.
Table 4: Spectral results and economic indicators for Austria and Switzerland
Numb
er of
years
with
negati
ve
deviati
on

Number
of years
with
positive
deviation

5
2

5
8

-15.2
-1.3

1960-1969

-0.5

1970-1979
1980-1989
1990-1999
2000-2004

2
7
8
2

8
3
2
3

-2.8
-1.2
-1.2
-1.7

Numb
er of
years
with
negati
ve
deviati
on

Number
of years
with
positive
deviation

1940-1949
1950-1959

1920-1929
1930-1939
1940-1949
1950-1959
1960-1969
1970-1979
1980-1989
1990-1999
2000-2004

1
7
7
2
3
4
5
8
4

9
3
3
8
7
6
5
2
1

Negative
deviations
(average)

Negative
deviations
(average)

-4.5
-3.7
-3.1
-4.1
-0.2
-3.4
-2.1
-2.3
-1.6

Austria

Positive
deviations
(average)

Unemplo
yment
rate
(average)

Average
number
of strikes

Average
number
of strike
days/hea
d

12.9
5.0

2.9
6.1

n.a.
n.a.

14
10

2.5

2.8

n.a.

15

2.4
2.0
0.7
4.3
0.7
6.5
0.3
6.6
Switzerland

13
6
2
5

3
1
1
40

Unemplo
yment
rate
(average)

Average
number
of strikes

Average
number
of strike
days/hea
d

2.2
8.9
1.6
0.4
0.0
0.2
0.7
3.4
2.7

n.a.
24
24
5
2
8
2
4
6

n.a.
13
11
2
1
1
0
1
3

Positive
deviations
(average)

3.3
2.1
12.5
2.4
1.6
0.9
1.0
0.4
0.7

Strike
frequency.
Explained
variance
(3-5 years)

Strike
frequency.
Explained
variance
(3-5 years)

10.4
4.8
5.2
6.0
8.2
8.5
11.0
13.8

Notes: Average growth rates for Austria (1940-2004) and Switzerland (1920-2004). Columns 1 and 2:
years with negative and positive deviations from the average growth rate. Columns 3 and 4: average
negative and positive deviations for each decade. Column 5: average unemployment rate per decade.
Columns 6 and 7: average number of strikes and strike days per head. Columns 8 to 11: average explained
variance and in-phase proportion (percentage of explained variance) of strike frequencies (three to five
years) and strike days (three to five years) per decade.
Source: see section data and Mitchell (2004) for unemployment rates.
27

For an overview on Austrian post-war economic history, see Otruba (1968), Butschek (1985; 1999),
Chaloupek et al. (2005) and Eichengreen (2007).

41

Due to missing data, it is not possible to compare these developments to the situation in
the first half of the 20th century, but there is evidence for the temporarily occurring procyclical behaviour of strikes and business cycles during the First Republic (Botz, 1978).
For the explanatory value of the business cycle with regard to strike activity in
the post-war period, the same line of argument holds true as for Germany, with the
difference that the international economic crises of the 1970s had a lesser impact on
Austria. Although the 1970s show a higher number of negative average deviations
( 2.8 per cent) compared to positive average deviations (2.4 per cent), this difference
between negative and positive deviations is not markedly pronounced for this decade
(Column 3). Moreover, the Austrian government and its economic policy of AustroKeynesianism managed to keep the unemployment rate low, at about 2 per cent on
average for this decade.28 Economic turbulence accompanied by rising unemployment
began notably in the 1990s. A potential explanation for these results might be the
particular format of Austrian industrial relations, which are strongly characterised by an
intensive type of social partnership.29 This social partnership, generally speaking,
helped to control strike levels as well as collective bargaining. In interaction with the
general economic conditions, it moreover seems also to have influenced the cyclical
pattern of strike activity. Taking the historical development of economic and social
conditions into account, the results derived from spectral analysis appear to be quite
sound. For the period from 1953 up to 1961, collective bargaining appears to follow
regular patterns, as demonstrated by the increasing proportions of the short cycle in the
proportion of total variance (not reported in Table 4). Strike activity is characterised by
a three to five year cycle and is moreover increasingly correlated with the short business
cycle (Column 10). As in Germany, this pattern might have been triggered by coordinated bargaining about the distribution of productivity increases, which is adjusted
with regard to the overall economic situation. The declining explanatory proportions of
the dominant short cycle in the univariate as well as the multivariate case after 1961
may be explained by the challenge to the social partnership on the part of employers
(Tlos and Kittel, 1992: 118). This change was reflected by a slight increase of strike
levels (see Colum 7, average number of strike days/head), but with regard to business
fluctuations, in a somewhat anti-cyclical pattern of strike occurrence. Strikes as a means
of enforcing social partnership collaboration thus seems to be described by this kind of
pattern. However, the 1970s shows a considerable increase in the correlation between
strike activity and the business cycle. Comparatively more favourable economic
conditions than those in other countries, along with a re-strengthening of social
collaboration under the socialist government, probably facilitated this change in the
cyclical pattern.
A look at the actual strike events indicates an additional explanation. Similarly
to Germany in the early years of this decade, labour disputes during the 1970s in
Austria increasingly adopted the form of wildcat strikes. As explained in the discussion
of the German results, these strikes tend to strongly correlate with the business cycle.
The majority of wildcat strikes, however, were ordinary bargaining strikes and basically
28

A brief overview on Austro-Keynesianism and its characteristics can be found in Tichy (1995). A
description and evaluation of the economic development as well as the economic policy between 1975
and 1981 is given by Butschek (1985, pp. 156-169).
29
Major contributions to the analysis of Austrian corporatism are Tlos (1993; 2008), and Karlhofer and
Tlos (1999). A brief overview on industrial relations and collective bargaining in Austria is by Traxler
(1993; 1998). For social partnership and wage policy, see particularly Guger (1993).

42

were not intended as opposition to union policies. To the contrary, the analysis of case
examples shows that the workers involved assumed that unions would support their
initial actions (Karlhofer, 1983: 133). Due to the centralised nature and macroeconomic
responsibility of the GB, it restrained in many cases from such support, which would
have privileged particular interests in favour of the general outlook. The diverging
perceptions of appropriate wage claims again resulted in the fact that in the prevailing
form of class collaboration, demands were adjusted to account for overall economic
aspects.30
The adjustment of strike activity to business conditions seems not to be a
conscious act, but a result of the fact that the rank and file experiences economic ups
and downs in terms of the living and working environment. As previous research has
stressed, the periodicity of the occurrence of strikes is determined to some extent by
business fluctuations. The scope of the rises and falls of a strike cycle, however, is not
predictable solely by taking business fluctuations into consideration. Analysis so far has
demonstrated that the significance of social partnership impacts on all aspects of strike
activity during the investigation period. The impact, however, varies along with its
significance to overall political and economical development over time. The gradual
deterioration of economic and social conditions in the 1980s placed the corporatist
tradition and its institutions under pressure. Tightened economic conditions, which also
narrowed the distribution leeway, made co-ordinated bargaining between different
social forces less important. This went hand in hand with a lessening of the involvement
of labour representatives in political decision-making processes.
The macroeconomic climate put labour on the defence, which further reduced
the overall strike levels so no major strikes occurred during this decade. The rate of
change of wages in Austria is particularly determined by unemployment. A rise in
unemployment is usually accompanied by the waiving of measures to enforce even
achievable wage rate increases (Rosner, 1999: 88). The largest protest movement was a
solidarity demonstration in 1986, in connection with the nationalised industry in Linz
and involving about 40,000 participants. Considering the cyclical pattern of strike days,
it shows an unprecedented correlation with business fluctuations. One explanation
might be that the declining significance of social partnership during this period was
associated with an upward revaluation of ordinary tariff policies without the
interference of social partners (Traxler, 1993: 286). It is very conceivable that although
conflicts did not increase, the omission of the protected area social partnership forced
the small number of strikes to follow the business pattern even more closely, in order to
increase the probability of success.
In the 1990s, there was a definite collapse in the relationship between strikes and
business cycles. This went hand in hand with a further deterioration in the importance of
social partnership. Meetings of the Parity Commission were reduced, its council for
wages sank into insignificance and its involvement in collective bargaining in the
meantime became a mere formality (Kittel and Tlos, 1999: 910). Moreover, the entry
of Austria into the European Union in 1995 posed a challenge to interest groups, as the
European Union already had its own structures in place.
30

Lohnforderungen werden im Rahmen der Sozialpartnerschaft gleichsam indexgebunden an


makrokonomischen nderungsraten formuliert; sie werden zudem zentralistisch formuliert, nicht als
Ergebnis horizontal aggregierter Mitgliedervorstellungen, sondern als Ergebnis eines vertikalen
Filterungsprozesses der heterogenen Mitgliederinteressen (Karlhofer 1983, pp. 133-134).

43

The bitter wind challenging social partnership can be mainly attributed to the
side of business, which sought to enforce more political and economic autonomy when
facing economic problems. This is also reflected in the conflict issues. Increasing the
resistance to collective agreements triggers conflicts that aim for the general
enforcement of contracts. Unlike the previous decades, where wage claims dominated
conflicts, there was a shift towards labour law related issues, such as employment
protection (Rosner, 1999: 91).
However, union representatives tended to invoke the opposite of the described
developments in Austria, as in the case of the former president of the railway union,
Wilhelm Haberzettel. He continued to consider the Austrian social dialogue, with its
particular emphasis on social peace, as an international exception. In other countries,
according to Haberzettel, it was necessary to call a strike in order to start negotiations.31
Commenting on the German experience, the former president of the influential metal
workers union, Rudolf Nrnberger, stated that collective bargaining in Austria seemed
to work much better than elsewhere. Unlike in Austria, workers in Germany had to
strike regularly to reach decent settlements.32 Apart from doubts that these descriptions
actually apply to all the mentioned countries, these statements from the late 1990s
obviously lag behind recent developments. It is true that strike levels are still low, but
the responsiveness of the employers side in exchange for social peace reduced
noticeably. At the beginning of the new millennium, the social partners were no longer
involved in the assessment of a planned drastic pension reform and unions felt
compelled to take strike action in order to defend their role within this tradition. This
resulted in a massive strike in 2003, which occurred in an economic dip and was not
successful. Contrary to the 1960s, where social partnership had also been challenged,
strike issues in recent years have not been common bargaining demands, but have been
concerned more with basic issues, such as the complete restructuring of retirement
arrangements without the involvement of the social partners.
These developments do not point to the existence of a strong and harmonious coordination between social forces. On the contrary, the time of compromises and cooperation also seems to be under challenge in Austria. Due to the particular design of
Austrian corporatism, as well as the course of special developments in the economy, the
collaboration between social forces, however, appears to have been stable for a longer
period than it did in Germany. The former chancellor Kreisky once explained the
strength of social partnership by its non-institutionalised character: The left-wing
people are opposed to social partnership and the right-wing people are opposed to it too,
both for different reasons. What we have done in Austria is a process of sublimation
of class values. We found a much more sublime way for our divergences. And it works
without institutionalizing it. If we institutionalize it, we would kill the system. You
know why? Not because I am against institutions, but because this system has only one
real sanction - to say that we shall leave the table. We are not obligated. We are not
compelled. This threat, to leave the table, is the strongest reason for remaining at all
31

Wilhelm Haberzettel, cited in Pellar (2003):Grundstzlich kann man sagen, dass es in allen Lndern,
wo Kampfmanahmen ergriffen werden, den klassischen sozialen Dialog wie er in sterreich Usus ist
nicht gibt. Die Franzosen und Belgier mssen streiken, damit sie erst einmal an den Verhandlungstisch
kommen.
32
Rudolf Nrnberger, cited in Pellar (2003): Bei den Kollektivvertragsverhandlungen waren in
sterreich gleichwertige, teilweise sogar bessere Ergebnisse mglich als in Deutschland, wo die
Metallarbeiter seit Jahren auf die Strae gehen mssen, um einen Abschluss zu bekommen.

44

(Kreisky, cited in Katzenstein, 1984: 144). The development of social partnership in


Austria, particularly in recent decades, shows that macroeconomic changes affected the
willingness to remain at the negotiating table. This in turn found an expression in the
development of the cyclical pattern of strike activity as well as its correlation with the
business cycle, both of which seem to be affected by the various phases of social
partnership.
Switzerland
Switzerland is the only country with complete series of all applied strike activity
measurements starting from 1927. Table 4 reveals that average numbers of explained
variance, particularly for the days not worked series, did not reach the levels of
Germany and Austria in this country. Strike frequencies reveal higher proportions of
explained variance over the decades. The corresponding in-phase proportions, however,
tend to remain at comparatively low levels, particularly in decades with higher numbers
of explained variance such as the 1930s. Considering economic development, it can be
seen that the years with positive deviations from average growth tend to prevail during
the second half of the 20th century. Only the last two decades exhibit more years with
negative deviations. The average negative deviations (Column 3), however, generally
tend to be higher than the positive ones (Column 4), except for the 1960s. This might
indicate that the Swiss economy was particularly affected by certain crisis years. The
negative growth in these years obviously resulted in high numbers of average negative
deviations for the entire decade, as in the 1970s. However, unemployment remained at
low levels, particularly between 1950 and 1990. The Swiss pattern does not point to a
pro-cyclical relationship between strikes and business fluctuations throughout the
investigation period, whereas the Austrian and German results point to such a pattern,
especially in periods of economic prosperity. Compared to these countries, Switzerland
moreover tends to be characterised by a low level of overall strike activity, particularly
after 1945 (Columns 6 and 7). Against the general trend, however, recent decades have
been characterised by a unique increase in strike numbers, which was not observed in
the other two countries examined.
One explanation for these deviating developments and results might be the fact
that Swiss unions were too fragmented and decentralised to pursue efficiently coordinated actions. This concerned not only the general adjustment of wage claims in
line with macroeconomic criteria, but also potential for collective actions to enforce
such claims. Switzerland is generally counted among the corporatist countries, because
it exhibits an institutionalised co-ordination of interest groups as well as the
involvement of these groups in political matters. However, the prevailing corporatist
institutions and habits are less pronounced than in the other two countries under
analysis, and labour relations in particular tend to be the weaker strand within the
general corporatist system (Oesch, 2007: 339).33 The Swiss system is sometimes
referred to as liberal corporatism, which is characterised, among other factors, by strong
and centralised employers organisations but weaker and less integrated unions and state
authorities (Katzenstein, 1984; 1985). This underlines an important difference between
Switzerland and Germany, and especially Switzerland and Austria, where labour
institutions exhibit a high degree of centralisation. Unions in Switzerland are politically
33

An overview of Swiss industrial relations can be found in Kriesi and Trechsel (2008, Chapter 7), and
Fluder and Hotz-Hart (1998). A widely-cited study comparing democratic corporatism in Austria and
Switzerland is by Katzenstein (1984).

45

divided, which makes the co-ordinated representation of workers interests more


difficult. Union organisation along political lines was also the prevailing characteristic
of Germany and Austria before World War II. After 1945, both these countries
established united federations along industrial principles, which facilitated robust and
co-ordinated union politics in the post-war period. In this respect, the Second World
War did not represent a turning point for Swiss trade unions, because they also kept
their political divisions in the second half of the 20th century.
Another reason for the low level of strike activity, in addition to the indistinct
cyclical pattern, might be the fact that a central part of the Swiss workforce, the metal
workers, was not allowed to strike during large parts of the 20th century as a result of the
peace accord (Friedensabkommen). The Friedensabkommen was signed in the year
1937, between the metal and clock workers unions (Schweizerischer Metall- und
Uhrenarbeiterverband, SMUV) and the corresponding business association, and applied
only to this branch. Strictly speaking, the peace accord was not a collective agreement,
because it lacked any normative regulations with regard to wages and working
conditions. The contents of this agreement were arranged on a bona fide basis and the
only binding stipulation concerned an absolute ban on strikes during the term of the
agreement, which was about five years.
The organisational fragmentation of the Swiss trade unions is also reflected in
the landscape of collective agreements, characterised by low overall coverage and many
firm contracts. This points to the fact that bargaining was also not shaped in very
centralist manner. As a matter of fact, these characteristics differ considerably from the
prevailing area-wide bargaining in Germany and Austria, where area-wide collective
action was easier to co-ordinate. As a result of the favourable economic conditions in
the 1950s and 1960s, it was often not necessary to reinforce wage claims with strike
action. Other issues concerning employment conditions, such as working hours, were
not approached in a centralised and co-ordinated way as the example of the plasterers
strike revealed.
These factors seemed to contribute to the overall low levels, as well as to the
cyclical pattern, of strikes. Both features were further strained by the weak position of
the unions at the plant level. Workplace representation and co-determination is
traditionally weak and fragmented in Switzerland. Despite the existence of works
committees (Betriebskommission), at least in larger companies, their influence and
rights are not comparable with the strong works councils in Austria and Germany.
Although these bodies play a regulating role in conflicts, they also facilitate union
action and co-ordination at plant level, as access to this level becomes easier through
works councils that often include union members. This is particularly important for the
organisation of strike action.
The massive influx of foreign workers during the boom period was moreover not
conducive to the strengthening of the coherency of the Swiss working class. It
supported the process of upward mobility of many Swiss workers and helped in the
regulation of unemployment rates, as the regularly occurring deportations revealed.
However, although many strikes occurred with the considerable involvement and
leadership of foreign workers (like a series of wildcat strikes in the early 1970s), the
general disintegration of this part of the workforce also hampered efficient procedures
with regard to strike action.34
34

For a study dealing with foreign workers and union politics in Switzerland, see Riedo (1976).

46

The 1970s experienced a short period of increasing numbers of strikes. In the


course of the severe economic crisis, most of the struggles tended to be defensive in
nature, especially in the notably shaken metal and clock industry, and were confined to
the French-speaking part of Switzerland (Autorenkollektiv, 1980: 404). Although these
disputes were contested with unprecedented severity, including factory occupations,
there were no similar activities in the rest of the country. Territorial fragmentation
tended to be an additional factor in the weakening of union coherency. Table 4 reveals
that defensive struggles in this period increased strike levels, but not the number of
strike days. The cyclical pattern, however, did not undergo changes in this period and
remained uncorrelated with business fluctuations, as was the case for most of the time
during the period under investigation.
7. Conclusion
The aim of this paper is to identify the cyclical interrelationship between business
cycles and strike activity in Austria, Germany and Switzerland during the 20th century.
Considering the classical ranges of the business cycles, the main results revealed the
existence of a cyclical pattern in strike activity of about three to five years. This cycle is
correlated with cycles of business activity of the same length, especially in the period of
prosperity between 1950 and 1973. This result particularly applies to Germany and
Austria, whereas industrial conflicts in Switzerland appear to be less correlated with
business fluctuations in general.
It turns out that the combination of economic prosperity and a particular
institutional setting, characterised by corporatist arrangements and institutionalised
collective bargaining, has been conducive to the promotion of a conflict pattern
characterised by increasing co-movement between labour conflicts and economic
conditions. During recent decades, characterised by economic slowdown, a weakening
of corporatism and the relaxation of collective bargaining traditions, the discovered
correlation between strikes and business cycles attenuated. The general three to five
year pattern of strike activity, however, remained fairly stable.
It has been argued that this increasing correlation between strikes and business
cycles reflects the peacefulness and regulation of industrial relations during the
extraordinary phase of economic growth following World War II. Favourable economic
conditions are a prerequisite for peaceful labour relations, as pressures relating to
structural distribution conflicts become relaxed. Conflicts between capital and labour
did not disappear, however. Consistent economic growth facilitated the dialogue
between social opponents regarding the distribution of national income. As a result of
extraordinary growth, collective bargaining adopted the character of positive sum
situations, where the gains for all involved parties were greater than zero due to the
enlargement of the general economic output. This in turn lowered the overall extent of
conflicts and fostered regulated structures for collective bargaining, which were not
superimposed by severe class antagonism. The general enforcement of collective
bargaining in this period might have contributed to the dominating conflict pattern of
three to five years. This indicates that strikes occurred in regular intervals and might be
attributed to the regularly occurring negotiations.
Moreover, during this extraordinary growth period the cyclical pattern of the
strike series coincided with the ups and downs of business fluctuations, particularly in

47

Austria and Germany: higher growth promoted higher demands on the part of workers,
which accordingly increased the intensity of strike action. The trough phases were
characterised by modest wage claims, which in turn provoked less strike action
compared to the peak phase of a given cycle. In Switzerland, the cyclical pattern of
industrial conflicts was not strongly affected by economic fluctuations, although the
economic environment tended to reveal similar characteristics to the other two
countries.
Country-specific development of industrial relations might explain these
deviations to some extent. Within the framework of favourable economic conditions, a
general trend to deal with social, as well as economic issues such as wage and price
setting on a macro-level, characterised the collaboration between social forces.
Particularly in Germany and Austria, these developments led to an increasing
adjustment of labour politics along the lines of macroeconomic criteria. These two
countries were particularly characterised by a high degree of centralisation of labour
institutions that was also reflected in the collective bargaining framework.35 Both
countries established a trade union system along industrial lines after 1945, whereas in
Switzerland the unions exhibited partisan structures. As a result of these institutional
prerequisites, co-ordinated actions and proceedings were easier to implement in Austria
and Germany than in Switzerland. This affected not only general collaboration with
employers, but also the conflict pattern that was to some extent also determined on a
central level. Switzerland, on the contrary, used to be characterised by highly
decentralised industrial structures, which also affected industrial relations. Collective
bargaining as well as strike activity was never co-ordinated in a similar way to Austria
and Germany, which might have contributed to the relative absence of the pro-cyclical
pattern between strikes and business fluctuations.
The pro-cyclical pattern which emerged as the result of this particular
combination of economic conditions and institutional settings, however, was not always
found in the previously described pure form. The German results as well as the Austrian
findings indicate that a considerable part of pro-cyclical strike action is attributable to
unofficial strikes. At the end of the 1960s in Germany, and the 1970s in Austria, procyclical wildcat strikes dominated the pattern of conflicts, because workers were
dissatisfied with union demands and settlements. There is also evidence that the
consideration of strike days as an indicator of strikes might cause some distortion.
Basing the analysis on this measurement, anti-cyclical strikes gain more weight because
they reveal higher numbers of strike days compared to strikes in boom periods.
However, the general trend of the findings points to a pro-cyclical pattern of
strike activity during the boom years. Compared to the 1950s and 1960s, economic
development in recent decades has been characterised by a slowdown of growth in all
three countries, albeit with different timings. In parallel with economic slowdown,
unemployment rates have increased in all three countries. Within this new framework,
the cyclical pattern of strike activity changed again, and the correlation of the strike
cycle with the business cycle has diminished.
35

Germany is usually not classified among the countries with highly centralised industrial relations. This
is due to the relative autonomy of the unions affiliated to the DGB in the collective bargaining framework
(Flanagan 1999). Any classification, however, depends on the reference values, which vary considerably
in most of the related studies. Considering the high coverage rate of labour agreements compared to that
of Switzerland, it is fairly safe to classify Germany as a country with a comparatively high degree of
union centralisation in the present context.

48

It has been argued that the decreasing correlation of the two indicators reflects a
general weakening of the position of labour in the social fabric, due to worsening
economic conditions and higher unemployment. The collaboration between social
forces and corporatist arrangements has been lessened, particularly on the side of
employers, as business has sought to act more autonomously in the face of an economic
downturn. These developments also found a reflection in the changing character of
strike causes. While the pro-cyclical pattern of labour conflicts during the boom years
was dominated by wage strikes, recent decades have shown an increase in defensive
struggles against privatisations and relocations, as well as contract violations.
Recent decades have shown a general decline in overall strike activity in Austria
and Germany compared to the 1950s and 1960s.36 Although it is apparent that strikes
did not correlate with business fluctuations in Switzerland, strikes occurred regularly
and, contrary to Austria and Germany, have even intensified in recent decades. Due to
structural characteristics like bargaining at branch level, German and Austrian unions
are able to mobilise large units of workers for industrial disputes. As a result of the
decentralised nature of industry and labour relations, strike activity in Switzerland is
predominated by conflicts in smaller units such as individual firms. The effects of these
structural differences between the countries have to be taken into consideration when
evaluating Switzerlands strike intensity based on the number of workers involved and
the number of strike days when compared to other countries. Taking these differences
into account, Switzerland emerges as a surprisingly dynamic country with regard to
industrial conflict.
The present analysis has confirmed previous findings, which suggest that short
business cycles in particular structure the occurrence of strikes. It has been shown that
short business cycles played a role in the entire investigation period, at least in the case
of Austria and Germany, although the intensity of its influence and the strength of the
correlation varied over time. Business cycles were able to explain the timing of strikes
in the specific periods characterised by a particular combination of economic
development and institutional setting. However, economic fluctuations alone cannot
determine the extent of the increase or decrease of these periodic movements. For a
complete picture of strike activity, it is important to consider short-term as well as longterm fluctuations.

36

The comparatively high number of strike days per head in Austria between 2000 and 2004 can be
attributed to the pension strike in 2003, which should be considered as an outlier.

49

References
A'Hearn, B. and U. Woitek (2001), More International Evidence on the Historical
Properties of Business Cycles Journal of Monetary Economics, vol. 47, pp. 299319.
Altmann, G. (2004), Aktive Arbeitsmarktpolitk. Entstehung und Wirkung eines
Reformkonzepts in Deutschland, Stuttgart, Franz Steiner Verlag.
Ashenfelter, O. and G. E. Johnson (1969), Bargaining Theory, Trade Unions, and
Industrial Strike Activity American Economic Review, vol. 59, no. 1, pp. 3549.
Aubin, H. and W. Zorn, Eds. (1976), Handbuch der Deutschen Wirtschafts und
Sozialgeschichte, Stuttgart, Ernst Klett Verlag.
Autorenkollektiv (1980), Schweizerische Arbeiterbewegung. Ergnzungsband 196879.
Dokumente zu Lage, Organisation und Kmpfen der Arbeiter. Zrich, Limmat Verlag.
Bauernfeind, W. and U. Woitek (1996), Agrarian Cycles in Germany 1339-1670: A
Spectral Analysis of Grain Prices and Output in Nuremberg, Explorations in Economic
History, vol. 33, pp. 459478.
Bergman, U. M. et al. (1998), 'Historical Evidence on Business Cycles: The
International Experience', Working Paper Series in Economics and Finance Stockholm
School of Economics.
Bergmann, J. et al. (1975), Gewerkschaften in der Bundesrepublik. Gewerkschaftliche
Lohnpolitik zwischen Mitgliederinteressen und konomischen Sachzwngen, Frankfurt
und Kln, Europische Verlagsanstalt.
Birke, P. (2007), Wilde Streiks im Wirtschaftswunder. Arbeitskmpfe, Gewerkschaften
und soziale Bewegungen in der Bundesrepublik und Dnemark, Frankfurt und New
York, Campus Verlag.
Bispinck, R. (1993), 'Bundesrepublik Deutschland', in Tarifpolitik und Tarifsysteme in
Europa, eds. R. Bispinck and W. Lechner, Kln, Bund-Verlag pp. 4879.
Board, T. C. (2009), Total Economy Database.
Boleslawski, R. (1919), Die Streikbewegung in sterreich whrend des Krieges im
Vergleiche zur Friedenszeit, Statistische Monatsschrift, vol. 1, pp. 146160.
Borchardt, K. (1982), Wachstum, Krisen, Handlungsspielrume der Wirtschaftspolitik,
Gttingen, Vandenhoeck & Rupprecht.
Botz, G. (1978), 'Streik in sterreich 1918-1975. Probleme und Ergebnisse einer
quantitativen Studie', Bewegung und Klasse. Studien zur sterreichischen
Arbeitergeschichte, eds. G. Botz, H. Hautmann and H. Konrad, Wien, Europaverlag, pp.
807831.

50

Brandner, P. and K. Neusser (1992), Business Cycles in Open Economies: Stylised


Facts for Austria and Germany, Weltwirtschaftliches Archiv Review of World
Economics, vol. 128, no. 1, pp. 6787.
Burg, J. P. (1967) Maximum Entropy Spectral Analysis, Proceedings of the 37th
Meeting of the Society of Exploration Geophysicists.
-------------- (1968), 'A New Analysis Technique for Time Series Data', NATO Advanced
Study Institute on Signal Processing with Emphasis on Underwater Acoustics.
-------------- (1975), Maximum Entropy Spectral Analysis, Stanford University.
Butschek, F. (1985), Die sterreichische Wirtschaft im 20. Jahrhundert, Wien, Gustav
Fischer Verlag.
---------------- (1999), Statistische Reihen zur sterreichische Wirtschaftsgeschchte. Die
sterreichische Wirtschaft seit der industriellen Revolution, Wien, sterreichisches
Institut fr Wirtschaftsforschung (WIFO).
Carlin, W. (1996), 'West German Growth and Institutions, 194590' in Economic
Growth in Europe since 1945, eds. N. Crafts and G. Toniolo, Cambridge, Cambridge
University Press, pp.455497.
Chaloupek, G., et al., eds. (2005), sterreichische Industriegeschichte 1955-2005. Die
ergriffene Chance, Wien, Ueberreuter Verlag.
Childers, D. G. (1978), Modern Spectrum Analysis, New York, IEEE Press.
Croux, C. et al. (2001), A Measure of Comovement for Economic Variables: Theory
and Empirics, Review of Economics and Statistics, vol. 83, pp. 232241.
Dribbusch, H. (2007), Industrial Action in a Low-Strike Country. Strikes in Germany
1968-2005, in: S. v. d. Velden 2007, pp. 133154.
Eichengreen, B. (2007), The European Economy since 1945. Coordinated Capitalism
and Beyond, Princeton and Oxford, Princeton University Press.
Feinstein, C. et al. (1997), The European Economy between the Wars, Oxford, Oxford
University Press.
Fischer, W. et al. (1987), Handbuch
Sozialgeschichte, Stuttgart, Klett-Cotta.

der

Europischen

Wirtschafts- und

Fluder, R. and B. Hotz-Hart (1998), 'Switzerland: Still as Smooth as Clockwork?' in


Changing Industrial Relations in Europe, eds. A. Ferner and R. Hyman, Oxford,
Blackwell, pp. 262281.
Franzosi, R. (1989), One Hundred Years of Strike Statistics: Methodological and
Theoretical Issues in Quantitative Strike Research, Industrial and Labor Relations
Review, vol. 42, pp. 348326.

51

--------------- (1995), The Puzzle of Strikes. Class and State Strategies in Post-war Italy,
Cambridge, Cambridge University Press.
Glyn, A. (2006), Capitalism Unleashed. Finance, Globalisation, and Welfare, Oxford,
Oxford University Press.
Goldner, W. (1953), Strikes and Prosperity. Comment on Levitt 1953, Industrial and
Labor Relations Review, vol. 64, pp. 579581.
Guger,
A.
(1993),
'Lohnpolitik
und
Sozialpartnerschaft',
in
Sozialpartnerschaft:Kontinuitt und Wandel eines Modells, ed. E. Tlos, Wien, Verlag
fr Geselschaftskritik, pp. 227242.
Hagelstange, T. (1979), Der Einfluss der konomischen Konjunktur auf die
Streikttigkeit und die Mitgliederstrke der Gewerkschaften in der BRD von 1950 bis
1975, Stuttgart.
Hansen, A. H. (1921), Cycles of Strikes, American Economic Review, vol. 114, pp.
616621.
Harrison, A. and M. Stewart (1994), Is Strike Behaviour Cyclical? Journal of Labour
Economics, vol. 12 pp. 524553.
Harvey, A. C. (1992), Forecasting, Structural Time Series Models and the Kalman
Filter, Cambridge, Cambridge University Press.
Hibbs, D. A. (1978), On the Political Economy of Long-Run Trends in Strike
Activity, British Journal of Political Science, vol. 82, pp. 153175.
Hillinger, C. and M. Sebold-Bender (1992), The Stylized Facts of Macroeconomic
Fluctuations, in Cyclical Growth in Market and Planned Economies, London, Oxford
University Press, pp. 63110.
Hirter, H. (1988), Die Streiks in der Schweiz in den Jahren 1888-1914: Quantitative
Streikanalyse. Arbeiterschaft und Wirtschaft in der Schweiz 1880-1914, in Soziale
Lage. Organisation und Kmpfe von Arbeitern und Unternehmern, politische
Organisation und Sozialpolitik, eds. A. Balthasar, E. Gruner and H. Hirter, Zrich,
Chronos 2/1, pp. 8371529.
Hobsbawm, E. J. (1964), Labouring Men. Studies in the History of Labour, London,
Weidenfeld and Nicolson.
Hodrick, R. J. and E. C. Prescott (1997), Post-war U.S. Business Cycles: An Empirical
Investigation, Journal of Money, Credit and Banking, vol. 29, pp. 116.
Hyman, R. (2001), Understanding European Trade Unionism: Between Market, Class
and Society. London, Sage.

52

International Labour Office, I. Yearbook of Labour Statistics.


--------------------------- (1993), Sources and Methods: Strikes and Lockouts.
Jurkat, E. and D. Jurkat (1949), Economic Function of Strikes, Industrial and Labor
Relation Review, vol. 24, pp. 527545.
Kablitz, R. (1972), Der Arbeitskampf in der BRD - Aussperrung und Streik 1948-1968,
Bochum, Ruhr-Universitt.
Kaelble, H. and H. Volkmann (1972), Konjunktur und Streik whrend des bergangs
zum Organisierten Kapitalismus in Deutschland, Zeitschrift fr Wirtschafts- und
Sozialwissenschaften, vol. 92, no. 5, pp. 513544.
Karlhofer, F. (1983), 'Wilde' Streiks in sterreich. Entstehungs- und
Verlaufsbedingungen industrieller Konflikte in den Siebziger Jahren, Wien und Kln,
Braumller.
Karlhofer, F. and E. Tlos, Eds. (1999), Zukunft der Sozialpartnerschaft:
Vernderungsdynamik und Reformbedarf, Wien, Signum.
Katzenstein, P. J. (1984), Corporatism and Change. Austria, Switzerland, and the
Politics of Industry, Ithaca and London, Cornell University Press.
---------------------- (1985), Small States in World Markets. Industrial Policy in Europe,
Ithaca and London, Cornell University Press.
Kennan, J. (1986), The Economics of Strike, in Handbook of Labour Economics, eds.
O. Ashenfelter and R. Layard. Amsterdam, Elsevier Science Publishers B.V., vol. 2, pp.
10911137.
Kittel, B. and E. Tlos (1999), Interessenvermittlung und politischer
Entscheidungsprozess. Sozialpartnerschaft in den 1990er Jahren, in Zukunft der
Sozialpartnerschaft: Vernderungsdynamik und Reformbedarf, eds. F. Karlhofer and E.
Tlos, Wien, Signum, pp. 95136.
Kittner, M. (2005), Arbeitskampf. Geschichte, Recht, Gegenwart, Mnchen, Beck.
Klenner, F. and B. Pellar (1999), Die sterreichische Gewerkschaftsbewegung: Von den
Anfngen bis 1999, Wien, GB-Verlag.
Koller, C. (2008), Ladehemmungen, in Kriegswirtschaft und Wirtschaftskriege, eds. V.
Groebner, S. Guex and J. Tanner, Zrich, Chronos, pp. 213229.
Kondratieff, N. D. (1926), Die langen Wellen der Konjunktur, Archiv fr
Sozialwissenschaft und Sozialpolitik, vol. 56, pp. 573609.
Kriesi, H. and A. H. Trechsel (2008), Switzerland. Continuity and Change in a
Consensus Democracy, Cambridge, Cambridge University Press.

53

Kropf, R. (2004), Streiks in sterreich von 1945 bis zur Gegenwart, in Streiknachlese.
Geschichten vom Widerstand in Obersterreich, eds. E. Gumplmaier and C. Janko,
Wien, GB-Verlag.
Levitt, T. (1953), Prosperity versus Strikes Industrial and Labor Relation Review, vol.
62, pp. 220226.
Lyddon, D. (2007), Strike statistics and the problems of international comparison. in S.
v. d. Velden ed. 2007, pp. 2439.
Maddison, A. (1991), Dynamic Forces in Capitalist Development. A Long-Run
Comparative View, Oxford, Oxford University Press.
----------------- (2004), The World Economy: Historical Statistics, Paris, OECD.
Martens, H. (1979), Der Streik um die 35-Stunden Woche in der Stahlindustrie.
Arbeitskampf um Arbeitszeit, in Kritisches Gewerkschaftsjahrbuch 1979/80, eds. O.
Jacobi, W. Mller-Jentsch and E. Schmidt, Berlin, Rotbuch Verlag, pp. 924.
Marx, K. (1853), Russian Policy against Turkey/Chartism, New-York Daily Tribune,
New York. No. 3819.
Mitchell, B. R. (2003), International Historical Statistics. Europe. 1750-2000,
Houndsmills, Basingstoke, Palgrave Macmillan.
Mller-Jentsch, W. (1973), Entwicklungen und Widersprche in der westdeutschen
Gewerkschaftsbewegung. Gewerkschaften und Klassenkampf. Kritisches Jahrbuch
1973, eds. O. Jacobi, W. Mller-Jentsch and E. Schmidt, Frankfurt am Main, Fischer
Taschenbuchverlag, pp.150162.
Oesch, D. (2007), Weniger Koordination, mehr Markt? Kollektive Arbeitsbeziehungen
und Neokorporatismus in der Schweiz seit 1990 Swiss Political Science Review, vol.
133, pp.337368.
Otruba, G. (1968), sterreichs Wirtschaft im 20. Jahrhundert, Wien und Mnchen,
sterreichischer Bundesverlag.
------------- (1981), Streikbewegung, Ausstnde und Militrassistenzen in der
sterreichischen Reichshlfte 1890 bis 1914, in Wechselwirkung von Konjunkturverlauf,
Preis- und Lohnentwicklung sowie gewerkschaftlicher Organisation. Wirtschaftskrfte
und Wirtschaftswege, ed. V. J. Schneider, Stuttgart, Klett-Cotta, vol. 8, pp.439504.
Pellar, B. (2003), Die schrfste Waffe, Arbeit und Wirtschaft vol. 57.
Perry, L. J. and P. J. Wilson (2004), Trends in Work Stoppages: A Global Perspective,
International Labour Organisation, ILO.
Ravn, M. O. and H. Uhlig (2002), On Adjusting the Hodrick-Prescott Filter for the
Frequency of Observations, Review of Economics and Statistics vol. 84, pp.371376.

54

Rees, A. (1952), Industrial Conflict and Business Fluctuations, The Journal of


Political
Economy, vol. 60, pp.377382.
Riedo, R. (1976), Das Problem der auslndischen Arbeitskrfte in der schweizerischen
Gewerkschaftspolitik von 1945-1970, Bern, Herbert Lang.
Rosner, P. (1999), 'Lohnbewegung und Bewegung der Lohnpolitik', in Zukunft der
Sozialpartnerschaft: Vernderungsdynamik und Reformbedarf, ed. F. Karlhofer and E.
Tlos, Wien, Signum, pp.7594.
Scheiblecker, M. (2002), Konjunkturzyklen in der sterreichischen Wirtschaft, WIFOMonatsberichte, vol. 7.
Schneider, M. (1987), 'Hhen, Krisen und Tiefen. Die Gewerkschaften in der Weimarer
Republik 1918 bis 1933', in Geschichte der deutschen Gewerkschaften von den
Anfngen bis 1945, ed. K. Tenfelde et al., Kln, pp. 279446.
Screpanti, E. (1987), Long Cycles in Strike Activity: An Empirical Investigation,
British Journal of Industrial Relations, vol. 25, pp. 99124.
Scully, G. (1971), Business Cycles and Industrial Strike Activity, The Journal of
Business, vol. 44, pp. 359374.
Shalev, M. (1980), Trade Unions and Economic Analysis - The Case of Industrial
Conflict, Journal of Labour Research, vol. 1, pp. 133174.
Silver, B. J. (2005), Forces of Labour. Arbeiterbewegung und Globalisierung seit 1870,
Assoziation A, Berlin.
Snyder, D. (1975), Institutional Setting and Industrial Conflict: Comparative Analysis
of France, Italy and the United States, American Sociological Review, vol. 403, pp.
259278.
Solomou, S. N. (1998), Economic Cycles. Long Cycles and Business Cycles since 1870,
Manchester and New York, Manchester University Press.
Soskice, D. (1978), 'Strike Waves and Wages Explosions, 1968-1970: An Economic
Interpretation' in The Resurgence of Class Conflict in Western Europe since 1968,
Volume 2: Comparative Analyses, ed. C. Crouch and A. Pizzorno, London, Macmillan,
pp. 221247.
Spode, H. et al., Eds. (1992), Statistik der Arbeitskmpfe in Deutschland. Quellen und
Forschungen zur historischen Statistik von Deutschland, St. Katharinen, Scripta
Mercature Verlag.
Tlos, E. (2008), Sozialpartnerschaft. Ein zentraler Gestaltungsfaktor in der Zweiten
Republik, Innsbruck, Wien und Bozen, Studienverlag.

55

Tlos, E., Ed. (1993), Sozialpartnerschaft:Kontinuitt und Wandel eines Modells, Wien,
Verlag fr Geselschaftskritik.
Tichy,
G.
(1995),
'Austrokeynesianismus,
Ein
Konzept
erfolgreicher
Wirtschaftspolitik?', in sterreich 1945-1995, eds. R. Sieder, H. Steinert and E. Tlos,
Wien, Verlag fr Gesellschaftskritik, pp. 213222.
Tomandl, T. (1965), Streik und Aussperrung als Mittel des Arbeitskampfes, Wien und
New York, Springer Verlag.
Tracy, J. S. (1986), An Investigation into the Determinants of U.S. Strike Activity,
The American Economic Review, vol. 763, pp. 423436.
Traxler, F. (1993), 'sterreich' in Tarifpolitik und Tarifsysteme in Europa, eds. R.
Bispinck and W. Lechner, Kln, Bund-Verlag, pp. 283313.
------------- (1998), 'Austria: Still the Country of Corporatism', in Changing Industrial
Relations in Europe, eds. A. Ferner and R. Hyman, Oxford, Blackwell, pp. 239261.
Velden, S. v. d. et al. (2007), Strikes around the World. Case-studies of 15 Countries.
Amsterdam, Aksant Academic Publishers.
Visser, J. (2009), The ICTWSS Database: Database on Institutional Characteristics of
Trade Unions, Wage Setting, State Intervention and Social Pacts in 34 countries
between 1960 and 2007, Amsterdam Institute for Advanced Labour Studies AIAS,
University of Amsterdam.
Vroman, S. B. (1989), A Longitudinal Analysis of Strike Activity in U.S.
Manufacturing: 1957-1984, American Economic Review, vol. 79, no. 4, pp. 816826
Woitek, U. (1996), The G7-Countries: A Multivariate Description of Business Cycle
Stylized Facts. Dynamic Disequilibrium Modelling: Theory and Applications,
Cambridge, Cambridge University Press, pp. 283309
Yoder, D. (1940), Economic Changes and Industrial Unrest in the United States The
Journal of Political Economy, vol. 48, no. 2, pp. 222237.
Zarnowitz, V. (1992), Business Cycles. Theory, History, Indicators, and Forecasting.
Chicago and London, University of Chicago Press.

56

Appendix
Multivariate autoregressive models (VARs) are fitted in the time domain and
transformed into the frequency domain.37 Spectral analysis requires stationary
processes, and thus the data has been filtered using the Hodrick-Prescott filter (Hodrick
and Prescott 1997) to make it stationary (smoothing weight = 6.25 , Ravn and Uhlig
[2002]). The starting point of the analysis is a VAR of order p :
p

xt = c + A j xt j + u t =
j =1

= (c A A p ) 1 xTt1 xTt p T + ut =
A

(1) .

Z t 1

= AZt 1 + ut ; ut ~ iid (0, H )


The VAR is estimated with real GDP per capita and the strike days series for each of the
countries under analysis. If the VAR model is rewritten as in Equation 1, it can be
reformulated as a state space model. This is a necessary step in order to obtain time
dependent parameter matrices.
The state space model requires a measurement equation which is vectorised. The
parameter matrix A and made time dependent is:
xt = (Zt' 1 I) vecA + ut

(2) .

The transition equation determines the pattern of the respective parameters t

t = T t 1 + t ; t ~ iid (0, Q )

(3) .

The model is estimated using Maximum Likelihood.38


The spectrum of the VAR is defined as
F( ) =

1
A( ) 1 HA ( ) ; [ , ]
2

(4) ,

where H is the error variance-covariance matrix of the model, and A( )t is the Fourier
transform of the matrix lag polynomial A( L)t = I A 1,t L A p,t Lp .39

37

This method was pioneered by John Parker Burg in the 1960s (Burg 1967; 1968; 1975). His
contributions in this time can also be found in Childers (1978). Examples for applications in macro
econometrics and economic history are Hillinger and Sebold-Bender (1992), Bauernfeind and Woitek
(1996), and AHearn and Woitek (2001).
38
To obtain the likelihood function, the Kalman filter has been used. For details, see Harvey (1992).

57

Once the time-dependent matrices of the VAR are estimated, it is possible to obtain the
spectral density matrix at each point in time:
F( ) t =

1
A( ) t1 HA ( ) t ; [ , ]
2

(4') .

This is also important for the cross-spectral analysis, which examines the
interrelationship of two different series: strikes and business cycles. For this purpose,
the cross spectra (off-diagonal elements of the spectral density matrix) together with the
univariate spectra are used to calculate the squared coherency sc( ) , a measure that is
similar to the correlation coefficient:

sc( ) =

f xy ( )

f x ( ) f y ( )

(5) .

; 0 sc( ) 1

The squared coherency can be used to decompose the fraction of overall variance in this
interval into explained and unexplained parts:
2

f y ( )d = sc( ) f y ( )d +

fu ( )d

(6) .

explained variance unexplained variance


To identify the lead-lag structure, Croux et al. (2001) suggested measuring the
correlation between the in phase components of two series at a frequency . The socalled dynamic correlation ( ) is given by

( ) =

cxy ( )

f x ( ) f y ( )

; 1 ( ) 1

(7 ) .

Using

sc( ) =

f xy ( )

f x ( ) f y ( )

cxy ( ) + qxy ( )
2

f x ( ) f y ( )

(5') ,

explained variance as described above can be further decomposed into explained


variance in-phase and explained variance out-of-phase components:

L is the backshift operator; the superscript denotes the complex conjugate transpose of the matrix
A( )t .

39

58

f y ( )d = sc( ) f y ( )d +
1

c xy ( ) + q xy ( )
2

f x ( ) f y ( )

c xy ( )

f x ( ) f y ( )

f y ( )d +

f y ( )d +

(6') .

f u ( )d
f u ( )d

q xy ( )

f x ( ) f y ( )

f y ( )d +

f u ( )d

explained variance (in-phase) explained variance (out-of-phase) unexplained


variance

59

Strike volume in Ghana. Trajectory of labour strikes in


Ghana
Edward Fokuoh Ampratwum
1. Case Study
Between 1968 and 1970 workers in Ghana's gold mines were involved in
a series of large-scale strikes, riots and demonstrations. In the course of
these events workers were shot by police; managers and union officials
were threatened and severely injured by striking mineworkers, and at
Ghana's richest mine the army was asked to intervene to restore order
and save the mine from closure.
Source: Crisp (1979:267)
The Ghanaian labour front has had a long history of industrial disputes. In most cases,
these industrial disputes culminated into strikes, lockouts, go-slows, boycott, overtime
bans etc., as the above cited case study shows. Unions and organized workers have, at
different times in history, used this tool to achieve various socio-economic and political
ends. Data on this phenomenon is however, largely inadequate, incomplete and focuses
mainly on strike volume variables. This chapter attempts a trajectory of strikes in the
industrial front in Ghana. The paper analyzes strikes from the pre-independence era
through to the return to democracy in 1992 and beyond. Particular attempt is made to
examine the propensity of strikes as well as the impact of strikes on Ghanas
development. The chapter further assesses the processes of collecting and analyzing
strike data in Ghana. Finally, strike data sets compiled from the Department of Labour
in Ghana is analyzed to assess the strike volume in Ghana.

2. Definition
A strike is a very powerful weapon used by trade unions and other labor associations to
get their demands accepted. It generally involves quitting of work by a group of workers
for the purpose of bringing pressure on their employers so that their demands get
accepted. When workers collectively cease to work in a particular industry, they are said
to be on strike. The 2003 Industrial Relations Act of Ghana defines a strike as any
action by two or more workers acting in concert which is intended by them to restrict in
any way the service they normally provide to the employer or diminish the output of
such service with a view to applying coercive pressure upon the employer (section
174:64). According to the Act, a strike can be permitted where the parties fail to agree
to refer a dispute to voluntary arbitration or the dispute remains unresolved at the end of
the arbitration proceedings. In that case, either party intending to strike or institute a
lock out shall give written notice of intention to the other party and the commission
within 7 days. A party who gives such a notice to resort to strike may do so only after
the expiration of the 7 days from the date of notice and not at any time before the end of

60

that notice. If a strike is declared without recourse to the above processes, it will be
illegal and participation in an illegal strike can be a basis for termination without notice
or may forfeit remuneration for the period of engagement in strike. Although these
procedures are in conformity with core International Labour Organization (ILO)
conventions which Ghana has ratified, including a procedure for calling a strike, the
complexity of such procedures as described above has made it almost impossible to
have a fully legal strike in Ghana (ICFTU, 2001).
3. Short History of strikes in Ghana
There is a long history of strikes in Africa. During the early part of the 20th century,
workers across Africa particularly those in British colonies began protests against poor
working conditions. In the Gold coast (Ghana), for instance, mineworkers and
fishermen embarked on strikes in 1919 and 1925 respectively to back their demand for
improved working conditions. These were followed by a series of other protests and
strike actions (Baah, 2000). Indeed, the first increase of the minimum wage in 1939 in
the Gold coast was a direct outcome of a strike in May 1939 by railway employees
(Ewusi, 1971).
Strikes were also used as a political tool to support the struggle for selfdetermination (Baah, 2000). The potential power of workers was actually greater than
was suggested by their numbers because workers' actions quickly combined with the
simmering discontent found in all African cities to cause an avalanche of popular
protest (Sandbrook, 1977). In 1950, for instance, a general strike was declared
throughout the Gold Coast to coincide with Dr. Nkrumah's call for a Positive Action
campaign of non-violent civil disobedience. This protest, no doubt, marked the birth of
political unionism in Ghana.
Between 1968 and 1970 workers in Ghana's gold mines were involved in a series
of large-scale strikes, riots and demonstrations. In the course of these events workers
were shot by police; managers and union officials were threatened and severely injured
by striking mineworkers. The strikes tended to "articulate the interests of the broadest
stratum of labour, the lower-paid and minimum wage earners" (Kraus, 1979; Herbst,
1993). Again, beyond wage concerns though, many scholars believe that the labor
unions used these strikes to exercise at least some political power. Crisp notes in his
study of mine-workers that "the history of Ghana in colonial and post-colonial periods is
a testament to the susceptibility of the Ghanaian state to the threat of popular unrest and
protest. Crisp makes clear, however, that because of a number of factors relating to
internal organization and ideology, mineworkers have not been able to go beyond being
an episodic political force in the country (Crisps, 1984; Herbst, 1993).
There was an equally high strike frequency between 1971 and 1980 the period
of economic downturn in Ghana. But the strike frequency between 1981 and 1990 was
remarkably low. This was the period the government of Ghana embarked on a wide
range of programs to reform state-owned enterprises. Central to this mission was the
reduction of surplus workers. For instance, in the mid-1980s, the government undertook
an evaluation and redeployment exercise that reduced the size of the Cocoa Board's
payroll from 100,000 employees to 50,000. Other state enterprises underwent similar
programs, although none could claim quite the extravagance of waste that the Cocoa
Board achieved. The government also reduced its own work force, and the Economic

61

Recovery Program included plans to eliminate approximately 36,000 positions from a


total civil service of approximately 540,000 (Herbst, 1993). Given the gravity of these
reforms and its associated sharp deterioration in the standard of urban living, there was
the potential for strikes and labor unrest which could ignite more general popular
protest. However, since the massive reform program was first announced in 1983,
Ghana did not experience significant popular unrest ignited by organized or
unorganized labor (or anyone else for that matter). Why was labor acquiesced to these
economic reforms? Many scholars have identified various reasons for this contradiction:
the Provisional National Defense Council (PNDC) which had come to power in 1981
through a coup dtat had destroyed much of the unions' traditional leadership and
replaced it with people who owed their political survival to Rawlings, the Chairman of
the PNDC. In January 1982, for instance, sixteen general secretaries of various unions
of the Trades Union Congress (TUC) were "reported to have gone into hiding for fear of
molestation by the [military Junta] and also workers, some of whom cursed and cried
for their blood"(Echo, 1982; Herbst, 1993). Again, on April 24, 1982, it was reported
that the existing union leadership was deposed by the Workers Defense Committee and
the interim Coordinating Committee - WDC/ICC and the National Defense Committee
(NDC), all functional organs of the PNDC, parading as Association of Local Unions,
and the TUC was placed under the control of an appointed Interim Management
Committee made up of radical supporters of the new regime (U.S. Department of Labor,
1988; Herbst, 1993). Besides these political maneuvers, strikes were also brutally
quelled using the military and revolutionary forces. During the strikes of workers at
Assere Manufacturing Company, and the Ghana Broadcasting Corporation in 1986 and
1988 respectively, for instance, heavily armed security personnel were deployed against
the workers and suspected ring leaders were arrested detained and branded as economic
saboteurs (Ghana Districts, 2006).. The state-controlled media was also used to
marginalize, undermine and repress labour (Ghana Districts.com, 2006). However, labor
acquiescence in this period was not based only on political and military repression but
also on the government's at least occasional ability to adopt strategies that avoid outright
political conflict (Herbst, 1993). According to Blunch and Verner, the fact that the
public workers permanently laid off from the notoriously overstaffed Cocoa Board
(Nelson, 1991) as part of an Economic Recovery Program (ERP) were compensated
with the equivalent of roughly two years total compensation indicated a concern for
wages over employment among the public workers unions (2001). Again, the cooptation of the leadership of labour unions into government structures also undermined
the capacity of organised labour to resist the anti-labour practices of the PNDC
government. A case in point was the appointment of the TUC Secretary-General as a
member of the National Commission for Democracy, contrary to the TUCs official
stance on the district level elections organised by the Commission in 1988 (Ghana
Districts.com, 2006). It is however noteworthy that, this was the first time in the nation's
history that a regime had gone so far as to try to supplant actual union organization on
the shop floor (Hansen, 1987).
The return to constitutional rule in the 1990s saw increased freedoms and
increased strike activities among working groups and unions for varied reasons:
political, economic etc. The freedoms bestowed on Ghanaians by the 1992 Constitution
as well as the resumption of multi-party democracy explains to some extent the
increased strike frequency beyond 1992 to date which will be reviewed at the latter
sections of the paper.

62

4. The Propensity to Strike


The debate in industrial sociology about the term strike-proneness no doubt epitomizes
the search for causes of industrial conflicts. The propensity to strike has been studied
from a number of different viewpoints. One of these is the organizational perspective,
which states that strikes are related to such structural factors as the extent of
unionization and the degree of centralization and institutionalization in collective
bargaining (Britt and Galle, 1972; Snyder, 1975; James, 1996).
Another perspective attributes strike propensity to such economic factors as
unemployment, inflation, and real wage changes (Yoder, 1940; Rees, 1952; Farber,
1978; James, 1996). In addition to these perspectives, a number of theorists have
suggested that a relationship exists between the incidence of strikes and the level of
industrial development (Ross and Hartman, 1960; Hibbs, 1976; Korpi and Shalev,
1979). Most theorists agree that the propensity to strike increases during early periods
of industrialization. Low levels of developments are characterized by severe social
problems, such as eighteen-hour workdays, poverty, and erratic business cycles. These
problems tend to fuel the propensity for workers to strike for better working conditions
(Haas and Stack, 1983; James, 1996). However, widespread disagreement exists about
the effect of advanced economic development on strike activity. Some theorists expect
strike activity to fall in late industrialization, for reasons such as the decline in
unionization and the identification of workers with management (Blum, 1964); the
separation of ownership from control and the emphasis of growth rather than profit
goals (Haas and Stack, 1983); and the creation of new job skills from technology and
the increased occupational mobility of workers (Dahrendorf, 1959; James, 1996). These
latter arguments associating reduction in strikes with advances in industrialization have
been critiqued by the KerrSiegel hypothesis which offers an alternative analysis of
inter-industry differences in strike-proneness, and argues that high strike-rates among
geographically or socially isolated, cohesive, homogeneous groups of workers (such as
longshoremen, miners, and sailors) were a consequence of their alienation from the
wider society and the unpleasant nature of their jobs. Thus, in the view of the Kerr
Siegel, industrialization may contribute to worker's feeling of alienation, and
consequently raise strike propensity (Haas and Stack, 1983; James, 1996).
The problem with analyzing strike propensity from largely economic
perspectives is that it underplays the political underpinnings or the real politique of
workers motivation to strike. The political consciousness and action of African workers
in large scale foreign and state enterprises has no doubt been a subject of lively debate
among radical scholars (Konings, 1978). Many scholars believe that labor unions can
exercise at least some political power. Freund for instance observes that "a sensitive
analysis of developments in a country such as Ghana also shows that the unions are
conduits at times for shocks that can present difficulties for regimes" (1988:108)
making Bates argue that it is precisely because workers can exercise power through
both organized and unorganized means that African governments find it particularly
difficult to suppress those (1981 as cited in Herbst, 1993).

63

5. Impact of strikes
Industrial conflicts and strikes have tremendous impact on public sector effectiveness
and efficiency as well as the implementation of public policies. Strikes in the education
sector in Ghana have both social and economic implications, for instance, in 1995 due
to a lecturers strike, no doctors passed out locally during that year (Dovlo & Nyonator,
2003). A strike at the Internal Revenue Service, for example, has implications for
government revenue generation, provision of social services and the implementation of
public policies (Baah, 2006). In the private sector, strikes may discourage investment
(Casely, 2006) as increased losses in revenue may be directly related to the duration of
the strike and the period the strike occurs. Crisps (1979:267) accounts of the strikes in
the gold mines (cited in the introductory case study) lasted for weeks and cost the
companies thousands or millions of US dollars.
In many developed and developing countries workers lose their pay when they
go on strike. Unions may pay their striking members in full or partially. But in Ghana,
in the past, workers have generally not lost their pay (or part of their pay) when they
embark on strike. There are many reasons for this. First, many of the employers,
(usually led and motivated by government who also doubles as the largest employer)
realize and accept their responsibility for the occurrence of the strike action and so they
feel morally obliged to pay the workers even for the period they are on strike. When
university teachers went on a 9-month strike in 1995/96 (April 1995 to January 1996),
they received their monthly pay even during the period they were on strike. When
public authorities threatened to eject them from their bungalows they were occupying
on university campuses, some members of the University Teachers Association
(UTAG) went to court. This trend is fast changing though. During the September 2006
strike by members of the National Association of Graduate Teachers (NAGRAT), the
Government did not pay the striking graduate teachers for the period they were on
strike. The difference in the governments handling of the two industrial disputes
emanated from the existence of a strict regime of enforcement in Ghanas 2003 Labour
Act as well as the institutionalization of the National Labour Commission of Ghana, a
labour dispute regulatory and mediating body.
6. Strike data in Ghana
Collecting and measuring strike data poses great challenges. In Africa, statistics on
strikes are sparse and highly dependent on the data source (which is usually incomplete)
(Alby, Azam & Rospab, 2005). Data difficiency issues in Ghana are not very different
from that on the African continent as a whole. For instance, strike data on Ghana found
on the International Labour Organization (ILO) labour statistics (LABORSTA) database
did not have any information from 1992 to date and makes analysis of strike data on
Ghana beyond 1992 almost impossible or at best unreliable and incomplete (See.
http://laborsta.ilo.org/STP/). Stack and Haas (1985) however contend that no evidence
exists that substantiate the argument that undeveloped countries cannot (or will not)
collect accurate strike information.
In Ghana, the Department of Labour is the primary source of strike data. This data is
generated mainly from reports of employers on strike actions and the Labour Department in
turn is required to report the occurrence of a strike in addition to information on number of

64

strikes, causes, duration, and workdays not worked to the Ministry of Employment and
Social Welfare (MESW). More recently, the National Labour Commission also started
collecting some limited data on strikes in Ghana, probably as by-products of its labour
dispute settlement and mediating processes. Unlike the Labour Department though,
variables such as the industry or sector, the duration of the strike in number of days or
hours; as well as the cause of the strike have formed the basis for collecting strike data by
the National Labour Commission. According to Hibbs, though strike volume (strike size
times duration times frequency, or man-days lost due to industrial disputes) is not affected
by differing methods of measurement or calculation, it is a preferred measure of strike
activity because it represents the "net impact of a nation's overall strike profile"
(1976:1036). However, critics have suggested that comparisons between nations are
inappropriate for two reasons: countries at low levels of development are unskilled in the
techniques of measuring strike data (Cameron, 1985; James, 1996), and researchers use
improper measures of strike activity--size (strikers per strike), duration (man-days lost per
striker), frequency (strikes per 1,000 civilian wage and salary workers) or a combination of
the three (Stern, 1978; James, 1996). In the case of the strike data obtained from the Labour
Department in Ghana, the method for compilation and, the variables covered differ from
period to period. According to Baah (2000) some of the earlier data did not capture the
sector in which the strike occurred. There was also the problem of high frequency of
missing values for duration, number of persons involved in the strike and person-days.
7. Analysis of strike volume in Ghana
Strikes have played varied roles in the socio-political development of Ghana and this is
clearly portrayed in the frequency of strikes as seen in Table 1 below. Between 1944-1950,
an average of 22 strikes was recorded. This period marked the creation of British styled
trade unionism in the then Gold Coast. The tempo of recorded strikes doubled in the 19501960 period owing to the struggle for political independence. The improved relations
between the Nkrumah government and the unions saw the period of 1961-66 the time with
the lowest strike levels in the country. Again, the economic downturns of the 1970s to
1980s accounted for the increased strike levels in the period. This was followed by the
culture of silence of the Rawlings regime where the military and pro-government bandits
were used to suppress freedom of association and expression. Strikes frequency has since
increased with the restoration of constitutional rule in the 1990s.

65

Table 1: Strikes in Ghana (1950-2004)


Period
Average number of strikes
1944-1950
22
1950-1960
47
1961-1966
14
1967- 1970
46
1971-1980
43
1981-1990
15
1990-2000
30
2001-2004
34
Source: Calculated from Obeng-Fosu P (1991). Industrial Relations Practice in Ghana:
The Law and Practice 2nd edition Ghana Universities Press, Accra. Page 89, Table 1;
Labour Department Strike Register 1995-2004.

The majority of the recorded strikes took place at the enterprise level. There have been
cases of sector level strikes, particularly in the education sector. General strikes have
not been common. Only two general strikes have been recorded in Ghana the first
took place in 1949 and the second occurred in 1961 (Baah, 2006). Further analysis of
the strike data shows that strikes are more frequent in the private sector, compared to the
public sector. Out of 479 strikes in Ghana between 1980 and 2004, 62% (297) occurred
in the private sector [see Table 2 ].
Table 2: Distribution of Strikes by Sector (1980-2003)
Sector
Private
Public
Total

Frequency
479

297
182

Percent
100

62.0
38.0

Source: Baah (2006) calculation based on strike data obtained from Labour
Department Annual reports.

Analysis of the distribution of strikes by industrial sector indicates that about a quarter
of all strikes that were recorded between 1980 and 2003 occurred in the manufacturing
sector. Construction and community, social and personal services sectors also
experienced considerably high strike frequencies during the period. The sectors that
recorded the lowest strike frequencies were utility, finance and trade sectors as seen in
Table 3 below.
Other aspects of strikes which are essential in accessing the impact of strikes are
the number of workers involved, duration of strikes, and the person-days or hours lost
due to strike actions. Using the 1980 to 2004 data sets, the duration of strikes in Ghana
ranged from one day to 55 days [see Table 4]. Majority of the strikes i.e. 70 percent
lasted between one and three days. 423,260 workers were involved in strike activities in
the 479 strikes recorded and the economy lost 2,836,274 person-days of productive
activity. On average, over 17,000 workers were involved in strikes every year and
Ghana lost 118,178 person days per year (See Table 5 below). It is worth stating that
though the average duration of strike is less than three days, the high numbers of
workers involved resulted in very high losses of person days per year, which could
affect investment, productivity as well as mobilization of revenue.

66

Table 3 Distribution of strike by industry (1980- 2003)


Industry
Agriculture
Mining
Manufacturing
Utility
Construction
Trade
Transport
Finance
Community & Personal services
Total

Frequency

479

50
42
124
6
88
23
40
19
87

Percent of Total

100

10.4
8.8
25.9
1.3
18.4
4.8
8.4
4.0
18.2

Source: Baah (2006) based on strike data obtained from Labour Department Annual reports.

Table 4: Duration of strikes in days (1980 to 2003)


Duration of strike (in days)
1 to 3 days
4 to 10 days
11 days and Over
Total

Frequency
recorded)

(Number

479

of

Strikes % of total number


of strikes
333
69.5
109
22.8
37
7.7
100

Source: Baah (2006) calculation based on strike data obtained from Labour
Department Annual reports.

Reasons underlying strikes in Ghana have ranged from demand for wage increases or
increases in non-wage benefits to discontent with wide disparities in the rate of increase
of salaries/wages between senior staff and junior staff etc. The strike data shows that
almost half of strikes in Ghana are wage-related (i.e., workers use strike actions to back
their demand for wage increases. Out of the 479 strikes that were recorded between
1980 and 2004, 226 were embarked upon to back demand for increases in wages and/or
non-wage benefits. Ninety-seven (97) of them representing 20.3 percent were for
working conditions and 36 representing 7.5 percent were solidarity or sympathy
strikes that were embarked upon mainly to demand the reinstatement of dismissed coworkers.

67

Table 5: Person-Days Lost in Strikes (1980-2003)


Year
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003

Total
AVERAGE FOR THE
PERIOD (1980-2003)

Number of Workers Involved Person-Days Lost


36,095
103,558
34,131
206,126
3,154
6,608
17,532
33,946
457
809
3,746
7,953
7,627
23,752
7,534
16,493
2,756
7,135
4,418
6,249
3,642
6,427
4,202
7,188
8,095
111,129
2,086
5,286
584
1,752
9,238
100,107
15,816
67,745
13,814
28,744
12,846
56,139
163,337
1,779,028
15,448
93,089
13,216
46,535
19,810
36,042
9,925
36,616
409,509
2,836,274
17,063
118,178

Source: Baah (2006) calculation based on strike data obtained from Labour
Department Annual reports.

Table 6. Major causes of strikes in Ghana


Cause
Wage or benefits
Solidarity
Working conditions
Others
Total

Frequency

479

226
36
97
120

Percent

Source: Baah (2006) based on strike data obtained from Labour Department
Annual reports.

100.0

47.2
7.5
20.3
25.1

68

8. Conclusions
Ghanaian workers have used strikes over a long period for various reasons as shown by
the limited data available and the summary statistics presented in the tables above.
Strike data has been collected at the micro level by the Labour Department and more
recently the National Labour Commission with emphasis on quantitative variables such
as the frequency, duration, number of man hours lost etc., from employers reports on
strikes and reports filed by parties at National Labour Commission. A register on strikes
in Ghana from 1995 to 2004 emerging from compilation of micro data has been
published by the renowned International Institute of Social History (IISH) data base on
strikes and industrial conflicts. This register with micro data is a unique and new source
for students of international labour because the ILO has not published Ghanaian data
since 1992 and the data published by Baah (2006) is only at the aggregate level. Table 7
shows the data from the register in a tabular form. This data differs slightly from the
data published by Baah (2006) due to the incompleteness of the micro data. In many
years the number of workers on strike and man-days lost is not mentioned in the file.
We also publish the number of cases available for summing up the data.
Table 7. Strikes in Ghana (1995-2004)
Number
strikes
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004

27
41
36
50
53
29
26
35
37
38

of Number of workers on Strike


Number

9.836
16.544
12.407
13.493
154.177
15.085
14.556
14.243
12.157
13.271

Cases

Source: https://collab.iisg.nl/web/labourconflicts/datafiles

22
41
31
47
42
29
22
35
21
37

Man-days lost
Number

102.031
63.414
31.365
56.390
1.779.731
94.989
46.161
36.592
16.089
53.131

Cases
21
35
30
46
41
29
21
35
19
33

69

References
Alby, Azam & Rospab (2005), Labor institutions, labor-management relations, and
social dialogue in Africa, World Bank.
Anyemedu, K (2001), Trade union responses to globalization: Case study on Ghana,
International Institute for Labour Studies, Geneva.
Baah A.Y. (2006), Strikes in Ghana: Facts and Figures, TUC Policy bulletin, vol. 2,
No 2, pp 69.
------------ (2000), Economic and Working conditions in Ghana in Agbesinyale (Ed)
Democratic workers participation for economic and social development: the case of
Ghana, Ghana Trades Union Congress.
Bates R. (1981), Markets and States in Tropical Africa, Berkeley: University of
California Press.
Blunch N and Verner D (2001) Asymmetries in Union Relative Wage Effects in
Ghanaian Manufacturing An Analysis Applying Quantile Regressions. Working Paper
01-07. Centre for Labour Market and Social Research. March 2001.
URL: www.cls.dk/workingpapers/docfiles/87.pdf Accessed: 22/07/2008 06:52
Casely, C. (2006), A Season of Strikes in Ghana. A Rapid Analysis
http://www.ghanaweb.com/GhanaHomePage/features/artikel.php?ID=105855
Crisp, J. (1979), Union Atrophy and Worker Revolt: Labour Protest at Tarkwa
Goldfields, Ghana, 1968-1969, Canadian Journal of African Studies / Revue
Canadienne des tudes Africaines, Vol. 13, No. 1/2, pp. 265+267293 Stable URL:
http://www.jstor.org/stable/484647 Accessed: 22/07/2008 04:52
------------ (1984), The Story of an African Working Class: Ghanaian Miners' Struggles,
18701980, London: Zed Books.
Dovlo & Nyonator (2003) Migration by Graduates of the University of Ghana Medical
School: A Preliminary Rapid Appraisal. URL:
http://www.who.int/hrh/en/HRDJ_3_1_03.pdf Accessed: 22/07/2008 09:52
Drake and Lacy (1966), "Government versus the Unions: The Sekondi-Takoradi Strike,
1961," in Politics in Africa: Seven Cases, ed. Gwendolen M. Carter (New York:
Harcourt, Brace & World, 1966.
Freund B. (1988) The African Worker (Cambridge: Cambridge University Press, 1988.
Hansen E. (1987) The State and Popular Struggles in Ghana, 1982-1986, in Popular
Struggles for Democracy in Africa, ed. Peter Anyang Nyongo (London; Zed Books,
1987)

70

Herbst, J (1993). The Politics of Reform in Ghana, 1982-1991. Berkeley: University of


California Press.
http://ark.cdlib.org/ark:/13030/ft2199n7n7/
ICFTU (2001) Internationally-Recognised Core Labour Standards in Ghana: Report for
the WTO General Council Review of Trade Policies of Ghana (Geneva, 26 And 28
February 2001) International Confederation of Free Trade Unions (ICFTU)
ILO (1983) Collective Labour Agreement, Strike and
ilo.org/public/english/region/eurpro/ankara/legislation/act2822.htm

Lock-Out

Act

Jeffries R, (1978) Class Power and Ideology in Ghana: The Railwaymen of Sekondi,
Cambridge: Cambridge University Press.
------------ (1975) "The Labour Aristocracy? Ghana Case Study," Review of African
Political Economy, no. 3 (MayOctober): 6869.
Konings P (1978) Political Consciousness and Political Action of Industrial Workers in
Ghana: A Case Study of Valco Workers at Tema. URL:
Kraus J, (1979) "Strikes and Labour Power in Ghana," Development and Change 10,
no. 2 (April ): 281. URL:
Labour ACT (2003) Assembly Press, Accra, Ghana.
Rimmer D (1961) The New Industrial Relations in Ghana .Industrial and Labor
Relations Review, Vol. 14, No. 2, (Jan., 1961), pp. 206-226 Published by: Cornell
University, School of Industrial & Labor Relations Stable URL:
http://www.jstor.org/stable/2520269 Accessed: 22/07/2008 04:59
Sandbrook, R (1977)"The Political Potential of African Urban Workers," Canadian
Journal of African Studies 11, no. 3: 42526.
Trachtman L. (1962) The Labor Movement of Ghana: A Study in Political Unionism.
Economic Development and Cultural Change, Vol. 10, No. 2, Part 1, pp. 183-200
Published
by:
The
University
of
Chicago
Press
Stable
URL:
http://www.jstor.org/stable/1151910 Accessed: 22/07/2008 05:09
U.S. Department of Labor (1989) Foreign Labor Trends: Ghana (Washington, D.C.:
U.S. Department of Labor.

71

The structure and dynamics of the workers protest


movement at the beginning of the 20th century in Russia.
Database analysis1
Leonid I. Borodkin, Irina M. Pushkareva and Irina V. Shilnikova
1. Introduction
Born at the dawn of the labour movement, the strike has so far been one of the most
efficient forms of workers protest. This paper analyses the dynamics and structure of
the strike movement in Russia from 1895 to 1904; the decade preceding the First
Russian Revolution of 1905 that substantially determined the countrys history during
the 20th century. Recent years have witnessed little attention being paid to the study of
strikes in pre-revolutionary Russia. Nevertheless, many more opportunities to study the
subject have emerged. These opportunities are the result of new techniques and
technologies for collecting, studying and storing information.
In the second half of the 1890s, the strike movement was the basic form of
working-class protest in Russia. Despite growing numbers, the strike movement as a
whole remained small. Each year strikes involved only about one per cent of all
industrial enterprises. The working-class movement during this period already covered
many of the countrys industrial centres, but protest actions were isolated as a rule.
Strikes were mainly economic in character. Political actions, short of illegal
mayovkas (rallies), were individual and mostly consisted of strikes in solidarity with
exiled social democrats and active workers. Participants in political strikes made up
annually no more than five per cent of the total number of strikers.
The landmark was the year of 1895, which was marked by a sharp increase in
the number of labour conflicts recorded in official statistics. The number of strikes in
1895, as compared to 1894, increased five-fold and the number of strikers increased by
a factor of two and a half (Istoriia Rabochego, 1989: 417, 487). At the same time,
political organisations switched from pure propaganda to active agitation among
workers. In 1900, due to the economic crisis, the number of workers protest actions and
the number of strikers were notably low compared to the previous year. In 1901 the
number of strikes and their participants rose again and in 1903 both indicators reached
their maximum for the entire pre-revolutionary decade (1894 to 1904). Besides the
expansion, the movements character also changed. Workers motives moved from
purely economic to political aims. This transition was caused by a number of factors,
including an aggravation of the social and economic contradictions in the country, a
growth of discontent among the proletariat and the expansion of the activities of
revolutionary organisations within enterprises (Kiryanov, 1987: 63).
The protest movement of Russian workers in the period from 1895 to 1904
attracted steadfast attention from various political groups, both within the country and
abroad, including representatives of the imperial bureaucracy. Though the events of
1905 "have as though covered" the preceding era, contemporary authors constantly
1

The research is supported by the Russian Foundation for Humanities (RFH). Project No. 09-01-12119.

72

came back to it both in legal and proscribed literature, trying to clarify the course of the
working-class movement in the country. V.I. Lenin in his works (Lenin, 1910; 19111912; 1914; 1917) repeatedly addressed the experience of the workers struggle during
the pre-revolutionary period. He especially underlined the importance of finding out the
influence of various social, economic and political factors (economic conjuncture,
government policy, war, etc.) on the working-class movement. Lenin also advocated the
study of preconditions, ways and methods, or the "mechanism" of the transformation of
the country on the eve of 1905.
K.A. Pazhitnov, who studied workers living conditions in pre-revolutionary Russia,
also considered the period from 1895 to 1904 to be a separate stage in the development of
the working-class movement, which prepared the revolutionary explosion of 1905. He
explained the appreciable shifts in the working-class movement that occurred in this decade,
by the growing number of not only factory workers but also other groups of workers, the
concentration of a labour force, the increase in continuity of factory work and socialdemocratic activity (Pazhitnov, 1906; Pazhitnov K.A., 1908).
The problem of periodisation of the working-class movement has also been
repeatedly discussed by Soviet historians. They considered the period before the 1905
Revolution as a rather separate stage, both in a wide chronological framework (1895 to
1904) and in a narrow one (1901 to 1904).2
We also concentrate our study on the decade which preceded the First Russian
Revolution, because during the years from 1901 to 1904 the strengthening of political
components in the working-class movement had only just begun, but economic
requirements prevailed. The subsequent period (from 1905 onwards) is quite different:
it includes the First Russian Revolution and the emergence of political parties in the
Duma. Our purpose is to analyse what workers themselves wanted and this study is best
undertaken with regard to a period when strikes by workers were aimed mainly at
improving their living conditions. Political slogans which appeared on the eve of 1905
were mainly the result of the activities of revolutionary organisations and do not provide
an adequate insight into what the workers themselves wanted. Because of the purely
economic character of the strike movement, this period allows us to analyse parameters
of the labour movement during positive and negative phases of the business cycle.
A few further remarks must be made about the strike movement. During the
period under consideration, trade unions did not exist, so strikes were quite frequently
ended due to the intervention of army and/or police units. The Factory Inspectorate
(initiated in the 1880s) played a significant role in the period from 1895 to 1904 during
workers protest actions. During the subsequent period their role decreased, partly due
to the growth of the strike movement.
Looking at the dynamics of the pre-revolutionary labour movement from the
viewpoint of resource mobilisation theory, we can consider collective actions in terms
of a groups capacity to mobilise resources and organise (Franzosi, 1995: 10). On the
basis of this approach, Shorter and Tilly analysed the long-term patterns of strikes in
terms of the increased organisational capacities of workers in France (Shorter and Tilly,
1971). As they concluded, the greater organisational capacities of workers have deeply
transformed the nature of strikes, increasing the number of strikers and frequency of
strikes and reducing their duration on average. Of course, they also accepted economic
and structural factors as being important. One of the specific features of Russian strikes
2

See, for example: Kiryanov, Yu. I. 1987, Perekhod k massovoi politicheskoi borbe (rabochii klass
nakanune Pervoi rossiiskoi revolutsii), Moscow.

73

in our research period was the active role of illegal radical parties in the mobilisation
processes. Counter mobilisation was reflected in the rigid behaviour of authorities, who
tended to use army and police units to suppress strikers and to arrest and punish
organisers and agitators.
Our research is based on the new unique publication series called Labour
Movement in Russia. 1895 - February of 1917. Chronicle (hereafter referred to as the
Chronicle).3 The Chronicle was the first publication to include reliable informative and
homogeneous micro data on thousands of labour conflicts in the territory of the Russian
Empire at the turn of the 19th to 20th centuries. Being a secondary source based on
archival data, the Chronicle includes a substantial amount of statistical and narrative
information. This has helped us to create a database absorbing primary data about each
conflict, and to carry out a multi-aspect analysis of the strike movement in prerevolutionary Russia for the period from 1895 to 1904.
Thanks to their micro nature, the Chronicle and the database offer opportunities
to analyse various aspects of the strike movement in the Russian empire during the
stated period. This analysis allows us to provide new characteristics of the strike
movement in Russia (and of the workers protest movement as a whole). Previous
researchers have not been able to discern these characteristics, because they had only
aggregated data at their disposal.
To show what can be studied using the new data from the Chronicle, we will
analyse four aspects of the Russian strike movement during the decade commencing in
1895. First, we present the dynamics of various types (forms) of labour conflicts. We
now have the opportunity to estimate the share and role of strikes in the general
progress of the workers movement. Second, monthly strike dynamics will be
considered, taking into account general strikes and strikes at individual companies.
Traditional historiography does not contain such comparative analysis.
Third, we will consider the actions of local authorities and factory management
during strikes, since they had the most direct impact on the courses and outcomes of
conflicts. Previously, this issue was considered on the basis of case studies; analysing
sources concerning specific, large strikes, which did not allow the detection of some

. 1895 - 1917 .. . I "1895 ". : "


" 1992.- 175 .; . II. "1896 ". .: - "", 1993. 247 .;
.III. "1897 ". .: - "", 1995. -352 .; . IV. "1898 ". .: - 1997, 352
.; .V. "1899 ". .: . 1998, 391 .; . VI."1900 ". .:
. 1999, 411 .; . VII. "1901 ". .: - "". 2000,
605 .; . VIII. "1902 " 2- . .: 2002, 407 .; . IX.
"1903 " 4- . .: . 2005, 1303 .; . "1904 " 3-
. .: . 2008, 811 .
The working-class movement in Russia. 1895 - February, 1917. The Chronicle. Issue I. "1895". Moscow:
"The first printing house" 1992, p, 175; Issue II. "1896". SPb.: publishing house "Blitz", 1993, p. 247;
Issue III. "1897". SPb.: publishing house "Blitz", 1995, p. 352; Issue IV. "1898". SPb.: publishing house
"Blitz" 1997, p. 352; Issue V. "1899". Moscow: Publishing Center IRI of the Russian Academy of
Sciences. 1998, p. 391; Issue VI. "1900". Moscow: Publishing Center IRI of the Russian Academy of
Sciences. 1999, p. 411; Issue VII. "1901". SPb.: publishing house "Blitz". 2000, p. 605; Issue VIII.
"1902" in two vol. Moscow: Publishing Center IRI of the Russian Academy of Sciences 2002, p. 407;
Issue IX. "1903" in four vol. Moscow: Publishing Center IRI of the Russian Academy of Sciences
2005, p. 1303; Issue "1904" in three vol. Moscow: Publishing Center IRI of the Russian Academy of
Sciences. 2008, p. 811.

74

commonalities in the actions of authorities and the management of enterprises


throughout the whole decade.
Fourth, we will analyse for the first time some characteristics of the strike
movement at different phases of the business cycle, focusing on the questions: 1) what
demands prevailed; 2) what was the ratio of won and lost strikes; 3) what were the
outcomes of strikes in various industries; and 4) what factors influenced the outcome of
strikes.
After this introduction, the first part of the paper outlines the Chronicle: the main
source of our research. The second part contains a description of the database developed
on the basis of the Chronicle and gives the results of the analysis of different types
(forms) of labour conflicts in the period under investigation. In the third part we analyse
the monthly dynamics of strikes by industry sector and the actions of local authorities,
management and the Factory Inspectorate in the course of strikes. The next part gives
the results of the analysis of the main demands of workers in labour conflicts and the
outcomes of these conflicts during the decade preceding the First Russian Revolution of
1905. In the concluding part we give an overview of the preceding sections.

2. The Chronicle: a unique dataset on workers protests


Labour Movement in Russia. 1895 February of 1917. Chronicle has no counterparts
in Russian historical science and is a new mass source on the history of the labour
movement. In 2008 the last of ten volumes (16 books) were published, covering the
period from 1895 to 1904. The Chronicle provides information about both economic
labour disputes and labour protests with a political character.
The creation of the Chronicle started in the 1980s and was part of an
international project initiated by the Feltrinelli Foundation (Milan), the Library of
Contemporary Documentation (Paris) and the International Institute of Social History
(Amsterdam). The project aimed at comparing data on labour movements in the USA
and the industrialised countries of Europe from the end of the 19th century to 1920. One
of the objectives was to create a corresponding databank. The ten volumes mentioned
above were published under the aegis of the History Department of the Russian
Academy of Sciences (RAN), the RAN Institute of Russian History, the Federal
Archival Agency of Russia, the State Archive of the Russian Federation (SA RF) and
the Russian State History Archive (RSHA), all of which carried out methodological and
organisational management of data collection and its preparation for publication. The
material was collected by dozens of research institutes and higher education institutions,
central and regional archives and museums, and the leading libraries of the former
USSR in Moscow, Leningrad, the Ukraine, Byelorussia, Baltic countries, the Caucasus,
Middle Asia, Kazakhstan, Moldavia, regions of Central Russia, Ural, the Volga Region,
the Northern Caucasus and Don, the Russian North, Siberia and the Far East.
Information about the labour movement in pre-revolutionary Russia has been
deposited in many and varied sources. The main sources are material from the Factory
Inspectorate of the Ministry of Finance, the Trade and Industry Department, the Ministry of
the Interior and other institutions, the press (mainly illegal) and from a later period, the
recollections of participants in strikes. The data collected by the Factory Inspectorate was
published at the beginning of the 20th century (Codes, 1903, 1904, 1906, 1907; Varzar,
1905) and includes information about industrial enterprises (such as the number of workers,

75

power of steam engines, cost of production, etc.) and data on the annual number of strikes
and strikers. This data was the basic source used by Vladimir I. Lenin when analysing the
dynamics and structure of strikes in Russia (see Appendix, Table 11).
Many documents and much material have been published both in journals and
books since the 1920s. In the 1950s, four large volumes of documents called Labour
Movement in 19th century Russia were published (Pankratovoi, 1951-1963). Without
belittling the importance of this work one must state that it was framed ideologically, as
were many others of the period. One further drawback of these publications was the choice
of the documents, reflecting primarily the largest protests and strikes. This could not help
but influence the appraisal of the character of the entire labour protest movement. This
drawback was even mentioned by critics in the 1950s (Chumalova, 1957: 129).
Compared to the Code of the Factory Inspectorate (hereafter called the Code),
the Chronicle gives researchers the opportunity to work with micro data. It has the most
complete data on strikes for each year in all industries. It also provides twice as much
data on strikes and strikers than the Code. For some years this ratio is even higher.
Data from the Factory Inspectorate compiled in the Code covered only enterprises
subordinated to the inspectorate and located in European Russia. Thus, enterprises located
in Siberia and Central Asia, together with some districts of Caucasus, were excluded. It
should also be noted that only industrial enterprises that employed 15 workers or more, or
used steam-powered machines were subject to supervision by the Factory Inspectorate.
Therefore, data about smaller enterprises (for example, small craft workshops) was not
compiled by inspectors. Further, the Code does not contain data on mining or on state
enterprises which were subordinated to the Ministry of Agriculture and State Property, the
Ministry of Railways, the Ministry of Defence or the Ministry of the Navy. This explains
the substantial difference in the total number of strikes recorded in the Chronicle and in the
Code (see Table 1 and Appendix, Table 11).
The Chronicle consists of three parts: Mass Labour Movement, Party and Labour
Organisations and Leaflets. This material, covering the period from 1895 to 1904, presents
documents from 466 repositories of 112 archives that are located in the territories of the
former USSR republics. The richest are the archival repositories of the Higher State
Institutions of tsarist Russia, located in Moscow and Saint-Petersburg. Much work has been
done by libraries to collect data on labour disputes, the protest labour movement and the
activities of labour and party organisations that were published in the press. As a result, 120
newspaper and magazine titles were used, including foreign ones published by Russian
political parties, as well as a great number of local publications with small circulations.
Among periodicals, the most important are those by labour organisations such as the
Worker from 1896 to 1899 and the Workers List from 1896 to 1899, as well as foreign
publications such as the Social-Democrats Iskra, the Socialist Revolutionaries
Revolutionary Russia, and the Bunds Poslednie Izvestiya.
Events and facts are arranged chronologically within the parts of the Chronicle.
The only exception is the data on general city strikes (where subsections with their own
chronology are presented) and data on the activity of certain large party and labour
organisations, which have subsections covering the work of groups and circles within
these organisations.
The compilers of the Chronicle followed the concept that, at the time, the
Russian working class comprised those wage labourers who, making surplus value,
were occupied in material production and who were the main breadwinners. This means
factory and plant workers, miners, mining plant workers, transportation workers (blue-

76

collar workers employed in the railway, sea, river and city transport sectors), people
employed in small urban industry and agriculture, felling, construction workers,
agricultural labourers and unskilled workers. Events in which middle-class labourers
dominated were not taken into account. These included professionals (medical doctors,
teachers, etc.), white-collar workers (civil servants, clerks, railwaymen, post and
telegraph employees, etc.) and domestic servants. The degree of accuracy of many
sources made it difficult to differentiate clearly between different types of industrial
enterprises and whether to include some workers in the factory and plant category or
into other categories. Therefore, the compilers accepted the principles adopted in 1895
by the Factory Inspectorate. Under these principles, a factory or a plant was an
enterprise that had 15 workers or more, or an enterprise with a steam (or equivalent)
engine even if the number of employees was less than 15. Other enterprises were
grouped as pre-factory (such as workshops, artisan undertakings and similar).
The initial point for the approach to collecting data on the strike is its definition
as one of the basic forms of proletarian struggle which, in contrast to unrest, dispute,
etc., was characterised by the cessation of work to assert political or economic demands,
or as a sign of solidarity in the course of a protest movement. When data was being
collected, each particular protest was detailed on a record card, which initially included
16 main points. Among these were the form of protest (strike, unrest, etc.), the start and
end dates, the place where the event occurred (province (guberniya), district (uezd), city
or settlement), the name of the enterprise, the number of participants, the cause (or
occasion), the demands and the protest result.
The compilers of the Chronicle critically analysed the origin and content of
sources and checked the reliability of the data. They gave more precise information
about the time and place of events, revealed inaccuracies in documents, etc. When
gathering factual material, researchers took into account such factors as the causes and
goals of actions and an opportunity having been taken to conceal or misrepresent
information, quite often related to differing social origins of sources, political or party
struggle, etc. Such factors as the ways and methods of getting information out of a
source and the means of data acquisition were also considered. Primarily, information
from institutions whose documents and materials were the most reliable was chosen. All
the variant readings are provided as a note with each article.
Taking all these factors into consideration, the Chronicle is the most complete
and reliable source for the study of the strike labour movement in the Russian Empire at
the end of the 19th and beginning of the 20th centuries.
3. A first analysis of the database
The way each conflict is described in the Chronicle determined the data model
presented in the database, together with its structure. To create the database, a sourceoriented approach was adopted, which makes it possible for it to be used for the study of
various aspects of a strike. As the best way to present the information from the
Chronicle, we used a relational data model realised in MS Access. To optimally present
the data of each recorded labour conflict, we built 15 tables within the database
structure. The table Main has the key field (conflict number) which helps to link records
in different tables. Main includes 7886 records: one for each conflict. To formalise the

77

data from the Chronicle, a special system of codes for a number of tables has been
developed (see the codebook at https://collab.iisg.nl/web/labourconflicts/codebook).
List of database tables
1. Main
2. Industry sector
3. Number of labour conflicts
4. Type of conflicts in the course of a strike
5. Repeatability of conflict
6. Workers' occupations
7. Cause of conflict
8. Demands
9. Attendant (alternative) demands
10. Actions of factory management (or owner) and authorities
11. Clashes with police/army units
12. Actions of workers
13. Workers ethnicity/nationality
14. Agitation
15. References to archives and literature
Let us now look at the structure and the dynamics of the labour conflicts. In this
section, we differentiate between strikes and general strikes, next to other forms of
labour conflict (unrest) and overt protests (of which we have chosen gatherings,
meetings and demonstrations). The difference between strikes and general strikes
is a matter of scale. A strike is confined to one particular enterprise, while a
general strike embraces a number of enterprises in the same area at the same time.
Figure 1. Dynamics of different forms of labour conflicts (number of cases).
1895 to 1904.

1200
1000
800

1
2

600

3
4

400
200
0
1895

1896

1897

1898

1899

1900

1901

1902

1903

1904

Legend: 1 strikes, 2 unrest, 3 general strikes, 4 demonstrations and meetings

78

Figure 1 clearly shows that the smallest number of strikes was recorded during the
period from 1895 to 1896 (under 300 each year) at the initial stage of the labour
movement in Russia, and in 1904, when after an increase in 1903 nearly all the
labour protests fell in number except for meetings and demonstrations. The latter
fact is mainly due to the mobilisation of workers when Russia entered the war
with Japan in 1904. This paper is virtually the first to show that the number of
demonstrations and meetings started to grow in 1899, increased to 200 in 1903 (a
year of general strikes and mass increase in the labour movement) and continued
to rise in 1904. The objectives of a small victorious war unleashed by the
government to stop the mass movement in the country had not been achieved by
the end of 1904.
The number of demonstrations and meetings as a percentage of the total
number of labour conflicts increased over the whole decade preceding the First
Russian revolution: from almost zero in 1895 to 24.8 per cent in 1904. This indicator
shows that though strikes remained predominant in the protest movement, by 1904
the proportion of strikes, unrest, complaints and demands had decreased, while
general strikes, gatherings and demonstrations had increased in number.
Table 1. Number and percentage (in brackets) of different forms of labour
conflicts and protests from 1895 to 1904.
Year

Strikes

Unrest

1895

264 (73.3)

24 (6.7)

12 (3.3)

1 (0.3)

59 (16.4)

360 (100)

1896

278 (67.5)

54 (13.1)

11 (2.7)

5 (1.2)

64 (15.5)

412 (100)

1897

441 (63.9)

93 (13.5)

64 (9.3)

10 (1.4)

82 (11.9)

690 (100)

1898

441 (65.5)

107 (15.9)

45 (6.7)

11 (1.6)

69 (10.3)

673 (100)

1899

553 (68.4)

85 (10.5)

53 (6.6)

38 (4.7)

79 (9.8)

808 (100)

1900

382 (59.2)

62 (9.6)

76 (11.8)

46 (7.1)

79 (12.3)

645 (100)

1901

506 (59.7)

110 (13.0)

90 (10.6)

65 (7.7)

77 (9.1)

848 (100)

1902

426 (54.7)

100 (12.8)

105 (13.5)

73 (9.4)

75 (9.6)

779 (100)

1903

1,014 (55.7)

125 (6.9)

308 (16.9)

190 (10.4)

184 (10.1)

1,821 (100)

1904

336 (39.5)

41 (4.8)

134 (15.8)

211 (24.8)

128 (15.1)

850 (100)

Total

4,641
(58.9)

898
(11.4)

650
(8.2)

896
(11.4)

801
(10.2)

General
strikes

Demonstra
tions and
meetings

Other
forms of
labour
conflicts

Total
number of
labour
conflicts

7,886
(100)

4. Monthly dynamics of strikes


In the previous section we separated strikes at individual enterprises from general
strikes. General strikes covered workers from two or more enterprises, with one or
several branches located in one or several nearby areas. They were almost always

79

planned in advance, instead of breaking out spontaneously. Those workers with the
greatest initiative (often connected with revolutionary groups or representatives of the
local organisations of left-wing parties) were engaged in their preparation. During
general strikes the overall aims were usually more sophisticated. The authorities
therefore considered collective strikes as more dangerous than conflicts at an individual
enterprise. They aspired to suppress them in the shortest time by using the army and
police force. Because of the differences between the two types of strike we consider
them separately, which allows a more thorough comparison.
4.1. Monthly dynamics of individual strikes and general strikes.
The number of individual strikes for each month aggregated from 1895 to 1904 is
shown in Figure 2 on the next page. The chart demonstrates the irregularity of strike
activity in the course of the year.
Figure 2. Monthly dynamics of the total number of strikes at individual
enterprises. 1895 to 1904.
1000
900
800
700
600
500
400
300
200
100

r
em
be

ec

m
be

r
ov
e

ct
ob
e
O

be
r
tem

Se
p

ug
us
t
A

Ju
ly

Ju
ne

ay
M

pr
il
A

ar
ch
M

ry
ru
a

Fe
b

Ja
n

ua
r

On the whole, the Figure 2 gives a clear view: a peak in May, a decrease in June and a
new peak in July. The chart in Figure 3 is roughly similar, demonstrating the dynamics
of general strikes. On the whole, both individual and general strikes were the least
numerous in February and March and at the end of autumn for some years. Such
dynamics are due to the fact that in April and May, many industrial enterprises renewed
their contracts with workers at the end of one operational year and the beginning of the
next. This allowed factory management to lower wage rates for certain types of work,
change penalty schemes, etc., resulting in labour protests which, since 1900, were
kindled by the May Day holidays celebrated in Russia under the influence of
revolutionary agitation on 18 April.

80

Figure 3. Monthly dynamics of the total number of general strikes. 1895 to 1904.
160
140
120
100
80
60
40
20

us
t
Se
pt
em
be
r
O
cto
be
r
N
ov
em
be
r
D
ec
em
be
r

ug
A

ly
Ju

ne
Ju

ay
M

pr
il
A

ar
ch
M

Ja
nu

ar
y

Fe
br
ua
ry

The July peak of strikes shown in the figures was most likely related to the fact that at
the time, many workers who had kept ties with agriculture left an enterprise during the
summer to take up work in the fields. It is recorded in the Books of workers
vacations (Knigi otpuska rabochikh), which are kept in the archives of enterprises.
Those who remained had to perform the duties of those who had left. This resulted in
discontent, as the remaining workers considered it a good opportunity to make demands
for pay rises etc.
4.2. Individual strikes and general strikes by industry sector
At the end of the 19th century and the beginning of the 20th, most workers were
employed in metalworking and textile manufacture. That is why we compare the strike
activity in these industries for the decade under study. Figure 4 shows the dynamics of
strike activity in these two industries from 1895 to 1904.
Both lines clearly demonstrate a similar peak in 1903. However, the number of strikes
at textile enterprises in 1903 was lower than in 1897. The latter witnessed 173 strikes,
which is the highest rate in textile manufacturing for the entire period. In metalworking
we see two peaks: 1903 and 1899. From 1899 the number of strikes in metalworking
exceeded that in the textile industry.

81

Figure 4. Number of strikes in the textile and metalworking industries. 1895 to


1904.
300
250
200
150
100
50
0
1895

1896

1897

1898

1899
Textile

1900

1901

1902

1903

1904

Metalworking

If we look at the proportion of strikes in the textile and metalworking industries out of
the total number of strikes (Figure 5), we may conclude that before 1898, i.e. at the
initial stage of the development of the Russian labour movement, strikes in the textile
industry dominated (40 per cent of all the strikes in 1897). Starting from 1899, the
proportion of strikes among metalworkers exceeded that among textile workers. More
noteworthy is the fact that from 1900 the proportion of strikes in both industries out of
the total of strikes in Russia remained fairly stable, regardless of fluctuations in the
overall strike activity.
The number of general strikes in the two industrial sectors showed a similar
profile. The only difference is that the lead of textile workers in general strikes passes
on to metalworkers in 1901 instead of 1899 as was the case for individual strikes.

82

Figure 5. The number of strikes in the textile and metalworking industries as


percentages of all strikes. 1895 to 1904.
45
40
35
per 30
cent
25
20
15
10
5
0

1895

1896

1897

1898

1899

Textile

1900

1901

1902

1903

1904

Metalworking

4.3. Actions of local authorities and factory management during strikes


Factory management and local authorities took various steps, either to calm or suppress
arising labour conflicts. Measures such as the dismissal of the active participants in a
strike or all workers in an enterprise, calling in army or police units and the invitation of
the Factory Inspectorate to regulate the situation had the greatest influences.
Factory inspectors could also intervene in a strike on their own initiative or at
the request of workers. The calling in of police and army units at places of work
occurred more often at the behest of local authorities or factory owners. The
management of large enterprises also quite often hired additional staff of police and
Cossacks at their own expense, supervising a situation at the place of work and in its
district. The arrival of armed forces sometimes sufficed to calm workers. However there
were also conflicts between workers and police and army units, usually ending in
numerous arrests.
Figure 6 shows the actions of owners and factory management as well as local
authorities in cases of strikes. We have chosen the most common variants: dismissals,
arrests, calling in the police or army, and intervention by a factory inspector. As can be
seen, six years out of the ten (1895, 1897, 1900, 1902, 1903 and 1904) show calling in
the police or army as the highest column, while for three years (1896, 1898 and 1899)
the highest is intervention by a factory inspector. Columns for arrests and intervention
by a factory inspector are nearly equal for 1901.

83

Figure 6. Actions of local authorities and factory management against strikers at


individual enterprises. 1895 to 1904.
140

number of strikes

120
100
80
60
`

40
20
0
1895
dissmissals

1896
arrests

1897

1898

1899

1900

police or troops called in

1901

1902

1903

1904

factory inspector's intervention

Source: Appendix, Table 3

Comparing the actions of owners and local authorities during individual strikes or
general strikes one thing is clear. Dismissals and intervention by a factory inspector
occurred more often during individual strikes, while arrests and the call for police or
army support were more common during general strikes. Local authorities and factory
management evidently took tougher measures during general strikes.
General strikes nearly always had not only initiators and instigators, but
organisers from amongst radical workers often connected with illegal organisations
party committees or party groups. That is why general strikes were characterised by
more arrests. Strikes at individual enterprises that started spontaneously were better
controlled by the intervention of a factory inspector.
Monthly dynamics of the total of all the actions taken by local authorities and
factory management throughout the decade show that May was the peak for all four
actions and July the peak for power methods to suppress strikes (calling in the police
or army). Both peaks correspond with the overall number of strikes, which also peaked
in May and July. This proves that when there were more strikes, the authorities tended
to resort to finding a solution by force. However, the Chronicle shows an inverse
correlation in a number of Russian provinces. Frequent suppressions of labour protests
by the use of force caused gatherings, meetings and demonstrations, as well as short
strikes against such measures taken by authorities.

84

5. Analysis of strike demands and results


In this section we compare significant strike indicators at two points in time: 1897 and
1903. Why these years out of the period from 1895 to 1904? From 1897 onwards, the
working-class movement in Russia began to grow and was at its peak in 1903. The
industrial boom in Russia also started in 1897, and ended with the economic crisis of
1903. Our comparison of two years allows us to understand in greater depth the
dependence of workers protests on a number of factors, on the eve of the revolution of
1905 to 1907.
Table 4 of the Appendix and Figure 7 show the change to the industrial structure
of the strike movement for the two years, 1897 and 1903. The greatest number of strikes
in 1897 was in the textile industry (n = 173), which is the largest number throughout the
whole ten-year period. The second highest is metalworking, with 75 strikes. This is only
43 per cent of the frequency in the textile industry (metalworkers included metallurgists
and workers from major railway workshops like Aleksandrovsk in Saint-Petersburg).
The third highest sector is the processing of animal products, with 48 strikes.
For 1903 the picture changes dramatically. The number of strikes in
metalworking increased to 259, thus outnumbering that of 1897 by a factor of over three
times. The second highest sector in 1903 was the food industry with 144 strikes (an 8.5
factor increase from 1897). The textile industry now was only the third highest, with
110 strikes (64 per cent of the number in 1897). The number of strikes increased in
mineral substances processing, woodworking and mining, but notably among workers
in the metalworking and food industries. They were not only greater quantitatively, but
their proportion also increased (see Figure 7).
Figure 7. The percentage of strikes in various industrial sectors in 1897 and 1903.
45
40

39,2

35
30

25,5

20

7,3
2,9
1,1

2,9

4,4

mineral
substances
processing
animal
products
processing

4,1 4,7

railway
construction

textile
industry

metalworking

4,14,2

woodworking

3,9

1903

10,9

10,8

mining

10

1897

14,2

food industry

15

17

4,8
2,5

mixed
production

25

In 1903 the 259 strikes at metallurgical and metalworking enterprises accounted for
25.5 per cent of the total number of strikes. The proportion of strikes in the textile

85

industry fell sharply (from 39.2 to 10.8 per cent) which means that textile workers,
having been the most active participants of the strike movement at its initial stage, gave
way to metalworkers at the beginning of the 20th century. This can be explained by the
increasing number of metalworking enterprises and associated workers, but also for
reasons of mobilisation. It is believed that on the eve of the First Russian Revolution,
metalworkers were the best-organised section of the Russian proletariat. They were the
main target for agitation and propaganda carried out by radical Social Democratic
organisations.
5.1. Main demands of striking workers and the overall results for 1897 and 1903.
Table 1 of the Appendix shows the demands made by workers during strikes, if these
are known. From the data it is clear that for 57 per cent of the strikes in 1897 and 42 per
cent of those in 1903, workers demanded a pay rise. In 1903 the proportion of strikes
with demands concerning workday duration also decreased, from 26 to 22 per cent.
These two negative developments had their logical counterpart in a considerable
increase in social demands. During strikes, workers in 1903 more often than in 1897
demanded an improvement in living conditions, better quality of food in shops, as well
as the introduction, preservation and/or increase of social payments and free medical or
educational services. A special place was held by demands for courteous treatment on
the part of white-collar workers and legitimacy of dismissals, as well as protests against
managements and white-collar workers rudeness, extortion, bribery, searches and
cheating in accounts. All these demands prove the development of the selfconsciousness of a certain section of workers.
Figure 8 shows two plots (for 1897 and 1903) representing the monthly proportion of
strikes where workers made wage-related demands (increase, preservation of previous
rates, equality between workers in different departments, etc). The plots show a similar
profile for only the last four months of the year. The monthly dynamics of this indicator
have a wide range, with 1897 showing more extreme values than 1903. In other words,
by 1903 the dynamics of the strike category under study become smoother, without the
marked peaks characteristic of 1897.

86

Figure 8. The proportion of wage-related strikes (percentage of the total


number of strikes per month) for 1897 and 1903.
35
30
25
20

1897
1903

15
10
5

Au
gu
st
Se
pt
em
be
r
Oc
to
be
r
No
ve
m
be
r
De
ce
m
be
r

Ju
ly

Ju
ne

ay
M

Ap
ril

ar
ch
M

ru
ar
y

Fe
b

Ja
nu
a

ry

The 1897 plot again shows a peak in April, because at the beginning of a new
operational year workers demanded an increase in wages or their preservation at the
same level (in the event of factory management announcing a reduction). After a fall in
May, the 1897 plot shows a rise in June, July and especially August, where we see the
second peak. A peculiar feature of the 1903 plot is a curve, rising from March to June
and going down from July for which we do not have an explanation at present. This
requires a special study of Russian industry peculiarities.
Both plots interestingly peak in October. These peaks probably relate to
workers protests against a switch to lower, winter wage-rates, as historical sources
suggest this practice was characteristic of the time. Anyhow, the correlation coefficient
representing the degree of consistency of a seasonal strike course in 1897 and 1903 is
nearly zero (r = 0.05). This proves the important role of an economic situation factor at
different phases of business activity, because the seasonal strike dynamics are not
correlated to positive and negative phases of the business cycle.
A conflict is always interesting in terms of its result. Thus, Table 2 of the
Appendix and Figure 9 show the number and percentage of strikes where demands were
completely met, partially met, or not met. There were also strikes in which demands
were not met, though the management promised this. This strike result often caused
duplicated strikes, because of the unfulfilled promises. We must keep in mind that the
outcome is unknown for a large number of strikes (33 per cent for 1897 and 41 per cent
for 1903) which puts the following results in perspective. If more outcomes were
known, the results might change considerably.

87

Figure 9. The results of strikes (percentage of the total number of strikes) for
1897 and 1903.
30
25,3

25
20
15

20,9

19,7

20,4 20,0

11,9

1897

10

1903

5,0

2,0

0
met

unmet

partially met

promised to
met

It is apparent that strikes with a negative outcome were the largest category and
their proportion in 1903 increased by 4.4 per cent compared with 1897. The proportion
of strikes with partially met demands is nearly the same. The number of strikes won by
workers showed a considerable decrease from 1897 to 1903 (of 7.8 per cent). The same
is true for strikes where the factory management promised to meet workers demands (
3%). These decreases can possibly be partly explained by the overall rise in the number
of strikes in 1903 resulting in a more determined attitude by employers not to give in to
strikers demands. Possibly it is also a result of the fact that employers at the beginning
of the 20th century had learned how to react to labour protests. Moreover, the industrial
crisis at the beginning of the 20th century gave employers the opportunity to take
tougher measures to cope with labour protests, as the threat of dismissal was a good
way in which to scare workers. One further thing is that the proportion of strikes
resulting in employers promises could have decreased because such an indefinite
answer no longer suited workers in 1903 and often could not pacify them.
Figure 10 shows the monthly dynamics for successful strikes in 1897 and 1903
throughout the two years studied. Both years show the largest percentage of strikes won
by workers falling in October (29.2 per cent in 1897 and 31 per cent in 1903). A sharp
decrease in the number of strikes where workers demands were met can be seen in July
for both years. July 1903 shows the lowest figure of all won strikes in 1903, at only 3.2
per cent. This, of course, can also be explained by the rise in the total number of strikes
in July.

88

Figure 10. The proportion of won strikes (percentage of the total number of
strikes). Monthly dynamics for 1897 and 1903.
35
30
25
20
15

1897
1903

10
5
ug
us
t
Se
pt
em
be
r
O
cto
be
r
N
ov
em
be
r
D
ec
em
de
r

Ju
ly

ne
Ju

ay
M

pr
il
A

ar
ch
M

ua
ry

Fe
br

Ja
nu

ar
y

Strikes lost by workers show a less similar pattern, sometimes moving in the opposite
direction (see Appendix, Tables 9 and 10). While won strikes per month in the two
years under study correlate positively with each other (r = 0.59), lost strikes have a
negative correlation (r = -0.52). The opposite directions are especially evident in
February and September. In September 1897, the proportion of lost strikes made up
nearly the half (47.6 per cent) while this proportion was only 12 per cent in 1903. The
differences are related to the seasonal economic situation in 1897 (a successful year)
and 1903 (a crisis year).
Figure 11 shows the percentage of strikes won by workers in different industries.
Several interesting effects can be observed here. The proportion of strikes won in textile
enterprises increased while it decreased in metalworking. This can again be explained
by the fact that the overall number of strikes in the textile industry in 1903 was smaller,
and in metalworking much larger, than in 1897. The more strikes, the smaller the
proportion of won strikes. This supposition is to a certain extent supported by the data
on the food industry. The number of strikes there showed (similar to metalworking) a
marked increase in 1903, and as with metalworking we see a decrease in the proportion
of won strikes from 23.5 to 6.9 per cent.
The reverse is true for lost strikes. There we can also observe an increase in the
proportion of lost strikes in the metalworking and food industries as the overall
incidence of strikes increased. The textile industry shows a different picture. There we
see an increase in won strikes and lost strikes at the same time. Similar dynamics can be
observed in mixed production, where some enterprises were related to the textile
industry.

89

Figure 11. Won strikes in different industries (percentage of the total number of
strikes in those industries) for 1897 and 1903.
40
35
30
25

1897

20

1903

15
10

mixed
production

mineral
substances
processing
animal
products
processing

railway
construction

woodworking

food industry

metalworking

textile
industry

mining

5.2. The outcome of labour conflicts


The outcome of labour conflicts was determined by many factors. Notably: the
participation of all the workers employed in an enterprise; calling in the police and/or
armed forces to suppress workers and disturbances; workers actions during strikes (in
particular anientisement (damage to property) and destruction of factory property);
intervention by a factory inspector and; political agitation.
We present the results of correlation analysis, based on Tables 9 and 10 of the
Appendix. Again the years 1897 and 1903 are compared. Statistics show that the role of
the factors listed in meeting workers demands had significantly changed by 1903 (at
the negative phase of the economic cycle) compared to 1897.
First we investigate the influence on the strike outcome of the participation of all
the workers employed in an enterprise. In 1897, the monthly dynamics of won strikes
and strikes where all workers participated have a negative correlation coefficient (r = 0.3). In 1903 the correlation has the same direction but with a lower value (r = -0.12).
This can be explained by the supposition that workers demands related, for example,
only to a narrow professional group of workers, but due to the specifics of production
the whole enterprise had to stop. Another explanation may be that employers will give
in more easily to smaller, and possibly more skilled, groups of workers. So the workers
chances of winning a strike were actually slightly weakened if all workers joined the
action. This conclusion is strengthened by data on strikes where mass participation had
some negative influence, because employers felt forced to call in the police or armed
forces. This conclusion can be confirmed if we use data related to any other year from
the decade under consideration. The correlation coefficient in all cases is negative
(values of r vary from -0.44 to -0.01). So the more general a strike, the more negative
will be the result from the workers point of view. This conclusion is, however, only
true at a very low level; the explained variance (r2) is never higher than 19.6.
We also calculated correlations between the won strikes and the calling in of
troops or police for the years 1897 and 1903. The results showed an inverse
dependency: 1897 r = 0.23 and 1903, r = -0.68. This means that the more frequently
army or police units were called in throughout 1903, the smaller was the chance that

90

workers won their strikes and the opposite: The less frequently troops were called in to
face strikers, the more likely were the strikes to be won. This results in some opposite
indicators for 1903. The most extreme example is for October 1903 when no army or
police assistance was recorded and the percentage of won strikes was the highest during
that year (31 per cent). In 1897 such dependency is not observed. According to the
descriptions in Chronicle, the 1897 characteristics can be explained by the fact that
police rather than army forces were called in, in particular the chief of police who could
intervene in the talks together with a factory inspector to stop a strike. This meant a
softer ending of a strike, with a possible positive outcome even though the police
intervened.
From the data it is clear that there was also an influence on the outcome of
strikes by workers aggressive actions and their fighting spirit. This included actions
such as the destruction of factories and other premises such as food shops, the breaking
of windows in apartments of those hated white-collar workers who humiliated workers
by harsh treatment, the beating of hated foremen, etc. There exists an evident correlation
between strike outcomes and workers aggressive actions in 1897 (r = 0.54). This is
easy to explain, because preserving factory property was more important to employers
than refusing quite feasible demands; usually related to reasonable pay rises,
improvements in living conditions or the dismissal of hated white-collar workers and
foremen. The years at the close of the 19th century saw a new form of protest, by taking
foremen out of the premises of an enterprise in wheelbarrows. In 1903 there is only a
small negative correlation (r = -0.16) which is probably connected to the general figures
for mass strikes in southern Russia, which dissolved some instances of workers
aggressive actions. The latter, by the way, were less numerous than in 1897, because the
negative phase of the economic cycle also played a role, as workers were afraid of
losing their jobs.
When we study the correlation between strike outcomes and agitation carried out
by members of different political parties, labour organisations and workers initiative
groups, we can conclude that in 1897 the influence of agitation was considerable (r =
0.64). The more there were strikes with agitation the more strikes were won. This
dependency is different in 1903, when the positive correlation stops at 0.35.
The final analysis we made was concerning the influence of factory inspectors
actions on strike outcomes. Looking at the monthly dynamics for 1897, we see that
intervention by a factory inspector is closely related to the chances of strikers demands
being met. In cases where the intervention of an inspector was recorded, strikes were
more frequently won (r = 0.43). This situation was different in 1903 (r = -0.12) when,
as revolutionary 1905 was nearing, intervention by factory inspectors did not have any
significant influence on strike outcomes. This was also noted by the inspectors
themselves in contemporary reports analysing the development of the labour movement.
At the time of the movements rise in 1903, factory inspectors did not have time to
follow the situation in all the controlled enterprises and their intervention in labour
conflicts was barely noticeable and therefore ineffective. Earlier sources described
many occurrences when workers themselves demanded an inspectors intervention in a
conflict with employers. However, the proportion of such cases in the strike movement
decreased due to the aggravation of the situation and increasing mass protests in this or
that territory.

91

6. Conclusion
The Chronicle is a really unique secondary source that has placed a huge file of micro
data (strike by strike) at the disposal of researchers into the labour movement in
Russia at the end of the 19th century and the beginning of the 20th. Our database
presents the data from the Chronicle in electronic format and offers good opportunities
to analyse different aspects of the strike movement in that period. The results obtained
through an analysis of such a huge number of conflicts allows us to analyse quantitative
indicators of the workers strike movement and their dependency on phases of the
business cycle, as well as the influence of different factors on the outcomes of conflicts.
Researchers have previously been unable to do this, because no micro data on a monthly
or industrial level was available. The only available strike data was the annually
aggregated totals from the reports of the Factory Inspectorate.
Let us recapitulate the main conclusions of our analysis. Throughout the decade
under study the proportion of strikes at individual enterprises steadily decreased,
whereas general strikes, demonstrations and meetings increased in number. This proves
the growth of organisation and the increase in a political component of strike dynamics.
However, at the same time strikes as a means of solving labour conflicts did not lose
their importance.
Our research has shown that in 1899 to 1900 textile workers, who were the most
initiative strikers, conceded their leadership in the strike movement to workers in other
sectors of commerce, especially the metalworking industry. We found various
explanations for this shift. First, for the earlier period, textile workers achieved greater
benefits as a result of their struggle, so were subsequently less active. Second,
metalworkers, as more competent and qualified workers, involved more propagandists
(agitators).
Strike activity fluctuated within each year. The smallest monthly numbers of
strikes in 1895 to 1904 occurred during the period from February to March and at the
end of autumn. February to March coincides with the end of one operational year and
the start of the next. To struggle for a change to employment terms during this period
was useless, since after Easter those conditions would be reconsidered and participation
in a labour conflict on the eve of the renewal of contracts could lead to dismissal. Gaps
in strike activity were influenced also by workers links with agriculture. Many of them
yesterday's peasants requested leave from enterprises in order to work at their
households in the countryside during the summer to autumn period. The decrease in the
incidence of strikes at the end of autumn (when a number of workers came back to
factories) was in many respects connected with this factor. The same concerns the
apparent peak of strike activity in July, when in the situation of a shortage of hands, the
remaining workers aspired to achieve more favourable working conditions. All this
indicates that despite a remaining element of spontaneity in the protest movement of
workers, they also took the prevailing situation into account. Workers chose the most
appropriate times to achieve the satisfaction of their aims.
As shown in our analysis, calling in army and police units was the most
widespread measure used by local authorities to suppress labour conflicts. This is
especially characteristic of the beginning of the 20th century, when the strike movement
development occurred. Under Russian legislation, participation in a strike was
considered as a crime that justified rigorous actions by authorities. Our conclusion is
interesting: that after 1900 the number of strikes with a factory inspectors intervention

92

decreases. This is mainly due to the general increase in the intensity of struggle, when
both the disputing parties employers (factory management) and workers took up
tougher measures to achieve their goals.
A wage increase was the main demand throughout the entire decade. However, it
is notable that by 1903 we see a considerable increase in demands related to the
development of the social infrastructure of enterprises.
At the beginning of the 20th century the proportion of won strikes decreased. It
also decreased when the tendency to call in the police and/or army grew. The
dependence of a positive strike outcome on agitation, a factory inspectors intervention
and workers aggressive actions was greater at the beginning of the decade than at its
end. Participation of all the workers employed in an enterprise in a strike had a negative
impact on the strike outcome (although left-wing parties activists were oriented
towards the involvement of all workers).
The observed tendencies of the strike movements structure and dynamics in the
Russian Empire at the end of the 19th century and the beginning of the 20th century
were central in determining the character of the social outburst of 1905. The database
analysis offers an opportunity to verify hypotheses formulated earlier on the basis of
case studies or on the basis of incomplete aggregated data, by correlating variables
which characterise the different aspects of labour conflicts. Our work is only the
beginning of such analyses and in the future, more sophisticated statistical techniques
may be applied. The publication of the database and our first efforts to analyse the
micro data are only the beginning of a new era in Russian strike research.

93

Appendix
Table 1. Main demands of strikers (number of cases), 1895 to 1904.
Wages
Workday
duration
Penalties
and other
deductions
Living
conditions
and

1895 1896
1897
1898 1899 1900 1901 1902 1903 1904
98
129
266
197
376
173
267
255
607
134
6

55

123

94

169

50

113

74

314

38

22

14

22

19

19

29

70

10

nourishment

21
16

12
9

19
12

31
7

16
15

21
16

26
18

36
23

111
56

22
16

management

12

19

21

21

44

35

51

134

19

10

26

25

144

Dismissals
Discontent
with
factory
Social
payments
and
services
Total

164

229

466

374

625

327

504

493

1436

246

Table 2. The results of strikes (number of cases), 1895 to 1904.


Won
Lost
Partially
met
Promised
to be met

1895 1896 1897


1898 1899 1900 1901 1902 1903 1904
44
54
87
96
91
62
94
94
121
56
57
64
92
118
158
101
115
119
257
75
29

36

90

59

49

39

94

68

203

42

10

22

17

11

20

94

Table 3. Dynamics of the actions of local authorities and factory


management against strikers (number of cases), 1895 to 1904.
Dismissals
Arrests
Police or
troops
called in
Factory
inspector's

intervention

1895
29
27

1896
31
51

1897
62
67

1898
54
58

1899
55
82

1900
44
46

1901
49
57

1902
57
43

1903
102
75

1904
22
21

40

55

79

77

95

57

54

66

111

34

21

59

68

88

116

53

57

33

51

26

1904

Table 4. Strikes in different industrial sectors (number), 1895 to 1904.


Textile
industry
Metalwork

Food
industry
Mining
Woodwork
Railway
construction

Mineral
substances
Animal
products
Mixed
production

1895

1896

96
35

59
59

11
18
6

1897

1898

1899

1900

1901

1902

1903

173
75

125
84

74
172

64
83

60
117

58
110

110
259

38
85

5
32
6

17
18
18

26
26
19

37
37
28

53
23
16

52
52
18

54
21
29

144
43
48

23
28
13

27

14

13

11

14

11

15

11

13

41

59

31

19

25

74

21

43

48

15

38

31

71

21

45

23

14

21

25

28

19

36

15

25

18

95

Table 5. Won strikes in different industries (number of cases), 1895 to 1904.


1895 1896 1897 1898 1899 1900 1901 1902 1903 1904
15
9
16
21
14
13
7
16
23
3
7
16
21
22
20
12
21
24
20
8
0
1
4
5
5
8
9
9
10
5
3
1
1
4
6
2
5
6
8
9
0
0
3
2
3
2
4
8
5
2

Textile industry
Metalworking
Food industry
Mining
Woodworking
Railway
Construction

Mineral substances

17

15

24

Animal products

16

Mixed production

11

11

Table 6. Lost strikes in different industries (number of cases), 1895 to 1904.


1895 1896 1897 1898 1899 1900 1901 1902 1903 1904
15
19
38
42
24
21
19
22
38
13
9
8
16
19
51
22
24
23
65
21
4
2
4
9
10
8
11
19
42
6
4
5
7
7
10
6
22
6
12
7
2
3
3
8
9
2
2
10
12
2

Textile industry
Metalworking
Food industry
Mining
Woodworking
Railway
construction

Mineral substances

20

12

Animal products

12

12

Mixed production

Table 7. Monthly dynamics of the total number of strikes at individual enterprises,


1895 - 1904.
Month
January
February
March
April

Number of
strikes
324
199
193
427

Month
May
June
July
August

Number of
strikes
719
453
865
331

Month
September
October
November
December

Number of
strikes
287
228
224
228

96

Table 8. Monthly dynamics of the actions of local authorities and factory


management against strikers, 1895-1904.

January
February
March
April
May
June
July
August
September
October
November
December

Dismissals

Arrests

49
32
25
40
82
78
50
32
33
25
30
20

40
28
26
60
90
80
78
27
25
24
19
23

Factory
Police or troops
inspector's
called in
intervention
53
31
30
42
101
94
127
39
33
31
43
37

65
35
26
45
118
75
58
22
32
37
31
26

Table 9. Categorisation of monthly data on strikes in Russia, 1897.


Month
January
February
March
April
May
June
July
August

September

October

November

December
Annual
average

1.
35.9
27.3
22.0
22.9
32.4
30.8
45.2
12.5
28.6
16.7
6.3
37.0

26.5

2.

3.

4.

20.5
13.6
11.1
5.7
12.2
34.6
21.9
4.2
14.3
25.0
25.0
22.2

17.5

2.6
4.5
0
0
2.7
1.9
0
4.2
4.8
8.3
0
0

2.4

5.
38.5
27.3
16.7
34.3
20.3
23.1
16.4
29.2
19.0
33.3
18.8
11.1

24.0

6.

7.

8.

17.9
18.2
0
14.3
10.8
15.4
9.6
12.5
19.0
33.3
37.5
29.6

43.6
45.5
50.0
28.6
54.1
57.7
63.0
33.3
61.9
41.7
18.8
18.5

12.8
4.5
11.1
28.6
10.8
15.4
17.8
29.2
0
16.7
6.3
11.1

25.6
18.2
16.7
22.9
10.8
23.1
13.7
20.8
23.8
29.2
18.8
18.5

18.2

43.1

13.7

20.2

9.

Legenda: 1 = % of strikes in which all workers employed in an enterprise took part, 2 = % of


strike where army/ police were called in, 3 = % of strikes with aggressive actions taken by
workers, 4 = % of strikes with elements of agitation, 5 = % of strikes with factory inspectors
intervention, 6 = % of strikes with basic wage rise demand (zarplata), 7 = % of strikes with wagerates rise demand (rastsen-ki), 8 = % of strikes with demands met completely, 9 = % of strikes
with demands unmet

23.1
13.6
16.7
22.9
14.9
21.2
19.2
16.7
47.6
25.0
25.0
22.2

22.3

97

Table 10. Categorisation of monthly data on strikes in Russia, 1903.


Month
January
February
March
April
May
June
July
August
September
October
November
December
Annual
average

1.
2.
3.
15.8
15.8
31.0
6.9
23.8
16.7
10.9
6.3
36.0
5.4
24.4
7.7
53.0
14.2
36.8
19.7
58.7
5.4
37.9
0
27.9
7.0
31.0
10.3
32.3

Legenda: See table 9

9.6

5.3
0
2.4
0
0.9
0
2.7
2.6
0
0
0
10.3
2.0

4.

26.3
34.5
14.3
15.6
13.5
10.3
15.1
17.1
5.4
24.1
18.6
17.2

17.7

5.
6.
7.
8.
9.
15.8
52.6
21.1
15.8
0
48.3
20.7
20.7
4.8
52.4
2.4
7.1
4.7
50.0
9.4
20.3
8.1
19.8
13.5
13.5
12.8
50.0
15.4
12.8
3.0
61.6
14.5
3.2
1.3
25.0
10.5
14.5
6.5
45.7
5.4
19.6
6.9
41.4
20.7
31.0
2.3
34.9
9.3
25.6
0
31.0
10.3
20.7
5.5

42.7

12.8

17.1

28.9
37.9
23.8
18.8
15.3
32.1
27.2
30.3
12.0
31.0
27.9
31.0
26.4

Table 11. Strike indicators according to the Codes of Factory Inspectorate,


1895 to 1904.
Year
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904

Number of
strikes

Source: Lenin 1913

68
118
145
215
189
125
164
123
550
68

Percentage of all
enterprises
0.4
0.6
0.7
1.1
1.0
0.7
1.0
0.7
3.2
0.4

Number of
strikers

Percentage of all
workers
31,195
29,527
59,870
43,150
57,498
29,389
32,218
36,671
86,832
24,904

2.0
1.9
4.0
2.9
3.8
1.7
1.9
2.2
5.1
1.5

98

References
Chumalova G.G. 1957, Vyiavlenie i otbor dokumentov dlya pechati v izdanii "Rabochee
dvizhenie v Rossii v XIX v, in: Trudy Moskovskogo gosudarstvennogo istoriko-arhivnogo
instituta. V.VIII. oscow
Codes of Factory Inspectorate (Svody otchetov fabrichnykh inspektorov 1903, 1904, 1906,
1907, S.-Petersburg
Franzosi, R. 1995, The puzzle of strikes: class and state strategies in postwar Italy, Cambridge
University Press
Istoriia rabochego klassa SSSR (1989): Rabochii klass Rossii ot zarozhdeniia do nachala XX v.
Izd-e 2-e, Moscow
Kiryanov, Yu. I. 1987, Perekhod k massovoi politicheskoi borbe (rabochii klass nakanune
Pervoi rossiiskoi revolutsii), Moscow
Lenin, Vladimir I. 1910, statistike stachek v Rossii (Strike statistics in Russia, Collected
works, Progress Publishers, Moscow 1974, Volume 16, pp. 393-422)
-------------------- (1911-1912), Izbiratelnaya kompaniya v IV Gosudarstvennuiu Dumu
-------------------- (1913), Strikes in Russia, in: Sputnik Rabochego for 1914, Priboi Publishers,
Saint-Petersburg (see also: www.marxists.org/archive/lenin/works/1913/dec/31.htm)
-------------------- (1914), Iz proshlogo rabochei pechati v Rossii
-------------------- (1917), Doklad o revolucii 1905 goda
Pankratovoi, A.M. red. (1951-1963), Rabochee dvizhenie v Rossii v XIX v. Sbornik materialov i
dokumentov pod, Volume I. 18001860, oscow, 1951; Volume II 18611884, oscow, 1950;
Volume III 18851894, oscow, 1952; Volume IV 1895-1900, Moscow, 1963
Pazhitnov K.A. 1906, Rabochee dvizhenie v Rossii. Saint-Petersburg
------------------ (1908), Olozhenie rabochego klassa v Rossi., Saint-Petersburg
Shorter, E. and Tilly, Ch. (1971), The shape of strikes in France 1830-1960. In: Comparative
Studies in Society and History, Vol. 13, pp. 60-86
Varzar, V.E. (1905), Statisticheskie svedeniia o stachkakh rabochikh na fabrikakh I zavodakh za
desyatiletie 1895 1904 gg., Saint-Petersburg

99

To participate or not to participate? Demographic aspects of


the first mass strike in Sundsvall, Sweden in 1879
Maria Bergman
1. Introduction: Swedens first mass strike
The economic crisis of the late 1870s resulted in fluctuating prices and export demands,
that throughout the decade, caused severe rifts between workers and employers within
the Swedish sawmill industry (Cornell, 1982: 318-319; Bjrklund, 1976: 42). This
development culminated in May of 1879 in the Sundsvall district of the central province
of Medelpad, when disputes regarding wage cuts resulted in Swedens first mass strike.
Involving twenty-two of the districts twenty-three sawmills and thousands of workers,
this strike occurred during the infancy of workers movements, before the unions had
evolved, before workers rights had been firmly established and at a time when strikes
were still regarded as illegal by employers (Bjrk and Schnell, 1979: 17).
This strike influenced numerous others during the following summer, but despite
a continuous failure to reach their objectives, the strike wave of 1879 made it clear that
the collective voice of workers in Swedish society could no longer be ignored by
employers (Olsson and Lindstrm, 1953: 73; Nilsson, 1973: 76; Bjrk and Schnell,
1979: 76-79; Cornell, 1982: 319; Johansson, 1988: 368). It also marked the beginning
of a new form of organised strikes, which became an important turning point for the
Swedish workers movement and future strikes (Bjrklund, 1976: 48).
The sawmill strike of 1879 has been featured in most of the research regarding
the sawmill industry in 19th century northern Sweden, but none of these studies on
specific sawmills and their workforces have used the strike as their focal point (Olsson,
1949; Bjrklund, 1976; Norberg, 1980; Cornell, 1982; Ostergren, 1990). Only a few
publications have been directed more towards the strike (Larsson, 1972; Lundberg,
1979; Kmpe, 1979; Rahm, 2000), although not from a demographic perspective. The
reason that research is scarce is a lack of material. There are no exact numbers of how
many workers participated in the strike, nor are there any exact numbers of how many
sawmill workers were actually employed in the district, or in what capacity. Harald
Lundberg and Tage Larsson are among the exceptions regarding the study of individual
participation, having researched the strike in relation to religious and temperance
movements. Larsson especially so, as he focused on the legal proceedings that followed
the strike, where several workers affiliated with the religious movement were accused
of having acted as instigators of the strike. The legal documents from these trials are the
closest one can get to naming individual strike participants.
The only workers known not to have participated were those employed at
Wifstavarv sawmill in the parish of Timr. Not only did they choose not to participate,
they also publicly condemned the striking workers (Olsson, 1949; Rahm, 2000). On the
centenary of the strike, Sundsvall Museum published a book about it, including drafts
from articles, telegrams, parliamentary discussions and official correspondence, and
eye-witness accounts (Bjrk and Schnell, 1979). To date, this is the most all-inclusive
source of information regarding the event.

100

The lack of documentation has previously prevented demographic research on strike


participation. However, newly discovered material offers an opportunity to connect sawmill
workers to employment, residency and strike participation at an individual level. This
material, in the form of personnel lists from 1879, recorded sections of the employed
workforces at four different sawmills. By linking these lists to available church records,
reconstructing the workers lives is made possible, and therein demographic patterns that may
explain why some chose to participate in the sawmill strike of 1879 while others did not.
The purpose of this paper is to use the workers demographic differences to
explore the possible explanations for strike participation and non-participation. What
demographic aspects increased or decreased the possibility of workers choosing to
strike? Four comparative aspects will be studied: local birth connection, age, marital
status and household size.
The article is structured to give an overview of the economic and industrial
context in section two, followed by section three with a description of the historic
background of social protests and strikes in Sweden. Section four maps out the events of
the 1879 strike and section five discusses the question of participation. The sources used
are analysed in section five. Sections six through eight provide an overview of the
relationships between strike participation and place of birth, age and marital status. A
concluding section deals with the new insights into the 1879 sawmill strike provided by
the use of micro data.
2. The economic and industrial context
Production from Swedish sawmills prior to the 19th century catered mainly for local markets
and the majority of sawmills were state owned. Strict regulations monitored the industry
and limited both production and export. There were sawmills that produced for export, but
the industrys potential for expansion was never prioritised by the state (Carlgren, 1926). It
was not until the 1840s, when the social and political climate became more liberal, that state
control over the sawmill industry began to ease. Production limits were lifted, and increased
demand from abroad and new technology opened up new possibilities for expansion
(Schn, 2007: 86-88, 112). By 1863, all state influence over the sawmill industry had been
completely nullified (Olsson and Lindstrm, 1953: 16; Nordberg, 1979: 296; Carlgren,
1926: 121-122; Wik, 1950: 82; Sderlund, 1951: 4-5).
During the second half of the 19th century, steam power transformed the sawmill
industry. The construction of the first steam-powered sawmill in Sweden (at Tunadal in the
parish of Skn in Sundsvall, in 1849) is regarded as the beginning of the new sawmill
industry. The steam-powered sawmills showed the largest growth in the industry between
1871 and 1900 (Wik, 1950: 88, 124).
Sawmills emerged all along the north-eastern coast of Sweden, and the northern mills
jointly produced about 70 per cent of the countrys timber exports between 1875 and 1910
(Rondahl, 1972: 25). The Medelpad province and the northern province of ngermanland,
which together made up the county of Vsternorrland, would come to dominate the industry,
especially the Sundsvall district, which became known as the epicentre of the Swedish
sawmill industry (Hglund, 1957: 24). Between 1849 and 1900, forty large steam-powered
sawmills were constructed in the Sundsvall district alone (Hglund, 1957). This district could
not be compared to any other area in the country, or in northern Europe, when it came to the
number of sawmills within such a geographically limited area.

101

Figure 1: The number of steam powered sawmills constructed in the Sundsvall


district 1840-1900
14

No. of sawmills

12
10
8
6
4
2
0
1840-49

1850-59

1860-69

1870-79
Year

Source: Hglund 1957.

Map 1: The location of the Sundsvall district

Source: Demographic Data Base, Ume University.

1880-89

1890-99

102

Map 2: Geographical location of the steam powered sawmills in the Sundsvall


district 1871-1900

Source: Wik 1950, p. 234.

Vsternorrland County had several geographical advantages. Rivers and streams


connect directly to the Baltic Sea and provided good transport routes from inland
logging sites, and bays provided the mill sites with natural barriers against the more
rough conditions at sea. The geographical location of the district, along the central
coastline, also facilitated good communications with the southern parts of the country
(Wik, 1950: 128-129; Layton, 1979: 230, 239; Berglund-Lake, 2001: 86).
Increased demands for Swedish timber during the 1860s and 1870s enabled the
building of several larger sawmill complexes and intensified labour recruitment in the
district. Demand and prices peaked between 1869 and 1874, but the main trend of the
latter half of the 1870s was declining prices and demand, together with increased
embargos and tariffs (Rondahl, 1972: 23). This caused the sawmill owners in the
Sundsvall district to grow increasingly concerned about an impending economic crisis.

103

To diminish any possible financial losses they began to decrease production and
severely limited logging during the winter of 1877 to 1878 (Bjrk and Schnell, 1979:
28). Timber prices continued to fall, and as a result wages were cut and an increasing
number of workers were left unemployed (Olsson, 1949: 97). During the early spring of
1879, wages were cut once again (Bjrklund, 1976: 44). This had devastating effects on
the workers at the bottom of the wage earning scale. Since work was always scarce
during wintertime, employment during the summer and reasonable wages were
imperative for workers and their families to survive (Bjrklund, 1976: 45-46). Olsson
estimated that wages were cut by between 15 and 20 per cent, something that would
have directly affected around 7,000 workers in the district (Olsson, 1949: 98).
Bjrklund, however, assessed that in total, wages were cut by between 20 and 25 per
cent between 1878 and 1879 (Bjrklund, 1976: 44).
To help the sawmill industry get back on its feet the Swedish government
approved a loan of three million Swedish crowns. Having heard the rumours, many
workers believed this would restore their wages. They were to be gravely disappointed.
Instead, they witnessed the sawmill owners celebrating their good fortune with a large
banquet (Olsson, 1949: 97, 112; Bjrk and Schnell, 1979: 39). A few weeks later, the
sawmill workers in the Sundsvall district went on strike.
3. The history of social protests and strikes
Swedish workers have a long tradition of social and work-related protests dating back to the
late 18th century (Karlbom, 1967). Food riots were the most common type of protest during
the first half of the 19th century, while work-related protests occurred less frequently and on
a smaller scale. Rolf Karlbom described these protests as occurring spontaneously and
rarely with any specific, collective long-term goals in mind (Karlbom, 1967: 10).
By 1846, the labour market had changed. The state had abolished the guilds
(institutions which had monopolised and controlled crafts and their labour demands from
late medieval times) and had discontinued regulations hindering the free movement of
workers (Schn, 2007: 86-88, 112). The reform of 1846, however, contradicted other laws
concerning workers rights. Workers were, according to law, subordinated to the will of
their employer and workers without proof of employment were forced to find work or risk
being drafted into the army or sent into forced labour (Bjrk and Schnell, 1979: 16;
Bjrklund, 1976: 47; Olsson and Lindstrm, 1953: 64-65). The patriarchal context of the
sawmill communities in many respects also preserved the relationship between employers
and workers as one of master and servant (Cornell, 1982: 131; Tedebrand, 1999: 207). No
collective agreements regulated wages and written employment agreements were rare.
Employers were seldom willing to pay more than the absolute minimum and workers could
do nothing but accept the situation, or try finding employment elsewhere (Bjrklund, 1976:
45). Breaches of employment agreements, such as drunkenness or disobedience, resulted in
forfeiting earned wages and unemployment (Sgverksfolket, 1995: 88-89).
Work-related protests grew in size and frequency after 1855, as workers became
more aware of their working and living conditions, enabled by the emergence of larger
industrial communities (Karlbom, 1967). The communities brought several hundred of
workers and their families together, under similar conditions and within defined
geographical areas. Residency and the social environment were important tools in creating a
more homogenous workforce with similar goals and beliefs (Bergman, 2010: 255). The

104

importance of the communities were highlighted in times of strike according to Laura


Frader, as the collective support that arose clearly showed how closely bound the
communities were (Frader, 1981: 197-198).
The last legal restrictions against striking were officially abolished in 1864, but
employers continued to deny the workers the legal right to strike because it undermined
their authority (Lundh, 2006: 23). Strikes were regarded as unjust and unlawful, and
striking workers were seen as nothing more than thugs who displayed a complete disregard
for society and the social hierarchy (Bjrklund, 1976: 22-23).
4. The sawmill strike of 1879
The sawmill strike of 1879 lasted for ten days. It involved the largest number of
simultaneously striking workers ever seen in contemporary Sweden and caused twenty-two
of the districts twenty-three steam-powered sawmills to temporarily cease production. As
no official records exist regarding the exact number of employees or striking workers,
estimates vary as to how many participated. Contemporary newspapers reported that 3,000
men were headed for the strikers camp on 28 May 1879 (Olsson, 1949: 99). Bjrklund and
Cornell estimated 5-6,000 workers joining the strike during 28 May and 29 May
(Bjrklund, 1976: 47; Cornell, 1982: 319). On Friday 30 May when the County Governor
of Vsternorrlad, Curry Treffenberg, spoke in the strikers camp, 6-7,000 workers were said
to have been present (Bjrk and Schnell, 1979: 57). Olsson claimed that there were only
1,200 workers in the camp on 31 May (Olsson, 1949: 105). Henrik Rahm, however,
estimated that a total of approximately 5,000 workers had participated throughout the entire
strike (Rahm, 2000: 9).
The first signs of an impending strike occurred on Monday 26 May 1879, when
workers at Heffners sawmill asked for their wages to be increased and this request was
refused. Even though work at the mill temporarily stopped, threats of unemployment and
eviction from company housing made the workers return to work. Displeasure among
workers was not limited to one mill and on the following day, workers from neighbouring
sawmills arrived at Heffners and announced that a district-wide strike had broken out (Bjrk
and Schnell, 1979: 41-42). Together, the workers marched from mill to mill to gather more
workers and they set up a camp just south of Sundsvall town.
The County Governor of Vsternorrland, Curry Treffenberg, was called in to
mediate and he arrived at Sundsvall on the following day, 27 May, accompanied by a
company of soldiers. Over the subsequent days, Treffenberg and the districts sawmill
owners met at Sundsvall City Hall, hoping to devise a strategy for how to make the workers
see reason. Treffenberg even ventured out among the workers. However, he was
unsuccessful in persuading the workers to return to work, as they were holding out for their
demand of a return to the wages of 1878 (Bjrk and Schnell, 1979: 45, 64; Olsson, 1949:
98-99, 103). On Saturday 31 May, the workers were told that the sawmill owners would
start hiring blacklegs and on Monday 2 June, Treffenberg announced that the workers were
to be evicted from the strikers camp, as they were occupying the land illegally. By that
time, more military had arrived in the district and by Monday night Treffenberg
commanded six gunboats anchored in the bay and over a thousand soldiers, half of whom
surrounded the strikers camp.
On Tuesday 3 June, more threats were lodged against the workers: loss of
employment and eviction from company housing unless they returned to work. Many

105

workers succumbed to the demands. The escalating threats and the increased military
presence marked the beginning of the end of the strike. During the last days in the camp,
Treffenberg and the military began interrogating those workers that had remained in the hope
of uncovering the assumed strike leaders. On Thursday 5 June 1879, Treffenberg telegraphed
his superiors in Stockholm to state that the camp had been disbanded and that 36 men had
been arrested. The strike was over and the workers had lost (Bjrk and Schnell, 1979: 61-70).
5. The question of participation
A distinction is usually made between protests in pre-modern and modern times, where the
former is described as reactive and spontaneous, and the latter as proactive (Dill and
Aminzade, 2007: 305). In his article about conflicts in Lancashire between 1750 and 1893,
Frank Munger viewed collective actions, such as demonstrations, riots and strikes, as
contentious gatherings. These gatherings, he claimed, were the result of two processes. The
first process included individuals brought together over conflicts of control relating to
matters such as food, wages and taxes; the second process concerned which particular
participants became mobilised at a particular time and place. A contentious gathering
always consisted of a group of twenty individuals or more who were predisposed to
argument or conflict, where consciousness of the situation was well-established and
Munger implied that the gatherings, the protests, could be viewed both as the means to a
higher gain, as well as goals in themselves (Munger, 1981). Birgitta Skarin-Frykman
claimed that workers behaviour during the early years of industrialism was structured from
a rational principle. This influenced both actions and values. The workers action-plan was
therefore short sighted (Skarin-Frykman, 1990: 282).
Contentious gatherings could, through the aspect of a collective psychological state,
be linked to social identity theory. They both rely upon the conflicting nature of society and
the division of social categories to produce opposites, enabling conflict (Hogg and Abrams,
1988: 3, 14-15). Glimpses of contentious gatherings and of a growing solidarity could be
found among the participants during the first recorded protests in Sweden. Those gatherings
were, as Munger pointed out, not so much a way of reaching more long-term improvements
as they were a way of righting a wrong in the here and now.
It has always been difficult to explain the reasons behind human actions throughout
history. Finding collective motives has therefore been imperative in understanding
collective actions in historic times (French and Barry, 2004: 1). People who engage in a
collective action usually share a common goal and identity. One goal of such protests is to
achieve a greater proportion of the social product (Klandermans, 1997: 2). Despite the
existence of collective goals, motives concerning participation in social protests remain
highly individual (Klandermans, 2002).
The strike of 1879 offers a clear motive: those involved sought higher wages because
of the need to be able to support their families (Kmpe, 1979: 70). Still, the question of
participation remains complex, and clearly individual, when reviewing eye-witness accounts.
Many were of course angry, but the workers, as a rule, did not dare to complain. Many put
their faith in God, and suffered patiently, others slaved like animals without thinking about
anything. Some believed in the possibility of extorting better conditions, but they were too few
to actually do something.
Anders Mattson-gren
(Authors translation. Bjrk and Schnell, 1979: 224)

106

I was fifteen years old at the time of the strike. I had been working at the Strands sawmill in
Aln parish since the age of seven and I had good knowledge of the living conditions and
understood very well the reasons behind the strike [] they had the best intentions and just
wanted to receive fair pay for their labour so they could support their families. Curry
Treffenberg called that an uprising.
Johan Ferdinand Jacobsson
(Authors translation. Bjrk and Schnell, 1979: 226-227)

I was young and green at the time and did not really understand what was happening. We had
heard the rumours of a strike at Heffners and that the workers were on their way to us at
Sknvik. They came in hundreds and since we did not mind joining, the crowd doubled.
Erik Nilsson
(Authors translation. Bjrk and Schnell, 1979: 224).

Some individuals participated in social protests out of curiosity, some more naively than
others, while others had set goals they wanted to achieve (Karlbom, 1967: 237; Ericson,
1987: 16; Olsson, 1949: 112). Participants could be trembling and fearful, while others
just heedlessly joined the ranks (Klandermans, 1997: 2; Olsson, 1949: 112).
Bert Klandermans implies that a persons identity regulates their participation in
social protests and the more a person identifies with a group, the greater the chance of that
person participating in protests on behalf of the group. Group identification makes
participation in protests more likely, and in turn participation intensifies group identification
(Klandermans, 2002). Even though identity is an important determining factor throughout an
individuals life and in respect of the life choices taken, the demographic outcomes of these
choices make up an individuals life situation, which in turn influences any future choices.
Participation is contingent upon an individuals risk assessment, as it either has to outweigh
any possible risks or present acceptable losses (Klandermans, 2002; Klandermans, 1997: 3, 6,
24). Mancur Olson disused as determinants for participation in collective action selective
incentives that related to the cost of participation, be these either gains or punishments
(Olson, 1990: 48).
Participation is voluntary, as is non-participation, (Klandermans, 1997: 13) and
Olsen wrote that there will always be individuals standing on the sidelines, because it is
not rational for everyone in a group freely to make sacrifices for the collective good
(Olson, 1990: 48). Klandermans too discussed the aspect of non-participation and how
achieving a collective goal is not dependent on the participation of the entire collective
in protests (Klandermans, 2002). The benefits of collective action would, according to
Olsen, be provided even to those who chose not to strike, because of their shared
membership of the collective (Olson, 1990: 48).
However, non-participants were not viewed favourably. There was always a
negative feeling surrounding those who chose not to participate, as by their silence they
appeared to support the old structures and implicitly opposed social change (Lundberg,
1979: 60). These were workers whose loyalty or fear of losing their employment was
stronger than the solidarity towards their striking peers (Ericson, 1987: 161).
Therefore, much time was spent trying to persuade workers to participate.

107

6. Sources and method


The basis of this analysis is information gathered from the personnel lists of four
sawmills in the Sundsvall district, recorded during the strike of 1879: Kubikenborg
(n=93), Heffners (n=212), Klampenborg (n=138) and Svartvik (n=435). The lists
contain information about workers, including name, place of residence, the church
where they were registered, assets, marital status and the number of children in the
household. The lists also included comments on the workers possible involvement in
the strike. However, information in the lists is not complete for every recorded worker,
nor do the lists contain details of the entire workforces at these specific mills. Based on
the sawmill owners own estimates of how many workers had joined the strike, it can be
assessed that these lists recorded approximately two thirds of the employed workforces
(Bjrk and Schnell, 1979: 45). Further, it also has to be noted that the sawmill owners
seldom gave the authorities the correct number of workers they employed, as this would
have forced them to take responsibility for more workers than they were willing to
provide for (Olsson, 1949: 85).
The validity of the lists may be challenged from three aspects: why the lists were
drawn up; why they did not include all employed workers; and the suspicions regarding
strike participation noted in the lists. Especially since the purpose and origin of the lists
was linked to which workers were included and excluded. Since the lists contain both
registered and permanently settled workers, and temporarily employed and unregistered
workers, motives may at first be difficult to discuss, but there is no reason to believe
that the lists included or excluded workers capriciously. The lists can therefore be
described as a cross section of the employed workforces in 1879.
The comments on possible strike participation classified workers into categories.
Five workers in the lists had been marked as possible participants, 62 had been labelled
as non-strikers, 127 had been marked as returning to work after the strike and 110 had
been recorded as not returning to work after. The majority of the workers (564) had not
been marked in relation to strike participation. The assumption is that the lists may have
been a way for the sawmill owners to keep an eye on possible disruptive elements and
that they may have functioned as the basis for evictions. The categories, however, must
be regarded as speculative from the perspective of the sawmill owners, especially as it
would have been impossible for the employers to know exactly what the workers did
and did not do unless they had been specifically informed.
Since the purpose of this study focuses on the workers demographic
differences, it was important to connect the lists to other available sources. We have
Swedish church records, of which microfiches of the originals are available at the
regional state archives and at several state research facilities. These records contain the
demographic and personal information for the registered population in all Swedish
parishes as far back as the 17th century. The church records used to link with the
personnel lists were the digitised church records for the Sundsvall district (Indiko)
available at the Demographic Data Base at Ume University.
The linking process shows that the information in the lists was extremely
reliable with regard to marital status and household size. The only information in the
lists that was not in agreement with the church records was information on actual
residence and continued employment. Employment and residency were not necessarily
contingent upon each other, and this makes it difficult to know which workers were
sawmill workers, especially if they were not officially registered at a mill site. Some

108

workers were officially registered elsewhere in the district, even though they were
shown to have resided at a mill site according to the personnel lists. The linking process
also revealed that some of the workers marked in the lists as not having returned to
work actually remained employed, while the opposite was true for some workers
marked as having returned to work. This inconsistency applied to 22 workers.
However, there is nothing to suggest that these labels would have been wrong when
they were written, which is why the 1879 lists offer the unique opportunity to identify
individual sawmill workers and distinguish between strikers and non-strikers; connecting
employment, residency and strike participation. The demographic differences between
workers may better explain how they justified their motives for strike participation.
As the status of those not marked, and those marked as possible participants in
the strike, remains uncertain, they will be excluded. This study will focus on workers
that were marked as non-strikers and the strike participants who were either marked as
having returned to work or not.
7. Strike participation and place of birth
At the beginning of industrialisation, most industries found sufficient workers among
local populations (Rondahl, 1972: 29). Sawmill work was seldom more than a
supplementary income and most workers lived within walking distance of a mill,
without having to settle at the mill sites. Migration to the sawmills began from the early
1860s and grew exponentially during the following decade. While some workers
migrated a relatively long distance, others came from local parishes in the district. As
industrial development progressed during the 1860s and 1870s, migration and
permanent settlements created proper sawmill communities close to the mill sites
(Bengtsson, 1990: 193, 186).
The populations connected to the mills were diverse and the seasonal character
of the work resulted in the populations at sawmills, for at least parts of the year,
predominantly being workers with no prior connection to the district. The personnel
lists show that while the majority of workers recorded had an official registration at the
mill of employment or in the surrounding parishes in the district, slightly more than one
in five workers among the three groups (non-strikers, those who returned to work and
those who did not) lacked such a registration. A. Nilsson wrote that because so many
residents had no prior ties to the communities, this may have increased a notion of
rootlessness that could have easily led to friction between workers and employers.
That the 1879 strike occurred might therefore not be surprising (Nilsson, 1973: 36).
Creating a sense of belonging in these new communities was, for that reason, of
the greatest importance to the migrants who decided to settle permanently. The popular
movements that arose in Sweden during the 19th century were influential tools in
achieving this, and they became common fixtures at most sawmills in the Sundsvall
district. The movements primarily attracted migrants, while members of the established
local populations appeared less inclined to join. hngren suggested that this might have
depended on migrants being less influenced by existing norms in the agricultural
communities and not having felt the same pressure not to oppose the local hierarchy as
the locals might have (hngren, 1996: 71). It is therefore possible that place of birth and
attachment to the area were determining aspects influencing a workers decision
whether to join the strike.

109

Table 1: Place of birth among strike participants and non-participants during the
sawmill strike of 1879
Total

Strikers
Return to work

Not return to work

Non-strikers

94 (47.0%)

81 (40.5%)

25 (12.5%)

200

Sundsvall region

10 (43.5%)

7 (30.4%)

6 (26.1%)

23

County of Vsternorrland

23 (46.9%)

21 (42.9%)

5 (10.2%)

49

Central and South of Sweden

46 (51.7%)

35 (39.3%)

8 (9.0%)

89

Northern Sweden

12 (35.3%)

17 (50.0%)

5 (14.7%)

34

Nordic countries

3 (60.0%)

1 (20.0%)

1 (20.0%)

Source: Lists from Klampenborg, Kubikenborg, Heffners and Svartvik sawmill. Vsternorrlands
landskansli, Sundsvallsoroligheterna 2 vol, 1879-1880 Exxx 1:b, County Archive of Vsternorrland,
Hrnsand.
Indiko digitised church records, Demographic Data Base, Ume University.

Generally, few appear to have been born in the Sundsvall district, which confirms the
sawmill populations as migrant. Those workers that were marked as returning to work
display the largest proportion of workers born within the district, even though the actual
numbers are few for all three examined categories. Surprisingly, non-strikers display the
lowest proportion of locally-born workers, as well as the lowest proportion of workers
born within the greater area of the County of Vsternorrland. In fact the non-strikers
show, in both proportion and in actual numbers, that the majority had come to the
Sundsvall district from central and southern Sweden, effectively disproving a link
between non-participation and local attachment.
The majority of the workers marked as returning to work were, similarly to the
non-strikers, migrants from central and southern Sweden. Among the workers that were
marked as not returning to work, the majority were born in Vsternorrland County, but
also showed a strong connection to northern Sweden. The results thus show that striking
workers were more likely than non-strikers to have had a local attachment to both the
district and the county.
8. Strike participation and age
Previous research from Matfors sawmill in Sundsvall shows that the majority of the
workforce at the mill in 1846 were seasonally employed workers with a mean age of 38
years. The permanently employed workers had a mean age of 36. The majority of workers
were between 30 and 44 years of age. There were no substantial changes to the mean ages
of workers during the following years. The mean age of the seasonal workers displayed a
slight increase to 39 years, while the mean age of the permanent workers dropped to 36.
Sawmill work does not appear to have been an activity for younger men. After all, the
majority of the workers were married, had families and were well established in their local
communities (Ostergren, 1990: 59-64). Similar results have been found for mechanical
industries in southern Sweden between 1880 and 1909: workers younger than 20 and older
than 50 years were extremely rare (Ericson, 1987: 166).
While the mean age among sawmill workers remained in the high thirties, ages
among the male heads of households in Skn parish indicated that the mean age

110

increased during the last decades of the 19th century (Tedebrand, 1977: 267). More than
a third of all men in the parish records for 1890 were older than 50, while the same
figure for 1901 shows that 50 per cent were over 50 years of age, probably as the result
of the sawmill industrys decline. Lowered demand for labour resulted in a rise in the
mean age of workers already employed. Therefore, even though mean ages may have
dropped due to more dispersed age categories after 1860, the mean age of employed
workers in general would have increased.
Table 2: Mean age among strike participants and non-participants during the
sawmill strike of 1879 (n=233)

Mean age

Strikers
Return to work
39.9

Not return to work


32.4

Non-strikers
50.0

Source: Lists from Klampenborg, Kubikenborg, Heffners and Svartvik sawmill. Vsternorrlands
landskansli, Sundsvallsoroligheterna 2 vol, 1879-1880 Exxx 1:b, County Archive of Vsternorrland,
Hrnsand.
Indiko digitised church records, Demographic Data Base, Ume University.

The findings of previous research suggest that the workforces during the strike of 1879
would have had a fairly dispersed age distribution and a low mean age. The mean ages
at the individual sawmills ranged from 32.3 to 35.8 years, which is in agreement with
previous results regarding the age of workers in sawmills. The result from Table 2
shows that the differences were greater between strikers and non-strikers than between
individual sawmills, with the most obvious difference being that the non-strikers were
markedly older. There was also a notable difference between the two types of striking
workers, with the workers marked as returning to work showing a higher mean age. The
workers marked as not returning displayed the lowest mean age. A link between strike
participation and age would therefore seem to exist.
Despite the mean age, it is also important to review the age group distribution to
be able to understand age in relation to strike participation. Figure 2 shows that
Ostergrens results on a wide range of age distribution among sawmill workers were
also applicable to the workforce employed in 1879. Men in their thirties were still
predominant, but there was an increased presence of workers in their twenties,
especially among those workers marked as not returning to work, which would account
for the relevant low mean age. The majority of the non-strikers, 53.1 per cent, were 50
years of age or more. They displayed the widest age distribution among the workers,
including almost equally large groups of workers in their twenties through to their
seventies. Despite the wider age distribution in this group, strike participation may still
be implied to have been age dependent. It could be suggested that older workers were
less inclined to participate in the strike because of the higher probability of being
married and having a family to support.

111

Figure 2: Age distribution among strikers and non-strikers (n=233)


45
40
35

Percent

30
25
20
15
10
5
0
15-19

20-29

30-39

40-49

50-59

60-69

70-79

Workers ages
Non-strikers

Return to work

Not return to work

Source: Lists from Klampenborg, Kubikenborg, Heffners and Svartvik sawmill. Vsternorrlands
landskansli, Sundsvallsoroligheterna 2 vol, 1879-1880 Exxx 1:b, County Archive of Vsternorrland,
Hrnsand.
Indiko digitised church records, Demographic Data Base, Ume University.

9. Strike participation, marriage and household size


Marital frequency in Sweden decreased in the beginning of the 19th century, as more
individuals chose to remain unmarried. However, Sundsvall experienced an increased
frequency of marriages from the 1870s (Tedebrand, 1977: 261, 257). The simultaneous
occurrence of the postponement of marriage due to industrial work also led to a rise in
marital age (Klvemark, 1983: 17; Tedebrand, 1977: 261; Persson, 1992: 282).
Research shows that the workers connected to Matfors sawmill were on average
middle-aged and married (Ostergren, 1990: 59). Similar age structures and an increased
marital frequency would imply that a large proportion of the workers during the strike
would have been married. That the largest age groups among the strikers and nonstrikers were found among the ranges 20-29 and 30-39 years, suggests that the workers
during the strike of 1879 would display similar levels of marital status. Because of the
temporary workers brief employment, sawmill owners preferred to hire married
workers with dependants, as such men would be more likely to behave, comply and
return the following season (Berglund-Lake, 2001: 30, 35).

112

Table 3: Marital status among strikes and non-strikers during the sawmill strike of
1879
Strikers
Return to work
n
98 (42.2%)
Married
82 (46.3%)
Unmarried 16 (29.1%)

Non-strikers

Not return to work


85 (36.6%)
49(21.1%)
55 (31.1%)
40 (22.6%)
30 (54.5%)
9 (16.4%)

Total
232
177
55

Source: Lists from Klampenborg, Kubikenborg, Heffners and Svartvik sawmill. Vsternorrlands
landskansli, Sundsvallsoroligheterna 2 vol, 1879-1880 Exxx 1:b, County Archive of Vsternorrland,
Hrnsand.
Indiko digitised church records, Demographic Data Base, Ume University.

The majority of the workers included in this study, 177 (76.3 per cent), were married
and as Table 3 shows, the majority of the married men were marked as having returned
to work. In fact, only a third of the married sawmill workers were marked as not having
returned to work after the strike. The corresponding number among the unmarried
workers indicates that more than half of all workers, 54.5 per cent, were marked as not
returning to work. Few of the workers marked as non-strikers or as returning to work
were unmarried. It was perhaps not a surprise to find that the majority of the workers
marked as not returning were unmarried and displayed the lowest mean age as they,
without the economic responsibility of a family, would have been more inclined to
leave. The results indicate a link between strike participation, age and marital status.
Generally, late marriages and continuing high rates of child mortality meant that
family size by the 1870s generally remained moderate (Alter, 1988: 165-166; Persson,
1992: 284). Industrialisation may have stimulated both marital frequency and family
growth in Sundsvall, but a decline is visible during the 1870s and early 1880s. This
development was an adjustment to the industrial lifestyle and very large families would
not have been common within the sawmill communities in the Sundsvall district
(Tedebrand, 1977: 263, 266). There were differences though, between sawmill
communities and sawmill districts. Christina Fjellstrm showed in her research on
Stocka sawmill in Hlsingland province, that it was common for families to have four or
more children (Fjellstrm, 1990: 161). Recent research on the sawmill populations in
the Sundsvall district have shown that children made up the largest demographic group
within the registered populations in the communities after 1870 (Bergman, 2010).
The mean age of the non-strikers would suggest that even though they may have
had children, these would have left the households by 1879, while the workers marked
as not having returned to work would have been more recently married and have had
few children in their households. Regardless of the number of children present in a
household, having dependants would have made workers more cautious before
jeopardising their employment.

113

Figure 3: Percent of registered children born prior to 1879 per married workers
among strike participants and non-participants in relation to the strike of 1879
(n=187)
50

Percent

40
30
20
10
0
Married -no
children

1-2

3-4

5-6

Children per married worker


Non-strikers

Return to work

Not return to work

Source: Lists from Klampenborg, Kubikenborg, Heffners and Svartvik sawmill. Vsternorrlands
landskansli, Sundsvallsoroligheterna 2 vol, 1879-1880 Exxx 1:b, County Archive of Vsternorrland,
Hrnsand.
Indiko digitised church records, Demographic Data Base, Ume University.

Table 3 and Figure 3 make it clear that most strikers and non-strikers were married and
had children. The majority had become fathers in their late twenties, mainly between
1860 and 1879. The non-strikers consisted of slightly older workers who had all
experienced the birth of their firstborn child prior to 1870. In fact, most of these workers
became fathers for the first time prior to 1859. Non-strikers also had, more commonly,
more than two children living in the household.
The majority of the workers who returned to work had experienced the birth of
their firstborn during the 1860s, whereas the workers that did not return to work had
experienced their first child having been born in the 1870s. The results would therefore
show that the majority of the strikers had children, albeit not as many as the other
groups. Since the non-strikers and those workers marked as returning to work would
have had on average slightly larger families, it would suggest that there might have been
a connection between strike participation and family size.
10. Conclusion: Sawmill workers and demographic aspects of strike participation
Olson hypothesised that collective action correlated with income and established
position (Olson, 1990: 52). This was, however, not the case for the sawmill workers in
Sundsvall. The workers movement and unions had been established among skilled
labourers and craftsmen prior to 1879, but would not successfully become established
within the sawmill communities until the late 1880s (Olson, 1990: 52; Bjrklund, 1976:
52). This is important to keep in mind, as this means that irrespective of the distinction

114

usually made between pre-modern and modern (reactive and proactive) movements, the
sawmill strike of 1879 cannot be placed in either category without difficulty, as it
includes elements of both (Dill and Aminzade, 2007: 305). Instead, the 1879 strike is
placed in-between these categories. It may have occurred in an industrialised setting, but
the goal was still short-sighted; as in improving the here and now as opposed to
sometime in the future. It was a spontaneous affair, without previous organisation and
planning (Olsson, 1949: 106; Olsson and Lindstrm, 1953: 66; Bjrklund, 1976: 48;
Lundberg, 1979: 15; Larsson, 1972: 119). The workers may have had a collective goal,
but they did not act collectively in response to threats, as groups of workers
continuously left the strikers camp to resume work. There was little that could override
the fear of unemployment and eviction. The strike of 1879 was therefore more
reminiscent of the contentious gatherings as described by Munger (Munger, 1981) than
more modern, proactive strikes.
While the strike itself may have been spontaneous, participation would not, at least
on some level, have been a spontaneous act even though it was an individual choice. The
sawmill workers should have been well aware of their employers and societys views on
strikes. Participation could for that reason be regarded as a decision in a longer, on-going
process that dictated the workers conditions and their relationships with their employers.
Still, participation should not so much be considered a consequence of latent feelings, but
more a consequence of changing, external circumstances forcing attitudes and priorities to
alter, thus causing behaviours and decisions to change.
Research has indicated that strike participation among the sawmill workers in
the Sundsvall district was extremely high. Johansson concluded that the reason for this,
was that it would have been impossible for the residents within the sawmill
communities to remain outside the conflict because of the social pressure the wandering
masses of striking workers would have subjected the residents to (Johansson, 1988:
370). However, outside pressure alone does not explain why some chose to participate
or not. Sawmill workers at more remotely located mills also joined the strike, without
having been visited by mobs of strikers (Olsson and Lindstrm, 1953: 66).
Reviewing the results of this paper, patterns and similarities among the different
workers have emerged, suggesting that there was a connection between strike participation
and certain demographic aspects. The different demographic aspects studied have
separately shown indications of having had an influence on the decision whether to join the
strike. It is also apparent that they were all tangled up, affecting one another.
The question of the workers connection to the local area provides somewhat
inconclusive results, partly due to the small group of included workers and the high
inward migration to the district. It could be determined though, that a local connection
was slightly more common among striking workers than non-strikers. hngrens
assumption that locals were less inclined to oppose local hierarchies could prove to be
valid in relation to the popular movements as well as strike participation (hngren,
1999: 71). More studies are required and should also include the variables of official
registration and actual place of residence.
The question of the workers ages provided a stronger link between workers and
strike participation, as it showed that participation was age dependent. The results show
that younger workers were more prone to strike than their older colleagues. This is
linked to the assumption that older workers were more likely to have been married with
families to support, and would have had more reasons not to oppose the sawmill owners
for fear of losing their employment. This conclusion is also supported by witness

115

accounts from Sknvik sawmill, which stated that older workers displayed an
unwillingness to strike, claiming that they were satisfied with their wages, and that
many of them hid in the woods when the striking mobs came to collect workers to join
in the protests (Olsson, 1949: 110-111).
Marital status in itself did not show the same apparent connection to
participation, because the majority of the workers, strikers and non-strikers alike, were
married. While married workers may have had other considerations in relation to strike
participation than unmarried workers, there is nothing to suggest that marital status
alone would have been more important than a local connection. Even though it is more
or less impossible to measure, the authority of wives over their husbands must also be
considered, as this would have influenced the workers on some levels.
It can be stated that married workers had more to lose than the unmarried, but also
more to gain by striking. The married man had to consider his entire family, while the
unmarried worker only had himself to consider. It is therefore not difficult to understand
why so many men chose to return to work after the strike. Olson and Klandermans both
argued that workers would have had to assess the risk of participation, both in connection
with gains and losses, before making a decision (Olson, 1990: 48; Klandermans, 1997).
While some workers may have chosen not to strike because of their responsibilities towards
their wives and children, other married workers may readily have participated, believing
that the strike was their chance to improve their families lives. This is not to say that
younger, unattached and seasonal workers may not also have struggled with similar
decisions. Unable to save money to make do during winter may have conjured up feelings
of having nothing to lose and everything to gain. Still, not all married workers with families
were non-strikers or returned to work after the strike, and not all unmarried participated and
left the area afterwards.
The micro data on the strike of 1879 has shown that it is possible to connect the
demography of the workers to strike participation. The results indicate that older
married workers with families were the most cautious and hesitant in making the
decision of whether to join the strike. They seem to have been more disposed to either
refuse to participate in the strike or to have been more likely to return to work,
accepting defeat after the event. Using micro data has also shown that participation was
complex and not as straightforward as the witness accounts portray it to have been.
Strike participation may have been individual, but it was also contingent on several
unspoken aspects of the workers lives.

116

References
Alter, G. (1988), Family and the female life course. The women of Verviers, Belgium,
1849-1880, University of Wisconsin Press, Madison.
Bengtsson, T. (1990), Migration, wages and urbanization in Sweden in the nineteenth
century, Urbanisation in history. A process of dynamic interactions, van der Woude,
A.S, de Vries, J. and Hayamui, A. (eds.), Clarendon, Oxford.
Berglund-Lake, H. (2001), Livet ger rum. Frsrjning och platstagande i norrlndska
sgverkssamhllen, Etnologiska avdelningen, Uppsala.
Bergman, M. (2010), Constructing communities. The establishment and demographic
development of sawmill communities in the Sundsvall district, 1850-1890, Ume
universitet, Ume.
Bjrk, B. and Schnell, J.B. (eds.) 1979, Sundsvallsstrejken 1879: Samtida dokument och
historisk belysning, Sundsvalls museum, Sundsvall.
Bjrklund, J. (1976), Strejk frhandling avtal: Facklig aktivitet, arbets- och
levnadsvillkor bland sgverksarbetare i Vsternorrland 1875-1914, Ume University,
Ume.
Carlgren, W. (1926), De norrlndska skogsindustrierna intill 1800-talets mitt, Almqvist
& Wiksell, Uppsala.
Cornell, L. (1982), Sundsvallsdistriktets sgverksarbetare 1860-1890: Arbete,
levnadsfrhllanden, rekrytering, Gteborgs universitet, Gteborg.
Dill, B. and Aminzade, R. (2007), Historians and the study of social protest,
Handbook of social movements across disciplines, Klandermans, B. and Roggeband, C.
(eds.), Springer, New York.
Ericson, H.O. (1987), Vanmakt och styrka. Studier av arbetarrrelsens tillkomst och
frutsttningar i Jnkping, Huskvarna och Norrahammar 1880-1909,
Studentlitteratur, Lund.
Fjellstrm, C. (1990), Drmmen om det goda livet: livskvalitet och matvanor i ett
uppvxande industrisamhlle: Stocka sgverk 1870-1890, Almqvist & Wiksell
International, Stockholm.
Frader, L. (1981), Grapes of Wrath, Class, conflict and collective action, Tilly, A. and
Tilly, C. (eds.), Sage Publications, Beverly Hills.
French, H. and Barry, J. (eds.), (2004), Identity and agency in English society, 15001800 Introduction, Identity and agency in England, 1500-1800, Palgrave Macmillan,
Houndmills.

117

Hogg, M. and Abrams, D. (1988), Social identification. A social psychology of


intergroup relations and group processes, Routledge, London.
Hglund, H. (1957), Sg vid sg jag sg, Boktryckeri Aktiebolaget, Sundsvall.
Karlbom, R. (1967), Hungerupplopp och strejker 1793-1867. En studie i den svenska
arbetarrrelsens uppkomst, Gleerups, Lund.
Klandermans, B. (1997), The social psychology of protest, Blackwell Publishers,
Oxford.
-------------------- (2002), How group identification helps to overcome the dilemma of
collective action, American Behavioural Scientist, 2002:45.
Klvemark, A.S. (1983), Kommer familjen att verleva? Den utsatta familjen,
Norman, H. (ed.) Uppsala universitet, Kristianstad.
Kmpe, A. (1979), Den stora strejken, Barrikaden, Stockholm.
Larsson, T. (1972), Vckelsen och Sundsvallsstrejken, Gummessons, Motala.
Lundberg, H. (1979), Sundsvallsstrejken 1879, CEWE Frlaget AB, Bjsta.
Layton, I.G. (1979), Hamnar och sjfart i vre Norrland. En studie av utvecklingen
frn 1700-talets mitt och dess industriella bakgrund, SCA 50 r. Studier kring ett
storfretag och dess fregngare, Utterstrm, N. (ed.), SCA, Sundsvall.
Lundh, C. (2006), Arbetskraftens rrlighet och arbetsmarknadens institutioner i
Sverige 1850-2005, Arbetsrtt, rrlighet och tillvxt, Rauhut, D. and Falkenhall, B.
(eds.), Institutet fr tillvxtpolitiska studier, Stockholm.
Munger, F. (1981), Contentious Gatherings in Lancashire, England, 1750-1893, Tilly,
L. and Tilly, C. (eds.) Class Conflict and Collective Action, Sage publications, Beverly
Hills.
Nilsson, A. (1973), Sgverkshistoria i Hlsingland, Sgverk och mnniskor,
Lundberg, B. (ed.), Rabn & Sjgren, Stockholm.
Norberg, A. (1980), Sgarnas . Aln och industrialiseringen 1860-1910, Almqvist &
Wiksell, Uppsala universitet, Uppsala.
Nordberg, P. (1979), Vattenkraftens utnyttjande i Ljungan, SCA 50 r. Studier kring
ett storfretag och dess fregngare, Utterstrm, N. (ed.) SCA, Sundsvall.
Olson, M. (1990), How bright are the northern lights? Some questions about Sweden,
Lund University, Lund.
Olsson, R. (1949), Norrlndska sgverksliv under ett sekel, Nordiska Museet,
Sundsvall.

118

Olsson, R. and Lindstrm, R. (1953), En krnika om sgverksarbetare, Sundsvall.


Ostergren, R.C. (1990), Patterns of seasonal industrial labour recruitment in a
nineteenth century Swedish parish. The case of Matfors and Tuna, 1846-1873,
Demographic Data Base, Ume University, Ume.
Persson, C. (1992), Jorden, bonden och hans familj. En studie av bondejordbruket i en
socken i norra smland under 1800-talet, men srskild hnsyn till jordgande,
sysselsttning och familje- och hushllsbildning, Stockholms universitet, Stockholm.
Rahm, H. 2000, Svensk sakprosa. P plankbrarnas villkor. Tre dagstidningars
bevakning av Sundsvallsstrejken 1879, Lunds universitet, Lund.
Rondahl, B. Emigration, folkomflyttning och ssongsarbete i ett sgverksdistrikt i sdra
Hlsingland 1865-1910. Sderala kommun med srskild hnsyn till Ljusne
industrisamhlle, Lromedelsfrlagen, Uppsala universitet, Uppsala.
Skarin Frykman, B. 1990, Arbetarkultur Gteborg 1890, Etnologiska freningen i
Vstsverige, Gteborg.
Sgverksfolket. Alnbygd i frvandling, Ankarsvik 1995.
Sderlund, E. 1951, Svensk trvaruexport under hundra r, Uppsala.
Tedebrand, L.G. 1977, Demografisk stabilitet och frndring under det industriella
genombrottet, Tedebrand, L.G. (ed.), Historieforskning p nya vgar. Studier
tillgnade Sten Carlsson 14.12.1977, Studentlitteratur, Lund.
Tedebrand, L.G. 1999, Historia och demografi valda texter, Ume universitet, Ume.
Wik, H. 1950, Norra Sveriges sgverksindustri frn 1800-talets mitt fram till 1937,
Almqvist & Wiksell, Uppsala.
hngren, B. 1996, Elit och massa - Spnningar i frden mot ett nytt samhlle.
Folkrrelse och samhllsutveckling 1850-1990, Sundsvalls historia D. 3, Tedebrand,
L-G. (ed.), Stadshistoriska kommittn, Sundsvall.

119

Why do workers strike? Looking for an answer using micro


data on Leiden strikers in 1914
Sjaak van der Velden
1. Introduction
The central theme of strike research is the question of why workers occasionally decide
to carry out a cessation of work. An answer to this question is essential for politicians,
employers and union leaders, but also for historians, sociologists, economists and other
students of the social sciences. Most research into the field has been based on
aggregated strike data as collected by statistical bureaus. Aggregated data has been
published since 1927 by the International Labour Office (ILO) and gives information
from many countries on the number of strikes per year, the number of workers involved
and the number of working days of production lost. National statistical bureaus often
publish more detailed data on regions, professions, causes and the month of outbreak of
a conflict. In the beginning, official statistics had the character of criminal statistics,
because strikes and unions were forbidden. After the right of association was
recognised, strike statistics dealt principally with the economic consequences of labour
conflicts (ILO, 1926: 5-6). Early statistics were very rich in detail, which was real micro
data published by or on behalf of labour ministries. Unfortunately those days have long
gone and all data is now at an aggregated level (Franzosi, 1982: 3). To give an example
that is important for this chapter: the last year that micro data was published in the
Netherlands was 1913, but from that date on the amount of data diminished almost
yearly. The initial reason was budgetary during World War I, but later the focus of
Statistics Netherlands shifted from a social to an economic angle. The main reason to
collect strike data from 1927 concerned the economic costs of strikes. From the 1980s
data was restricted further for reasons of privacy: a user of the data should not be able to
identify the company where strikes occurred. As a result, the level of aggregation rose
again.
Using the above-mentioned data, researchers can try to answer questions about
the propensity to strike of workers in certain countries, professions or regions. Patterns
of strike activity over time can also be discerned. Unfortunately, the character of the
data makes it impossible to answer the question of why individual workers go on strike.
For this, strike data on an individual basis and data on other characteristics of the
relevant individuals is needed.
In this contribution I will try to answer the question of whether using micro data
concerning strikers who struggled for higher wages in a Leiden cotton company in 1914
opens new insights into workers behaviour. The outcome of my research is only
valuable in respect of the company in question, but this may be a starting point for
further research. It may be a new piece of the puzzle.
In section 2, I will give a short overview of the history of the Leiden cotton
printing company, LKM. Section 3 provides a history of strikes at the company. The
section that follows is dedicated to a description of the micro data on the strikers and
non-strikers involved in the industrial action that took place in 1914. Hypotheses about

120

workers behaviour regarding strike activity are formulated in the next section. In
section 6, an effort is made to falsify the hypotheses using available data, after which
conclusions are drawn in section 7.
2. A critique on aggregated strike data
From the early days of capitalism, researchers have wondered why workers strike. This
question invited many to publish material about strikes, but it took some time before
researchers started to use more sophisticated statistical techniques. Maximilian Meyer
published a book in 1907 on strikes in seven countries (Meyer, 1907). Although the
work done and statistics published still show the Sisyphean task he undertook, all he did
was to collect statistics from seven countries. It took five more years before statistical
calculations entered the field of strike research. In 1912 the French researcher C. Rist
published an article relating strike activity and economic indicators (prices and
unemployment) using the coefficient of dependency (Rist, 1912). The coefficient of
dependency used by Rist was developed by the Frenchman L. March in 1905. This
coefficient is somewhat similar to the correlation coefficient developed by Galton and
Pearson, although the latter was not generally known in those days.
Rists article inspired the Dutch statistician Van Dam van Isselt to calculate the
coefficient of dependency of the number of strikes to unemployment in six countries for
the years 1901 to 1912: Belgium, Germany, England, France, the Netherlands and
Austria (Van Dam van Isselt, 1914). His outcome differs slightly from the coefficients
that Rist calculated, but the differences are the result of the greater extent of annual data
Van Dam van Isselt had at his disposal. In general, both authors saw their expectations
regarding the relationship between strikes and unemployment as having been
confirmed. According to their expectations, higher unemployment would lead to
diminished strike activity, and similarly lower unemployment results in less fear among
workers of being dismissed or otherwise penalised, and therefore in more strike activity.
Since those early days of the statistical analysis of strike trends, a vast number of
studies have been published. Often mentioned in overviews is John I. Griffins book on
strikes. Griffin published the first thorough statistical book on strikes in the United
States, in which he investigated the relationship between strikes and lockouts on the one
hand and the business cycle on the other (Griffin, 1939: 182). He did not calculate
correlation coefficients, but concluded from the graphical movements of the number of
strikes and the price index that both coincided frequently (Griffin 1939: 69).
In the greater part of statistical studies there is not a living soul included. All the
theories are based on aggregated data and therefore cannot explain why individual
workers decide to join a strike or, on the contrary, join the ranks of the scabs.
Aggregated data is only useful in explaining group behaviour or abstract developments
in strikes related to (even disputed) economic long waves (Kelly, 1997). If researchers
discover that strike propensity moves in some way with the economic cycle they can
only observe the fact, but will not be able to make valid remarks about the reason why
individual workers make the decision to join or abstain. Trying to use the discovered
correlations between strikes and economic indicators in explaining the individual
behaviour of workers is an example of confusing ecological and individual correlations
that was coined the ecological fallacy by W.S. Robinson in 1950 (Hackett Fischer,
1971: 119120). This may be explained by giving a reversed example of this fallacy.

121

Although it is clear from epidemiological research that nicotine can cause lung diseases,
the entrenched smoker will always point to a very old person who has smoked for their
entire life without becoming ill. We all understand that this approach is wrong.
Let us return to strike research. Since Griffins publication, a vast number of
strike studies have been published. These have been put in order, according to the
several approaches by different authors. Roberto Franzosi in his illuminating study
mentions five different approaches (Franzosi, 1995: 10-12). The Business-cycle
explanation is best represented by the Ashenfelter and Johnson model, which argues
that the state of the labour market modifies the bargaining position of the workers. This
modification influences the propensity to strike.
Economic hardship theories have sought to explain strikes as resulting from the
level of grievances. When this level becomes intolerable, workers are likely to go on
strike. Set apart from these two economic angles are the political-exchange and
institutional theories. Their advocates try to explain strikes as resulting from the
duration of collective agreements or the position of labour (parties) in the political
power structure. The fifth approach tries to explain strikes as resulting from the capacity
of workers to organise themselves into stable unions. Franzosi adds a new approach to
these theories, in which in explaining strikes, the strikes themselves are not only a
dependent variable but also an independent variable. In other words: strikes explain
strikes. Somewhat apart from these theories stands the Marxist notion that strikes are an
expression of the capitalist relations of production, which are expected to end in the
overthrow of capitalism per se. In this view, workers are almost forced into strikes by
the class nature of society. According to many Marxists, intervention by a wellorganised political party is, however, needed to guide the striking workers into
socialism. It may be because of this perspective that Marxist researchers are almost
completely absent in quantitative strike research. They consider it futile to study strikes
as long as the envisaged party does not exist.
The mentioned approaches have one thing in common: they are very general and
need aggregated data. The reasons why an individual worker will or will not go on strike
cannot be answered via these angles. Klandermans (2004) developed a typology of motives
for people to engage in protest. Even if people agree on the goals of a protest movement
(which is what a strike is) they need more motives to actually participate. The types
Klandermans distinguished are instrumental motives, ideological motives and collective
identity. To study these motives, modern researchers use questionnaires; a technique we
cannot use in historical research. In the rare cases when workers from the past were
questioned, we can use the answers.
In the majority of cases we will have to use other kinds of data. Most of the time
this data will be demographical in character, such as sex, age or education. The findings
of most of these studies are not very consistent (Gallagher and Strauss, 1991: 9) but are
better than nothing. In this chapter I will therefore try to discover the personal motives
for people joining or abstaining from strikes, by using data at an individual level. Where
the strikers and non-strikers lived, how they lived, what their earnings were, etc. Are
there any differences between strikers and non-strikers that can be attributed to a certain
stage in their life course? Are young workers, unmarried workers or workers without
children more strike prone than their older, married colleagues or those with children?
Or are there any other characteristics of their life cycles that may help to explain why
workers go on strike?

122

Looking for an answer to the question of why individual workers participate in


strikes, I was lucky to discover a special historical source in the records of a cotton
printing company in the municipal archives of Leiden. The data from this source will
perhaps give some clues to help answer the question of why the individual workers
from this company did or did not engage in a strike.
3. Hypotheses
Given the notion that the search for workers motives to go on strike may benefit from
micro data, there is still the problem of how to measure the motives of strikers and nonstrikers. Can we deduce the intentions of people from their actual acts? The actor-oriented
approach tries to discover the circumstances under which individuals are willing to enter
into collective action (Kelly, 1997: 23). This goes beyond the study of the behaviour of a
group or profession. The best known example of this type of study was published in 1954
by Kerr and Siegel. The authors compared the willingness to strike of workers in eleven
countries during the period from 1915 to 1939 and categorised by industry. They concluded
that miners, seamen and dock workers were by far the most strike prone (Kerr and Siegel,
1954: 190-212). This high willingness was explained from these workers isolation in
society, but of course this mass isolation hypothesis does not answer the question of why
some workers participate in strikes and others do not. The problem that needs to be solved
is that of finding out why individuals are or are not mobilised to engage in a conflict, given
the expectation that they will all benefit from a certain collective goal. All potential strikers
have to cross a threshold. Will they join the strike or other collective behaviour or not
(Granovetter, 1978)? The considerations by workers in making the decision to join an
action or not are influenced by the behaviour of their peers; it is safer to join a strike than to
start one. Crossing the threshold is a risky act. The intentions behind any decision can be
discovered in modern sociological research by various methods such as interviews, surveys
and observing participation. An example of this is provided by Falk, Grimes and Lord
(1982), who studied a strike of teachers and discovered some differing personal qualities
between strikers and non-strikers. Their findings showed that striking teachers were more
professionally oriented than their non-striking colleagues. The strikers they studied desired
more authority, autonomy and control in the workplace. This modern conclusion cannot be
confirmed in historical research, because we cannot ask workers from the past about their
attitudes. Therefore we have to resort to the characteristics of the individual workers. Are
they male or female, old or young, skilled or semi-skilled etc?
The Dutch historian Theo van Tijn wrote about the willingness of young
workers to join labour unions: It is assumed that men of about 20 to 35 are best in a
position to undertake meaningful, responsible and if need be bold collective action
(Van Tijn, 1976: 233). These young men are supposed to have left the guidance of their
more careful parents and are not yet in the position of being parents themselves. From
the same idea we can expect married workers to be less strike prone than single
workers, who will for the most part also be younger than their married colleagues.
Following this line of thought it seems likely that workers with big families, meaning
more children, will not be very willing to go on strike. This may be expected to be the
case especially in families with young children who are unable to generate family
income. The Netherlands has been for a long time a society divided by religious
boundaries, contrary to for example Italy, where most of the inhabitants are Roman

123

Catholics. According to the census of 1909, 35 per cent of the Dutch population was
Roman Catholic, while 52.7 per cent belonged to the Dutch Reformed Church; in
Leiden these figures were 25.8 per cent and 51.9 per cent respectively1. Keuning (1970:
80) has stated that Roman Catholics are regarded as being more docile than Protestants,
because their church is more centralised than the protestant communities. This
supposition is supported by opinion research among more than 300 textile workers from
Twente region in 1920. More protestant respondents than Roman Catholic workers said
that they did not think it necessary for the minister or priest to give political guidelines
during a sermon (Heerma van Voss, 1987: 51-52).
Clark Kerr and Abraham Siegel in their 1954 comparison of the propensities to strike in
eleven countries concluded that workers living as an isolated mass and performing
unpleasant jobs are the most strike prone. This conclusion was heavily criticised on
methodological, logical and empirical grounds by Edwards (1977), but is often still quoted.
It is not necessary to follow their conclusion, but their train of thought may serve as a
guideline for constructing a hypothesis. The appreciation of workers labour and their job
satisfaction may be expected to play a role in their willingness to strike. Skilled workers
probably have greater job satisfaction than do the unskilled, and in line with the Kerr and
Siegel conclusion, are therefore less strike prone. Contrary to this view stands the dual
commitment approach. In this view, workers satisfied with a company are also satisfied
with their union and therefore more committed to union activities such as strikes (Gallagher
and Strauss, 1991: 8). If we follow the Keer and Siegel approach, we would expect that
workers who have been employed for a long time at a company are less willing to strike
than their more recently employed colleagues. They are more loyal to the company and
management. Labour turnover is a well-known alternative to a strike (Knowles, 1960: 301).
Therefore workers may see more opportunities in looking for another employer rather than
in engaging in a struggle. Because the data covers the entire careers of LKM workers, it is
also possible to study the propensity to strike of workers who were self-employed for part
of their careers. Are people who have been self-employed during their working life more, or
less, willing to strike than the real proletarians who only ever worked as paid employees?
Because these half-workers might see an opportunity to escape from working-class life, I
expect them to keep a distance from working-class activities, such as unionism or strikes.
We can also look at the wages earned by LKM workers. Higher-skilled workers are likely
to earn higher wages and therefore the wage levels of groups of workers may also provide a
clue. Are better-paid workers more, or less, willing to strike than their worse paid fellow
workers? If workers who get higher wages are indeed more skilled, it is in accordance with
hypothesis five that these workers are less strike prone than their lower paid colleagues. We
can argue this behaviour from the idea that they lack the drive to improve their life because
it is already better than that of the other workers.
From the above mentioned considerations we can formulate a number of hypotheses
to test:
1. Young workers are more strike prone than older employees.
2 Single workers are more strike prone than married workers.
3. Workers from big families are less willing to strike than those from small families
or with no children.
4. Protestant workers are more strike prone than Roman Catholic workers.
5. Skilled workers are less willing to strike than their unskilled colleagues.
1

www.volkstellingen.nl/nl/volkstelling/jaarview/1909/

124

6. Workers with a long career at the company are less strike prone than more recent
employees.
7. Semi-proletarian workers are less strike prone than real proletarians.
8. Higher paid workers are less strike prone than workers who earn lower wages.
4. The data
The company records of LKM, kept in the communal archives of the city of Leiden,
consist of a card box file. This file was compiled to offer a good insight into the lives of
the LKM workers. This insight was necessary because of the Workmens Compensation
Act of 1901, which forced employers to insure their employees. Employers had to
register their workforce and the card box is such a register (Klein, 2003: 3). The box
contains about 500 cards with individual data about almost all the workers from the
printing department who worked at LKM from 1900 until the end of 1932. Each card
gives an insight into the past and present of these workers: their date of birth, their
marital status and children, their profession before they joined LKM, their contract with
LKM, the tasks they performed, whether they joined the strikes at LKM, how much
they earned and if and why they left the company. This card box is a good source for
reconstructing the lives of workers at LKM, especially if we combine it with data from
the Leiden population register. Unfortunately the Leiden register is one of the few in the
Netherlands that suffers from being incomplete.3 From the card box it is clear which
workers went on strike in 1882, 1902, 1914 and 1922 and which workers did not. For
the 1882 and 1902 strikes, the lists from the card box are very incomplete, because
many of these strikers no longer worked at the company when the card box was
compiled. Elsewhere in the archive, complete lists of strikers during these early
conflicts are available; so much more work can be done there. However, at this stage of
the research I decided to study only the 1914 strike, because data from the card box and
the population register are the most complete for this period. Since the data from LKM
is not a sample, but includes the entire population of the workers at the factory during
the 1914 strike, this means that all results of calculations hold true and there is no need
to give the statistical significance. Of course the strike under research is only a small
one and the outcome of the calculations holds true for this strike only.
A short history of the company and an overview of the strikes at LKM during
the years before 1914 will precede the examination of the hypotheses.
5. The Leiden Cotton Company and cotton printing
Leiden is a city in the Netherlands, situated halfway between Amsterdam and
Rotterdam. In the seventeenth century it was the second largest city in the Netherlands
after Amsterdam. In that century Leiden was known for its textile and especially its
woollen industries. Although the textile industry declined in the eighteenth century due
to French protectionism, textiles were a feature of the city until the 1950s. During the
2

Regionaal Archief Leiden, Archief Leidsche Katoenmaatschappij,


In 1929 an enormous fire destroyed the town hall where the population register was kept. Only parts of
it were saved and can be used for historical research. Most information from the pre-1890 register and
everything from the post-1923 years was lost.

125

later part of the nineteenth century, textiles regained some of their prominence. Cotton
also then entered the Leiden economy.
One of the leading textile companies was the Leidsche Katoen Maatschappij
(LKM, or Leiden Cotton Company). Although LKM was one of the biggest companies
in Leiden, employing some 900 workers around 1890, no extensive company history
has been written so far. This is probably a result of the fact that the economic crisis of
the 1930s and the cheap production of printed cotton in Asia forced the factory to close
its gates in 1932. Most company histories were written after the Second World War and
often, the existence of a company for a century or so motivated the board of directors to
commission a memorial book. The only scientific history is an article by G. Verbong
about the first ten years of LKM (Verbong, 1987).
Originally LKM was a company based in what would become Belgium. In 1835,
after the separation of the United Kingdom of the Netherlands into two separate
countries, the factory (named De Heyder & Co. and established in 1757) was relocated
from Lier in Belgium to Leiden in the Netherlands. The reason for this move was the
Dutch possession of vast colonies in what is now Indonesia, called the Dutch EastIndies in those days, to where printed cotton could be shipped on beneficial terms. The
city of Leiden even donated to the company the land on which to build a factory.
Cotton printing was performed on a massive scale for that part of the empire,
and production was promoted and subsidised by the Dutch government as a measure
against poverty in the homeland. Labour had been regarded as a good solution for
pauperism since the end of the eighteenth century.4 This attitude also had negative
results on productivity, because promoting labour resulted in the neglect of laboursaving developments. Even worse, after ten years in Leiden the once technologically
advanced factory was in a state of decline. Labour was cheap and the state bought the
product, so the owner felt no incentives to modernise. He even used all the profits for
his own pleasure and refused to invest (Verbong, 1987: 29). In 1846 two Dutch
entrepreneurs (Van Wensing and Driessen) bought the factory and gave it a fresh start.
The new management had better instincts for the needs of modern capitalism
and purchased new machinery. In 1896 the company was renamed Leidsche
Katoenmaatschappij v/h De Heyder & Co., but this new name was not a guarantee of
success. One year after the name was changed the factory was ruined by a major fire,
which made rebuilding and further modernisation necessary. Unfortunately, hard
economic times followed, and especially after World War I the company was unable to
compete with cheap Japanese products. The depression of the 1930s finally caused the
company to close down in 1932.
6. Strikes at LKM between 1882 and 1922.
Of the 30 strikes that occurred in Leiden textile companies between 1882 and the
closure of LKM in 1932, 16 took place at LKM. This results in an over-representation
of the company under research in terms of the strike history of Leiden. The workers at
LKM were obviously more strike prone than their colleagues in other factories, but we
cannot offer an explanation for this.
4

Many people were living on welfare and an endless discussion started about the reasons for pauperism
and how to fight it. Most agreed that the poor had to learn to work and that employment should be
promoted (Van der Velden, 2001).

126

The first strike that took place in the Leiden textile industry occurred in 1882 at
LKM, then still named De Heyder & Co. For an overview of the strikes that occurred at De
Heyder/LKM and the sources used, see appendix 1. In 1882 a number of strikes took place.
On 19 May of that year, 75 male and 57 female printers refused to continue working. The
reason was a reduction in wages. After three days the workers were victorious and resumed
work. In June of the same year the weavers, inspired by the success of the printers, also
went on strike. After almost two weeks of strike, riots and police intervention, the 300
workers gave in. A few strikers were not allowed to return. One of the spokesmen of the
May strikers was sacked in July. Almost one hundred workers stopped production in
solidarity with their colleague, but after police intervention they left the factory. The next
day the strike was over. In September of the same year, printers of bed-spreads calculated
that their wages had been lowered. They demanded a restoration of the former wage-level
but this demand was rejected. More than two hundred men and women went on strike. The
150 female strikers went back to work after four hours, because the board of directors
threatened to lock them out. The men did not give in and were actually locked out. This
lockout caused the women to go on strike again five days later. After two weeks of strikes
and lockouts, the strikers were allowed to gradually enter the factory again, but eight of
them were victimised. The strike was a complete failure and also caused the decline of the
recently formed workers organisation.
The next strike at De Heyder & Co, now known as the Leidsche Katoen
Maatschappij, took place in 1895. On 13 June the weavers stopped working. Again, the
underlying reason was a recent reduction of wages, although the immediate cause was
the suspension of a fellow weaver. The strike was supported by the carpenters union of
Amsterdam. However, this support was in vain and after four days the 400 weavers
decided to go back to work, but another grievance almost immediately caused a new
strike. During a meeting between the management and a delegation from the workers
over complaints about the bad quality of the yarn that had to be woven, a strike broke
out5. This strike was instigated by a member of the weavers union. Intervention by
union leaders and a promise by the directors that better quality yarn would be bought
ended the strike after 15 days. In 1902, 24 printers went on strike for 56 days. They
demanded smaller and lighter printing plates and a rise in wages. The strikers refused
any intervention by the labour unions. This wildcat strike was lost. In 1907, washers
complained about bad yarn. One of their colleagues, a member of the works council,
told them to go back to work. In 1914 a further strike occurred at LKM. In the autumn
of 1913 the workers had already asked for a wage increase, payment for overtime and
more holidays. When these demands were rejected, the unions presented an ultimatum
signed by 88 workers. A strike broke out, but almost nothing is known about the events
although the strikers won at least the demanded pay rise.
In 1919 two strikes occurred that were both won by the strikers. This fits in with the
overall pattern of this period. At the end of World War I and during the following years
many strikes were victorious because employers gave in easily out of fear of revolution.
The last strike at LKM took place in 1922, when 265 weavers went on strike against a wage
reduction. Wage reductions were common in this period. Employers and right-wing
politicians tried to regain what they had lost during the revolutionary years and were
supported in their ambitions by a declining economy. The strike was lost after four months.

Unlike most other cotton-printing companies LKM did not spin its own yarn.

127

Of the 16 labour conflicts that were detailed, only three were won by the
workers. The rest were settled (four) or lost (nine). So even if we add the strikes that
were settled to the conflicts with a positive outcome, workers at LKM were not very
successful in their efforts to fight for a more decent living. Only 44 per cent of the
strikes were other than completely useless.
7. Characteristics of strikers and non-strikers during the 1914 strike
In 1914, 87 workers went on strike including the five members of the action committee
that initiated the action. There seems to have been a small group of union activists at
LKM and according to mobilisation theory, this may have played a role in the fact that a
strike occurred (Kelly, 1997: 26). However, in this chapter I am looking for an
explanation as to why individual workers joined or refused to join this specific strike.
The question is, who was influenced by the activists and who was not?
On their personal cards in the card box file, 21 workers were explicitly recorded
as being non-strikers. Four of them were afterwards rewarded with a secret pay rise for
not striking. Even two of the strikers were allowed such a pay rise, but there is no
reason given for this. The one striker who always received a secret bonus before the
strike lost this allowance afterwards, probably because he was also a member of the
action committee. After connecting the incomplete data from the population register to
the personnel cards, we can note a few things about the two groups of workers. There
are a number of issues that may be important in highlighting the differences between the
strikers and non-strikers. Is there a difference in age, religion or gender? Matrimonial
status and having or not having children may also play a role according to the
hypotheses. Former career and wage level are also possible determinants of strike
behaviour.
7.1. Age, gender and matrimony
Hypothesis: Young workers are more strike prone than older employees.
The average age of the 108 workers involved in this research was 45.1 years.
The average for the strikers was 44.5 years, and that of the workers who refused to
participate in the strike was 47.8 years. This means that strikers were slightly younger
than the non-strikers but the difference is not extreme. The youngest striker was 21
years old, whereas the three youngest non-strikers were 17, 17 and 15. This means that
they were still living with their parents. An example of one such young non-striker is
Willem Taffijn. He was 17 years old and lived with his father after the death of his
mother in 1911. The father of Willem, Petrus Taffijn, decided not to strike and neither
did his son. We can suppose that they influenced one another.
The oldest non-striker was 78 years old and the oldest striker 72. The 78-year
old non-striking worker, Mooten, was the father-in-law of another non-striker. The
highest and lowest ages are thus to be found in the non-striking group which coincides
with the higher standard-deviation for this group. The decision whether to join or not
was possibly subject to some influence from relatives. All 108 workers, strikers and
non-strikers, were male, so it is not possible to discover a gender perspective in this
strike.

128

Table 2. Age characteristics of strikers and non-strikers


Average
Minimum
Maximum
Median
Standard
deviation

Strikers (n=87)
44.5
21
72
43
12

Non-strikers (n=21)
47.8
15
78
47
20.3

Hypothesis: Single workers are more strike prone than married workers.
The average age of all the workers was around 45, equating to being born around
1869. At that age most workers could be expected to have been married, because the
1909 census indicated that 88 per cent of the male population born in the years from
1865 to 1869 was married. Matrimony, and especially having children, will influence
the willingness to strike according to this second hypothesis. 77 per cent (n=67) of the
strikers and 71 per cent (n=15) of the non-strikers were married, which contradicts the
hypothesis, although the difference is not very great.
Hypothesis: Workers from big families are less willing to strike than those from small
families or with no children.
If we look at having children with regard to both groups then we may still expect a
difference. After all, children need to be fed and taken care of every day. When
calculating a correlation coefficient between the family size of workers and their
willingness to strike, this results in a negative value of R= -0.57. There seems to be a
relationship between the two. Workers with more children are less strike prone than
workers with small families. This coincides with the slightly higher average age of nonstrikers, because workers with bigger families tend to be older than workers with
smaller families.
7.2 Religion
Hypothesis: Protestant workers are more strike prone than Roman Catholic workers.
The two most widespread religions in the Netherlands around 1900 were Roman
Catholicism and the Dutch Reformed Church. In 1891 the pope, in his encyclical letter
Rerum Novarum, warned Catholics that strikes and socialism were in conflict with the
Bible. They were not allowed to desire the riches of their fellow men, including their
patrons. The Protestants organised a Social Christian Congress in the same year, which
stipulated the same rule, although leaving some room for resistance against very harsh
living conditions. It is important to realise that both denominations opposed the use of
strikes as a means to improve living conditions for workers.
Of the LKM workers whose religion is known (n=78), 24 per cent were Roman
Catholics, which is roughly the same as the percentage in Leiden as a whole (26 per
cent). 71 per cent of the employees at LKM belonged to the Dutch Reformed Church
(Nederlands Hervormd) whereas only 52 per cent of the city population was Dutch
Reformed. Lutherans and Calvinists counted for approximately two per cent each. Of
the Lutheran and Calvinist denominations, all four workers went on strike, while 79 per
cent of the Catholics and 80 per cent of the reformed joined the strike. One of the
Catholic workers had even once worked as a Zouave in the papal army, but this did not
keep him from striking.

129

In short, it seems that there was no religious influence on the behaviour of workers,
but we must keep in mind that we know the religion of only 57 per cent of the nonstrikers and 76 per cent of the strikers.
7.3 Career
Hypothesis: Skilled workers are less willing to strike than their unskilled colleagues.
Of course the career path of a worker and the kind of work he or she performs
may influence their willingness to engage in a strike. Let us first look at the level of
skill of strikers and non-strikers.
It is common knowledge in the historiography of strikes and labour unions that
unskilled workers may go on strike faster, or at least more spontaneously, than skilled
labourers. The latter group is, however, easier to organise into unions.
The historical sources available do not mention membership of the textile
workers unions at LKM. The only thing we know is the membership of five workers of
the action committee mentioned earlier. The jobs these five performed were evenly
distributed over the qualifications of unskilled, semi-skilled and skilled as indicated by
Wiegersma (Wiegersma: 97-108).
However, there is a clear difference between the complete groups of strikers and
non-strikers. Table 3 indicates this. Although it was impossible to qualify between one
third and almost forty per cent of the professions, we may so far conclude that the
strikers were mainly semi-skilled while the non-strikers were unskilled or skilled.
Table 3. Qualifications of strikers and non-strikers out of the total working
population at the LKM (%)
Unskilled
Semi-skilled
Skilled
Unknown

Strikers (n=87)

3.6
44.6
13.3
38.6

Non-strikers (n=21)

20.0
13.3
33.3
33.3

Hypothesis: Workers with a long career at the company are less strike prone than more
recent employees.
One might expect that workers who had worked all their life at LKM were so
embedded in the companys culture that out of loyalty they would be less strike prone
than those workers who had started their careers elsewhere. This seems to be the case.
55 per cent (n=48) of the strikers started their career at LKM, as did 70 per cent (n=15)
of the non-strikers. This indicates the tendency that loyalty to LKM promoted an
attitude of not engaging in strikes.
Hypothesis: Semi-proletarian workers are less strike prone than real proletarians.
Real proletarians are defined as people who are, and were, paid employees during
their entire careers. Semi-proletarians are workers who may, if necessary, switch from being
paid employees to self-employed labourers or vice versa. A small group (n= 10) of workers
who started elsewhere, or worked temporarily somewhere else although beginning at LKM,
were self-employed during that period. They ran a small shop or trade. Nine of them (90%)
went on strike in 1914, while 81% (n= 79) of the real proletarians did. So the difference in
attitude between the two groups is very small and even contradicts the hypothesis. This is
not true for workers who had once worked in the military. Of the strikers, 23 per cent

130

(n=20) had once served in the army, whereas 30 per cent (n=6) of the non-strikers had a
military career. In other words, workers at LKM who had once been soldiers lost some of
their class-consciousness and became less strike prone than their colleagues. The difference
may however also result from the pension paid to discharged soldiers.
7.4 Wages
Hypothesis: higher paid workers are less strike prone than workers who earn lower
wages.
Let us differentiate between the wages of strikers and non-strikers at the
beginning and at the end of the 1914 events.
Table 4. Average weekly wages of strikers and non-strikers before and after 6 May
1914, in guilders
Before the strike
After the strike
Rise in %

Strikers (n=87) Non-strikers (n=21)


7.345
7.870
7.1

6.405
7.140
11.5

From Table 4, two things are clear. Strikers were better rewarded than non-strikers
before the strike, and non-strikers benefited more from the strike than did those who
participated in the action. This seems a somewhat confusing conclusion, but it has to do
with the fact that the employer was seemingly forced to give in to the strikers. By better
rewarding the non-strikers, the management was able to give a sign to the workers that
striking was a dangerous affair. Some of the strikers lost the secret bonuses they
received before, and the non-strikers got an extra pay rise over that which was awarded
to the strikers. Just to give one example: on one of the cards from the card box we read:
because of not striking 10 guilders,6 whereas the normal rise would have been to 9.75.
The fact that the strikers were those workers earning higher wages probably
gives an insight into relations in the workplace. These better paid workers (but not
necessarily better skilled, see 6.4) may have felt more strongly than their colleagues.
This feeling must have come from their better position in the labour market. This
position resulted in higher wages, and the fact that these workers were more prone to
striking is an indication that strikes are not a sign of despair, but a proof of strength.
7. Conclusion
On the basis of literature about strikes and labour unions it is possible to put forward a
number of hypotheses about the influence of mainly demographic indicators of the
possibility that workers will go on strike. If we compare strikers to non-strikers during
the 1914 strike at the Leiden Cotton Company following these hypotheses, we may
conclude as follows.
Of the eight hypotheses four were falsified in the analysis. These conclusions are
of course provisional and only valid for the 1914 strike at LKM in Leiden, the
Netherlands. If we construct a profile of the 1914 striker at LKM it is of a man who
6

Normale verhoging naar 9,75, maar "wegens niet staken 10,-"

131

earned a little more than non-strikers before the strike and was semi-skilled. He was
married and young, while he and his wife had few children. The striker acted regardless
of his religious beliefs and was not very loyal to the company.
The results of course only show tendencies and are not absolute, because the
calculations on which they are based always show values somewhere between falsified
and not falsified, but never 100 per cent falsification or not. The results are also only an
indication of the possibility that certain workers can be expected to join a strike.
Because workers from a century ago cannot be asked about their motives for joining or
abstaining from the 1914 strike, we will never be able to answer the question of why
some individuals joined and others did not. However, we did make some progress.
From international literature (Kerr and Siegel, 1954: 209-210) and from my own
research on Dutch strikes (Van der Velden, 2000: 195) it is clear that textile workers
showed an average or medium-high propensity to strike during the period under
research. In the Netherlands during the period from 1900 to 1940 there was a tendency
for workers in bigger companies to be more prone to strike than workers from smaller
companies, but the textile industry was an exception to this rule (Van der Velden, 2000:
204-205). From these two general conclusions we could anticipate that workers at the
biggest textile company in Leiden would strike more readily than workers from other
industries. They actually did, which is shown by the fact that LKM was the most strikeprone company in Leiden (Van der Velden, 2002). The research in this chapter into the
demographic characteristics of the workers at LKM shows the probability as to why
some of them joined the strikes while others did not. However, the results remain
probabilities and we can never be certain about the willingness of individuals.
Table 5. Testing the hypotheses
1

Hypothesis
Young workers are more strike prone
than the elder employees.

Single workers are more strike prone than


married workers.

Workers from big families are less


willing to strike than those from small
families or with no children.
Protestant workers are more strike prone
than Roman Catholic workers.
Skilled workers are less willing to strike
than their unskilled colleagues.

4
5
6
7

Workers with a long career at the


company are less strike prone than more
recent employees.
Semi-proletarian workers are less strike
prone than real proletarians.
Higher paid workers are less strike prone
than workers who earn lower wages.

Result
There is a small tendency that
strikers were on average younger
than non-strikers.
Married workers were slightly
more strike prone than single
workers.
The bigger the family, the less
willing workers were to strike.

Falsified
No

Religion did not matter.

Yes

Most strikers were semi-skilled.


The non-strikers were mainly
unskilled or skilled.
Workers with a long career at
LKM were less strike prone.

No

The most proletarianised workers


and those who were self-employed
at some time show roughly the
same propensity to strike.
Strikers earned higher wages
before the strike than non-strikers.

Yes

Yes
No

No

Yes

132

The results of the LKM research open the way, however, to obtain more general
answers about groups. Answers are not able to be obtained by using aggregated strike
data, or even data about the strike behaviour of workers from certain regions or
professions. The way to get such answers is by doing more research in the manner
shown in this chapter (see also the chapter by Maria Bergman in this book). To get a
better view of the question as to whether or not we can explain strike behaviour from
the life course of workers, more research needs to be carried out. Luckily, there are still
hundreds of company archives waiting to be exploited by historians who can build
databases with micro data. After this work is done, we can get a better answer to the
question Why do workers strike?.

133

References
Dam van Isselt, E.W. van (1914), Over het verband tusschen werkstakingen,
werkloosheid en Prijzen, De Economist, pp. 19.
Edwards, P.K. (1977), A critique of the Kerr-Siegel hypothesis of strikes and the
isolated mass: a study of the falsification of sociological knowledge, The sociological
review no. 25, pp. 551574.
Falk, William W., Michael D. Grimes and George F. Lord (1982), Professionalism and
Conflict in a Bureaucratic Setting: The Case of a Teachers' Strike, Social Problems,
vol. 29, No. 5, pp. 551560.
Franzosi, Roberto (1982), One hundred years of strike statistics: data, methodology,
and theoretical issues in quantitative strike research, CRSO Working paper, no. 257,
Michigan.
--------------------- (1995), The puzzle of strikes. Class and state strategies in post-war
Italy, New York, Cambridge University Press.
Gallagher, Daniel G. And George Strauss (19910, Union membership attitudes and
participation, http://escholarship.org/uc/item/0571x3vf
Granovetter, M (1978), Threshold models of collective behavior, American journal of
sociology 83, pp. 14201443.
Griffin, John Ignatius (1939), Strikes. A study in quantitative economics, New York.
Hackett Fischer, David (1971), Historians fallacies. Towards a logic of historical
thought, London, Routledge and Keagan Paul.
Heerma van Voss, A.F. (1987), Van CAOs tot zonden; 330 Twentse textielarbeiders in
1920 over hun drijfveren ondervraagd, Textielhistorische Bijdragen vol. 27, pp. 31-52
ILO (1926), Methods of compiling statistics of industrial disputes, Geneva.
Kelly, J. (1997), Long waves in industrial relations: mobilization and countermobilization in historical perspective, Historical Studies in Industrial Relations, no. 4,
pp. 335.
Kerr, Clark and Abraham Siegel (1954), The interindustry propensity to strike - an
international comparison, in Industrial Conflict, ed. Arthur Kornhauser et al., New
York.
Klein, Marian van der (2003), The Widows of the Gasworks: Gendered Path
Dependency and the Early Dutch Welfare State, Social Politics: International Studies
in Gender, State and Society, vol. 10, no. 1, pp. 122.

134

Keuning, H.J. (1970), Het Nederlandse volk in zijn woongebied. Hoofdlijnen van een
economische en sociale geografie van Nederland, The Hague.
Klandermans, B. (2004), The demand and supply of participation: social psychological
correlates of participation in social movements, in The Blackwell companion to social
movements, ed. D.A. Snow et al., Oxford: Blackwell, pp. 360-379
Knowles (1960), Strike-proneness and its determinants in Industrial Conflict, eds. W.
Galenson and S.M. Lipset, New York.
Meyer, Maximilian (1907), Statistik der Streiks und Aussperrungen im In- und
Auslande, Leipzig.
Rist, C. (1912), Relation entre les variations annuelles du chmage, des grves et
des prix, Revue dEconomie Politique, vol. 26, pp. 748-758.
Silver, Beverly J. (2003), Forces of labour. Workers movements and globalization
since 1870, Cambridge, Cambridge University press.
Tijn, Th. Van (1976), A contribution to the scientific study of the history of trade
unions, International Review of Social History, vol. XXI, pp. 212239.
Velden, Sjaak van der (2000), Stakingen in Nederland. Arbeidersstrijd 1830-1995,
Stichting Beheer IISG/NIWI, Amsterdam.
---------------------------- 2001, Rijken en armen ontmoeten elkanderen. Een laat
achttiende eeuws geschrift over werkloosheid, Onvoltooid Verleden 12, pp. 3135.
---------------------------- (2002), Stakingen in Leiden 1868-2000, Jaarboek Dirk van
Eck-Stichting, pp. 172191.
---------------------------- (2004), Werknemers in actie. Twee eeuwen stakingen,
bedrijfsbezettingen en andere acties in Nederland, Amsterdam: Aksant.
Verbong, G. (1987), De Katoendrukkerij
Textielhistorische Bijdragen vol. 27, pp. 19-30

De

Heyder

&

Co.

1836-1846,

Wiegersma, S. (n.d.), De wereld der beroepen. Beschrijvingen van beroepen in de


hedendaagse samenleving, Haarlem

135

Appendix
Table 1. Strikes at LKM, 1882-1922

Date

Profession

Number of strikers Duration Strike days Cause


Total

57

male

female

1882, May 19 Printing

75 132

396

1882, June 29 Weaving

300

12

3300

1882, July 25 Printing

91

91

0.4

85

1882,
Printing
September 20
1882,
Printing
September 20

150
71

71

11

639

1882,
Printing
September 24

150

150

14

4500

62 212

1895, June 13 Weaving

400 400

1600

1895, June 19 Weaving

400 400

15

5200

23 24

56

1144

16

32

88
11
55

6
100

55
4154

1902, August Printing


4
1907,
Washing
December 12
1914, April 27 Other
1915, April 14 Weaving
1919,
Printing
September
1919,
Printing
November
1920
Printing
1922

Weaving

1
16

result

265

133

30210

Wage
reduction
Wage
reduction
Solidarity with
sacked
spokesman
Wage
reduction
These strikers
were locked
out
Solidarity with
locked out
male
colleagues
Wage
reduction,
Solidarity with
suspended
colleague
Bad quality of
yarn
Wage rise

Won

Bad quality of
yarn
Wage rise
Wage
Wage rise

Lost

Lost
Lost
Lost
Lost
Lost

Settled

Won
Lost

Settled
Lost
Won

Reduction of Settled
working hours
Other
Unknow
n
Wage
Lost
reduction

Source: https://collab.iisg.nl/web/labourconflicts/search-database
Note: This online database was built during a more than ten year long search of archives, magazines,
research reports, books and official data collected by Statistics Netherlands.

136

137

Beyond the average and the aggregate: Researching strikes in


Canada
Linda Briskin1
1. Introduction
Using Canadian data, this chapter considers two strategies for doing research on strikes. Part
One explores patterns of militancy using the data from Human Resources and Skills
Development Canada (HRSDC) on the 23,944 stoppages in Canada between 1960 and 2004.
It considers how the data set has traditionally been used, and based on interviews with the
provincial correspondents, it discusses how the data are collected. It also presents three reinterpretations of the data: on strike duration, strike size and strikes for first collective
agreements. Part One also highlights buried data on labour disputes which reveal
characteristics of strikers, rather than strikes in government surveys from Statistics Canada -the Labour Force Survey (LFS), the Survey of Labour and Income Dynamics (SLID), and the
Workplace and Employee Survey (WES).
Part Two explores the use of newspaper archives to enhance work stoppage microdata using two Canadian examples: the eight-month Mini-Skools strike in the 1980s, and the
three-month strike of women garment workers at Puretex in the 1970s. The narratives
demonstrate how a qualitative measure can enrich quantitative data, offer not only a window
on the lived experience and texture of strikes, and provide access to the voices of strikers.
The chapter is framed with a labour militancy perspective which offers an alternative
to the employer focus on time lost, the government concern to measure the economic impact
of stoppages, and the scholarly emphasis on strike determinants (Briskin 2006). A focus on
labour militancy supports re-interpretations which help make visible the significance of
stoppages for workers, unions and communities. It begins with similar assumptions to Godard
(2005: 340-1, also 1992) who argues for a strikes as collective voice rather than what he
calls the strikes as mistakes approach:
[S]trikes may indeed appear to be irrational from an economist's point of view. But
from the viewpoint of the parties themselves (especially workers and their
representatives), the decision to strike often involves a much broader rationality, one
which involves competing values, principles, and fairness beliefs, and often reflects
underlying sources of discontent in the workplace [...] [S]triking serves as primary
mechanism by which workers can voice this distrust and resentment collectively [...]
1

I would like to thank Donna Bernardo for her research on the provincial work stoppage correspondents;
Walter Geisbrecht, the data librarian at York for his help in finding sources; Leah Vosko from the Gender
and Work Database (www.genderwork.ca) for reading an earlier draft of this paper; Rachel Hurst for her
work in the newspaper archives, and Kristine Klement whose work on the HRSDC data has been
invaluable.
This research was partly funded by the Social Sciences and Humanities Research Council of Canada
(SSHRC) project on Restructuring Work and Labour in the New Economy, Initiative on the New
Economy (INE), housed in the Centre for Research on Work and Study (CRWS), York University. This
research has also been funded by the Small Grants Programme from SSHRC and the Faculty of Arts at
York University.

138

[T]o view strikes only as mistakes in negotiations is too narrow. Rather strikes should
be viewed, first and foremost, as mechanisms of collective voice, serving as a
means by which workers can collectively express discontent and distrust.

In emphasizing the broader rationality that inspires strike action by workers, Godards
approach underscores the reality that experiences of strikes differ significantly for
management and for workers.2
Although this chapter focuses on strikes, it does not assume that strikes are the only
form of labour militancy. Hebdon (2005) maps other forms of labour militancy. In his
discussion of workplace conflict, he distinguishes among covert collective actions (such as
sick-outs, slow-downs and work-to-rule), other collective actions such as claims of unfair
labour practices, and individual forms of militancy around grievances. Further, this chapter
starts from a broad distinction among labour, union and worker militancies (Briskin 2007).
Worker militancy speaks to the collective organization and resistance among non-unionized
and often marginalized workers, many of whom are women and workers of colour. Worker
militancies may be of increasing importance given the transformations wrought by
restructured labour markets. Union militancy focuses on the politics of unions themselves.
2. Canadian work stoppages data3
Although in-depth qualitative accounts of particular struggles are likely the most
illuminating studies of strikes, Part One considers what can be learned about the overall
patterns of Canadian labour militancy by mining the work stoppage statistics. Three
examples are considered: strike duration, strike size and strikes for first collective
agreements. This approach reveals the potential of the work stoppages data from Human
Resources and Social Development Canada (HRSDC) to highlight multiple forms of
labour militancy, and to enrich the picture that emerges from commonly-used aggregate
and average data. As part of re-examining the HRSDC work stoppage data from a
labour militancy perspective, the data sources are considered through interviews with
the provincial correspondents who generate them.
Data on every work stoppage in Canada is collected by the Workplace Information
Directorate of HRSDC. Work stoppages include both strikes and lockouts (although the
variable for lockouts was only added in 1976) which are a minimum of half a day in length
and involve ten or more person-days lost (PDL). Person-days lost (previously mandays and
sometimes referred to as time lost) are generally the duration in working days4 multiplied
by the number of workers involved. Workers indirectly affected, such as those laid off as a
result of a work stoppage, are not included in the data.
In the current Work Stoppages manual, a strike is defined as a concerted work
stoppage, by one or more groups of workers, aimed at forcing an employer to acquiesce to
the groups demands. Strikes are most commonly the result of a labour dispute between a
2

In elaborating his collective voice perspective, Godard relied on surveys completed by senior
management officials involved in industrial relations in the unionized Canadian companies he studied.
Union officials were not surveyed (1992: 164-165).
3
This section draws on some material from Briskin (2006).
4 The data distinguish between calendar days lost and working days lost. Calendar days lost include
weekends and holidays while working days are limited to those in which the establishment would
normally be in operation (five days a week). Duration figures in this paper use working days lost, given
that it is the variable used by HRSDC to calculate person-days lost (Renaud et al. 2005).

139

group of employees and their employer (Renaud et al. 2005: 3). A lockout is a work
stoppage declared by an employer or group of employers where negotiations concerning
wages or working conditions have not been able to bring about an agreement (3). Although
strikes and lockouts are coded differently in the HRSDC data and can be disaggregated, as a
result of the permeability between strike and lockout and the difficulty distinguishing
between them, the HRSDC coding for lockout is used if the stoppage involved only a
lockout or if both a strike and a lockout occurred. This means that strike and lockout are
not mutually exclusive categories in the data. Given this cross over, tables in this chapter
include data on both strikes and lockouts.5
Franzosi (1989) notes the problems with the reliance on official strike statistics:
scholars' almost exclusive reliance on official strike statistics, which convey only limited
information, has prevented them from investigating some important basic questions about
strikes (348). This section also suggests that certain aspects of strikes have been neglected,
although unlike Franzosi who is interested in strike determinants, it focuses on highlighting a
quantitative mapping of strike experience from the point of view of workers.
The full work stoppage data set from HRSDC permits disaggregated manipulation,
the importance of which Franzosi as well as others (Gramm 1986, for example) have
stressed.6 Since major exclusions do not exist in Canada, the data offer many possibilities for
illuminating patterns of militancy, unlike the official data available in many countries. For
example, in contrast to Canada where data on strikes involving ten or more person-days are
included in official statistics, in the United States, since 1982, only stoppages involving 1000
or more workers are included (Gunderson et al. 2005: 347-8).
Since work stoppages data focus on the characteristics of strikes, they reveal little
or nothing about those on strike. In Canada, major government surveys contain buried
data that highlight the demographic profile of those who go on strike, through a variable
on absence from work: one of the reasons for absence from work is strike or lockout.
Although the viability of these data is limited by sample size, they do suggest another
route for developing profiles of those who go on strike. Little published research exists on
these data (Briskin with Klement 2004), although some research has looked at the more
generic issue of absence from work (Anon 2009) .
The Labour Force Survey (LFS) collects data on the labour market activities and
demographic characteristics of the working age population in Canada. The LFS is a monthly
household survey of a representative sample of individuals. Since July 1995, 54,000
households have been included. In one section, respondents are asked to select the main
reason for an absence from work from a list which includes illness or disability, personal or
family responsibilities, maternity leave, vacation, work schedule as well as labour dispute.
LFS tables make visible some of the demographic characteristics of those involved in labour
disputes: age, education, type of economic family (dual earner, single earner, single parent),
hourly earnings, and employment form (part time/full time and private and public sector).
The Survey of Labour and Income Dynamics (SLID) captures longitudinal changes in
the economic well-being of individuals and families, and the determinants of labour market
5 Short periodic reports on work stoppages are published in the Workplace Bulletin at
http://www.hrsdc.gc.ca/eng/labour/labour_relations/info_analysis/index.shtml. An interactive database
provides the number of strikes and lockouts which have occurred in Canada since 1976.
http://www.hrsdc.gc.ca/eng/labour/labour_relations/info_analysis/work_stoppages/index.shtml.
6
The author has negotiated full access to the records of all Canadian stoppages from 1946 to 2004 and
wishes to thank the Workplace Information Directorate of HRSDC for providing the microdata. All HRSDC
data quoted in this paper are from the work stoppage data unless otherwise specified.

140

and income changes. The samples from SLID are selected from the monthly LFS and include
roughly 15,000 households. SLID also directs its attention to race, ethnicity and immigrant
status, glaring absences in the LFS. Like the LFS, the small sample size limits the value of
SLID data on labour disputes compounded by the fact that SLID data does not capture strikes
of less than five workdays.
The Workplace and Employee Survey (WES) emphasizes human resource practices,
labour requirements and the interaction of employers and employees, and includes questions
about labour actions. The WES differentiates work-to-rule, work slowdown, strikes, lockouts
and other labour-related actions, and offers a profile of the striker in the worker questionnaire,
and a profile of the firm in the employer questionnaire. Employers in public administration
are excluded so whatever picture emerges from WES data highlights private sector
businesses. The 1999 inaugural WES sampled 6350 locations and interviewed employers and
about 24,600 workers in these businesses.7
Although the value of all three surveys is limited by sample size, the inclusion of
labour dispute as a variable points to another important source of data for strike research.
Furthermore, as researchers make active use of this variable, it may be possible to encourage
statistical agencies to enrich the data by expanding what questions are asked.
Strikes in Canada
Workers have gone on strike to improve the conditions of and remuneration for their work,
and to defend their rights to union protection. They have used the strike weapon to resist not
only employer aggression but also government policy. Undoubtedly Canadian workers have
been militant. HRSDC records 23,944 work stoppages between 1960 and 2004.
Generally, industrial relations specialists identify the following trends in Canadian
strike activity: moderate until the mid-1960s, extremely high levels from 1970 to 1981,
moderate and declining levels throughout the 1980s, and a sharp drop in the 1990s and into
the 2000s (Gunderson et al. 2005: 348). In the mid-1960s strike activity begins to rise; in fact,
although 1966 is not the year of the most strikes, it is the year which marks the beginning of a
dramatic increase in person-days lost to the economy as a percentage of working time: .34%
of working time compared to .17% of working time in 1965 (Peirce and Bentham 2007: 304).
To a great extent, industrial relations scholars and state agencies have been concerned
with the relative degree of overall strike activity in the economy(Gunderson et al. 2005:
348). Time lost to strikes and lockouts has always attracted widespread attention because of
the economic and social upheavals that often accompany industrial disputes. Given increasing
economic globalization and trade liberalization, the interest appears to be gaining strength
since international differences can influence corporate decisions on plant or office location
(Akyeampong 2006a: 5).
Canadas poor strike record by international standards is often highlighted in the
mainstream industrial relations literature as a possible contributor to its poor productivity
performance and as a possible concern to foreign investors and importers(Gunderson et al.
2005: 365). Further, as Gunderson et al. (2005: 353) acknowledge: the macroeconomic
measurement of strikes as lost work time necessitates a view of strikes from an employer
perspective.
7

Guide to the Labour Force Survey is available at http://www.statcan.gc.ca/pub/71-526-x/71-526x2007001-eng.pdf. For a guide to the SLID, see http://www.statcan.gc.ca/bsolc/olc-cel/olccel?lang=eng&catno=75F0011X. For the WES, http://www.statcan.gc.ca/cgibin/imdb/p2SV.pl?Function=getSurvey&SDDS=2615&lang=en&db=imdb&adm=8&dis=2.

141

Yet Peirce (2003) points out that strikes have never cost the country as much as one
percent of total working time, even in tumultuous years like 1919 (marked by the Winnipeg
General Strike) and 1976 (marked by a nation-wide Day of Protest against federal wage
controls) (338).8 Godard also recognizes the significance of worker resistance:
while it may be argued that Canada's level of strike activity is too high and that
there are ways to reduce it without violating workers' rights, Canada's higher
level of strike activity may not only be expected, it may also not be entirely
unhealthy. Indeed, if one believes that underlying conflicts are fundamental to
labour-management relations, then a substantial decline in strike activity might
even be considered a worrying development (2005: 337).

In order to measure time lost, averages and aggregates have been highlighted, especially in
industrial relations textbooks, such as the average number of workers involved per strike,
the average days lost per worker on strike, and aggregate data such as the person-days lost,
particularly as a percentage of working time (Gunderson et al. 2005; Peirce and Bentham
2007). Data on person-days can be used to provide a common denominator to facilitate
comparisons across jurisdiction, industry, sector, and even across countries. However,
Peirce and Bentham (2007: 306) note some difficulties with aggregate data, especially
comparing strike rates over time. Since some of the shifts are a result of increase in the size
of the labour force, in union density, and the extension of the right to strike to public sector
workers, it is difficult, even using person-days lost and the percentage of estimated working
time, to accurately assess the data in aggregate terms.
It is also the case that removing a key strike from the aggregate data can
significantly change the overall patterns. For example, in 1976, a National Day of Protest
against the introduction of wage and price controls lasted only one day but involved
830,000 workers, 56.3% of all workers involved in stoppages in that year. Similarly,
removing a single strike involving a large number of person-days lost from the aggregate
data impacts the stoppages profile for that year. In 1980, almost 10 million person-days
were lost in Canada. Quebec teachers went on strike for twenty working days. In this single
strike of 75,500 teachers, only one of the 952 strikes that year, 1,064,500 person-days were
lost, 10.8% of total person-days lost in that year.
Aggregate data, then, can be problematic. This chapter suggests that moving away
from these average and aggregate figures increases the visibility of workers strike
experience. For example, it is interesting to compare the measure of person-days lost as a
percentage of working time, which is widely offered in overview discussions on measuring
strike activity, with the percentage of employed workers and the actual numbers of workers
on strike in any given year, as offered in Table 1.

Peirce (2003) also puts forward a persuasive argument that the higher strike rate is completely
understandable given the practices, policies and legislation around labour disputes. In contrast to Canada,
the highly centralized bargaining system in Germany means fewer sets of negotiations; more experienced
negotiators; collective agreements which cover fewer issues; and union access to financial information
about company performance (361-2). He concludes: Arguably, Canadian grievance arrangements have
helped make strikes longer and more bitter (363).

142

Table 1: Strikes, Workdays and Workers in Canada, 1960-2004

Start Year

of
Strikes

Percent of
all Strikes

Number of
Strikers

Percent
of NonAgricultu
ral Paid
Workers

1960-1964

1,474

6.2

372,311

7.8

32,799

4.2

1965-1969

2,685

11.2

1,305,427

22.2

59,326

7.6

1970-1974

3,457

14.4

2,072,328

29.4

87,362

11.1

1975-1979

4,755

19.9

2,944,937

31.0

139,688

17.8

1980-1984

3,735

15.6

1,690,340

17.2

135,454

17.3

1985-1989

3,049

12.7

1,823,849

17.0

121,163

15.4

1990-1994

1,906

8.0

799,642

6.9

84,548

10.8

1995-1999

1,507

6.3

1,044,438

8.8

68,264

8.7

2000-2004

1,376

5.7

816,067

6.1

56,504

7.2

Total

23,944

100.0

12,869,338

--

785,108

100.0

Number

Notes:

Number of
Working
days lost 5

Percent of
all working
days lost

1. The data on work stoppages include strikes and lockouts which last ten or more person-days. Person-days
are calculated by multiplying the number of workers by the number of work days lost.

2. Non-agricultural paid employment and labour force data are from Statistics Canadas Labour Force Survey

(LFS). No survey was conducted in 1979. Statistics Canada only began including union status in LFS data
in 1997 so the number of workers in this table includes all part time and full time employees whether or not
they are unionized. Although many of these workers could not go on 'legal' strike, the HRSDC data do
include some 'strikes' by unorganized workers.
3. Number of strikers includes all workers involved in work stoppages, including lockouts.
4. Stoppages which continue from one year to the next are counted only once, in the year they started.
5. Number of working days lost includes days lost to both strikes and lockouts.

Source: Work Stoppage Data, Workplace Information Directorate, Human Resources and Social
Development Canada

This Table demonstrates the decline in strike activity: in the number of strikes, strikers, and
working days lost. The detailed data (not available on the Table) show the high point for
strike frequency between 1974 and 1981. The 1990s witnessed a relative decline in the
number of working days lost, but the actual number of strikers varies more widely over the
whole time period. The highest percentage of worker involvement in strikes was in 1976
when strikes involved 18% of all employees. Since 1999, only about 1% of employees have
been on strike, although as Akyeampong (2006a) notes, 2004 sees a moderate increase to
1.8%. While this might seem a somewhat insignificant percentage, it amounts to more than
250,000 additional workers on strike.9 And even in the period of 2000 to 2004, 816,067
workers were on strike. The shift from person-days lost as a percentage of working time to the
numbers of employed workers on strike helps to embody the strike experience.
9

Akyeampong (2006a: 20) also notes the increase in person-days lost from 1.7 million in 2003 to 4.1
million in 2005.

143

Lockouts in Canada
HRSDC began collecting lockout data in 1976. From 1976 to 2004, 1839 lockouts
12% of all stoppages involved 7.4% of all workers who engaged in stoppages, and
represented almost 20% of workdays lost. The high point for lockouts was the period
from 1983 to 1986; the 473 lockouts during these years represent almost 26% of all
lockouts in the entire period. For an overview, see Table 2.
Table 2: Lockouts, Workdays and Workers in Canada, 1976-20041

Stoppages
Num
Percent
ber
of total
of
Lock stoppage
s
outs

Number
of
Workers

Workdays

Workdays lost
to Lockouts

Percent of
total
workdays

Workers
involved in
Lockouts

Percent
of total
workers

1976-1979

253

6.9

11,837

11.3

85,919

3.5

1980-1984

521

13.9

29,674

21.9

226,545

13.4

1985-1989

393

12.9

23,435

19.3

179,506

9.8

1990-1994

251

13.2

20,018

23.7

28,591

3.6

1995-1999

196

13.0

16,562

24.3

77,618

7.4

2000-2004

225

16.4

11,694

20.7

38,400

4.7

Total

1,839

12.1

113,220

19.8

636,579

7.4

Note:
1. HRSDC began collecting data on lockouts in 1976. A lockout is coded by HRSDC "if the stoppage involved only a
lockout or if both a strike and lockout occurred." From the Training Manual: Work Stoppages.

Source: Work Stoppage Data, Workplace Information Directorate, Human Resources and Social
Development Canada

Lockouts tend to be longer than other strikes. Between 1976 and 2004, 33.6% of all
strikes settled in less than one week compared to only 15.8% of lockouts. And despite
the relatively small number of lockouts overall, 30.4% of all stoppages lasting longer
than one year (84/276) were lockouts. Overall, lockouts involve fewer workers;
however, seven lockouts involved 10,000 workers or more.
Although one might expect that lockouts would occur during the re-negotiation
of agreements, thirty-six lockouts occurred during the term of the agreement, eight of
which were in the public sector. The data on lockouts by industry and sector are also
revealing. Although 88% of lockouts occur in the private sector, from 1976 to 2004,
there were 220 lockouts in the public sector, 162 of them clustered in three key
industries: Educational Services (fifty lockouts); Health Care and Social Assistance
(forty-four lockouts); and Public Administration (sixty-eight lockouts). In 1980 alone,

144

twenty-two public sector lockouts involved 50% of all workers locked out between
1976 and 2004.
To fully understand labour militancies, they need to be set against patterns of
employer aggression, that is, pro-active initiatives on the part of employers to
undermine and often prevent the functioning of the union-management relationship.10
Lockouts are one form of employer aggression, especially when employers continue to
operate with replacement workers. Forms of employer aggression also include pursuing
profits regardless of the effects on workers, families, communities, and countries;
sabotaging the functioning of the union-management relationship; limiting worker input
into and control over the labour process; and increasing employment instability by
undermining standard and secure jobs in favour of more precariousness (Briskin 2006).
Briskin (2005) examines the HRSDC data on strike issues related to employer
aggression (contract violation, disciplinary action, failure to negotiate, delay in
negotiations, etc), and differentiates employer aggression from employer resistance to
the introduction of a union, what Ewing et al. (2003) call union avoidance.
How the work stoppage data are collected
As part of re-examining the HRSDC work stoppage data from a labour militancy
perspective, it is revealing to explore how these data are collected, that is, whose voices
are reflected in the data. All statistical data are necessarily limited in their accuracy and
what they can represent. Interviews were conducted with many of the provincial
correspondents who collect the information for HRSDC in order to assess the validity
and viability of the HRSDC data.11 These interviews underscore the political nature of
data collection (what is seen to be germane and not), data presentation (what is made
visible and what is not), and data sources (whose voices are heard). They also highlight
the importance of crossing methodological borders to juxtapose statistical data with
qualitative sources.
In Canada, neither employers nor unions is required to record information about
work stoppages. Instead, HRSDC works with a correspondent in each provincial labour
ministry who collects information about work stoppages. Although HRSDC indicated
that many of the choices around coding of the work stoppage data were made by these
correspondents, conversations with correspondents in seven provinces revealed that
none was involved in coding. In all cases but one, they had never seen the coding lists.
One correspondent who collected the data for years said in reaction to interview
questions about coding choices: I was kind of taken aback to see the emphasis on
10

Labour militancies also need to be situated in the context of the growing intervention of the state into
the management of labour relations, especially in the public sector. Statutory incomes policies are
complemented by back-to-work legislation and the increased designation of public sector workers as
essential, thereby removing their right to strike (Panitch and Swartz 2003). In Canada 170 pieces of
legislation have restricted, suspended or denied collective bargaining rights (Fudge and Brewin 2005).
For more information about the campaign to defend free collective bargaining, go to
http://www.labourrights.ca.
11
Labour relations is regulated provincially, and so each of the ten Canadian provinces and three
territories has a distinct system. The author had contact with correspondents from seven provinces: British
Columbia, Alberta, Manitoba, Saskatchewan, Ontario, New Brunswick and Prince Edward Island and
appreciates the time they spent on this project. Many of then were very reluctant to speak to the author;
some were stopped by their supervisors. The most informative respondent had recently retired.
Correspondents filled out a survey questionnaire; in some cases they participated in a telephone interview. In
all cases, the correspondents requested anonymity. All quotations are from these interviews. Since HRSDC
would not provide the list of correspondents, each provincial ministry was called by a research assistant.

145

coding because this is stuff we never did. Nevertheless, conversations with these
correspondents did provide some interesting background to, and insight into the
HRSDC data.
Correspondents were asked how the work stoppage data are collected. Questions
included: How do you find out about the occurrence of a work stoppage? The number
of workers involved? The issues involved? Accessing information about the occurrence
of a work stoppage depends on a variety of factors, not the least of which are the
regulatory processes embedded in provincial labour relations acts. For those provinces
like Ontario where conciliation is mandatory, the Minister of Labour must issue a notice
that no settlement could be effected and no Conciliation Board will be appointed
(colloquially known as a no board report) prior to a strike, or Alberta where notice of
strike or lockout must be given to their mediation services, or Prince Edward Island
where parties must participate in conciliation prior to a strike/lockout, the
correspondents work closely with the conciliation and mediation services (although the
actual conciliator reports are confidential). For some provinces like British Columbia,
such a process is not mandatory. The trouble is the Labour Relations Board didn't
really know officially about any strike unless somebody went to them. And you didn't
have to go to them. So that meant that there was a lot of strikes that they never dealt
with because they got resolved in another way (Work Stoppages Correspondent). In
such cases, the correspondent relies on provincially-gathered media summaries and on
employer and union web sites.
Some correspondents have regular contact with employers, and some with
unions, or use union websites. For example, in Ontario, The employer, and very rarely,
the union, are contacted directly to obtain detailed information regarding the following:
a) type of work stoppage, b) number of employees involved, c) start date and start time,
d) termination date and termination time, e) number of days out, f) number of work days
per week. Another correspondent reported: Most of the time I would start talking with
the union. It would depend what kind of a strike it was. But I found if it was a small
company, they really didn't want to waste any time talking to me. They were quite
annoyed whereas the union was pretty happy to talk [...] If I could phone both, I would.
I'd write down all the information I could get. In general, the correspondents did not
suggest that they received divergent information from union, employer and media. The
Saskatchewan correspondent indicated that we rely on our conciliators to provide us
with an unbiased version of events. The Manitoba correspondent indicated: For the
most part, there is consistency between the sources as to the issues at hand.
Most provinces rely, to some extent, on the media to alert the correspondent to the
fact of a stoppage, especially in those where a conciliation/mediation process is not
required. For example, in British Columbia, We had a communications department for
every Ministry and each communications department would go through the news sources
and put together a little email that had clippings related to our Ministry. The media
summary was pretty extensive [...] [although] some strikes are so small that the news media
don't pick them up. This correspondent did acknowledge that the media reports could
sometimes be misleading.12
12

Another part of HRSDC which deals with employment insurance has a different view on the media. The
following statement from the HRSDC Digest of Benefit Entitlement Principles which guides the assessment
of entitlements to employment insurance in the context of work stoppages is worth quoting. It is intriguing
that provincial governments often depend on the media to provide information about stoppages, while at the
same time, another part of the state apparatus is emphasizing the difficulties of relying on the media.

146

Interviews with the correspondents suggest that the data collection across
Canada is not standardized, a point confirmed by Rene Poulin from HRSDC (email
correspondence 29 Sept 2010). Significant variations emerged among the provinces. In
addition to the above examples which demonstrate differences in the sources of data,
not all provinces collect the same information about work stoppages. For example,
unlike Ontario, Alberta does not collect data on what they consider illegal work
stoppages.
Alberta excludes any strike that does not conform to the definition of legal
work stoppage as stated in the Labour Relations Code. A work stoppage can
only legally occur when a collective agreement is expired, after a mediator has
been appointed and after notice to strike or lockout has occurred... A wildcat
strike would likely be handled by the parties themselves without us even being
aware. If not handled by themselves, the Labour Relations Board would
legislate them back to work (Alberta Correspondent).13

Interviews with provincial correspondents who collect work stoppages information for
HRSDC shed light on both the limits and possibilities of the data set. Certainly,
understanding more about the source of the data and the collection process is a reminder
of the hidden qualitative and subjective aspects of statistics.
In the next section, other aspects of the data collection process come into focus through
discussions of strike size, duration, and first contract disputes. Although considerations of
strike duration (how long a strike lasts) and strike size (how many workers are
involved) have attracted extensive scholarly attention, little quantitative research is
available on the issue of strikes for first collective agreements. On all three issues, this
research finds that a labour militancy lens enriches the empirical mapping of strikes.
Certainly, re-configuring aggregates and averages illuminates workers experiences of
strikes.
Strike size: Number of workers
One of the real problems was trying to figure out how many people were
involved. And the hours. That could be an absolute nightmare. [M]ost of the
time the union was more helpful.... but they couldn't always tell me, especially
if you had an operation like a donut shop or something, where nobody's fulltime, and nobody's even sort of, regularly half-time or quarter-time. And there
you're really just doing guestimates. And there's just no way around it. Because
Strictly speaking, newspaper articles are no more than a written form of hearsay. Such published
information may in principle be considered but as with any other forms of information, its credibility must
be assessed at the time the decision is actually made. This assessment is especially important in the case of
labour disputes; by their very nature, such disputes involve contradictory claims by the parties involved
[Section 8.1.10 of the Digest]. This section of the Digest concludes with this contradictory statement: One
should not lose sight of the fact that the claimant bears the onus of proof to entitlement to benefits. For
example, with respect to exempting conditions, claimants or their representatives are required to refute all
unfavourable facts that appear in a news item or release [Section 8.1.10 of the Digest]. A discursive
analysis of 'unfavourable facts' from a 'hearsay' source is warranted. The Digest is available at
http://www.hrsdc.gc.ca/en/ei/digest/table_of_contents.shtml.
13
Given this definitive guideline, it is curious that during the period 1960 to 2004, data are recorded for
eighty wildcat strikes in Alberta. The author has not been able to ascertain from HRSDC the source of these
data.

147

no one's going to sit down and say 'well, Bob works three hours a week and
Janet works 24 and [...] And there were some strikes, for example, a
community services one, where it was impossible. You had a gazillion
employers that had these small operations, where they had, say, group homes
for people and there were two employees and you had hundreds of these things.
And they were not going to sit down and work out the hours [...] So there were
strikes where I ended up putting down: 'information is not available' [...] You
know, because partial information was useless [...] and misleading (Work
Stoppages Correspondent).

Although the HRSDC data report the number of workers involved in any given strike,
these comments make clear that collecting such information is no easy task, and what
the numbers actually mean remains somewhat opaque. HRSDC has confirmed (email
29 September 2010) that each province is calculating the number of workers
differently. There are no central guidelines, so each may be counting part time workers,
shift workers, or workers who work outside the work days (on weekend)
idiosyncratically. Furthermore, some provinces use full time equivalents (FTE) to
calculate person-days lost, that is, they aggregate several part-time workers into a single
worker. Measuring FTE rather than embodied workers and not specifying the number of
full-time and part-time workers involved in strikes is misleading. It certainly erases the
involvement of part-time workers in militancy, a group already marginalized in the
workplace.
Canadian work stoppages data would be significantly enriched by information
on work status (information already collected by correspondents) especially as labour
market restructuring increases precariousness and part-time work, and undermines the
standard work form. For example, Statistics Canada data (2008) show that 33% of fulltime workers are covered by a union compared to 24.3% of part-time workers; but the
work stoppages data do not take account of work status in the measurement of the
number of workers on strike.14
Nor do the Canadian data map those who cross picket lines and return to work
during a strike but only reference the maximum numbers of workers on strike at the
beginning of a stoppage (email 13 November 2009). Without this longitudinal tracking, the
data on workers on strike and person-days lost are distorted.
HRSDCs public reporting emphasizes major work stoppages, that is, those which
involve 500 or more employees. A number of researchers have highlighted problems with
limiting consideration to large strikes (Campolietti et al. 2005). For example, in their study
of US strikes, Skeels et al. (1988: 589) conclude:
Our evidence strongly suggests that conclusions about strikes in general
cannot safely be made on the basis of research investigating only strikes
above some given size cut-off. Strike samples thus selected do not appear to
meet the basic requirement of randomness, probably because bargaining unit
size has a systematic relationship to dispute issues, contract status,
geographic distribution, and industry. More simply put, our results indicate
that small strikes differ significantly from large strikes.

14

Work stoppages data would also be enriched by information on the gender of striking workers. Gender
patterns of union density have shifted dramatically. Since 2004, the unionization rate for women has been
slightly higher than for men (Statistics Canada 2009).

148

Like Skeels et al. (1988), Harrison and Stewart (1993) stress, in their study of the relation
between strike size and duration, the importance of considering all strikes and not only
large ones. In Canada, it is not surprising that between 1960 and 2004 strikes of over
500 workers account for a disproportionate percentage of person-days lost (69.3%).
However, the focus on large strikes means that 87.6% of strikes are not reported in any
detail in HRSDC publications. This invisiblizing of strikes involving fewer workers
underscores the states narrow interest in the economic impacts of strikes.
An alternative reading of the stoppages data from a labour militancy perspective
suggests a different profile of Canadian strikes. It counts actual workers on strike rather
than the aggregate reporting common in the industrial relations literature where strike
size is defined as the average number of workers involved per strike calculated as the
number of workers involved divided by the number of strikes (Gunderson et al. 2005:
351).
The breakdown for workplace size used in the Labour Force Survey (Statistics
Canada), that is, less than 20, 20-99, 100-500 and more than 500, a commonly used
standardization was first considered as a frame to examine the number of workers
involved in strikes. However, in order to make visible both very small strikes and very
large strikes, the following breakdown was used: 1-19, 20-50, 51-99, 100-250, 251-500,
501-1000, 1001-2500, 2501-9999, 10,000+.15
Table 3: Strike Size, Canada, 1960-2004
Number
workers

of

Number of
Stoppages

Percent

Cumulative
Percent

1-19

4220

17.6

17.6

20-50

5318

22.2

39.8

51-99

3799

15.9

55.7

100-250

5139

21.5

77.2

251-500

2495

10.4

87.6

501-1000

1458

6.1

93.7

1001-2500

890

3.7

97.4

2501-9999

469

2.0

99.3

10000+

156

0.7

100.0

Total

23944

100.0

Source: Work Stoppage Data, Workplace Information Directorate, Human Resources and Social
Development Canada

As Table 3 demonstrates, the more than 500 category represents 12.5% of all strikes. In
aggregated data of more than 500, very large strikes would remain buried. In fact, 6.4% of
15

Harrison and Stewart (1993) use the following breakdown: 1-24, 25-49, 50-74,75-99, 100-149, 150-99,
200-299, 300-499, 500-999, and 1000+ which makes visible small strikes but does not reveal much about
the range of very large strikes.

149

the more than 500 category involve more than 1000 workers: 890 strikes involved 10002500 workers, 469 involved 2501-9999 workers, and 156 strikes involved more than 10,000
workers, many of them in the public sector.
At the same time, it is noteworthy that between 1960 and 2004, 17.6% of strikes
involved 19 or fewer workers, 22.2% of strikes between 20-50 workers, and 15.9%
involved 51-99 workers. Over the period, then, from 1960 to 2004, 55.7% of strikes
involved less than 100 workers. Although it may be that strikes are more likely to occur in
large workplaces (see Godard 1992 and Gramm 1986),16 in Canada, the vast majority of
strikes actually occur in small workplaces. Expressions of militancy in small workplaces
suggest strategic directions for organizing campaigns to bolster dwindling union
membership. In 2008, 33% of all workers were employed in workplaces with fewer than 20
employees, and only 13% were unionized (Statistics Canada 2008).17
Sector and industry breakdowns by size are also revealing. In general, large strikes
tend to be in the public sector: 12.2% of public sector strikes involved more than 1000
workers compared to only 4.9% of private sector strikes during the period from 1960 to
2004. Some industries are more likely to have strikes involving few workers, for example,
Trade (81% of strikes involved less than 100 workers) and Finance, Insurance, Real Estate
and Leasing (85% of strikes). Educational Services has the lowest percentage of strikes
under 100 workers: only 33.7%.
Duration of strikes
Industrial relations experts often emphasize the average length of strikes. Gunderson et al.
(2005: 352) report that from 1901 to 1998, the average strike lasted 18.8 days, though it
dropped to 16.6 days in the 1980s, and 13.8 days in the 1990s. They point out that, with the
exceptions of both world wars, Canada has almost always had strikes of fairly long duration
compared to many countries. For Gunderson et al., the length of an average strike is
calculated, perhaps somewhat counter-intuitively, not by dividing the number of strikes by
the days lost but rather by calculating the average days lost per worker on strike. As pointed
out earlier, they acknowledge that the macroeconomic measurement of strikes as lost work
time necessitates a view of strikes from an employer perspective (353).
However, they also point out that calculating the average length of strikes by
dividing the number of strikes by the days lost highlights the point of view of individual
workers, and reveals that, in fact, strikes have been getting longer, not shorter (2005: 353).
This approach shows that, between 1960 and 2004, the average strike duration was about
thirty-three working days. However, from 1990, the average number of days began to rise
over forty. See Table 4.18

16

In their sample of 1363 Ontario strikes from 1984 to 1992, Campolietti et al. (2005: 620) found that
the odds of a strike for small bargaining units were statistically significantly smaller than for bargaining
units with more than 500 members.
17
Despite commonsense views which suggest that women workers are clustered in small workplaces,
30.6% of male workers are in workplaces of 20 or less compared to 34.6% of women. Only in these small
workplaces was union density higher for men than for women: 16.4% compared to 13%. For all other
workplace sizes, union density was higher for women (Akyeampong 2006b: 27-28).
18
Although the discussion of Gunderson et al. (2005) of average strike duration used calendar rather than
work days, a similar profile of increasing duration is demonstrated here using work days. However, as
Hebdon pointed out (email correspondence), calendar days are likely a measure closer to the experience
of those on strikes. Although workers may not picket on a weekend, they still experience themselves on
strike.

150

Table 4: Average Strike Duration, Canada, 1960-2004


Years
1960-1964
1965-1969
1970-1974
1975-1979
1980-1984
1985-1989
1990-1994
1995-1999
2000-2004
Total

Workdays1
22.3
22.1
25.3
29.4
36.3
39.7
44.4
45.3
41.1
32.8

Note:
1. The HRSDC data distinguish between work days and calendar days. "The days counted as working days are those
on which the establishment would normally be in operation (five days per week)" From The Training Manual: Work
Stoppages, 2005.

Source: Work Stoppage Data, Workplace Information Directorate, Human Resources and Social
Development Canada

The scholarly literature on duration has focussed on explaining or predicting strike


duration, in relation to many possible variables. Some research has considered the
relation between duration and the business cycle (for example, Harrison and Stewart,
1989), public policy relating to strikes (Gunderson and Melino 1990), or number of
unsettled issues at the start of the strike (Ondrich and Schnell 1993). Others have
developed multi-variable analysis which include attention to strike duration or severity
(for example, Gramm 1986; Godard 1992). This discussion tries to map what can be
learned about workerss experience of strike duration by disaggregating the strike data,
starting from the simple proposition that how long a strike lasts is critical to the workers
involved.
HRSDC specifies that all strikes of ten or more person-days are recorded in the
data. However, provinces vary in their attention to both very short and very long strikes.
One Work Stoppages Correspondent commented on data collection of very long strikes.
When a strike had gone on for two years, we just ended our recording of it
because we figured that there was just no point in going on. You know after two
years if a union was still picketing they weren't going to get anywhere. Nothing
was ever going to happen. And so, we stopped keeping records for things [...]
Quite often what happens in a situation like that is that the company closes down.
The company may not really be out of existence, they may open under another
name, but you got some union people picketing a site that doesn't really exist
anymore and so we decided it was pointless. And so, we didnt bother going on
phoning them. I mean, if we had phoned the union, they probably would have
said: 'Well, we're still out.' But so often these strikes involve twenty people, it
really didn't matter.

It is not so much that very long strikes don't matter. Rather, they may be very difficult
to settle. Gunderson et al. (2005: 342) describe the hazard rate, that is, the strike
settlement probabilities as the strike progresses:

151

Generally, as the strike progresses, the probability of settling the next day (the
conditional strike probability) declines, implying that the remaining life
expectancy of the strike actually increases as the strike progresses. Much of this
simply reflects the fact that the composition of the remaining strikes increasingly
consists of strikes that are hard to settle, the easy ones having been settled earlier
and dropped out of the sample.

Undoubtedly much can be learned about unions, stoppages and militancy by considering the
very long strikes in Canada. The HRSDC data from 1960 to 2004 record sixty-nine strikes
that lasted longer than two years, all remarkable instances of persistent militancy. Of these,
95.7% were in the private sector and 75.3% involved less than fifty workers. The fact that
30% were during the negotiation of a first agreement is also significant to understanding these
long strikes from the point of view of workers, especially given that only 13% of all stoppages
took place during first contract negotiations. Of the 307 private sector strikes lasting more
than one year, 19% involved less than twenty workers; in fact 14% involved ten workers or
less.
On the issue of very short strikes, one Correspondent noted: We had a policy of
ignoring a strike if it was less than half a day. It may be that from the point of
aggregate statistical data, strikes of few workers, short strikes or even very long strikes
do not matter. At the same time, as this correspondent also recognizes, the numbers on
strike or the length of the strike do not always reflect the significance of a strike to the
workers, to the union and to the community.
There were times when I thought, these numbers are useless. For example, we
had a nursess strike and the number of days lost was so small that I thought,
well, this doesn't make any sense. It was all over the media, everyone was up in
arms, and practically no time was lost because there's so many essential services
[...] [If] you're just looking at the numbers, this doesn't tell you what a huge
impact this strike had (Work Stoppages Correspondent).

Shifting from various average measures further highlights patterns of labour militancy not
revealed by the traditional aggregate figures. The HRSDC data show that 35.6% of strikes
between 1960 and 2004 lasted between one and five days; in fact, 21.3% lasted only one to
two days.19 These include political walkouts such as the 1976 Day of Protest around wage
and price controls (involving 830,000 workers), and the 1996 and 1997 Days of Protest in
Ontario. Equally interesting is the fact that 8.6% of strikes lasted seventeen to fifty-two weeks
and 1.4% more than one year. See Table 5.
Not surprisingly, there are significant industry and sector differences. For
example, in construction, 55% of strikes settled in five days or less; however, in trade
the figure is only 22%, and in finance, only 20.7%. In the private sector, 34% of strikes
settle within a work week; in the public sector, 43% of strikes settle in this short time
frame.

19

The high incidence of very short strikes may have been missed in some other research. For example
although Ondrich and Schnell (1993) focus on strike duration, their shortest duration category is one to
thirty days. Even in their relatively small sample of 320 US strikes from 1975 to 80, about 30% of strikes
fell in this short category; however, without disaggregation, it is not possible to know how many lasted
five days or less.

152

Table 5: Strike Duration, Canada, 1960-2004


Strike duration

Total number

Cumulative
Percent

Percent

1 week

8,527

35.6

1-2 days

5,103

21.3

21.3

3-5 days

3,424

14.3

35.6

>1 to 2 weeks (6-10 days)

2,968

12.4

48.0

>2 to 4 weeks (11-20 days)

3,475

14.5

62.5

>4 to 7 weeks (21-35 days)

2,965

12.4

74.9

>7 to 16 weeks (36-80 days)

3,651

15.2

90.2

>16 to 52 weeks (81-260 days)

2,020

8.4

98.6

338

1.4

100.0

23,944

100.0

>52 weeks (261+ days)


Total

Notes:
1.

2.

HRSDC data indicate both calendar days and working days. Calendar days refers to the number of
calendar days in the month, while "The days counted as working day are those on which the
establishment involved would normally be in operation (five days per week)". From the Training
Manual: Work Stoppages, 2005. The data presented here are based on 'working days'.
'1 week' includes stoppages which lasted less than or equal to one week, or 5 working days. '>1 to
2 weeks' includes stoppages which lasted longer than one week, but not longer than two weeks, or
6 to 10 working days.

Like the average duration figures, the disaggregated figures show that strike duration is
increasing. Between 1960 and 2004, 35.6% of strikes lasted less than one week;
however, in the period 2000 to 2004, only 25.8% of strikes were resolved within five
working days. And whereas over the whole period, only 15.2% of strikes lasted between
seven and sixteen weeks, between 2000 and 2004, the percentage had increased to 19%.
What patterns are visible when the data on duration and size are combined? In
their study of 1363 strikes in Ontario between 1984 and 1992, Campolietti et al. (2005)
found a relation between duration and bargaining unit size: Smaller bargaining units
were slower to settle strikes than were bargaining units with 500 or more members
(621). This finding holds true for the 23,944 work stoppages between 1960 and 2004.
Of all strikes, 35.6% settle within five days. The more workers involved, the greater the
likelihood of such a quick settlement: for example, 43% of strikes of 251 or more
workers settle within five days. In contrast, of strikes which last more than one year,
71.3% involve under fifty workers. Almost 49% of strikes which last between
seventeen and fifty-two weeks involve less than fifty workers. These numbers suggest
that some leverage comes with large collectivities of workers, although a full analysis
would need to take into account many other factors. Campolietti et al. (2005: 626) offer
the following hypotheses, although noteworthy for its absence is the possibility that
large collectivities have more bargaining power:

153

Our finding of longer strikes in smaller bargaining units is consistent with the notion of greater moral commitment to work by employees in small plants or units
[] . We also interpret this finding as supporting the industrial relations view
[] that workers in smaller units exhibit greater solidarity or cohesion which
would make them more likely to experience longer strikes than workers in larger
bargaining units. Similarly, smaller firms may be found in more competitive
industries and may have less pricing power than larger firms. Consequently, they
might be less likely to settle than larger firms once there is a strike, since settling
might adversely affect their financial position. These considerations about firm
behavior and the smaller bargaining units' greater solidarity might explain why
we see longer strikes in smaller firms.

Contract status: First agreements


Although Canadian workers have the formal right to be in a union, access to union
representation and recognition depends upon an adversarial certification process. Once
certified, the new union faces the challenge of negotiating a first collective agreement
with an often-resistant employer, and may turn to strike action.
HRSDC data record the contract status for each strike. The following options are
available: negotiation of first agreement, renegotiation of agreement, during term of
agreement, in other circumstances, and no signed agreement. Strikes during the
negotiation of first agreements represent 13.3% of the total, a relatively small and
perhaps insignificant percentage. As Gunderson et al. (2005: 358) point out:
Recognition or first-agreement strikes occur quite often; however, they do not involve
many workers and hence do not contribute much to the total person-days lost because of
strikes. Although these strikes only involved 1.9% of all workers on strike, this
translates into more than 238,000 workers, and 21% of total workdays lost to strikes.
See Table 6.
Table 6: First Agreement Strikes, Workers and Workdays, Canada, 19602004
1960-1964
1965-1969
1970-1974
1975-1979
1980-1984
1985-1989
1990-1994
1995-1999
2000-2004
Total

Strikes
n
396
380
388
456
530
441
293
171
138
3,193

Strikers
%
n
26.9
41,842
14.2
31,187
11.2
28,947
9.6
32,186
14.2
22,847
14.5
22,685
15.4
13,601
11.3
15,350
10.0
29,615
13.3
238,261

%
11.2
2.4
1.4
1.1
1.4
1.2
1.7
1.5
3.6

Workdays
n
12,348
16,264
18,000
25,554
32,486
25,494
16,239
11,536
7,918
1.9
165,839

%
37.6
27.4
20.6
18.3
24.0
21.0
19.2
16.9
14.0

21.1

Source: Work Stoppage Data, Workplace Information Directorate, Human Resources and Social
Development Canada

A labour militancy perspective on first contract strikes recognizes that such strikes
epitomize key moments of workers struggling for union representation. And unlike the
irrelevance to the aggregate data of a first contract strike of fifteen women which lasts
for many months, such a struggle would be very consequential to the women, their
political consciousness and the communities in which they live and work. A labour

154

militancy lens also highlights continuing employer resistance to union recognition.


Bentham (2002: 159) investigated the prevalence of employer resistance to union
certification applications in eight Canadian jurisdictions and found that employer
resistance was the norm with 80 percent of employers overtly and actively opposing
union certification applications. She concluded that such opposition impacts on the
probability of the parties establishing and sustaining a collective bargaining
relationship. Bentham does not touch on first contract strikes but her data offer
additional evidence of employer resistance.
The time series data on first contract strikes suggest continuing employer
resistance. In the early period of 1960 to 1964, 27% of all strikes were for first
contracts. The high point for first contract strikes was the period between 1980 and
1984 during which time 16.6% of all such strikes occurred. Numbers of first contract
strikes have declined over recent decades perhaps in part because of the availability of
first contract arbitration.20 However, in the decade between 1995 and 2004, 309 such
strikes occurred in Canada, each of which would offer up a story of persistence and
militancy in the face of considerable employer resistance. Furthermore, almost 30,000
workers were involved in 138 first contract strikes between 2000-2004, a number not
matched since the period of 1975 to 1979. Unlike percentage data which can minimize
the reality of strikes for workers, absolute numbers offer a profile more closely related
to worker experience. So 309 strikes may represent only 10% of strikes in this period,
but also represent 309 instances of workers collectively organizing and striking
Little interest has been paid to the phenomenon of first agreement strikes in the
quantitative mapping and measuring of strikes in the literature. Walker (1987) is one
exception. Yet much can be learned about labour militancy and employer resistance
from these often heroic struggles for a first collective agreement. A brief profile shows
that 71.4% of first contract strikes involved less than fifty workers (compared to 35% of
all other strikes). Workers were locked out in 13.4% of first contract strikes. Perhaps
surprisingly, 12.2% of first contract strikes were in the public sector although these
tended to settle more quickly than in the private sector: 32.7% in five days or less. For
all first contract strikes, only 23.2% settled within one week, and 18.2% lasted more
than sixteen weeks. For other strikes, 37.5% settled within one week and only 8.6%
lasted more than sixteen weeks. First contract strikes, then, often involve small groups
of workers struggling over long periods of time. In fact, 32% of all strikes which lasted
longer than one year were struggles for first agreements. See Table 7.

20

In 1974, British Columbia became the first jurisdiction to adopt first contract arbitration. Currently
seven jurisdictions covering 80 percent of the Canadian workforce provide coverage (Rose 2006: 200).
Rose points out that one of the purposes of such arbitration is to bring an end to a first contract work
stoppage. He reports that from 1974 to 2000, of 28,000 certifications granted in jurisdictions with first
contract arbitration, only 5.9% drew on such arbitration (201).

155

Table 7. Snapshot of a Few First Agreement Strikes Lasting Over One Year,
1984-2004.
From June 21, 1999 to March 31, 2002 (714 workdays), 46 workers (Fdration du commerce)
at Alimentation Picard in Quebec were on strike. An agreement was reached.
From Oct 14, 1988 to June 30, 1994 (1439 workdays), 32 workers (United Steel Workers) at
the Wittke Iron Works in Alberta were on strike. The strike was abandoned.
From Sept March 23, 1987 to June 30, 1990 (824 workdays), 7 workers (Service Employees
International Union) at the Royal Canadian Legion Branch #56 in Saskatchewan were on
strike. An agreement was reached.
From April 15, 1987 to April 27, 1988 (263 workdays), 276 workers (United Food and
Commercial Workers) at Zellers in Quebec were on strike. An agreement was reached.

Source: Work Stoppage Data, Workplace Information Directorate, HRSDC

3. Voices and textures of striking workers in newspaper archives


In the HRSDC data, the detailed (but not publicly available) record for each stoppage of ten
or more person-days lost contains a wealth of information: contract status, result, sector,
province, metro/city, industry code, jurisdiction, affiliation, union status (various, single,
unorganized), and information on lockouts and rotating strikes. Yet even this micro-data is
limited, and sometimes misleading.
Here is a brief comparison of the HRSDC record on a significant strike with the
abbreviated but qualitative account provided by Peirce (2003). In 1999, workers at the
Canadian Forces Base in Goose Bay, Labrador went on strike about the contracting-out of
support services to Serco, a British-based multinational company. Community support
which included a one-day general strike, was instrumental: After the strike began, the
company attempted to hire local contractors to take over tasks normally performed by the
strikers, such as garbage collection and grass-cutting, but all refused. In addition, the
contractors refused to rent the company equipment that would allow it to do the work itself
(353). After a 46-day strike, the Public Service Alliance of Canada (PSAC) signed a fouryear agreement which provided the 260 remaining workers with job security for the
duration of the agreement and also increased wages, benefits, and shift premiums for fire
fighters. The HRSDC data indicate that the issues were not reported and that an agreement
was reached.
This account is a reminder of the limits of statistical data. Certainly the most
illuminating studies of strikes are found in qualitative in-depth accounts of particular
struggles. However, short of finding and interviewing strikers, rarely feasible for strikes
which occurred decades earlier, this chapter suggests that newspaper coverage of strikes
offers not only a window onto the lived experience and texture of strikes, but also access to
the voices of strikers. Such coverage enhances the profile of particular strikes embedded in
the aggregate data, and enriches the statistical data. As Martin (2005: 144) points out,
given the lack of readily available data on social movement activity, scholars have relied
on media sources, primarily daily newspapers, to study diverse forms of collective
behavior.
For example, the HRSDC micro-data report that, for almost three months, between 13
November 1978 and 7 February 1979, 225 workers at Puretex Knitting, members of the
Canadian Textile and Chemical Union (CTCU), were on strike. The strike took place during

156

the renegotiation of an agreement, the issues were wages and fringe benefits, and the result of
the strike is that workers returned to work (a coding which indicates unsettled issues).
Newspaper accounts enrich this data enormously. Two key strike issues emerge in
the coverage: gendered wage discrimination, and workplace and washroom surveillance,
also gendered. The strikers almost all of whom were women, asked: Why do the women
cutters get so much less than the men? Maria Renda, for example, with the company 28
years, earns $4.31 an hour as a cutter the men working next to her earn $5.60 and $6.00.
On the issue of surveillance, two years prior to the strike, the company installed eight
security cameras, one beamed at the door of the womens washroom. No camera was
placed outside the mens washroom. Why, they [the women strikers] keep asking dont the
men have a camera outside their washroom door? Why, every night before they leave, do
they have to open their purses and show the contents to the uncle of the firms president?
The men never have to turn out their pockets or open lunchboxes (Bullock, Toronto Star
16 November 1978).
The Ontario Human Rights Commission refused to investigate, which, some
argued, actually triggered the strike. Madeline Parent, the head of the CTCU union and a
well-known woman trade union activist said: Its worse than a prison. Its a precedent.
Were fearful that if we give in, anti-labour employers and their legal advisers will resort to
using cameras anywhere they can get away with it (Toronto Star 15 November 1978). She
later commented: We intend to settle this thing during the strike. It is degrading, it is
offensive, and as humans we feel we shouldnt have to put up with it (Toronto Star 5
December 1978).
In February 1979, the union won an increase in wages and a qualified victory on
the camera issues (Globe and Mail 7 February 1979). The washroom camera was removed,
and in June of 1979, labour arbitrator Ellis directed the company to remove five of its eight
cameras after describing electronic surveillance of workers as fundamentally anti-human.
[E]lectronic surveillance poses a conflict between the right to privacy and to the
preservation of the fragile concept of human dignity and simple considerations of
efficiency. The Arbitrator also noted that apart from one incident, the evidence before him
did not establish that there had been a theft or pilfering problem of serious proportions at
Puretex (Globe and Mail 1 June 1979), this despite claims by Puretex president Gary Satok
that the company was being robbed blind (Toronto Star 1 June 1979). A final footnote:
in 1981, Maria Scarpell Iori, the president and founder of the union local at Puretex
received a Women of Distinction Award from the YWCA (Toronto Star 29 Oct 1981).
A second example. Seventy-four childcare workers at two Mini-Skools (part of a
large American chain of commercial for-profit child care providers) in Mississauga and
Toronto, members of the Ontario Public Service Employees Union (OPSEU), were on
strike from 8 Oct 1982 to 2 June 1983. A third group of childcare workers who also went on
strike faced the closing of their Hamilton, Ontario school in November 1982. The HRSDC
micro-data indicates that the strikes occurred during the renegotiation of the agreement,
wages were the primary issues, and an agreement was reached.
Newspaper coverage does suggest that wages were the central issue in this strike:
Mini-Skools pays mini-wages was a key slogan, about half the Toronto average for daycare workers (Globe and Mail 16 October 1982). Sean Flynn, then-president of the 75,000
strong OPSEU, joined the picket lines with busloads of other OPSEU members, and called
the pay of between $3.60 and $4.43 an hour earned by the strikers starvation wages
(Toronto Star 25 November 1982).

157

What also emerges is a struggle among workers, parents and employer over the care
of children during the strike. In Mississauga, the workers set up alternative care for the
children at a rate cheaper than what parents had been paying. Many parents and children
have been supporting the strikes by opting for the day care arrangements offered by the
strikers [] There was no question of where Eric [Chubbs] wanted to go when the strike
started, according to his mother Linda. My son said he goes where his teacher goes [...] I
know hes in the best of hands (Toronto Star 16 October 1982).21 In response to this
initiative by the strikers, Mini-Skools offered 50% off the regular rates using replacement
workers, thus encouraging some parents to bring their children across the picket lines on a
daily basis, and provoking some ugly picket line incidents which were covered repeatedly
by the Toronto Star. In a letter to the Star editor, Collin Gribbons (16 December 1982)
comments:
Mini-Skools has provoked incidents [...] by offering half price day care during
the strike. How concerned can these parents be for the welfare of their children,
to take them through a picket line every morning? How can the parents who use
Mini-Skools support a policy of paying less that $4 an hour to teachers who have
to have a college degree to qualify? [] To the parents who cross their picket
lines, I say: Shame on you!

Debbie Young, a parent who was taking her son across the picket line said in a letter to
editor (Toronto Star 25 November, 1982):
The Mini-Skool has been my sons home away from home []While I may
sympathize with the low-paid staff, I will not sacrifice my sons home for
union supporters [...] My primary concern is for the well-being, care and
education of my child. I only wish the so-called day-care supporters on the
picket line had the same concern. If they did care about children, Im sure the
harassment and abuse would never have taken place (Toronto Star 16 Dec 1982).

The coverage also gives some voice to the childcare workers. Lori Joy, 21 commented
I didnt realize how much joy there was in the world until I started working with the
children [] It was then that I realized I had found my lifes work (Globe and Mail 12
March 1983). On 15 February 1983, charges of dangerous driving and intimidation
were laid against Joy and Lucy Lucia after an incident when they followed a MiniSkool van that, they say, backed into one of their cars [] They have been arrested,
fingerprinted, photographed, handcuffed, placed in jail overnight and are defending
themselves against criminal charges. And they were fired [...] Naturally, my attitude
has changed in light of everything, said Joy. Despite everything, I would go back []
just for the children and to keep our union strong. If we don't stick together, they'll just
bring in more young girls and it will start all over again [...]We can't throw away what
we've worked so hard for. Arlene Gunn, union steward at the Mississauga school said:
It's not too difficult to see how one can become a radical [] First it was the school
(officials) against us; then the parents; and now even the law seems to be opposed to us
(Globe and Mail 12 March 1983). After an eight-month strike, the Globe and Mail (1
June 1983) reported that the strike ended with wage increases of 10-15%.
21

Coverage in the Globe and Mail (10 May 1983) wrongly indicated that the alternative centre had been
set up by mothers, and on May 11 apologized for the mistake.

158

Community interest, involvement and concern also emerge in the newspaper


coverage. For example, on 10 May 1983 (Globe and Mail), high-profile labour leaders
joined the picket line, part of a strategy of singling out for mass picketing companies
that have been operating with replacements and strikebreakers [which] was developed
by the Ontario Federation of Labor as part of its campaign for anti-strikebreaking
legislation. During the International Womens Day activities in Toronto in March
1983, a fund raising dance was held for the Mini-Skools strike (Toronto Star 24
February 1983). The Globe and Mail (1 June 1983) also reported that during the strike
demonstrations of support on the pickets lines were organized by many labour,
community and womens groups. The two examples of the strikes at Min-Skools and
Puretex demonstrate the potential of newspaper archives for enriching quantitative data
on strikes, and enhancing the profile of the strikers.
The scholarly literature on strikes and the media covers a range of issues, some
of which are illuminating for this chapter. In the first instance, numerous studies
indicate that the most significant problem is silence: the type of bias likely to occur in
mass media consists more of silence and emphasis rather than outright false
information (Franzosi 1987: 7). Martins 2005 study of the New York Times and the
Daily Labor Report found that only 8% of all strikes received any media attention
(171), although he also notes that other forms of protest receive no more coverage.
Despite these limits, Martin concludes that newspapers are an invaluable source of data
on industrial conflict (Martin 2005: 172). Franzosi (1987: 6) remains optimistic about
the potential of newspapers as a source of historical data as they often constitute the
only available source of information and further no data source is without error,
including officially collected statistics. As earlier sections of this chapter demonstrate,
both error and bias affect official work stoppage statistics.
Unlike my use of newspaper coverage on particular strikes a qualitative
measure to enrich quantitative data, many researchers have addressed the criteria for
coverage used by the newspapers. For example Martin (2005) found that size, length,
and disruptiveness were key variables. In addition, Erickson and Mitchell (1996) found
that occurrence of federal intervention (in the US context), key industry status, and
locational bias (relevant to the readership of particular paper) determined the extent of
news coverage. On the latter point, numerous studies indicate that national and
regional/local papers offer different kinds of coverage. This was certainly true for the
two strikes considered here. The national Globe and Mail included less coverage and
fewer human interest angles, and thus offered less access to the voices of the strikers.
Quantitative research on media coverage of strikes consistently finds that the
number of workers involved impacts on the newspaper coverage. However, there was
considerable coverage of the Mini-Skools and Puretex strikes, both of which involved
small numbers of women in marginal industries with little national interest.
Undoubtedly, newsworthiness is a complex and permeable issue.
Another finding of relevance is that most coverage of industrial relations issues
focuses on strikes. Vermas 1988 study of media coverage of industrial relations events
in two Canadian dailies found that 41% of the stories were about strikes. He concluded
that the media tend to trivialize industrial relations news by stressing events at the
expense of issues (118) and that much of strike coverage continues to be driven by the
drama of events (120). In her study of media coverage of labor unions between 19461985, Schmidt (1993) found that that coverage became increasingly concentrated on
strike activities and exaggerated the frequency of strikes. She concludes: [E]ven

159

though strike frequency had declined, public information about labor unions had
become increasingly dominated by stories about strikes (159). The most troubling
finding in her study is that disapproval of labor unions increased as the percentage of
strike-related coverage increased over time (160).22
The extent of bias against unions and striking workers is also addressed in the
literature. Farrow and OBriens 2005 study of coverage of the strike of mental health
nurses in New Zealand in 2001-2, found blame for the strike was apportioned to
nurses, with little analysis of the role of service management in the industrial action
(187) and nurses were depicted as greedy, lazy, and militant (192). Perhaps counterintuitively given the corporate domination of media, Vermas 1988 study found
coverage portraying both negative and positive assessments of union and management
and [] the conduct of each side during bargaining (119). In another qualitative study,
in this case of television news coverage of nurse strikes, Kalisch et al. (1983) concluded
that nurse unions receive more positive television news coverage when they project an
image of solidarity, maintain unity over time, and receive the support of other types of
health-care workers (175).
Maney and Oliver (2001: 164-65) conclude: There are no perfect records of
collective events, nor are there perfect methods for gathering all of the collective events
in any given source. Each record source and data collection method has its own logic
and selectivity. Newspaper archives are one such imperfect but invaluable resource
which help researchers juxtapose statistical data with qualitative sources. Further, given
the fact that work stoppage data are not gendered, archives will be of particular value in
helping to gender the strike experience.

4. Conclusion
Using Canadian data, this chapter considers two strategies for doing research on strikes,
each of which is framed by a labour militancy perspective that highlights the point of
view of workers (Briskin 2006). A labour militancy frame presents an alternative to the
employer focus on time lost, the government concern to measure the economic impact
of stoppages, and the scholarly emphasis on strike determinants.
The first strategy re-interprets the data from Human Resources and Social
Development Canada (HRSDC) on the 23,944 stoppages in Canada between 1960 and
2004. A detailed examination of strike duration, strike size and strikes for first contracts
within a labour militancy frame enhances the picture that emerges from aggregates and
averages. The shift from person-days lost as a percentage of working time to the
numbers of workers on strike helps to embody the strike experience. A shift from strike
size averages to the detailed data show that 55.7% of all strikes between 1960 and 2004
involve less than 100 workers and 6.4% involve more than 1000 workers. Furthermore,
although on average measures, Canada has always had long strikes in comparison to
other countries, in fact, the disaggregated data show that 35.6% of strikes between 1960
and 2004 lasted only between one and five days. And although first agreement strikes
22

Flynn (2000) also demonstrates the significance of media coverage. He found that pre-strike media
attention of disputes involving 10,000 or more workers increased strike duration. The author uses a social
psychological approach to explain these data: An increase in public attention will reduce the likelihood
of an individual or a group compromising a stated position (144).

160

do not involve a significant proportion of person-days lost, it is revealing that 21% of


total workdays lost to strikes involved such struggles. Data on first agreement strikes
support other research on employer resistance to union certification (Bentham 2002).
Through a consideration of lockouts, labour militancies are also set against patterns of
employer aggression.
The labour militancy approach enriches understandings of Canadian work
stoppages and helps to illuminate workers experiences of strikes. It makes visible the
local and the particular, and supports new ways of looking at overall patterns of labour
militancy, highlighting very long and very short strikes, strikes of a few workers or
those involving many thousands of workers, and strikes for first agreements. Although
such an approach to statistical data deepens understandings of Canadian strikes,
undoubtedly numbers can never adequately represent the anger, risk, struggle and
solidarity involved in the strike experience.
Interviews with the provincial correspondents who collect work stoppages
information for HRSDC reveal that data collection is not standardized across Canadian
provinces. Examining the data in light of these interviews highlights the often-hidden
qualitative and subjective aspects of statistics, and underscores the political nature of
data collection (what is seen to be germane and not), data presentation (what is made
visible and what is not), and data sources (whose voices are heard).
The second strategy explores the use of newspaper archives to enhance work
stoppage micro-data using the examples of the Mini-Skools and Puretex strikes. A
qualitative measure to enrich quantitative data, this archival material offers not only a
window onto the lived experience and texture of strikes, but also access to the voices of
strikers.
These strategies for doing research on strikes support re-interpretations which
help make visible the significance of such stoppages for workers, unions and
communities. They offer an alternative to detailed historical studies of particular strikes,
on the one hand, and the commonly-used averages and aggregates, on the other.

161

Abbreviations
HRSDC
LFS
PDL
SLID
WES

Human Resources and Skills Development Canada


Labour Force Survey
Person-days lost
Survey of Labour and Income Dynamics
Workplace and Employee Survey

References
Akyeampong, E. 2006A, Increased work stoppages, Perspectives on Labour and
Income, vol. 7, no. 8, pp. 5-9.
Akyeampong, E. 2006B, Unionization, Perspectives on Labour and Income, vol.
7, no. 8, pp. 18-42.
Anonymous. 2009, Work absence rates, Perspectives on Labour and Income,
vol. 21, no. 2, pp. 59-62.
Bentham, K. 2002, Employer resistance to union certification: A study of
Canadian jurisdictions, Relations industrielles/Industrial Relations, vol. 57, no. 1,
pp. 159-188.
Briskin, L. 2007, Public sector militancy, feminization, and employer aggression:
Trends in strikes, lockouts, and wildcats in Canada from 1960 to 2004 in Strikes
around the world, eds. H. Dribbusch, D. Lyddon, K. Vandaele and S. van der
Velden, Amsterdam, Aksant.
Briskin, L. 2006, From person-days lost to labour militancy: A new look at the
Canadian work stoppage data, Relations industrielles/Industrial Relations, Vol.
62, No.1, pp. 31-65.
Briskin, L. 2005, The work stoppage data from Human Resources and Skills
Development Canada: A research note, Just Labour, vol. 5, 80-89 [Online],
available at http://www.justlabour.yorku.ca/.
Briskin, L. with Klement, K. 2004, Excavating the labour dispute data from
Statistics Canada: A research note, Just Labour, vol. 4, pp. 83-95 [Online],
available at http://www.justlabour.yorku.ca/.
Campolietti, M., et al. 2005, Strike Incidence and strike duration: Some new
evidence from Ontario, Industrial and Labor Relations Review, vol. 58, no. 4, pp.
610-630.
Erickson, C. and Mitchell, D. 1996, Information on strike and union settlements:
Patterns of coverage in a newspaper of record, Industrial and Labor Relations
Review, vol. 49, no. 3, pp. 395-407.

162

Ewing, K.D., et al. 2003, Unfair labour practices: Trade union recognition and
employer resistance, London, Institute of Employment Rights.
Farrow, T. and OBrien, A. 2005, Discourse analysis of newspaper coverage of
the 2001/2002 Canterbury, New Zealand mental health nurses strike,
International Journal of Mental Health Nursing, vol. 14, pp. 187-195.
Flynn, F. 2000, No news is good news: The relationship between media attention
and strike duration, Industrial Relations, vol. 39, no. 1, pp. 139-160.
Franzosi, R. 1989, One hundred years of strike statistics: Methodological and
theoretical issues in quantitative strike research, Industrial and Labor Relations
Review, vol. 42, no. 3, pp. 348-362.
Franzosi, R. 1987, The press as a source of socio-historical data: Issues in the
methodology of data collection from newspapers, Historical Methods, vol. 20,
no. 1, pp. 5-16.
Fudge, D. and Brewin, J. 2005, Collective bargaining in Canada: Human right or
Canadian illusion? Ottawa, NUPGE and UFCW.
Godard, J. 2005, Industrial relations: The economy and society, 3rd edn, Toronto,
Captus Press.
Godard, J. 1992, Strikes as collective voice: A behavioral analysis of strike
activity, Industrial and Labor Relations Review, vol. 46, no. 1, pp. 161-175.
Gramm, C. 1986, The determinants of strike incidence and severity: A microlevel study, Industrial and Labor Relations Review, vol. 39, no. 3, pp. 361-376.
Gunderson, M., et al. 2005, Strikes and dispute resolution in Union-management
relations in Canada, 5th edn, eds. M. Gunderson, A. Ponak and D. Taras, Toronto,
Pearson Addison Wesley.
Gunderson, M. and Melino, A. 1990, The effects of public policy on strike
duration, Journal of Labor Economics, vol. 8, no. 3, pp. 295-316.
Harrison, A. and Stewart, M. 1993, Strike duration and strike size, Canadian
Journal of Economics, vol. 26, no. 4, pp. 830-848.
Harrison, A. and Stewart, M. 1989, Cyclical fluctuations in strike durations, The
American Economic Review, vol. 79, no. 4, pp. 827-841.
Hebdon, R. 2005, Toward a theory of workplace conflict: The case of U.S.
municipal collective bargaining, in Advances in industrial and labor relations:
Volume 14, eds. D. Lewin and B. Kaufman, Amsterdam, Elsevier.

163

Kalisch, B., et al. 1983, Television news coverage of nurse strikes: A resource
management perspective, Nursing Research, vol. 32, no. 3, pp. 175-180.
Maney, G. and Oliver, P. 2001, Finding collective events: Sources, searches,
timing, Sociological Methods & Research, vol. 30, no. 2, pp. 131-69.
Martin, A. 2005, Addressing the selection bias in media coverage of strikes: A
comparison of mainstream and specialty print media, Research in Social
Movements, Conflicts and Change, vol. 26, pp. 141-78.
Ondrich, J. and Schnell, J. 1993, Strike duration and the degree of disagreement,
Industrial Relations, vol. 32, no. 3, pp. 412-431.
Panitch, L. and Swartz, D. 2003, From consent to coercion: The assault on trade
union freedoms, Aurora, Ontario, Garamond Press.
Peirce, J. 2003, Canadian industrial relations, 2nd edn, Toronto, Prentice Hall.
Peirce, J. and Bentham, K. 2007, Canadian industrial relations, 3rd edn, Toronto,
Prentice Hall.
Renaud, L., et al. 2005, Training manual: Work stoppages. Work Stoppage,
Labour Organization, and Collective Agreement Analysis Section of the
Workplace Information Directorate, Ottawa, Human Resources and Social
Development Canada
Rose, J. 2006, Collective bargaining performance by newly created unions in
Canada: Process and Outcomes, in Union recognition: Organizing and
bargaining outcomes, ed. G. Gall, London, Routledge.
Schmidt, D. 1993, Public opinion and media coverage of labor unions, Journal
of Labor Research, vol. 14, no. 2, pp. 151-64.
Skeels, J., et al. 1988, The importance of strike size in strike research, Industrial
and Labor Relations Review, vol. 41, no. 4, pp. 582-591.
Statistics Canada. 2009, Unionization, Perspectives on Labour and Income, pp.
27-35.
Verma, A. 1988, Good news and bad news: Print media coverage of industrial
relations events in Canada, Some Aspects of International Industrial Relations:
Proceedings of the 25th meeting of the Canadian Industrial Relations Association,
Laval University, pp. 113-121.
Walker, J. 1987, First agreement disputes and public policy in Canada, School of
Industrial Relations Research Essay Series No. 12. Kingston, Industrial Relations
Centre, Queens University.

164

165

Building a repository for strike data. The search for micro


data
Sjaak van der Velden
1. Introduction
The strike, or the fundamental statement of the humanity and intelligence of the working
class (Cronin, 1979: 195), will probably persist. Labour conflicts are an important
phenomenon in the history of capitalism, but also in other societies in which unequal labour
relations exist. Through labour conflicts, social, economic, political and legal relations have
been greatly altered. Researchers concerned with social and economic history, but also
those involved with institutional or political history, cannot ignore these conflicts. However,
this fact is not reflected in the subjects chosen by most social historians over the last
decades. The study of strikes as a research object has been out of fashion for quite some
time. Such a downswing happened earlier in history, after which an upswing followed the
growth of strike incidence (Franzosi, 1995: 1). The resurgence of class conflict in the
emerging economies of China and India will possibly cause new interest among scholars in
this statement of human intelligence.
Those studying strikes, lockouts and others forms of labour conflict need data.
National statistical bureaus generally compile three annual main indicators: number of
stoppages, employees/workers involved and days not worked. It is commonplace in
research on strikes to highlight the inadequacies of officially collected data. The general
consensus is that figures for days not worked pick up the bulk of strike volume; by contrast,
figures for the numbers of stoppages usually seriously underestimate the actual picture.
In the history of historical strike research, several investigators have therefore built
their own databases. In 1974 Tilly and Shorter researched the history of strikes in France
(from 1830 to 1968). They collected micro data on which they based their often quoted
study: Strikes in France, 18301968 (Shorter and Tilly, 1974). Having started in the midseventies, I have created a database for the Netherlands, which covers the whole period
from 1830 to 2005 (see van der Velden, 2000; 2003). Research conducted by the World
Labor Research Working Group (WLG) of the Fernand Braudel Center in New York also
shows that it is possible to build large databases containing data on labour conflicts (Silver,
1995, 2003). This database is distinctive when related to most other data, because the WLG
collects information about all kinds of labour conflicts on a global scale from newspaper
clippings.
The mentioned data is an extension of official data, but is also collected from
periods for which official data is nonexistent. In Russia, a database was built from Tsarist
archives, in which almost 8,000 strikes were recorded by Leonid Borodkin and his coworkers. Russian researchers used this data to apply chaos theory when analysing the labour
movements struggles in pre-revolutionary Russia (Andreev, 1997). This database is dealt
with in chapter 3 of this book.
Databases have also been built that contain information on strikes in specific
professions or regions. In 1997 Quentin Outram published data on almost 8,000 strikes in
the coal mining industry in the United Kingdom from 1893 to 1940, compiled from official

166

sources. Outram used the data to write a book together with Roy Church (Church, 1988)
and deposited the data at the Economic & Social Research Council (ESRC) UK Data
Archive at the University of Essex,1 with a detailed guide.
David Green collected data about 294 labour conflicts that occurred in London
between 1790 and 1870, from accounts derived largely from London newspapers (Green,
1998).
The compilers of all the mentioned databases used different methods of
accumulating, and particularly publishing, their data. These differences make it difficult to
connect the data from all databases. However, such a connection is necessary if researchers
want to compare strike movements in a spatio-temporal environment.
2. Comparisons over time and place
When researchers want to compare strikes that occurred in different times and places,
they encounter a statistical problem. In general, they have the data collected by national
bureaus and published by the International Labour Organisation (ILO) at their disposal.
There are quite a few problems with this data, because collectors of national data have
refrained from following the recommendations of the ILO (1993). If national statistics
are not compiled in a standardised way they are hardly comparable. The ILO
encountered this problem from the early days of its existence. Therefore the office
published an overview of national methods in order to outline where practicable the
standards on the basis of which some degree of international comparability may be
secured (ILO, 1926: 6).
The major problem for social scientists with regard to the ILO data from 1926 to
date is the fact that this data is compiled and published to deal with the economic
aspects of labour conflicts. It is for this reason that the ILO in 1926 regarded industrial
conflicts from the point of view of risk. Starting from this viewpoint, two methods of
risk measurement were proposed: the calculation of its frequency and of its severity.
Since then, most national and especially international statistics on strikes have used the
frequency rate and the severity rate, or the number of disputes per 100,000 full-time
workers and the number of man-days lost per 100,000 full-time workers. In order to
avoid too small numbers, the severity rate can also be calculated per 1,000 workers
(ILO, 1926: 48). This last calculation has been the standard for a long time, expressed in
the modern phrase total days not worked per 1,000 employees (Chernyshev, 2003).
Strike research to answer social questions, instead of the purely economic
outcomes, is in need of a measurement in which people and the frequency of labour
conflicts play a more prominent role. The severity index may seem fair enough, but
causes a number of new problems. The number of strike days (D) is a mathematical
function of the number of strikers (S) and the duration (d) of strikes (D = S x d). Strike
research has shown that (d) is largely a function of employer behaviour. The more eager
employers are to resume production quickly, the more prepared they might be to give in
to the demands of striking workers. As a result, fewer working days will be lost. My
own research has shown that in the period from 1970 to 2000, when strikes were shorter
than in earlier years, over seventy per cent of all strikes in the Netherlands ended either
in victory for the strikers or in a compromise. The corresponding figure for all other
1

www.data-archive.ac.uk/

167

years, when strikes were longer, supports the idea that lengthy strikes often work
against those striking. In other words, a high number of working days lost mostly
indicates the strength of capital.
Another and even greater problem with the severity index is the fact that when
using the number of working days lost, the number of conflicts and the number of
participants are completely left out of sight. This problem has worried researchers in the
past, such as P. Galambos and E.W. Evans who proposed a more sophisticated index
which included the number of conflicts, the number of workers affected and the number
of days not worked. They compared the last two indicators to the total number of
workers (Galambos and Evans, 1966). Others have built on their work to improve this
index. One of the grandfathers of strike research, K.G.J.C. Knowles, wondered why
Galambos and Evans did not also relate the number of strikes to the number of
employees (Knowles, 1966). In my thesis I have also tried to connect the strike
indicators (Van der Velden, 2000) because I also sensed problems with the simple ILO
index. As a social historian, the number of participants in a social conflict in particular
seems to me to be of the utmost importance. If we want to solve these problems, it
seems necessary to construct an index that contains the number of conflicts (N), the
number of workers involved (S) and the number of lost working days (D). These strike
indicators will then have to be measured against the background of the development of
the total number of workers (W) and the labour volume (V, the number of working
days).
This index may be calculated according to the following formula (Van der
Velden, 2003):
I = { (Nt / Navg) + ((St / Savg) / (Wt / Wavg)) + ((Dt / Davg) / (Vt / Vavg)) } x (100 / 3)
t = any given year, avg = the average of all years
Compared to the ILO index, the above is rather complex and we may wonder whether
the statistical robustness is such that the calculation is worth the trouble. Wessel Visser,
from Stellenbosch University in South Africa, and I used this formula in our
comparative research for the development of strike activity in the Netherlands and
South Africa (Visser and Van der Velden, 2006) from which we concluded that the
developments were much more similar than one might expect. The comparison between
the Netherlands and South Africa showed that over a long period of time, a rough view
indicates that the ILO index and the more complex one exhibit similar movements. This
view does not hold when one takes a closer look though. First, there are a number of
years in which the indices move in opposite directions. Second, the ILO method of
calculating shows more extreme movements in most years. The more complex
calculation thus dims the movement. The third point is that the correlation coefficient
shows that the two lines exhibit the same movements for only just over fifty per cent (R
= 0.75, R2 = 0.57).
Given these three comments, it seems wise to use the more complex index
instead of the lost days per thousand workers. Of course this is only possible when all
the necessary data is available. To a lesser extent the same is true of the ILO index.
Using the index we can calculate an international index. I did this for a number
of countries, mainly from the core of capitalism:

168

Figure 1. Strike index of 16 countries, seven-year moving average, 1880 to 2000

Countries: USA, Canada, Australia, New Zealand, Belgium, Germany, Denmark, France, GB, Italy,
Norway, Austria, Spain, Sweden, Switzerland and Japan
Source: Van der Velden, 2003: 400.

It is not surprising that the three big peaks of strike activity coincide with international
literature on the subject: the two post world war periods and the 1970s. Figure 1 shows
that there is a possible relationship between the Kondratiev cycle and strike activity. It
also shows that, given this measurement, strike activity in the countries under research
at the end of the period was plummeting back to the level of activity at the beginning.
No matter what index a researcher uses, the foundations for the calculation have
to be solid. These foundations are the collection of data, and as mentioned before there
are many problems with this. To this we might add another problem that is typical for
aggregated data and pointed out by Beverly Silver (Silver, 2003: 188-189). Aggregated
data equates all different forms and situations in time and space. Given the same
number of strikers and the same duration, a strike by well-paid football players, as
occurred in 1987 in the USA, gets the same weight as a strike by underpaid house
cleaners. This problem is true for all quantitative ways of looking at labour conflicts
over time and can only be solved by having the statistics accompanied by a qualitative
story. This means stepping down from aggregated data to micro data.
To address the previously mentioned problems of using aggregated data, highly
comparable micro data is also needed. Only with this micro data can we calculate
aggregated series of all the needed indicators. This data should be made available to all

169

researchers in order to promote truly comparative studies. The data should be stored in a
central repository providing access to the international community of strike researchers.
3. An international data repository
The International Institute of Social History (IISH) in the Netherlands has taken the
initiative to establish a moderated list of data files concerning labour conflicts. To
obtain such a list, we have called on researchers who have access to digital data about
labour conflicts to make these available to the IISH, or connect them to the collection
via a link.
The databases should cover strikes, lock-outs and other labour conflicts. The
reason for collecting this data is, as has been said before, that the official data from the
various national statistical offices and derivatively that held by the ILO is demonstrably
incomplete, often not easily comparable, only starting from 1926 and often available
only at a high level of aggregation (annual data). Many old card-box files on labour
conflicts have obviously not been transferred to digital storage. However, these will also
have a place in the list, through a description or reference to the source in a monograph
or periodical.
Preferably the data should be standardised. More than 80 years of collecting data
by the ILO has, however, proved that this demand is unrealistic. The IISH therefore will
accept all databases regardless of the format, but we have nevertheless developed and
made available standardised database, text file and spreadsheet formats. These are free
to use by anyone who wants to download them. The only condition is that the user
mentions the name of the IISH and makes his or her data available. When we
constructed the standard, we started from the recommendations made by the ILO.
4. Starting from ILO
Since the start of the ILO, efforts have been made to streamline the data on strikes and
lockouts as gathered and published by national statistical bureaus. However, these
numerous efforts have been in vain. After the publication of the first one, more than
eight decades ago (ILO, 1926), there are still many national bureaus that go their own
way. This may be a result of political developments, methodological insights or lack of
funding.
To give an example of these arguments: Nazi Germany stopped publishing strike
data because the class struggle was considered to have come to an end. In fact strikes
did occur in those dark years, and afterwards historians published material about strikes
under the Hitler regime (Roth, 1977: 172). In the United States the minimum size of
strikes in order to be entered into statistical records was set at 1,000 workers in 1980,
because of financial constraints (Galenson, 1996: 89), to mention just another example.
In his overview of all the statistics on which the yearly publications of the ILO
are based, Chernyshev (2003) made it clear that comparability was still a major issue.
Because of the long-term experience of data collecting, it seems nevertheless wise not to
start from scratch, but to use the expertise built up by the ILO. Historians may assert
that using fixed definitions will inevitably lead to anachronistic approaches to the

170

phenomenon of labour conflicts. This risk is of course not negligible and therefore we
have to make it possible to use shifting and vague definitions or descriptions.
The following definitions are used in the proposed database of micro data. They
have been derived from the Resolution concerning statistics of strikes, lockouts and
other action due to labour Disputes (ILO 1993) but have been slightly adapted:
A Labour dispute is a state of disagreement over a particular issue or group of issues
over which there is conflict between workers and employers, or about which grievance
is expressed by workers or employers, or about which workers or employers support
other workers or employers in their demands or grievances.
A strike is a temporary work stoppage effected by one or more groups of workers with a
view to enforcing or resisting demands, or expressing grievances, or supporting other
workers in their demands or grievances.
A lockout is a total or partial temporary closure of one or more places of employment, or
the hindering of the normal work activities of employees, by one or more employers
with a view to enforcing or resisting demands, or expressing grievances, or supporting
other employers in their demands or grievances.
Other action due to labour disputes consists of action effected by one or more groups of
workers or by one or more employers, with a view to enforcing or resisting demands, or
expressing grievances, or supporting other workers or employers in their demands or
grievances, in the course of which there is no cessation of work.
Workers involved in a strike: Workers directly involved in a strike are those who
participated directly by stopping work. Workers indirectly involved in a strike are those
employees of the establishments involved, or self-employed workers in the group
involved, who did not participate directly by stopping work but who were prevented
from working because of the strike.
Workers involved in a lockout: Workers directly involved in a lockout are those
employees of the establishments involved who were directly concerned by the labour
dispute and who were prevented from working by the lockout. Workers indirectly
involved in a lockout are those employees of the establishments involved who were not
directly concerned by the labour dispute but who were prevented from working by the
lockout.
Workers involved in other action: Workers directly involved in other action are those
who participated directly in the action. Workers indirectly involved in other action are
those employees of the establishments involved or self-employed workers in the groups
involved who did not participate directly in the action but who were unable to perform
their work in the usual manner or prevented from working as a result of the action.
Duration counts the calendar days from the beginning of the conflict to its end.
Duration includes days on which work would usually be carried out by the groups of
employees concerned and the weekly rest-days and public holidays, etc., on which work
was not scheduled for the groups of employees involved. If work is organised in shifts,

171

one shift should be considered as one workday. The normal hours of work for the
groups of workers concerned should be defined in accordance with the most recent ILO
standards.
An establishment involved is one in which one or more groups of employees are directly
involved in action due to a labour dispute, or in which such action is effected by the
employer. The definition of an establishment should be in accordance with the most
recent guidelines of the United Nations.
The secondary effects of action due to labour disputes are the effects on other
establishments where workers are prevented from working or their work is disrupted, or
the effects on other groups of self-employed workers who are prevented from working
or whose work is disrupted.
With these definitions in mind, data on labour conflicts can be classified into the
following structure.
5. The structure of data collection and publication
Basic data to be collected includes:
a. Strikes, lockouts and other expressions of labour conflicts.
b. The number of workers involved in these actions.
c. The duration of these actions in calendar days regardless of year.
d. The number of working days not worked by workers directly, indirectly or
secondarily involved in these actions.
e. The number and names of the companies involved, including any
conglomerates to which they belong.
f. The demand(s) that caused the actions.
g. The outcome of the actions and method of settlement.
h. The calendar date of the actions.
i. The geographical position of the action (if possible the longitude and latitude).
j. The profession of the workers consistent with the Historical International
Classification of Occupations (HISCO, see: http://historyofwork.iisg.nl).
k. The economic sector according to the International Standard Industrial
Classification (see: http://unstats.un.org/unsd/cr/registry/regcst.asp?Cl=27).
l. The workers and/or employers organisations concerned.
m. Was the action official or unofficial (wildcat).
n. Special groups of workers (e.g. women, children, non-indigenous workers).
o. An account (if possible in English, otherwise in the national language).
p. Sources where the data is found (magazines, archives, etc.)
This structure is used to build a database which is accessible at
http://socialhistory.org/en/projects/labour-conflicts. As previously stated, anyone is free
to download the database and use it for their own purposes if credits are given to the
maker and the International Institute of Social History. Although the database is userfriendly we also provide a short manual.

172

First results
In 2007 the IISH appointed me to start collecting strike data on a micro level from all
around the world and construct the moderated list mentioned above. So far, a number of
databases have been forwarded to the list at http://socialhistory.org/en/projects/labourconflicts.
The list includes the following databases: the Netherlands 1350-2009, France
1830-1968, British coal mining 1893-1940, London 1790-1940 and Russia 1895-1904.
The publication of the coal mining set inspired Outram to write a new article
about industrial relations in the coal mining industry (2008). The London database was
used by Agustin Santella from Argentina to write about a remark by Friedrich Engels in
his Die Lage der arbeitenden Klasse in England from 1845 2(Santella, 2008: 22).
During a workshop held at the IISH in September 2008, Edward Fokuoh
Ampratwum from Ghana presented his micro data on strikes between 1995 and 2004 in
an illuminating paper (see this book, chapter 2). His data can also be found in the
moderated list, together with data from Portugal (Raquel Varela), Germany (Heiner
Dribbusch and Sjaak van der Velden), Argentina (Nicolas Inigo Carrera) and Sweden
(the Swedish Mediation Office).
Apart from the data on Portugal from Varela, no one has delivered in the
proposed format; all have been put on the list in the original format used by the
constructors. There is still a lot of work to be done, but the start has been promising.

. "Una investigacin contrasta esta clsica asercin de Engels (La historia de estas uniones es una larga
serie de derrotas obreras, interrumpidas por pocas victorias aisladas). De un total de 294 huelgas
registradas entre 1791-1865, la base provee informacin sobre los resultados en 132 de ellas. 45 huelgas
fueron ganadas contra 25 perdidas, y en 12 terminaron en compromiso. Si nos fijamos en el perodo 17911845 momento del libro de Engels), en 80 huelgas no hubo informacin sobre resultados, pero se
registraron 24 huelgas ganadas contra 20 perdidas y 3 compromisos. Elaboracin propia a partir de la
base construida por David Green y puesta a disposicin del proyecto Conflictos Laborales dirigido por
Sjaak van der Velden en el Instituto Internacional de Historia Social de Amsterdam."

173

References
Andreev, Andrey, Leonid Borodkin and Mikhail Levandovskii (1997), Using methods
of non-linear dynamics in historical social research: application of Chaos Theory in the
analysis of the Workers Movement in pre revolutionary Russia Historical Social
Research, no. 22, 1997, pp. 6483.
Barkin, S. (1975), Worker militancy and its consequences, 196575: new directions in
western industrial relations, New York, Praeger.
Chernyshev, Igor (2003), Decent work statistical indicators: strikes and lockouts
statistics in the international context, Bulletin of labour statistics no. 3, XIIIXIV.
Church, Roy A. and Quentin Outram (1988), Strikes and Solidarity: Coalfield Conflict
in Britain, 1889-1966, Cambridge University Press.
Cronin, J., (1979), Industrial conflict in modern Britain, London, Croom Helm.
Crouch, C. and A. Pizzorno (1978), The resurgence of class conflict in Western Europe
since 1968. Volume 1, national studies, and Volume 2, comparative analyses, London,
Macmillan.
Galambos, P. and E.W. Evans (1966), Work-stoppages in the United Kingdom, 19511964: a quantitative study, Bulletin of the Oxford Institute of Economics & Statistics,
volume 28, pp. 3355.
Galenson, Walter (1996), The American Labor movement, 1955-1995, Westport.
Green, David R. (1998), Lines of conflict: labour disputes in London 1790-1870,
International Review of Social History, vol. 43, pp. 20334.
ILO (1926), Methods of compiling statistics of industrial disputes, Geneva, ILO
---- (1993), Resolution concerning statistics of strikes, lockouts and other action due to
labour disputes, adopted by the Fifteenth International Conference of Labour
Statisticians, at http://www.ilo.org/public/english/bureau/stat/download/res/strikes.pdf
Kelly, J. (1997), Long waves in industrial relations: mobilization and countermobilization in historical perspective, Historical Studies in Industrial Relations, no. 4,
pp. 335.
Kerr, C. et al. (1973), Industrialism and industrial man: the problems of labour and
management in economic growth, Harmondsworth, Penguin (2nd edn; first published in
the US, 1960).
Knowles, K.G.J.C. (1966), Work-stoppages in the United Kingdom: a comment
Bulletin of the Oxford Insitute of Economics & Statistics, volume 28, pp. 5962.

174

Outram, Quentin (2008), Early British strike statistics, Historical studies in industrial
relations, no. 25/26, pp. 17796.
Ross, A. M. and P. T. Hartman (1960), Changing patterns of industrial conflict, New
York, John Wiley.
Roth, Karl-Heinz (1977), Die andere Arbeiterbewegung und die Entwicklung der
kapitalistische Repression von 1880 bis zur Gegenwart. Eind Beitrag zum
Neuverstaendnis der Klassengeschichte in Deutschland, Mnchen.
Santella, Agustin (2008), Trabajadores, sindicatos y conflictos en la industria
automotriz argentina. Un estudio de caso (1989-2006), Tesis de Doctorado,
Universidad de Buenos Aires.
Shalev, M. (1992), The resurgence of labour quiescence, in The future of labour
movements, ed. M. Regini, London, Sage, pp. 102132.
Shorter, E. and C. Tilly (1974), Strikes in France, 18301968, London, Cambridge
University Press.
Silver, B. J. (2003), Forces of labor. Workers movements and globalization since 1870,
New York, Cambridge University Press.
Silver, B. J. et al. (eds.) (1995), Labor unrest in the world economy, 18701990, Review
(Fernand Braudel Center), Special Issue, vol. 18, no. 1.
Velden, S. van der (2000), Stakingen in Nederland: Arbeidersstrijd 18301995,
Amsterdam, Stichting beheer IISG.
---------------------- (2003), Strikes in global labor history. The Dutch case, Review
(Fernand Braudel Center), vol. 26, no. 4, pp. 381405.
---------------------- (2006), Lockouts in the Netherlands: why statistics on labour
disputes must discriminate between strikes and lockouts, and why new statistics need to
be compiled, Historical Social Research / Historische Sozialforschung, vol. 31, no. 4,
pp. 341362.
Velden, Sjaak van der and Wessel Visser (2006), Strikes in the Netherlands and South
Africa, 1900-1998: a comparison, in: South African Journal of Labour Relations, vol.
30, no. 1, pp. 5175.

175

Contributors
Mr. Edward Fokuoh Ampratwum is a Researcher in Governance and Social Development
and Policy at the Center for Democratic Development (CDD-Ghana), a non-governmental
research and policy Think Tank in Ghana. He is a Sociologist and Development policy
analyst by training and has worked extensively as researcher and analyst on Labour Market
Policies and social protection in Ghana and other West African Countries; Governance and
Development, Corruption in Developing and Transitional Economies, HIV/AIDS and
Reproductive Health etc. Prior to his current work at CDD-Ghana, Mr. Ampratwum
worked as a Labour Market researcher at the Labour Policy and Research Institute of the
Ghana Labour College.
E-mail: fokuoh2001@yahoo.co.uk
Dr. Maria Bergman has a PhD in History from Ume University, Sweden. She did her
undergraduate research on women conditions and population development in 19th century
Sweden at the Department for Historical, Philosophical and Religious Studies at Ume
University, earning her a Master in History in 2006. She conducted her PhD research at the
Centre for Population Studies, also at Ume University. Her research centered on the
population development in the sawmill communities that evolved during 19thcentury
Sweden; Constructing communities. The establishment and demographic development of
sawmill communities in the Sundsvall district, 1850-1890 (2010). This reaearch lead her to
include sources from one of the first major strikes in Sweden, the Sawmill strike in
Sundsvall in 1879. Other publications include an article on Ethics and Demography in;
Forska rtt. Texter om etik och historisk forskning (2007) and participation in a publication
of demographic study material produced for upper secondary school, Bland pigor,
utvandrare och vanliga ddligai 1800-talets Sverige. Historiska problem i digitala kllor.
Ett arbets- och inspirationsmaterial fr gymnasieskolan (2006). Dr. Bergman completed a
Degree of Master of Education in 2011 and wrote her thesis on comparative child and youth
literacy in a rural parish during the 19th century. She is currently teaching in upper
secondary school.
E-mail: maelbe@hotmail.com
Dr. Linda Briskin is a Professor at York University (Toronto). In addition to numerous
articles, she has authored Equity Bargaining/Bargaining Equity (2006); co-edited
Women's Organizing and Public Policy in Canada and Sweden (1999), Women
Challenging Unions: Feminism, Democracy and Militancy (1993); and Union Sisters:
Women in the Labour Movement (1983); and co-authored Feminist Organizing for
Change: the Contemporary Women's Movement in Canada (1988), and The Day the
Fairies went on Strike (for children) (1981). She is currently researching union
leadership, strategies for ensuring equity representation inside unions, womens
participation in collective bargaining and social dialogue, and worker militancies with a
special focus on nurses on strike. She has been a union activist for many decades.
E-mail: lbriskin@yorku.ca

176

Prof. Leonid Borodkin is a Head of the Department for Historical Information Science
and Director of the Center for Economic History at the History Faculty of Lomonosov
Moscow State University. He is the author of a number of books in the field of Russian
labor history, Russia's economic history of the 19th - 20th century and methodology of
social and economic history.
E-mail: borodkin@hist.msu.ru
Dr. des. Julia Casutt obtainded a PhD in Economic History at the University of Zurich.
The article is based on research results of her thesis which deals with the influence of
business cycles on strike activity in Austria, Germany, and Switzerland. She now works
at the Department of Economic Affairs of the Canton of Zurich.
E-mail: julia.casutt@gmx.net
Dr. Irina Pushareva is a senior researcher of the Institute of Russian history (Russian
Academy of Sciences). She is the author of a number of works in the field of Russian
pre-Revolutionary labor history (especially workers' protest movements) and co-editor
of the unique publication series called "Labour Movement in Russia. 1895 - February of
1917. Chronicle" published in 1990s-2000s.
pushkarev@mail.ru
Dr. Irina Shilnikova is Associate Professor at the Moscow Higher School of Economics.
She is the author of a number of articles and co-author of two books on Russia's labour
history and workers' protest movements in pre-Revolutionary Russia.
E-Mail: irinash_yar@mail.ru
Dr. Sjaak van der Velden is an independent Dutch researcher of history also affiliated
with the International Institute of Social History (IISH). He wrote his Phd. thesis on
Strikes in the Netherlands in 2000 and numerous articles since. Together with Heiner
Dribbusch, Dave Lyddon, and Kurt Vandaele he edited Strikes around the world. Casestudies of 15 countries in 2007. In 2011 Van der Velden was co-organizer of the Lisbon
conference Strikes and social conflicts in the XXth century. In 2012 he will publish
together with Jim Docherty the Historical Dictionary on organized labor.
E-mail: sjaakvdvelden@gmail.com

177

IISH Research Papers


1.
2.
3.
4.
5.
6.
7.

8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.

Tony Saich, Frank Pieke, The Chinese Peoples Movement Spring 1989: Some
Initial Impressions. Amsterdam, 1989
Ursula Langkau-Alex, Der Kampf fr die Demokratie und den Frieden. Die
Debatte in der Sozialistische Arbeiter-Internationale 1938/1939. Amsterdam,
1991. Zweite, um Literatur erweiterte Auflage 1992.
Jan Lucassen, Dutch Long Distance Migration. A Concise History 1600-1900.
Amsterdam, 1991
Jan Lucassen (red.), Symposium Racisme en Arbeidsmarkt: IISG september 1991.
Amsterdam 1992.
C.H. Wiedijk (in samenwerking met L.J. Altena, J.M. Peet, G.J. Schutte en
H.E.S. Woldring, Kalendarium Honderd jaar sociaal 18911991. Amsterdam,
1992.
Marcel van der Linden en Jan Willem Stutje, De Nederlandse vakbeweging, haar
basis en de staat. Een lange-termijnperspectief. Amsterdam, 1992.
Tjebbe van Tijen, Je bevrijden van de drukpers. Jongeren en hun eigen pers in
Nederland: 1945-1990. Met een bibliografisch aanhangsel over de tijdschriften
van Provo, Kabouter, de culturele underground- en kraakbeweging, vrije
stadskranten en punkfanzines. Amsterdam 1993.
Emile Schwidder, Selected Bibliography on Labour and the Law in Historical
Perspective. Amsterdam 1993.
Jan Gielkens, Books and articles on German labour law. Selected Bibliography.
Amsterdam 1993.
Larry Peterson, The Free Labor Unions and Arbeiter-Unionen in RhinelandWestphalia, 1920-1924: Statistical Sources. Amsterdam, 1993.
Gijs Kessler, Vakbonden in verandering. Een verkennende studie naar de
vakbondsontwikkeling in Rusland na 1985. Amsterdam 1994.
Ursula Langkau-Alex, Asiel en ballingschap in Nederland. Amsterdam 1994.
Marcel van der Linden. Social Democracy and the Agrarian Issue, 1870-1914:
Notes for discussion. Amsterdam 1994.
Reinier Deinum, Verenigd door Vaart. Gids van de bronnen betreffende
watertransport en havenbedrijven in het IISG en NEHA. Amsterdam 1994.
Jacques van Gerwen and Jan Lucassen, Mutual Societies in the Netherlands from
the Sixteenth Century to the Present. Amsterdam 1995.
Sander Vis, Survey of the Archival Sources Concerning Migration and Settlement
Held at the IISH. Amsterdam 1995.
Gijs Kessler, Trade Unions in Transition. Moscow 1994, a case study.
Amsterdam 1995.
Patricia Kennedy Grimsted, Displaced Archives on the Eastern Front: Restitution
Problems from World War II and its Aftermath. Amsterdam 1995.
Kees Mandemakers, Negen classificaties voor 19e en 20e eeuwse beroepstitels.
Amsterdam 1995.
Marcel van der Linden, Marx and Engels, Dutch Marxism and the Model
Capitalist Nation of the Seventeenth Century. Amsterdam 1995.
Adam Conroy, Christiania: the Evolution of a Commune. Amsterdam 1996.
Flemming Mikkelsen, Working-class formation in Europe: In search of a
synthesis. Amsterdam 1996

178

23.
24.
25.
26.
27.
28.
29.

30.
31.
32.
33.
34.
35.
36.

37.
38.
39.
40.
41.
42.

Gijs Kessler, The schools of communism under neo-liberal reform. Russias


traditional trade union movement in the transition to a free market. Amsterdam,
1996.
Alfons Fransen, Verzekering tegen Seerovers en Godts weer. Een onderzoek naar
de geschiedenis van de zeevarende beurzen, circa 1635-1815. Amsterdam, 1996.
Port Reports prepared for the Conference Comparative International History of
Dock Labour, c. 1790-1970, Amsterdam, 1315 November 1997. 3 vols.
Amsterdam, 1997.
Patricia Kennedy Grimsted, Archives of Russia Five Years After: Purveyors of
Sensations or Shadows Cast to the Past? Amsterdam, 1997.
Leo van Rossum, The Former Communist Party Archives in Eastern Europe and
Russia: A Provisional Assessment. Amsterdam, 1997.
Het Italiaanse complex. Crisis in de Europese politiek: de gevallen Itali en
Belgi. Amsterdam, 1997.
Ursula Langkau-Alex, The International Socialist Labor Movement and the
Elimination of the German Problem. A comparative view on ideas, politics,
and policy of the French, English, Swedish and US Labor Movement.
Amsterdam, 1998.
Klaus Misgeld, Trade Union Neutrality? The Swedish Trade Union
Confederation (LO) and the Trade Union International at the Beginning of the
Cold War. Amsterdam, 1998.
Marcel van der Linden, Producer Cooperatives:The Historical Logic of Workers
Organizations (I). Amsterdam, 1998.
Marcel van der Linden, Consumer Cooperatives:The Historical Logic of
Workers Organizations (II). Amsterdam, 1998.
A Working Guide to Sources on Historical Utopian Experiments in the Western
World at the Internationaal Instituut voor Sociale Geschiedenis, Amsterdam.
Compiled by Nienke van Wijk. Edited by Huub Sanders. Amsterdam, 1998.
Marcel van der Linden, Metamorphoses of European Social Democracy.
Amsterdam, 1998.
Jan Gielkens, Maranga mai te hunga mahi. De Internationale internationaal.
Amsterdam, 1998.
Simone Goedings, EU Enlargement to the East and Labour Migration to the
West. Lessons from previous enlargements for the introduction of the free
movement of workers for Central and East European Countries. Amsterdam,
1999.
Nicola Hille, Zur Darstellung und dem Wandel von Gewalt auf russischen und
sowjetischen Plakaten der Jahre 1917-1932. Amsterdam, 1999.
Hein Wiedijk, Het nieuwe socialisme van de jaren dertig. Frans en Nederlands
neo-socialisme gedurende de grote depressie. Amsterdam, 2000.
Jan Lucassen, In Search of Work. Amsterdam 2000
Free Love and the Labour Movement. Papers presented at the workshop 'Free
Love and the Labour Movement', International Institute of Social History,
Amsterdam, 6-7 October 2000. Amsterdam, 2001
Marcel van der Linden and Jan Lucassen, Work Incentives in Historical
Perspectives. Preliminary Remarks. Amsterdam, 2001.
Patricia Kennedy Grimsted, The Odyssey of the Turgenev Library from Paris,
1940-2002. Books as Victims and Trophies of War. Amsterdam 2003.

179

43.
44.
45.
46.
47.
48.
49.
50.

Jelle van Lottum, Immigranten in Nederland in de eerste helft van de 19e eeuw.
Een onderzoek op basis van de Utrechtse volkstellingen van 1829 en 1839.
Amsterdam 2004.
Sergey Afontsev, Gijs Kessler, Andrei Markevich, Victoria Tyazhelnikova and
Timur Valetov, Urban Households in Russia and the Soviet Union, 1900-2000.
Size, Structure and Composition. Amsterdam, 2005.
Klaus Misgeld, A Complicated Solidarity. The Swedish Labour Movement and
Solidarno. Amsterdam, 2010.
Jan Lucassen and Leo Lucassen, The mobility transition in Europe revisited,
1500-1900: Sources and methods. Amsterdam, 2010.
Patricia Kennedy Grimsted, Reconstructing the Record of Nazi Cultural Plunder.
A Survey of the Dispersed Archives of the Einsatzstab Reichsleiter Rosenberg
(ERR). Amsterdam, 2011.
Erika Kriukelyt, The Creation of Modern Prisons in the Russian Empire:
Research Report. Amsterdam, 2012.
Peter Waterman, An Emancipatory Global Labour Studies is Necessary! On
Rethinking the Global Labour Movement in the Hour of Furnaces. Amsterdam,
2012.
Striking Numbers: New approaches to quantitative strike research. Edited by
Sjaak van der Velden. Amsterdam, 2012.