You can't start a fire without a spark: Strikes and
class struggle in the Basque Country, 1914-1936.
Stefan Houpt and Juan Carlos Rojo Cagigal
Universidad Carlos III de Madrid · Instituto Figuerola
March 27, 2012
Abstract
What drove social conflict in Spain’s industrial areas in the period before its Civil War? This paper
is concerned with contrasting the determinants of working class conflictivity in northern Spain at the
beginning of the twentieth century. We concentrate on the region of Biscay, one of Spain’s emerging
industrial districts. Labour relations in Biscay have been extensively examined, but the economic
factors behind the processes of bargaining have never been formally revealed. We seek to discover
the forces behind and the consequences of the shifting power in negotiations over the period
preceding the Spanish Civil War. Using monthly data from 1914 until 1936 the empirical analysis is
especially concerned with testing the importance of economic factors in labour conflict. The variables
to be included in the contrasts are related with business cycles, real incomes, family cost of living,
delinquency, deprivation, and the different actors involves in labour conflicts. Our results suggest that
in addition to institutional, political and ideological elements that have been considered the main
drivers of conflict over this period, relative material deprivation and redistribution also played an
important role.
Paper to be presented at the EHS meeting, Oxford, March 31st, 2012.
Preliminary version. Please do not quote without permission of the authors.
Contact Address: [email protected]
1
Strikes are merely symptoms of more fundamental maladjustments, injustices, and
economic disturbances [...]. Strikes are always regrettable; but not always reprehensible.
Until the social millennium is attained, they will continue to occur and will sometimes be
necessary both as a direct defence against injustice and oppression and as the only way of
compelling the public to give its attention to hidden evils in industrial relations.1
(C.W. Doten, 1921)
Introduction
Wildfires and social violence require both combustion material to feed upon and initial ignition to
set off the process. Labour strikes as well as arson damage wealth and neither are Pareto optimal
dispute solutions, although, no doubt, they benefit someone and serve an end. This article seeks to
discover the logic behind the interwar blazes of social disruption in one of Spain’s rapidly progressing
industrial districts. The succession and the failure of different models of labour relations over the
interwar period in establishing social dialogue and providing institutions to resolve conflicts was
surely an element which contributed to the outbreak of the Civil War. A closer understanding of the
economic factors involved in these labour conflicts may enhance our perception of how the
conflagration came to be.
We will concentrate our efforts on strikes —one of the key manifestations of labour conflict. Work
on strikes, as strikes themselves tend to do, come in waves. Following the high incidence of work
conflict in the late 1960s and 1970s, labour academics made important efforts to recollect time series
on strikes and related variables and to apply statistical methods to contrast theories and hypotheses
on bargaining models and labour conflict. But the decline in labour union power in the 1980’s
removed labour unrest from the mainstream economics research agenda.2 However, the topic
resurfaced with important shifts. The organizational-political models proposed by Tilly and Shorter,
Snyder and Shalev added an interest in long-term cycles3, given the often contradictory results
obtained from macro data academics turned to micro data analysis and there was a surge in the
application of the newer developments in micro-economics (game theory, collective action and new
institutional economics). The study of labour conflict in Spain has remained dissociated from many of
these efforts and trends over long periods.
But since the return to democracy, notable progress in collecting data and approximating the issue
of Spanish labour relations has been made. Work on labour relations in Spain received important
boosts from the extensive empirical research of Tuñon de Lara (1972) and Soto Carmona (1989).
Labour relations in the emerging industrial region of Biscay were among the first to be examined with
a less ideological, more academic approach. Two monographs are considered path-breaking. Fusi
(1975) provided a thorough traditional British labour history approximation, whereas Olábarri (1978)
1
Knowles (1956: xii)
Franzosi (1989), pp. 358-360.
3
See Kaufman (1982) for an exposition on the debate over economic versus organizational-political models.
2
2
connected to the structured theoretical approach marked by Walter (1970) or Clegg (1976). Their
work laid the foundations in defining methodology, historical sources and the academic study of
labour relations in Biscay. Later contributions have concentrated in more detail on different periods
and have extended the analysis to other regions. Rivera (1985) analyzed the labour movement in
Vitoria, Luengo (1990) in Guipúzcoa, and Castells (1993) put together a synthesis for all of the Basque
regions of Spain. Sanfeliciano (1990) studied the socialist trade union UGT in Biscay during the Second
Republic, and Mees (1992) analysed the Basque Nationalist labour movement. Other important works
include Cabrera and Rey’s (2002) survey on employer’s associations. Miralles (1992) provides a solid
review of most of this literature.
In the pages to follow we will relate the existing narrative of what we know about labour conflicts
and the extensive data we have collected in order to contrast the importance of economic factors in
labour relations over this period. The first section of the paper will introduce a general background on
strikes and some simple comparisons between strikes in Biscay and Spain. This is followed by a
summary of Biscayan labour relations over the period we are examining. The next two sections
present different theories of labour conflict which have been applied to strikes over the past and the
data series we have collected and constructed for our contrasts. Section five presents our empirical
analysis of labour conflict. This is followed by a closer examination of the corporatist bargaining
model established during the later part of the period. We close with a presentation of our preliminary
conclusions.
What we know about strikes in general
Kennan (2008) situates the economic origins of labour conflict in that “employee’s labour is
generally greater than the wage paid by the employer [...]. This gives rise to a surplus to be divided
between the worker and the employer. [...] Under ideal competitive conditions, the employment
surplus [would be] negligible.”4 Under perfect competition, many employers would compete for
labour and bid up the wage until it equals the worth of the employee´s value added. And likewise,
many workers competing for jobs would bid down wages until they equal their alternative use of
time. Thus, wage disputes should occur only in non-competitive markets. In a two agent Hicks model
of negotiation, in which each party were rational and well informed, each would know the other
party’s concession curve and could agree on distribution of surplus without the waste implied by a
strike. This holds true even in an environment of shifting bargaining power. But strikes happen!
In search of an answer to why strikes happen, Ashenfelter and Johnson (1969) offered an
alternative bargaining model to explain industrial dispute. They propose that the root of the problem
is that actually more than two parties are involved in the bargaining process. Workers and capital
interests are at the heart of the contention, but it is firm management and labour unions that defend
4
Kennan (2008). Other issues at debate such as work safety, working conditions, etc. could be treated as disamenities to
be compensated in the wage – labour effort balance.
3
their interests. The representatives of capital and labour include other considerations in their long
term utility functions. Reputation, profit-based bonuses, work discipline, the survival and growth of
their trade unions, their personal survival and power as union leaders and managers are among the
many aspects representatives will include when negotiate the redistribution of surplus. Strikes are
therefore not a paradox to a Hicks world in which correct negotiation and adequate knowledge would
make a peaceful settlement possible. They are more likely the result of agency problems and the logic
of collective action. What Ashenfelter and Johnson (1969) suggest, is that union leaders have
incentives to lead their ranks and files to strike when labour’s expected wage increase is beyond what
management is willing and able to concede. Strikes will signal employer´s resistance and thereby
lower worker’s expectations. Or it will visualize threat and raise capitalist´s fear of regime change or
help convince stockholders of the damage that holdup con cause to profits without, at the same time,
harming union leaders’ and managements’ reputations. Strikes thereby become a useful tool to
approximate deferring positions and maximize agent utility, rather than a manifestation of
irrationality and faulty negotiation. Strikes will happen.
For more than a century economist have strove to establish regularities between what happens in
economies and the frequency, duration and depth of strikes. They have found strikes to be more
frequent when general economic conditions have been good. More recently contract cycles have also
been important to take into account, as strike behaviour tended to increase when collective
agreements or wage contracts drew near to their end.5 Strike incidence has been found to be procyclical and strike duration countercyclical to the business phase.6 Rees (1952) found a resemblance
of the way strike frequencies move with those for the production of durable goods. For him this
suggested that many of the grievances of workers are ‘durable’ or at the very least ‘semi-durable’ in
nature. With this he referred to the fact that “grievances can be stored up for long periods [and]
[t]hey are most likely to boil over into strikes, or to be used by union leaders as strike issues, when
business conditions promise that strikes may be successful.”7 Strikes associated to real wage
decreases have shown an extraordinary sensitivity to the business cycle, resistance funds and
unemployment rates. But overall strikes have inclined to increase wages.
Strikes and trade unions are closely intertwined and their relationship is worth looking into.
Whereas strikes are “only a temporary withholding of labour, […] a union is a permanent association
of workers.”8 Over the course of time, strikes as ‘weapons of labour’ have fallen into the hands and
control of business unionism. In Ross’ (1954: 23) words, “The union did not create the community of
interest among workers; the factory system did that. The union did not originate industrial unrest but
5
More recent micro-level studies show that much of the variation in strike activity is related to the structure of collective
bargaining, particularly the timing of contract expirations. Shalev (1980) and Kaufman (1981) for the United States and
Swidinsky and Vanderkamp (1982) for Canada. See also Mauro (1982) and Gramm (1987) in Franzosi (1995: 38, fn9)
6
Card (1990); Cramton and Tracy (2003) in Kennan (2008)
7
Rees (1952: 382). “The same provocation which causes a strike when employment opportunities are rising might cause
only grumbling during depression.”
8
Kaufman (1982: 477)
4
merely sharpened and took advantage of it. The union did not invent the strike but only sought to
appropriate it for its own purpose.”
This brings us to a second strand of strike characteristics, more related to the collective action
they require. Somehow strikes were made possible by the discipline and cooperation demanded by
the logic of the productive system: the division of labour. In the same way workers could be organized
and applied collectively to the production system, they could be withheld collectively.9 Therefore
strikes will have a much higher incidence in urban factory labour environments. The urban centres
have been focal points for political movements of social change. Biscay is no exception. It became one
of the areas of operation of the both the socialist movement and Basque nationalism. Ross
emphasizes that for promoters of radical change, trade unions were schools and organizing centres of
revolution and strikes were a means of preparing for the violent overthrow of capitalism. For them
there was a great danger in settling strikes to accommodate the grievances of the working class, for
this would reduce their thirst of revolution.
In the late nineteenth and early twentieth century, especially after 1917, strikes may therefore
also be seen as a measure of the threat of revolution. Socio-economic instability during the interwar
years increased people’s social risks and their anxiety concerning the appropriate institutional
framework to provide their well-being. “Among various forms of social unrest such as riots,
demonstrations, and social disorder, strikes have been historically considered the most potent
menace to a regime because they interfere with the functioning of the capitalist economic system on
which the regime is based. [...] Under the threat of revolution driven by socioeconomic factors, the
rich elites […] agree to a regime change by extending the franchise to workers, which then leads to
the enactment of social programs that the workers want. The change in political regimes serves as a
commitment device to people, signalling that instituting social programs is a serious action by the
ruling elites.” As suggested by Elster’s (1985) abdication theory, the commitment problem, thus, is
solved by “tying their own hands.” By incorporating people into national politics and providing social
compensation to them, the ruling elites are able to undermine the political base of radical leaders
who would take advantage of revolutionary moments. Notice that without a high level of threat from
society, the ruling elites have no compelling reason to transform their political institutions. Instead,
they would create a ‘‘seemingly democratic institution’’ without granting a voting right to the people
so that they can just co-opt some segment of civil society by sharing rents with them (Gandhi and
Przeworski, 2006).” 10 In Spain regime change was heralded by the Commission for Social Reform
created in 1883 to fulfil an important role in implementing labour legislation and improving working
conditions. Universal male suffrage was introduced in 1890 and industrial accident law in 1900.
Strikes had been approved in 1903 by government decree and were legalised in 1909.11
9
Ross (1954: 24).
Kim (2007).
11
Van der Meer (1997).
10
5
Another interesting aspect of strikes is their seasonality. Griffin (1939:54) found that the seasonal
movement of strikes is “in general determined by two elements, first the already-existing pattern of
business, and second the greater capacity for resistance on the part of employees in the early
summer and later summer as opposed to spring and winter. Here, as elsewhere, the weather shapes
the affairs of men.”12 Perrot (1974: 111-2) expands on Griffin’s idea : “the difficulties of daily life, for a
working class barely living at a subsistence level, were greater during the winter —the greater
necessities for such items as clothing, heating and better nutrition, the greater likelihood of illness. As
the land lies idle, workers cannot rely on the garden patches through which many of them acquire a
modicum of ‘country independence’ during the summer months.” Severe economic hardship, which
the winter-time typically brought to a much larger segment of the working class population, reduced
the collective resources at the disposal of workers and therefore their capacities to carry on strikes.
We also find strong traces of seasonality in strike frequency in the data for Biscay.
Figure 1
Payne (2009: 3) has recently characterized Spain as “the only European country to introduce a
democratic regime during the Depression decade, a time when half the states of Europe turned to
one or another form of authoritarianism.” This democratization process had been made possible by
the accelerated growth and rapid modernization during World War I and the 1920’s. Strike frequency
in interwar Spain reflects this strongly in its periodisation. There are two strong waves of strikes on
both extremes and an important decline in between —rapid growth, consolidation and crisis. Biscay,
the industrial district we will be examining, shows a pattern similar to the general movement in Spain
for the first two periods but less of a resurgence in strike frequency in the third period.
Silvestre (2003) has examined annual strike data for the whole of Spain over the first third of the
century and finds that characteristics of strikes disaggregated into its three components, frequency,
magnitude and duration, are comparable or slightly higher than those reported by Shorter and Tilly
(1974: 306-34) for France, the US and other European countries. The pattern of Spanish strikes seems
most similar to those observed in France, Germany, Belgium and Italy. Union density (the percentage
of worker’s affiliation to trade unions) was extremely low in Spain. Strikes motivated by wage increase
demands were around 35 % of all strikes. Sector breakdown shows construction, mining, metal
working and textile as those with highest conflictivity. Strikes are closely correlated with geographical
concentration and high worker concentration; Biscay is one of these industrial nodes. Around ten per
cent of the strikes registered in Spain were concentrated in the Biscay which had approximately two
per cent of total population.
12
Kennan (1986: 1130).
6
We can observe a lower average strike duration in Biscay over the whole of the interwar period.
This could be indicative of a lesser degree of information asymmetry, better negotiation or less means
of resistance on behalf of the workers. As we have indicated before, strikes in Biscay show summer
seasonality, no such patterns can be found in the cost of living, unemployment or money wages. The
months of highest incidence are July and August. We have only found similarity in the returns on the
Bilbao Stock Market, but mark-up profits in AHV —the dominant iron and steel concern— show
opposite trends. These seasonal patterns changed in the late 1920s and early 1930’s, they evened out
and shifted to the month of May. Increasing real incomes in the 1920s could be at the origin of these
shifts. Union membership in the Socialist Trade Union UGT grew more than proportionally in Biscay
over the 1920’s and we find a close association between the movements of inflation and trade union
membership which will be discussed in more detail further ahead.13
The narrative of Biscayan industrial labour relations
In the interwar period —the twenty some years leading up to the Spanish Civil War— labour
relations in Biscay can be divided into three distinct regimes which coincide with the periods of strike
frequency we have shown before. The period between 1911/14 to 1923 is a phase of conflictive
bargaining, which has its foundation in the steady rise of the socialist trade union (UGT) power
combined with employer’s associations’ tolerant resistance during the economic boom and the
gradual weakening of socialist trade union power after 1920 when internal excisions and economic
slowdown allow for a crackdown on unions. As a result of Spain’s war neutrality Biscay’s industry had
the opportunity to accumulated important war profits, at the same time speculation, shortages and
economic mismanagement doubled the prices of basic needs with money wages lagging behind over
most of the war. These two circumstances combined to creating a period of negotiation between
employers and unions. In 1921 the long announced economic stagnation set in, this ended the
prolonged period of redistribution lingering from the extraordinary profits made during war neutrality
and increasing costs of living. Growing political and social instability within the country finally led to
the endemic solution in 1923: a military take-over.
The second period, between 1923 and 1929, was a time of corporatist-state dominated labour
relations. The escalation of political and social violence in Barcelona and Madrid, especially the
growing threat of anarchist power in Catalonia, and the lack of direction and authority on behalf of
the existing Restoration political factions, had led to a military coup d’état and an endeavour at
creating a new institutional framework. General Miguel Primo de Rivera had staged what was
intended to be a short transitional military take-over, but clung to power and attempted to create a
moderate corporatist State aimed at overcoming social chaos and the tensions and antagonisms
between the old and new forces of power and wealth. This involved the repression of the more
13
Ashenfelter and Pencavel (1969) find a strong causal relation between growing trade union membership and changes in
price levels as a measure of changes in real worker’s income.
7
radical trade unions and seeking the collaboration of the more moderate socialist union within the
dictatorship’s corporatist labour relations management.
As Primo de Rivera’s corporatist project faded off into disinterest and lack of momentum, the
forces of regeneration united to oust monarchy and establish a Republic. Thereby democracy was
back in place after April 1931. During this third phase of renewed democracy, the working class and
its representatives pursued the goal of positioning labour representation within the new institutional
framework. The end of post-war boom of the 1920’s had increased trade union activity in quest of a
higher participation in surplus for its ranks and files. Due to its suppression during the Primo de Rivera
dictatorship, the newly gained freedom of action in the interim period (1930-31) and during the first
years of the Republic led to an increasing union member and social agitation. But the State-promoted
corporatist labour management system put into practice during the preceding Dictatorship was to
remain intact with minor adjustments throughout the Republic.
We can take a more detailed look at the evolution of labour relations over these three periods. In
its early origin’s the labour movement in Biscay can be best described by three characteristics: the
war of attrition between socialist trade unions and employers, its short waves of social agitation,
disappearance and reappearance, and the dominance of socialist trade unions. On one hand, the ‘allout war’ nature of the early industrial conflicts can be explained by the ease at which their protesters
were exploited and replaced by scabs. Union activity concentrated on miners and dock workers, both
with low skill qualification, reliant on and prone to physical force and easily replaced with the
constant stream of immigrant labour. On the other hand, worker’s resistance societies were weak and
employer associations, aware of their leverage, were unwilling to negotiate and keen on hungering
out dissent. They used scabs and pushed on outbursts of violence to force interventions on behalf of
the armed forces and civil authorities, presuming that intervention would be in their favour. Closer
descriptions of the three intense waves of strikes in 1890-92, 1899-1903 and 1910-11 —with the
general disorganization and passivity in between and which precede the period we examine— can be
found in Olábarri (1978: 396-404) and Fusi (1975: 81-104, 118-52, 203-21, 230-42 and 318-31).
The increasing record of connivance of public authorities —more concerned with maintaining
social peace than supporting industrial oligarchy— and the growing power of socialist trade unions
was to initiate a new era of labour relations in Biscay after 1911. Worker’s resistance was to refrain
from provoking general strikes and moved to focused local walkouts and stoppages and, at the same
time, employers backed down from their uncompromising die-hard stance and considered
negotiations as a means of reaching labour agreement. It was during this phase that the socialist trade
union UGT attained the height of its power. Its hegemony and moderate path to social redistribution
were to be challenged by yellow unions —Catholics and Basque nationalist— and excisions within the
socialist ranks —left wing Anarchist Sindicatos Únicos and socialist youth Communists. This phase of
social bargaining was initiated by the blatant failure of the 1911 general strike, the labour agreement
signed between the Graphic Arts union and employer’s association in December of 1912 and the
opening of negotiations for minimum wages in the mining sector the following year. There were also
8
important changes taking place within UGT: on one hand, the creation of the powerful metal worker’s
trade union and, on the other, the final showdown between moderate and radical socialist leadership,
with the former prevailing but sowing the seed for secession of radical dissident within the socialist
movement in 1918-19. The rise in union membership had been concentrated in the yellow trade
unions until 1914, but the great majority of working population shifted towards UGT in 1916-17 as
they sought to protect themselves from the rapidly increasing prices of their basic needs.14
Spain’s war neutrality enclosed the explosive mix for strikes, a working class grieved by rising costs
of living, and a thriving war economy providing industrialists with extraordinary profits and therefore
determined to avoid stoppages and make concessions. But the policies adopted by employers in order
to stop the advance of union power were merely being postponed by their good business
perspectives. Strikes spread both in scale and scope and were preceded, accompanied and follow up
by negotiations. The violence of the initial years only reappeared in 1919, spiralled by assassinations,
clashes between rivalling trade unions and the rising political and revolutionary tone given to strikes.
Go-slow strikes and boycotts were introduced as new formulas of resistance and protest. Wildcat
strikes kindled by extremist minorities in their struggle to win workers out of socialist trade union
control, using violence and radical discourse to embezzle fellow workers. These changes were faced
head on by the employer association’s campaign to reduce salaries to pre-war levels. The climax in
this phase of cooperative labour relations was attained between 1918 and 1920 at the height of strike
frequency and followed with a slow decline in the search for negotiation, bargaining and agreements.
During this phase of rise and decline the large metalworking companies and the socialist unions
dominated labour relations in Biscay.15 Some of the factors which contributed to the deterioration of
relations in 1920 were the end of the war, socio-political revolutions in post-war Europe, the
consolidation of the Russian Revolution, continuously rising prices of basic needs, the worsening of
labour conflicts in Barcelona and Madrid, the 8 hour working day beginning of 1919, and the rise of
Anarchist Sindicatos Únicos in March of 1920. Concessions had also been strained beyond the means
of small and medium sized enterprises. The larger firms had excellent business perspectives as long as
their production processes did not come to a halt. They were more concerned with discipline than
with wage concessions. By July 1920 there were major divisions between the groups of large
businesses and small and medium size enterprises and violence began to reappear in strikes against
small and medium size firms — cabinetmakers, construction and dockworkers. Labour relations within
the large corporate was different and Olábarri (1978) sustains that only the pact between UGT and
large metal-working firms prevented labour conflict from escalating to levels like in Barcelona and
Madrid.16
With the radical wings of the labour movement preparing for revolution and a major labour
market crisis at hand, five general strikes with revolutionary character were declared between July
14
Olábarri (1978: 404-5)
Olábarri (1978: 406-8)
16
(arson naval cruise ship Alfonso XIII Naval Nov 20; assassination CEO AHV Manuel Gómez Jan 21)
15
9
1921 and September 1923. Even though employers refused to negotiate with communist and
anarchist trade unions the economic situation weakened socialist unions and the violent strikes they
were unable to control curbed social dialogue and their role as bargaining partners.
And so during the first years of Primo de Rivera Dictatorship union activism was slow to recover
due to the adversities of economic recession, internal clashes and working class deception. But
government arbitration in the metal sector strike in 1925 was followed in November of 1926 by the
government initiative to institutionalize the resolution of labour disputes through corporatist
arbitration boards —comités paritorios— with mandates to dictate terms of agreement and resolve
labour conflicts. Communists and sindicalistas únicos refused to participate in these organisms, but
socialist and yellow unions rivalled for worker’s representation on these bodies. It was nevertheless
the Catholic and Basque nationalist unions, the non-revolutionary worker’s unions, that increased
representation in the late 1920’s and consolidated their position into the 1930s. The socialist union
UGT came under heavy fire for the ‘sell-outs’ in the early twenties and their collaboration during the
Primo de Rivera dictatorship. Combined with their steady relative loss of membership, throughout the
1930s, their conflict bargaining would be strongly conditioned on the criticism they were to receive
from their left-wing rivals. Employer’s associations reorganized throughout the 1930s mainly to resist
what they felt to be a government offensive against them led under the auspices of the Socialist
Labour Minister, Largo Caballero. Until then the arbitration boards, now renamed ‘mixed juries’ in
reminder of the First Republic, had resolved social conflict satisfactorily in Biscay. During the
dictatorship and especially after its fall they were not always accepted as representative by the
working class. They excluded minorities and overrepresented the socialist trade union. But they also
generated opposition among the employers mainly because of the political mandates of their
presidents and their tipping vote in these juries. This was aggravated by the number the conflicts in
which workers refused to be represented by socialist union leaders and the growing nature of
revolutionary strikes in the very early 1930s. This intra-union rivalry lessened to some extent as
nationalists concentrated on the Basque Statute and job protection and socialist thereby gained a
stronger grip within these bodies, but the more radical left-wing unions continued to sabotage the
socialist union´s authority to represent workers. Employers opted for negotiating with the socialist
unions as the represented the majority of workers and in hope of receiving state aid to overcome the
industrial crisis in exchange. It was mainly Communists and Anarchists who continued their
revolutionary crusades with general strikes, assassinations, violence and wildcat strikes beyond the
control of mixed juries.
According to Sanfeliciano once the economic motives causing strikes had been deactivated by
negotiations and labour market improvements obtained by the socialist unions, the Republican strike
pattern came to reflect the political cycle. Strikes surge in June of 1931 coinciding with the Elections
of the Constitutional Congress; from August to September 1932, with no apparent labour grievances
but coinciding with the failed coup d’état attempt staged by General Sanjurjo; and resurged again
around the general elections in November of 1933. The strong repression exerted by the centre-right
10
government from 1934 to 1935 reduced conflictivity drastically, only to resurface in February of 1936,
with the victory of the Popular Front, but again tied to the political cycle.17 Díaz Freire holds a very
different view. Based on his work, he comes to the conclusion that the main driver of strike activity in
Biscay was unemployment. Between 1931 and 1933 half of the strikes were called for this reason.18
Cronin (1978) insisted that the “parallel movement of strikes in different industries suggests that
what matters most in determining strikes are not those aspects of technology and organisation that
differentiate one industry from another, but historical factors common to each other.” A sector
breakdown of strikes in Biscay between 1931 and 1935 concentrate the majority of them in the metal
sector, mining and construction, with a reduced relative weight in the metal sector compared to
previous periods. The construction sector gained importance in strike incidence and in 1936 44% of all
strikes were called in this industry. The growing loss of weight of the metal sector and mining can be
attributed to the dominance of socialist unions in these industries. Altos Hornos de Vizcaya, whose
conflicts had traditionally dragged other sectors into conflict, became the battlegrounds for interunion disputes, foremost UGT and Anarchists. But the Metal Sector Union always maintained a
moderate position seeking to alleviate the effects of the crisis. On occasions union leaders had to
refrain their ranks and file from wildcat strikes. Conflictivity in mining was rare. Most of them involved
the recognition of worker’s rights, almost always with a political connotation.19 (Sanfeliciano 1990:
312).
After November 1934 the victory of centre-right in the general elections and their threat to undo
many of the previously attained social advances swung UGT unto the revolutionary course. Hitler’s
ascent to power and the political changes in Austria unleashed the discourse of fascist threat and
labour conflictivity picked up new momentum. Labour relations collapsed completely between
October of 1934 and January 1936 after the violently repressed October revolution attempt. There
may have been massive lay-offs in Biscay as a consequence. Political uncertainty and outright
repression reduced conflictivity substantially until the general election won by the Popular Front in
February of 1936. But the election victory set off a most intense wave of strikes and demands.20
Revealing the nature of strikes
As with the blind men and the elephant, understanding labour conflict measured by strikes can be
confusing. This may be best visualized by the very different approaches to strikes taken by
economists, sociologists and political scientists, which have been succinctly summarized by Franzosi
(1995: 7-11)
17
Sanfeliciano (1990: 300-02).
Díaz Freire (1990: 87).
19
Sanfeliciano (1990: 307-12).
20
Olábarri (1978: 423-30)
18
11
A. Business cycle theory. This approach concentrates on the relationships between business cycles,
the bargaining position of labour and strikes. It proposes higher intensity of strikes during
economic upswings because of the redistribution opportunities generated by business
perspective. The preceding recession strengthens the bargaining position and collective action:
the relative deprivation caused by the downward pressure on real wages and strengthening of
worker’s organizations during downswing promoted by the increasing unemployment enhance
worker’s collective strength. During the upswing, workers improve their bargaining power as
labour becomes relatively scarcer and employers become more sensitive to hold-up. The situation
inverts during recessions as worker’s leverage decreases with excess labour supply and company
inventories which make them less sensitive to holdup.
B. Economic hardship and relative deprivation theory. This theory sustains that the deterioration of
economic living conditions and increasing social injustice are at the heart of conflict. Negative
amenities have a multiplicative effect. Lower real income decreases food intake and increases
morbidity. Unemployment increases family tensions and desperation. Social violence may blaze
out of control once ignited.
C. Resource accumulation and collective action. Collective action depends on a group’s capacity to
mobilize resources and organize. According to this theory, the success of labour protest depends
on organization and resources, this links labour conflict to the level of unionization and the
permissiveness of the political cycle. Labour mobilization increases when the political framework
allows for it. Strikes have a higher possibility of providing change in periods of greater worker
affiliation, more efficient trade union organization and significant political integration.
D. Institutional theory. Strike activity is determined by the calendar of collective contracts, especially
in the case of industry-wide or nationwide bargaining.
E. Political exchange theories. Labour oriented social-democratic parties provide alterative less costly
mechanism for achieving a more favourable distribution of resources. The objective to improve
redistribution is thereby shifted to capturing the State.
F. Franzosi (1995) adds Marxist theory whereby the rise of class conflict is linked with long-term
immiseration, rather than to prosperity. The recurrence of crisis, perceived as the falling hammer
of continuous downswings in the long-term economic cycle, heightens and escalates the levels of
class confrontation.
G. Modernization theory views social violence as a reaction to the social disruption caused by ruralurban immigration and its impact on the existing social ties in urban settings. Violence is seen as a
transitional state until new social ties and networks have been established.
It would be a fallacy to maintain any of these as a single explanation for strikes. In the same way,
we are aware of the fact that the complete understanding of labour conflict cannot be based on
workers and their actions alone. A complete understanding needs to include the other agents
involved, both their social and economic relations, and their interests and strategic interactions.
12
Special attention should also be paid to extreme conjectures. They may mark turning points and
moments of templing in the process of forging ideologies, alliances and new formulas in the sphere of
labour relations. Extreme events are moments during which the lethargic disinterested parties are
drawn into conflict, unborn beasts show glimpses of their natures through the radical tactics and
demands adopted, trust is undermined by rapid concessions followed by cruel repression and
positions are hardened and polarized.
Well aware of shortcomings, our analysis will concentrate on the economic aspects determining
strikes. In light of the weak evidence for the relevance of economic variables in explaining strike
frequency found by Silvestre (2003) using macroeconomic data for all of Spain, we would like to
contrast the importance of economic change and necessity in the explanation of strikes in a more
closely defined setting. Such findings could indicate that a regional approach concentrating on the
large industrial districts of Barcelona, Madrid, Valencia and Biscay may reveal very different patterns
from those obtained with macro data. It would also move the impact of relative deprivation on social
violence in the auspice of the Spanish Civil War into the foreground.
Data series used
One of the advantages of this study is that it concentrates of a small geographic area in which a
fairly rapid process of industrialization took place in an isolated form. Localism, the limitation of social
conflicts to a local level, is predominant in the political arena of the time.21 There were no competing
loci within a reasonable distance. The closest industrial centre was in Barcelona, some 600 km away.
As one of Spain’s major ports, it had well integrated commodity markets and as a consequence of the
late 19th century iron ore mining boom it had well-functioning labour markets. We have been able to
construct a very detailed monthly data base for Biscay over this period.
Therefore the empirical examination of labour unrest will introduce monthly data frequency.22 We
will use a recently calculated cost of living series for a breadwinner family in the iron and steel
industry.23 This has been constructed with monthly price series from the municipal statistical bulletin
of Bilbao. Real incomes are calculated from average money incomes for workers of the Baracaldo
21
Fusi (1975)
“Burns and Mitchell conclude after much study, ‘annual data are exceedingly crude materials for comparing the cyclic
behaviour of different activities in the same period or of the same activity in different periods. They obscure timing
relations, they make it impossible to trace cyclical patterns with confidence, often they obscure and sometimes they
obliterate cyclical fluctuations.’,” in Arthur F. Burns and Wesley C. Mitchell (1946): Measuring Business Cycles. New York:
National Bureau of Economic Research. P. 43. Quoted in Rees (1952: 372).
23
Rojo Cagigal and Houpt (2011). The money earnings for workers presented are average monthly income for a blast
furnace worker at Baracaldo factory of Altos Hornos de Vizcaya. It runs from January 1914 to December 1921. We have
chosen blast furnace workers because their income is close to the average workers’ income as shown in Figure 3. From
January 1922 to December 1927 we no longer have detailed data on a shop level and only have the average monthly
income of all workers at the same factory (again including bonuses and extra hours). And finally from January 1928 to July
1936 we have interpolated annual average incomes to obtain monthly averages because we have no monthly
observations.
22
13
factory of Altos Hornos de Vizcaya [AHV]24. Altos Hornos de Bilbao was created by merger of two of
the leading iron and steel companies with a smaller tinplate factory in 1901. They employed 5,905
workers in 1916 and 8,300 in 1930, approximately one third of the workers employed in the iron and
steel industry.
Unemployment rates are not available for this period but they have been proxied with inverse
nuptiality rates following Southall and Gilbert (1996).25 Nuptiality rates not only reflect the marriage
postponement due to unemployment but also growing underemployment and any reduction in
earnings and income. We believe that the inverse marriage rate is an excellent proxy for un- and
underemployment over this period. It also avoids the need for instrumentalisation which would have
been required if we had data series for unemployment. Unemployment series are codetermined with
money wages and profits.
Business profits and different measures of the business cycle have been used to measure the
variation of economic surplus over time. Monthly business profits can be approximated for Altos
Hornos de Vizcaya between 1914 and 1921 by calculating mark-ups for each product produced, and
multiplying those by the amount of each product sold. We have also used the monthly stock market
index for Bilbao and the monthly stock quotes for AHV, the dominant industrial firm in the metalworking sector. The different series will be discussed in more detail as we proceed.
Monthly strike frequency and days striked per month have been collected from Lazcano (1950)
and Sanfeliciano (1990).26 Both series are based on exhaustive scrutiny of local newspapers. The
Sanfeliciano series registers only the strikes in the 1930’s and will be used to examine the last period
of labour relations.
Figure 2
As a first glance at the relationship between incomes and conflictivity, we can use the annual data
on strikes provided by Olábarri (1978: 498) to compare the coincidence between the number of
strikes and tensions on family budgets. Figure 1 shows how our lower bound cost of living calculation
follows a sharp and constant increase throughout the First World War and the immediate post-war
years up to 1920, a downward trend between 1921 and 1923 followed by a very gradual increase up
to 1930. A second sharp increase is observed between 1930 and 1931, followed by stability and a slow
decrease between 1932 and 1934, further stability in 1935 and a renewed increase in 1936.
24
For a detailed discussion of the source see Fernández de Pinedo (1992). Annual averages have been taken from
González Portilla (1984), pp. 74 and 85 (proceden de los libros de cuentas de AHV Ejercicios 1902-1936, Consejo de
Administración y Carpeta Financiera) mirar signatura AFV.
25
“If marriage and trade union unemployment corresponded closely in timing and, in some senses, in amplitude of
variation, it is tempting to treat both as proxies for the unemployment rate [...] [especially because] the further back we
go [...] the less meaningful unemployment was for large sectors of the labour force.” Southall and Gilbert (1996: 55)
26
Olábarri (1978) has noted that Lazcano’s information is based on newspaper coverage and not always complete, but it
supplies us with monthly data and covers all the important stoppages.
14
When confronted with the money earnings for day-labour at the biggest iron and steel concern in
Bilbao, Altos Hornos de Vizcaya, we can make some interesting observations. Labour conflict,
measure here in the number of strikes, reacts strongly when nominal earnings decrease. Strikes
increase even more vehemently when average nominal family income falls below the lowest cost of
the family basket, i.e. they are not able to acquire the amount of calories required for work energy
and basic vital needs. This is the case in 1917 when the situation escalated into a revolutionary
general strike, but also in 1919-20 when nominal earnings came close to the lower bound anew and
briefly in 1922 when employers were trying to adjust money wages back down to pre-war levels. The
decrease in total earnings between 1929 and 1931 seems again correlated with increasing
conflictivity. If anger is related to hunger (or relative deprivation) over the period examined,
breadwinner families had numerous occasions on which their families hungered and these coincide
with moments of anger as expressed by labour conflicts. At the same time conflictivity disappears as
earnings situate above maximum costs of living over most of the 1920s until 1930 when average
workers’ earnings again fall below the maximum cost of living. In this case anger may be more related
to loss of purchasing power or relative deprivation.
The figure we present here is no more than a rough approximation. But the series for real earnings
and cost of living for an average family are based on very detailed data proceeding from a single
source. The monthly data introduces seasonality. This adds an important aspect to the issue of
nutrition. Studies on energy requirements in present day agricultural environments in East Asia today
have shown that during high workload seasons land workers consistently consume more energy than
they take in. This is sustainable during short intervals, but energy reserves must be built up during
idler seasons of the year. Factory work does not allow for an inter-annual energy reserve
compensation. Energy requirements are constant year round and workers will be very sensitive over
very short periods to sudden falls in real income. Families can compensate for this to some point, by
having the remaining members of the family consume less. But in the mid-term this would have
negative effects on their health and family harmony.27
Figure 3
According to our previous analysis on living conditions, real family incomes worsened between
January 1914 and December 1919. They improved from January 1920 to April 1930, from then on
they worsened until August 1932. They recovered to the levels attained in the mid-1920s by May
1935 but started a downward hike from then on. This is visualized below with welfare ratios, the
number of times a family can buy its basic consumption bundle which provides the necessary
nutrition for basic body functions, maintenance and work. Some of these results are surprising,
27
See Rojo Cagigal and Houpt (2011) for mortality, pawning and child abandoning response to real income variations in
Biscay over this period.
15
others expected. War and post-war economy deteriorated workers living conditions, dictatorship and
right-wing democracy improved them and left-wing democracy maintained or worsened standards of
living for these workers.
Model.
Our endeavour to reveal what is hidden in the high-frequency data pushes us to concentrate on
the short-term factors generating labour conflict, together with our interest to discern economic
factors contributing, the most obvious approach to choose is the business cycle model. According to
Snyder (1975, 1977) we should expect the economics to prevail in the explanation of strikes as we are
in a context of high labour institutionalization.
In the story we will tell, a large part of the social relations behind the economic bargaining will be
lost; the interests of the different actors and their strategic interactions will not be visible. Strikes are
after all only a part of a wider repertoire that workers use and to which the other actors react. Mass
meetings, rallies, protests, silent strikes, menaces, threats, street fighting, sabotage, absenteeism, goslow are other arms at their disposal. Employers do not simply sit back and watch; they hire, fire,
blacklist, lock out and reorganize the production process to erode worker’s power. The State also
intervenes as the guarantor of peace and order through the police, the military, courts, laws,
parliament, arbitration boards, the legal framework and political repression. And finally investors
pressure, reallocate their financial assets, condition, lobby and oppress.
Many of these aspects cannot be included in a thorough statistical examination, because we do
not have data to capture these dimensions, we lack knowledge of the functional relations and the
strategic dynamics. Within these limitations, we can still approximate the world of conflicting
interests and the struggle for redistribution in a simple model proposed by Ashenfelter and Johnson in
1969.
The Ashenfelter and Johnson model
Although Ashenfelter and Johnson developed their analytical framework to analyse strikes in postwar Western economies, many of their precepts will also hold in earlier phases of capitalism, such as
Biscay in its late phase of industrialization. The model includes seasonal dummy variables to account
for regularities in contract renewals in specific moments of the year, a variable which reflects the
cumulative effect of real earnings variation in the recent past, the level of unemployment, a measure
of business perspectives (some measure of business profits), a time trend to capture the structural
evolution of the bargaining system over time and an error term. What the model tries to detect is the
play of forces between what workers need to ask for and if they are in a position to ask for it and
what employers are able to concede and if they are willing to concede it.
=
+
+∙∙∙ +
+
∑
∆
+
+
+
+ ( )
16
In the original model St’ is the probability of a strike, which should be approximated by St/N, the
number of observed strikes over the number of contracts that are renegotiated. We don’t have data
on the contacts being negotiated in each month, so the probability must be proxied by the total
number of strikes registered (this is consistent with assuming contracts being renegotiated evenly
throughout the period).
Rt is the real worker’s earning. The variation of real income has been included as a two period
moving average. This is the statistically significant lag length with which pawning reacted to shocks in
real income in our previous work, which would be indicative of economic strife affecting working class
families.28 We have avoided using a distributed-lagged Almon polynomial as was included in
Ashenfelter and Johnson (1969). The original model includes the lagged moving average as a measure
of the difference between actual and expected change in real wages. Determining the lag length and
the degree of the polynomial of the Almon distributed lag distribution remains very arbitrary. They
also impose a specific functional form on the lag structure which later has shown very time and
country specific and may only be suitable a context of constantly increasing real wages.29
In the case of our short-term analysis we were more interested in seeing how short-term hardship
affected strike frequency, so we have followed the modification proposed by Stern (1978) using levels
and changes in order to capture both the effect of prosperity and relative deprivation. For the lagged
effects we have preferred to use our empirical findings on how quickly working class families react to
changes in income. In a parallel paper on living conditions in interwar Biscay we found the lag
between income decreases and a significant increase in pawning to be two months, child abandoning
increased after 4 months and mortality after 4 to 5 months.30 Hunger and anger should be providing
combustion material for strikes after two months of income reduction, so we have included a two
month lagged moving average of income variation in the regressions. The regressions in Table 1 show
different combination of the income variables. In the majority of the cases the coefficients show the
expected inverse relations and are statistically significant.
UNt is the current level of unemployment proxied by the inverse of nuptiality. Nuptiality rates not
only reflect the postponement of marriage due to unemployment, but also growing
underemployment and any other form of reduction in earnings and income. πt-6 is the six period
lagged Bilbao stock exchange index. We have applied a six month lag to the Bilbao Stock Exchange
index based on Moore (1983: 148) who calculated an average 6 month lead for troughs and a 5
month lead for peaks between the US Common Stock Price index and business cycle in the US
between 1873 and 1945. Profits, and real income and unemployment rates will interact to signal the
give and take dynamics between employers and workers, their willingness to grant concessions and
28
See Rojo Cagigal and Houpt (2011).
Shalev (1980) finds that the original Almon specification may only have empirical validity in periods when wages and
prices increase at a relatively low and stable rate, allowing expectations to be realistically based on past experience.
30
Rojo Cagigal and Houpt (2011)
29
17
mount collective action respectively.31 T is a time-trend variable introduced to approximate long-term
structural changes and εt is an error term.
The expected signs for the coefficients for real income and unemployment are negative and for
the time-trend is positive. According to the original model the previous level of profits has an
undetermined sign. An increase of profits increases probability of management to give in to wage
demands but also tends to increase wage demands made by workers —perhaps more than what
management would be willing to concede. The overall balance will determine if unions lead workers
to strike and so the effect of changing profits remains undetermined.
The table below shows the results of the OLS estimation in regression (1). The results provide
strong support for business cycle hypothesis of strikes. Economy matters! With the exception of the
time trend, the coefficients show the correct signs and are significant. The serial correlation in the
first regression is indicative of omitted variables and the residuals recommend a GLS estimation —
regression (2). This pushes unemployment out of the foreground but earnings variations and profits
remain significant variables. Both specifications reflect the seasonality of strikes in July and August,
very similar to what the original model found for the US. With other causes to be taken into
consideration both estimations show that economic variables are capable of explaining an important
part of the variability of strike frequency over the period. The residuals shown below reflect the three
periods of labour relations regimes in Biscay. If the differences in the coefficients estimated by OLS
and GLS can give us any hint on the variables omitted from the model, these would be such that they
lower the effect of changes in real income and that of unemployment. Higher job security and better
working conditions are possible candidates.
Table 1
Figure 4
The first period of conflictive bargaining stretches from 1914 to 1923. It includes important surges
in strike activity which are not explained by the movements of the explanatory variables included in
the regressions; with extraordinary levels of strikes taking place in July of 1917 (the revolutionary
strike of 1917), May of 1919 and August 1920 (the height of union power: UGT pacts with CNT to
oblige all workers admitted to the factories to be affiliated to either UGT or CNT). All the observed
strike peaks are replicated in the estimated values but they are underestimated. Economic variables
alone cannot explain the explosive nature of conflict in this stage although it does very well in
explaining when strike waves do occur within that particular economic bargaining framework.
During the Primo de Rivera dictatorship the fitted observations continue to pick up strike
occurrence until 1926 when the arbitration boards enter the scene. Thereafter until the end of the
31
Franzosi (1995: 32-33)
18
dictatorship strikes practically disappear. In the last phase of labour relations the Ashenfelter and
Johnson business cycle model fits the observed data well during the first two years (1930-31) but is
completely inadequate thereafter. We will therefore move on to examine other factors and processes
which may contributing to the new pattern of strikes.
For the socialist union UGT strikes stopped being a signalling device after 1926. They were no
longer part of a procedure of ‘rite’ful redistribution as it had been in the previous period. Before 1923
asymmetric information had been at the base of what seemed to be a non-Pareto optimal solution to
surplus disputes. Private information was held by the firm about the demand for its product. Strikes
became a device to make firms signal whether they had good enough business perspectives to
therefore grant higher wages immediately or whether they had poor perspectives and resisted the
strike to force lower wages demands. Ex post the private information on demand would become
available to labour by way of the audited accounts in the shareholder reports, dividend
announcements and the ´climate of prosperity’ workers breathed in their shops.32 The production
foregone by strikes may be considered the cost of extracting such private information at each
moment of time.
In any case it is worth noting that if anything remained in place during the following phases of
corporatist labour relations it is the high sensitivity of strikes to decreases in real earnings —see both
regression (4) and table 2. It might be interesting to insist on this before we move on to analyse
labour conflict during the Republic. As our title suggests, we postulate that hunger and anger are at
origin of labour conflict over the period we are examining. They are, so to say, the combustion
material which unions may ignite. Hicks (1963:163) has insisted that strikes are not a rational strategy.
The economic surplus to be redistributed shrinks during stoppage as capitalists and labour argue over
how it is to be divided. Anger arising from the growing awareness of relative deprivation may be at
the origins of such irrational behaviour. Knowles (1952: 314) found that strikes in interwar Britain
were mainly aimed at avoiding decreases in wages. Both Knowles for Great Britain and Griffin (1939)
for interwar US place particular importance on 1920-21, a period in which wholesale prices fell by
more than a third in both the US and in the UK. In that same period, strikes fell by 30% in 1921 and
67% in 1922 in the US and by 50 % in GB in 1921.33 Strikes acquire a completely different meaning in a
context of growing recession. They will no longer be used as a tool for extracting information on
business perspective (it is now general of general knowledge that they are bad). Rather than this, they
will be employed to curb the burden of economic adjustment on workers.
Some very convincing evidence on procyclical fluctuations of strike activity for the interwar period
was provided by Jurkat and Jurkat (1949) for the US between 1915 and 1938.34 The NBER US business
32
Corporation laws obliged firms to publish their annual accounts.
Kennan (1986: 1116)
34
It has been repeated by Rees (1952) and Weintraub (1966); although questioned by Scully (1971) who found differences
in the frequencies of both cycles using spectral analysis. Nevertheless it adds the interesting aspect of how inflation and
strikes relate to each other.
33
19
cycles were divided into nine phase, phases 1, 5 and 9 being trough, peak and trough, respectively.
The total number of months between trough and peak —or peak and trough— were divided into
three phases of equal length, expansion phases 2, 3 and 4 and contraction phases 6, 7 and 8,
respectively. For each phase we count the amount of strikes in that phase expressed as a per cent of
the average number of monthly strikes over the whole cycle. Seventy-three per cent in cycle one,
phase one means that the number of strikes observed in these three months is 73 per cent of the
average monthly number of strikes over the whole cycle (3 months times the monthly average). In our
exercise we use the cycles in the cost of living index. Table 2 shows there is a high level of coincidence
between these inflations cycles and the relative frequency of strike waves.
Table 2
Knowles (1952: 147) has stressed the cost of living as “a non direct indicator of economic pressure
on the worker’s life, in so far as workers live on narrow margins they might be expected to resist rises
in the cost of living by striking. Applying this antiquated methodology to measure economic pressure
on worker’s life, we find that the cyclic movement in the cost of living is very closely tied to the strike
cycle. Strikes have a high incidence at the peaks of basic-needs price increases and linger some time
after. Hunger or relative deprivation cannot be discarded as ultimate sources of labour conflict during
neither of the cycles. As to why strike frequency does not remit until some time after prices stop
increasing, perhaps Eric Hoffer has expressed the dormant nature and lagged timing of addressing
grievance best. “Discontent is likely to be highest when misery is bearable; when conditions have so
improved that an ideal state seems almost within reach. A grievance is most poignant when almost
redressed […] It is not actual suffering but the better taste of things which excites people to revolt.”35
With this in mind, we can return to the change in the bargaining scenario initiated by the dictatorship
in November of 1926.
Cat and mouse
The corporatist arbitration boards introduced by Primo de Rivera’s Labour Minister Eduardo
Aunós, opened the door for socialist trade unions to institutionalize bargaining and outrival
contenders. The new regime followed a two-tiered strategy of repressing radical left-wing unions and
at the same time offering power and overrepresentation to the moderate socialists and the yellow
unions. UGT merely followed a pragmatic strategy of gathering power, consolidating its organizational
prowess and converting itself nationwide into the hegemonic trade union it had been in Biscay. For
socialists, power and organization would pave the way to revolution. Arbitration boards were
implemented in November of 1926 and already after the fall of the dictatorship in January 1930, UGT
had become the strongest political worker’s representative in Spain and as such it was naturally
35
Eric Hoffer, The True Believer, New York, Harper, 1951, pp. 27-28. Quote taken from Rees (1954: 214).
20
predestined to hold the reins of labour policy in the future centre-left governed Republic. The political
architect of this strategy throughout and beyond dictatorship, UGT Secretary General Largo Caballero,
from his position in the Labour Ministry had only to refurbish the corporatist system to accommodate
a Republican flair.
Success was guaranteed as Anarchists refused to participate in these state-corporatist bargaining
organisms and they thereby self-excluded themselves. In Biscay the Basque nationalist and the
Catholic unions were strongly underrepresented by the proportional voting system. And so Socialist
pounced on the opportunity to consolidate their power from within the State and then promoted the
system on as a solid vehicle to continue accumulating power during the first years of the Republic.
Once they lost their grip on the captured state in November of 1933, they turned their organization
and discourse to revolution. How were the ranks and file accommodated so easily? Between its 16th
Congress in 1928 and its 17th Congress in October of 1932 UGT carried out a strong centralizing
reform within the trade union. This included the creation of 37 National Industrial Branches which
came under the full control of the national trade union hierarchy. Strikes and relations to employers
were relegated to the centralized bodies of the union.36 For Tilly (1989: 444) “to the extent to which
crucial decisions visibly take place at a national scale […] and the power position of labour is
uncertain, we can expect political strikes to become prominent.” In such a scenario the temporal
patterns of strikes should change accordingly in order to accommodate from the local productive
fluctuations to changes in policy, regimes and power on a national level.
Table 3
Table 4
In search for the political cycle which may have been introduced by UGT’s strategy of
consolidating more organizational power within the corporatist-state framework we have modified
the original model to include a dummy variable we have named ‘rivals’. This dummy variable will be
zero over almost all of the period and take value one for the months surrounding the publication of
agreements struck by the arbitration boards (comités paritorios until the Republic and jurados mixtos
afterwards) in the metal sector of Biscay.37 The hypothesis being tested here is that rival unions would
intensify strike activity whenever UGT-dominated arbitration boards obtained agreements on
holidays, minimum wages, overtime compensation, wages adjustments, etc. Two motivations are
behind this type of sabotage. On one hand, it sends a strong signal to employers that UGT does not
represent all workers and that other unions should have a higher presence in the arbitration boards. It
also undermines the authority of UGT as a negotiating party. On the other hand, the publication of
36
Ibañez and Pérez (2005: 145).
Publication dates are taken from Prat and Molina (2011: 14). Dummy variable takes value one for the month before
publication, the month of publication and the two months following publication.
37
21
such agreements can be used to convince workers that they have been sold out by UGT and rally
anger for protest. As we can see in table 3, the ‘rival’ variable is significant and positive for any of the
models we specify. The political struggle we observe here is therefore not one tied to national politics
but rather driven by the rivalry of competing unions at moments when Biscayan arbitration boards
publish the agreements they have negotiated.
Two other interesting results are, on one hand, the change in the seasonality of strikes, especially
the high incidence of strikes in the month of May —perhaps a political shift— and the other result
worth noting is that the economic variables remain relevant in this new regime of labour relations.38
Grievances have lost their spontaneous impact on strike frequency but firm profits and
unemployment continue to show the correct coefficients and are significant. The impact of
unemployment should be significant especially in the 1930’s. Edwards (1978) has suggested that the
relation between unemployment and proneness to strike can be either inverse due to the tightness of
labour markets during low unemployment or direct due to discontent during moments of high
unemployment. We have therefore also used trough and peak variables of unemployment which
emphasize the opposing impact that unemployment may have on strike frequency.39 As expected
unemployment plays a significant role when emphasising low unemployment phases —peaks.
Tightening labour markets encourage workers to strike in the thirties. The threat of job loss and rising
un- and underemployment discourage them from stoppages.
A different insight might be gained from crime statistics. This will also be a way of testing the
largely dismissed modernization theory developed in the 1950’s. Kerr et al. (1955) proposed a theory
of transitional violence. Early stages of industrialization are accompanied by high levels of
immigration, which accelerate the destruction of previous social ties and relations, produce stress,
strain and alienation and lead on to protest and rebellion.40 In the regressions in table 4 we can see
how political violence has an inverse relation to strike frequency —political violence is high when
strikes are low. No apparent explanation other than frustration over passivity can be associated. On
the other hand, thefts show seasonality and also a high negative relation to lagged earnings and an
inverse relation to profits. The lag for earnings we have chosen is the five month lag we determined
for variations in money earnings to have a proportional impact on mortality (double for child
mortality). Desperate situations lead to desperate solutions. The hypothesis that social disruption in
the form of theft may be driven by short term variation in earnings as a consequence of price
38
Knowles (1952: 157-60) finds the same May seasonality for the US and GB for the 1919-1939 period, specifically in the
metal, engineering, shipbuilding sector but also in transport and construction.
39
Edwards (1978: 376). “The measure of business cycle troughs is calculated as follows. The [...] series of unemployment
was examined and a number of troughs (defined as [months] in which unemployment was greater than in proceeding
or subsequent [months]) identified. For the [months] between each trough, the value of the index is the unemployment
rate of the previous trough, unless the rate in the current [month] exceeds the rate in the recession [month], in which
case the current rate is substituted. The reason for this is that the variable is designed to measure the impact of
recessions: if current unemployment is greater than that in the past, it is the present situation which should be taken into
account. The 'peak' variable is calculated analogously.”
40
Cronin (1978: 200).
22
increases cannot be rejected with these results. But the seasonality combined with the significant
inverse relation to income level should incline us to think that theft may be driven by short-term
economic hardship rather that displacement.
Down to detail – a closer look at the corporatist bargaining model
The introduction of a large-scale corporatist scheme during the Primo de Rivera´s Dictatorship
(1923-1929) changed the previous regime in which Spanish industrial labour relations had developed.
Although there was a gradual drift from one regime into another, it took some time for union power
to concentrate within these new institutions and for enforceability to be imposed by the state and
tight union control. In the early twenties the last liberal governments had organized comités paritarios
as tools of intermediation and resolution of disputes between unions and employers. However, the
use of these corporatist arbitration boards was slow and limited. One of the main reasons was that,
like their Italian counterparts, Spanish industrialists were sceptical about the effectiveness of
negotiating in such bodies as long as trade-unionization remained non-compulsory.41
A broader and more efficient corporatist system of labour relations was enacted in 1926, when
Labour Minister Eduardo Aunós published a decree creating the National Corporatist Organization.
The decree extended the use of comités paritarios to all industrial sectors and organized them
hierarchically into local representative bodies, industrial tribunals, and corporations. The corporations
integrated all the comités paritarios of a branch or industry. The Ministry of Labour was placed at the
peak of the organizational pyramid.42 Aunós’ corporatist architecture, sanctioned well before the
Carta del Lavoro (1927) by Mussolini, retained the more liberal dispositions but restructured them
hierarchically and increased the degree of government intervention. In contrast to other corporatist
experiences of the time, Aunós’ model allowed for pluralism both for unions and employer
associations. Primo de Rivera had created a single official party, but made no attempt at verticalizing
unions. Thereby moderate unions gained access to the arbitration boards. The socialist union UGT,
which was the dominant trade union in Biscay at the time, collaborated decidedly in the dictatorship’s
corporatist scheme.43
The improvement of industrial worker’s living conditions over the first half of the twenties, the
unions’ defeat in 1922 and repression on behalf of the dictatorship was undermining union power in
Biscay. All trade unions lost affiliates over the first half of the decade.44 This steady decline in
41
The liberal political elites rejected compulsory trade-unionization because it clashed with basic liberal principles like the
freedom of association. See Rey Reguillo (1983) and (1992) and Adler (2002).
42
Perfecto (1997).
43
Weak government intervention combined with trade-union pluralism induce Linz (1981: 381) to define this model as a
“social corporatism with pluralist features”. McIvor (1982) attributes the regime’s alliance with the socialist trade union as
an artefact conceived by the ILO. Its cooperation in developing the corporatist bargaining model provided the regime with
international legitimacy and at the same time it allow the ILO to exert pressure to include UGT in the scheme. See Molina
and Prat (2011: 6-7) for a summary.
44
Castells (1993: 161-62).
23
membership and UGT’s pragmatic stance pushed the organization to actively participate in Aunós’
corporatist system. From 1927 to 1930 the socialist union put into practice an intensive campaign to
position themselves in the comités paritarios contending with the yellow Catholic and Basque
Nationalist unions. The system by which delegates were elected was ruled by a majority system,
favouring the bias towards the dominant position of UGT in the dictatorship’s arbitration boards,
especially in the metal sector.45 This in turn fed back into higher level of affiliation. UGT had increased
its ranks and files from just 4,000 members in 1924 to almost 18,000 in 1928.46
The repressive nature of the dictatorship, the conception of a corporatist regime of labour
relations, socialist collaboration and the improvement of worker’s standards of living jointly
contributed to reducing the strike frequency in the second half of the twenties. Strikes were
infrequent and those occurring were usually solved within the corporatist bodies or by agreements on
a firm level. UGT assumed the role of moderator, steering strike impulses into bargaining processes.
In 1928 it undid a strike which broke out in Altos Hornos de Vizcaya on account of a reduction in work
bonuses by convincing its members that the proposal being made by the company was beneficial for
its workers.47
The social peace was disrupted in 1930 by the final decay of the dictatorship. This was the year of
greatest labour conflictivity in Biscay in the period between 1926 and 1936, coming close to the
highest levels attained in 1919-1920. The literature stresses three factors that explain this increase in
conflictivity. Firstly, the dominance and overrepresentation of UGT in the corporatist bodies kindled
rivalry and hatred in the ranks of the contending unions, especially in the Communist unions. The
articulation of this confrontation was the rejection of arbitration within the comités paritarios which
were controlled by UGT and their resort to strikes. Secondly, the fall of Primo de Rivera opened a
period of uncertainty in which the arbitration seemed doomed to disappear. In this setting unions,
including UGT, called strikes with a clear political or revolutionary motivation.48 Biscay was scene of a
‘Huelga General de Protesta’ and a ‘Huelga General Revolucionaria’ in 1930. The number of strikes
seeking the re-admittance of workers increased, especially in the metal sector, were a sign of growing
defiance and attrition. Finally, the increase of unemployment as a consequence of the economic
recession reduced the usefulness of these arbitration bodies in the eyes of the working class,
distancing the workers from bargaining and pushing them towards conflictive means of arbitration.49
Nevertheless, there are two circumstances that also need to be taken into consideration to fully
understand the conflictivity in 1930. Unemployment could not have had an important impact on
labour relations in 1930, it did not appear until the end of the year. The number of employees in the
45
Ibáñez and Pérez (2005: 122).
Olábarri (1978); Otaegui (1986); Castells (1993); Ibáñez y Pérez (2005). The Catholic and Basque Nationalist unions
united into a common block, Bloque de Unión Sindical (1928).
47
Ibáñez y Pérez (2005: 120, fn. 367).
48
Olábarri (1978: 427-28).
49
Lazcano (1950).
46
24
larger iron and steel concerns actually increased slightly between December 1929 and December of
1930.50 The second important trend that needs to be taken into account is that the important
increase in the cost of living and the resulting fall of real incomes motivated a greater number of
strikes calling for increases in salaries (Figure 2).
The proclamation of the Second Republic in April of 1931 modified the labour relation’s
framework anew. The Left-Republican governments developed a reformist social agenda aimed at
improving the standards of living of the working class. The strategy designed by the socialist Labour
Minister, Francisco Largo Caballero, was to maintain the basic structure of collective bargaining
created during the preceding dictatorship, imposing a stronger bureaucratic control by the Labour
Ministry through local and provincial delegations. Two laws in November of 1931 enacted the general
legal framework for collective bargaining. It promoted the creation of arbitration boards, renamed
jurados mixtos, in all sectors of the economy; it encourages detailed and lasting agreements,
denominated bases de trabajo.51 The new legislation and the firm leading hand from the Labour
Ministry developed a dense network of collective bargaining in Spain, which showed some degree of
similarity to the institutions existing during the latter half of Primo de Rivera’s dictatorship.52 Biscay
made good use of the jurados mixtos, with a large record of balanced settlements, much more than
other regions of Spain.53
Apart from the modification and expansion of collective bargaining institutions, unemployment
gained an increasing role in labour relations. The impact of the economic depression on industrial
employment in Biscay was particularly harsh. The iron and steel industry and metal working industries
were especially hard hit, followed by mining and construction. The three sectors added up 90% of
unemployment in Biscay.54 The increase in unemployment reached its heights between 1931 and
1932, and failed to drop substantially in the following years. In January of 1936 approximately 25% of
the payroll was laid off, another 25% was working reduced shifts. Close to fifty per cent of the metal
industry workers were suffering unemployment or underemployment. Biscay had the highest
unemployment rates in Spain, around 20%.55 The unemployed metal workers in Biscay were 38% of
all unemployed metal workers in Spain. Unemployment numbers did not vary substantially in the
Biscayan metal sector over the period, whereas they began improving by 1934 in other regions of
Spain.56
50
Miralles (1988: 117, Table 16).
Soto Carmona (1989).
52
For the institutional continuity of the collective bargaining system in Spain during the 1920s-1930s, see Rojo Cagigal
(2009: 97-98).
53
Sanfeliciano (1990: 337).
54
Rivera (2002: 128).
55
There are no reliable statistics, and given the difficulty of estimating the impact of reduced hours this is the magnitude
usually assumed by the literature.
56
Miralles (1988: 117). Biscay accounted for 2.06% of the Spanish population and 4.2% of the unemployment. Rivera
(2002: 128), quoting Díaz Freire, La República y el porvenir.
51
25
The coming of the Second Republic increased union membership. Socialist UGT affiliates grew
from 18.000 members in 1928 to more than 31.000 in 1934. The second most important union, the
Basque nationalist Solidaridad de Obreros Vascos, increased its membership from 6.200 in 1929 to
18.000 in 1935.57 Union pluralism re-legalized Anarchist and Communist unions, prohibited during the
dictatorship, although their activity remained limited to very occasional conflicts.
All in all, the scenario of labour relations we have described for 1931 is that of a region which is
increasing its organized collective bargaining, which has growing union membership, mainly socialist
and Basque nationalist, and which is facing a severe situation of unemployment —much higher than
any other part of Spain. Unemployment must be combined with the increasing cost of living which
causes a fall in real income for workers. The important political involvement with which the masses
reacted to the new regime, the high expectations for improvement harboured in the new government
by way of its socialist participation, the growing strength of unions, and diminishing living conditions
combined to an highly explosive cocktail. But quite contrary to what might be expected it generated
one of the lowest strike frequencies in Spain.
An important explanation for this low conflictivity was the strategy of moderation employed by
the two most important unions. In the beginning, in 1931, workers blamed employers for lay-offs.
They accused them of sabotaging the consolidation of the new political regime. The most important
strike motives during 1931 were the re-admittance of laid-off workers and the reduction of the
working day.58 But socialist union leaders conducted an important campaign of contention, convincing
workers that the cause of unemployment was exogenous and that the measures taken by employers
were justified by the severity of demand contraction. The operation included visits by government
officials, newspaper campaigns and wide-scale factory briefings.59 At the same time the Basque
Nationalist union developed a similar policy of finding a non-conflictive way out of the crisis. Both
unions promoted alleviation of the crisis through the jurados mixtos; the creation of a short-term
unemployment insurance; organizing fundraising to pay for unemployment benefits; increasing the
capacity of soup kitchens; or lobbying the government for stimulus packages, especially socialist
Public Works Minister Indalecio Prieto, as a means of reactivating the construction and iron and steel
industries.60 As a result of their combined efforts, strike activity died down between 1931 and 1934,
to the contrary of what happened in other regions of Spain.
57
Membership numbers from Larrañaga (1977), vol. 2; Olábarri (1978); Sanfeliciano (1990).
Data in Lazcano (1950).
59
Castells (1993: 167-68).
60
Olábarri (1978: 428-29). According to Sanfeliciano (1990: 301), the underemployment or the unemployment itself was
the main obstacle to sustain the workers´ demands via strikes.
58
26
Conclusions
In contrast to Silvestre (2003) findings for Spain, we find strong evidence for opposing economic
forces driving labour conflictivity in Biscay. This may be due to higher levels of fine-tuning. Monthly
data, real worker’s earning instead of GDP per head, a regional perspective for analysing a strongly
segmented economy are but a few of the elements that have contributed to finding very different
results.
Conflictivity in Biscay, until now considered the highest wage earning industrial area in Spain, is
driven by short-term variations in workers incomes, profit cycles and unemployment. This is indicative
of a re-distributional conflict spread throughout the entire period. Even when worker’s claims for
higher real incomes are repressed by authoritarian rule or refrained by moderate pragmatic union
policy, they show in the statistical analysis, they resurge with accumulated intensity and remain latent
and endemic. Only on occasions of moderation do political forces move to the foreground. But even
then strikes still fed upon economic grievance. Even when UGT had its ranks and files under tight
control and was negotiating collective agreements, we have been able to trace back some of the
political character of strikes to the inter-union struggle for worker representation. The receptiveness
of workers to more radical agitators is a sign of discontent or expected economic improvement which
has not been forthcoming, very different from what happened in the 1920s when living conditions
were steadily improving.61 The short-term reaction of strikes and theft to changing economic
conditions support the hypothesis of precarious living conditions and a high sensitivity to relative
deprivation. This, in turn, is coherent with some of the theories on the coming of the Civil War in
Spain. “Revolution growth is the product not of extreme oppression but of relatively rapid
improvement in conditions in countries were notable internal problems are followed by a downturn
or significant new frustrations.”62
When compared to the industrial forerunners, Biscay lived through the different stages of trade
union consolidation at a much higher pace than Great Britain, France or the United States. Economic
crisis put an end to the American style business union model developed during the war. Depression
and Gerschenkron backwardness (dominance of big business, mixed banks and State building)
installed a corporatist state-led arbitration system which failed due to unprecedented and persistent
levels of unemployment, stagnant living conditions, and a growing desire for the Promised Land
revolution.
61
For contemporaneous strikes activity in France and Britain, Taft (1952: 259-60) insists that —beyond ideology and
agitation— dissatisfaction on behalf of the mass of workers was the means by which they were mobilized and questions
that sharp rises in strike numbers may have occurred if real grievances had not existed.
62
Payne (2009: 4)
27
References:
Adler, Franklin Hugo (2002): Italian Industrialists from Liberalism to Fascism: The Political
Development of the Industrial Bourgeoisie, 1906-34. Cambridge: Cambridge University Press.
Ashenfelter, Orley and George E. Johnson (1969): “Bargaining Theory, Trade Unions and Industrial
Strike Activity,” American Economic Review, 50 (1), pp. 35-49.
Ashenfelter, Orley and John H. Pencavel (1969): “American Trade Union Growth: 1900-1960,” The
Quarterly Journal of Economics, 83 (3), pp. 434-48.
Buenacasa, Manuel (1977): El movimiento obrero español 1886-1926. Historia y crítica. Madrid: Júcar.
Cabrera, Mercedes and Fernando del Rey (2002): El poder de los empresarios. Política y economía en
la España contemporánea (1875–2000). Madrid: Taurus.
Card, David (1990): “Strikes and Bargaining: A Survey of the Recent Empirical Literature,” American
Economic Review, 80 (2), pp. 410-15.
Castells, Luis (1993): Los trabajadores en el País Vasco (1876-1936). Madrid: Siglo XXI Editores.
Clegg, Hugh A. (1976): The System of Industrial Relations in Great Britain. Oxford: Blackwell.
Crampton, Peter C. and Joseph S. Tracy (2003): “Unions, Bargaining and Strikes,” in J.T. Addison and C.
Schnabel (eds.), International Handbook of Trade Unions. Cheltenham: Edward Elgar.
Cronin, James E. (1978): “Theories of Strikes: Why Can´t They Explain the British Experience?” Journal
of Social History, 12 (2), pp. 194-220.
Darlington, Ralph (2006): “Agitator ‘Theory’ of Strikes Re-evaluated,” Labor History, 47 (4), pp. 485509.
Davies, James C. (1969): “The J-Curve of Rising and Declining Satisfactions as a Cause of Some Great
Revolutions and a Contained Rebellion.” in H.D. Graham and T. R. Gurr (eds.), Violence in America:
Historical and Comparative Perspectives. New York: Praeger. Pp. 690-730.
Díaz Freire, J. J. (1990): Expectativas y frustraciones en la Segunda República (Vizcaya, 1931-1933).
Bilbao: Universidad del País Vasco.
Dubin, Robert (1965): ”Industrial Conflict: The Power of Prediction,” Industrial and Labor Relations
Review, 18 (3), pp. 352-63.
Edwards, P.K. (1978): “Time Series Regression Models of Strike Activity: A Reconsideration with
American Data,” British Journal of Industrial Relations, 16 (3), pp. 320-34.
Elster, Jon (1985): Making Sense of Marx. Cambridge: Cambridge University Press.
Fernández de Pinedo, E. (1992): "Beneficios, salarios y nivel de vida obrero en una gran empresa
siderúrgica vasca, Altos Hornos de Vizcaya (1902-1927). Una primera aproximación", Revista de
Historia Industrial, 1, pp. 127-157.
Franzosi, Roberto (1989): «One Hundred Years of Strike Statistics. Methodological and Theoretical
Issues in Quantitative Strike Research», Industrial and Labor Relations Review, 42, 3, pp. 34-63.
Franzosi, Roberto (1995): The Puzzle of Strikes. Class and State Strategies in Postwar Italy. New York:
Cambridge University Press.
Fusi, Juan Pablo (1973): “Algunas publicaciones recientes sobre la historia del movimiento obrero
español,” Revista de Occidente, 123, pp. 358-368.
Fusi, Juan Pablo (1975): Política obrera en el País Vasco 1880-1923. Madrid: Turner
Gandhi, Jennifer and Adam Przeworski (2006): ‘‘Cooperation, Cooptation, Rebellion under
Dictatorships.’’ Economics and Politics, 18(1), pp. 1–26.
28
González Portilla, Manuel (1984): “Tecnología y productividad en la siderurgia española el caso de
Altos Hornos de Vizcaya, 1880-1936”. in Garcia Delgado , J.L. , España 1898-1936: Estructuras y
cambio. Coloquio de la Universidad Complutense sobre la España contemporánea. Madrid:
Universidad Complutense, pp. 71-89.
Gramm, Cynthia L. (1986): “The Determinants of Strike Incidence and Severity: A Micro-Level Study,”
Industrial and Labor Relations Review, 39 (3), pp. 361-76.
Granja Sainz, José Luís de la (2004): “1934, Un año decisivo en el País Vasco. Nacionlismo, socialismo y
revolución”. Sancho el Sabio - Estudios Vascos, 21, pp. 11-25.
Griffin, John I. (1939): Strikes a Study in Quantative Economics. New York: Columbia University Press.
Hicks, John R. (1963): The Theory of Wages. London: Macmillan.
Ibáñez, Norberto and José Antonio Pérez (2005): “Orígenes y desarrollo del socialismo en el País
Vasco (1890-1936)”, Bilduma, 19. URL: http://www.errenteria.net/es/html/40/1101.shtml
Jurkat, Ernest H. and Dorothy B. Jurkat (1949): “Economic Function of Strikes,” Industrial and Labor
Relations Review, 2 (4), pp. 527-545.
Kaufman, Bruce E. (1981): “Bargaining Theory, Inflation and Cyclical Strike Activity in Manufacturing,”
Industrial and Labor Relations Review, 34 (3), pp. 333-55.
Kaufman, Bruce E. (1982): “The Determinants of Strikes in the United States, 1900-1977,” Industrial
and Labor Relations Review, 35 (4), 473-90.
Kennan, John (1986): “The Economics of Strikes,” in O. Ashenfelter and R. Layard (eds.), Handbook of
Labor Economics. New York: Elsevier. Vol. 2, pp. 1091-1137.
Kennan, John (2008): "Strikes." In S.N. Durlauf and L. E. Blume. (eds.) The New Palgrave Dictionary of
Economics. Second Edition., http://www.dictionaryofeconomics.com/article?id=pde2008_S000304
Kerr, C., F. Harbison, J. Dunlop and C. Meyers (1955): “The Labour Problem in Economic
Development,” International Labour Review, 72, pp. 13-44.
Kim, Wonik (2007): “Social Insurance Expansion and Political Regime Dynamics in Europe, 1880–
1945,” Social Science Quarterly, 88 (2), pp. 494-514.
Knowles K.G.J.C. (1952): Strikes: A Study in Industrial Conflict. Oxford: Blackwell.
Kornhauser, Arthur (1954): “Human Motivations Underlying Industrial Conflict.” in A. Kornhauser, R.
Dubin, and A. Ross (eds.), Industrial Conflict. New York: McGraw-Hill. Pp. 62-85.
Korpi, Walter (1974): “Conflict, Power and Relative Deprivation,” American Political Science Review,
Vol. 68, No. 4, pp. 1569-78.
Larrañaga, Policarpio de (1977): Contribución a la historia obrera de Euskalherria. Zarauz.
Lazcano, J. (1950): Los conflictos sociales en Vizcaya. Universidad de Deusto: Tesis de Licenciatura.
Linz, Juan J. (1981): “A Century of Politics and Interests in Spain,” in S. Berger (ed.), Organizing
Interests in Western Europe. Pluralism, Corporatism and the Tranformation of Politics, Cambridge:
Cambridge University Press. Pp. 365-415.
Mauro, Vicenzo (1982): “Strikes as a Result of Imperfect Information,” Industrial and Labour Relations
Review, 35 (4), pp. 522-38.
Mees, L. (1992): Nacionalismo vasco, movimiento obrero y cuestión social, 1903-1923. Bilbao: F. S.
Arana
Miralles Palencia, Ricardo (1988): El socialismo vasco durante la II República: organización, ideología,
política y elecciones, 1931-1936. Bilbao: Servicio Editorial de la Universidad del País Vasco.
29
Miralles Palencia, Ricardo (1992): “Historiografía del movimiento obrero en el País Vasco: 1880-1936,”
Historia Contemporánea, 7, pp. 237-55.
Moore, Geoffrey H. (1983): “Security Markets and Business Cycles,” in Business Cycles, Inflation and
Forecasting. http://www.nber.org/books/moor83-1.
Olábarri, I. (1978): Relaciones Laborales en Vizcaya (1890-1936). Bilbao: Zugaza.
Otaegui, Margarita (1986): “Censo de comités paritarios de Guipúzcoa y Vizcaya,” in J.L. García
Delgado (ed.), La crisis de la Restauración. España entre la Primera Guerra Mundial y la Segunda
República. II Coloquio de Segovia sobre Historia Contemporánea de España. Madrid: Siglo XXI. Pp.
291-314.
Payne, Stanley (2009): Franco and Hitler. Spain, Germany and World War II. Yale University Press.
Perez Castroviejo, P.M. (1992): Clase obrera y niveles de vida en las primeras fases de la
industrialización vizcaína. Madrid: Ministerio de Trabajo y Seguridad Social
Perfecto García, Miguel Ángel (1997): “Regeneracionismo y corporativismo en la Dictadura de Primo
de Rivera,” in J. Tusell, F. Montero García and J. M. Marín Arce (eds.), Las derechas en la España
contemporánea. Barcelona: Anthropos. Pp. 177-196.
Perrot, Michelle (1974): Les ouvriers en grève (France 1871 – 1890). Paris: Mouton.
Piven, Francis Fox and Richard A. Cloward (1993): Regulating the Poor: The Functions of Public Welfare
Prat, Marc and Oscar Molina (2011): “State Corporatism and Democratic Industrial Relations in Spain,
1926-1935. A Reappraisal,” paper presented at the X AHEE Congress, Carmona, September 8-10.
Redero San Roman, Manuel (1992): Estudios de historia de la UGT. Salamanca: Ediciones Universidad
de Salamanca.
Rees, Albert (1952): “Industrial Conflict and Business Fluctuations,” Journal of Political Economy, 60
(5), pp. 371-82.
Rees, Albert (1954): “Industrial Conflict and Business Fluctuations,” in A. Kornhauser, R. Dubin and
A.M. Ross (eds.), Industrial Conflict. New York: McGraw Hill. Pp. 213-20.
Rey Reguillo, Fernando del (1983): “Actitudes políticas y económicas de la patronal catalana (19171923),” Estudios de Historia Social, 24-25, pp. 23-148.
Rey Reguillo, Fernando del (1992): Propietarios y patronos. La política de las organizaciones
económicas en la España de la Restauración, 1914-1923. Madrid: Ministerio de Trabajo y
Seguridad Social.
Rivera, Antonio (2002): “Obreros y organizaciones obreras vascas en la II República,” Cuadernos de
Alzate, 27, pp. 121-135.
Rojo Cagigal, Juan Carlos (2009): “Las instituciones para la negociación colectiva en España en el siglo
XX,” Revista Empresa y Humanismo, 12 (1), pp. 89-120.
Rojo Cagigal, Juan Carlos and Stefan Houpt (2011): “Hunger in Hell’s Kitchen. Family Living Conditions
during Spanish Industrialization. The Bilbao Estuary, 1914-1935. Working Paper # 11-04.
Universidad Carlos III de Madrid.
Ross, Arthur M. (1954): “The Natural History of the Strike,” in A. Kornhauser, R. Dubin and A. Ross
(eds.), Industrial Conflict. New York: McGraw Hill. Pp. 23-36.
Sanfeliciano López, María Luz (1990): UGT de Vizcaya: 1931-1936. Bilbao: Unión General de
Trabajadores.
Scully, Gerald (1971): “Business Cycles and Industrial Strike Activity,” Journal of Business, 44 (4), pp.
359-74.
30
Shalev, Michael M. (1980): “Trade Unionism and Economic Analysis: The Case of Industrial Conflict,”
Journal of Labor Research, 1 (1), pp. 133-74.
Shorter, Edward and Charles Tilly (1974): Strikes in France, 1830-1968. New York: Cambridge
University Press.
Silvestre Rodríguez, Javier (2003): “Los determinantes de la protesta obrera en España, 1905-1935:
ciclo económico, marco político y organización sindical” Revista de Historia Industrial, 24, pp.
51-80.
Snyder, David (1975): “Institutional Setting and Industrial Conflict: Comparative Analyses of France,
Italy and the United States,” American Sociology Review, 40 (3), pp. 259-78.
Snyder, David (1977): “Early North American Strikes: A Reinterpretation,” Industrial and Labor
Relations Review, 30 (3), pp. 325-41.
Soto Carmona, Álvaro (1989): El trabajo industrial en la España Contemporánea, 1874-1936.
Barcelona: Anthropos.
Southall, H. and D. Gilbert (1996): “A Good Time to Wed?: Marriage and Economic Distress in England
and Wales, 1839-1914.” Economic History Review, 49 (1), pp. 35–57.
Stern, Robert M. (1978): “Methodological Issues in Quantative Strike Analysis,” Industrial Relations,
17 (1), pp. 32-42.
Swidinsky, Robert and John Vanderkamp (1982): “A Micro-econometric Analysis of Strike Activity in
Canada,” Journal of Labor Research, 3 (4), pp. 455-71.
Taft, Philip (1954): “Ideologies and Industrial Conflict,” in A. Kornhauser, R. Dubin and A. Ross (eds.),
Industrial Conflict. New York: McGraw Hill. Pp. 257-66.
Tilly, Charles (1989): “Introduction,” in L.H. Haimson and C. Tilly (eds.), Strikes, Wars, and Revolutions
in an International Perspective, Cambridge: Cambridge University Press. Chapter 18, pp. 433-48.
Tuñón de Lara, Manuel (1972): El movimiento obrero en la historia de España. Madrid: Taurus.
Van der Meer, Marc (1997): Trade Union Development in Spain : Past Legacies and Current Trends.
Working Paper MZES # I-18.
Walter, Kenneth F. (1970): Australian Industrial Relations Systems. Harvard: Harvard University Press.
Weintraub, Andrew (1966): “Prosperity Versus Strikes: An Empirical Approach,” Industrial and Labor
Relations Review, 19 (2), pp. 231-38.
31
Table 1. Test of the Ashenfelter and Johnson model for Biscay.
Dep. Var.: Nr. Strikes
C
D1
D2
D3
D4
D5
D6
D7
D8
D9
D10
D11
(1) OLS
Full Sample
(2) Coch-Orc
Full Sample
(3) Phase 1
Jan14-Aug23
(4) Phase 2
Nov23-Dec35
1.473
(0.947)
-0.157
(0.541)
0.196
(0.545)
0.231
(0.549)
0.082
(0.541)
0.255
(0.542)
0.616
(0.541)
1.574***
(0.534)
1.513***
(0.543)
0.733
(0.543)
0.172
(0.543)
0.065
(0.535)
0.268
(1.694)
-0.125
(0.329)
0.149
(0.422)
0.164
(0.472)
0.091
(0.493)
0.227
(0.507)
0.607
(0.510)
1.644***
(0.508)
1.525***
(0.488)
0.742
(0.461)
0.178
(0.413)
0.042
(0.323)
-2.726
(4.449)
-0.445
(0.695)
0.065
(0.891)
0.419
(0.996)
0.031
(1.029)
0.472
(1.058)
0.988
(1.063)
3.578***
(1.066)
3.137***
(1.014)
1.597
(0.967)
0.629
(0.873)
0.175
(0.691)
-0.407
(0.932)
0.066
(0.240)
0.168
(0.288)
-0.074
(0.307)
0.079
(0.314)
-0.034
(0.316)
0.263
(0.317)
0.118
(0.371)
0.270
(0.317)
0.151
(0.303)
-0.087
(0.281)
-0.009
(0.232)
-1.565***
(0.475)
-0.096**
(0.041)
-1.200***
(0.335)
-0.025
(0.078)
-1.260**
(0.535)
0.048
(0.179)
-0.956
(0.609)
-0.004
(0.054)
MA-12 Y
MA-2 DY
UN
@
(5) OLS
Full Sample
(6) OLS
Full Sample
8.127***
(1.562)
-0.151
(0.517)
0.252
(0.521)
0.294
(0.525)
0.089
(0.517)
0.272
(0.518)
0.617
(0.516)
1.540***
(0.521)
1.549***
(0.517)
0.804
(0.517)
0.149
(0.517)
0.096
(0.517)
-1.138***
(0.219)
-1.768***
(0.457)
-0.121***
(0.040)
4.711
(2.935)
-0.175
(0.523)
0.237
(0.527)
0.290
(0.530)
0.089
(0.522)
0.276
(0.524)
0.619
(0.523)
1.528***
(0.527)
1.561***
(0.522)
0.809
(0.522)
0.147
(0.522)
0.092
(0.523)
-0.995***
(0.282)
-1.829***
(0.462)
16.637
(9.988)
-0.476
(1.083)
-0.950
(1.067)
-0.923
(1.065)
-0.134
(1.069)
2.359**
(1.075)
-0.121
(1.067)
-0.005
(1.089)
0.258
(1.111)
0.051
(0.047)
-0.307
(1.072)
-0.675
(1.062)
-0.326
(0.715)
1.135
(2.977)
0.007***
0.001
-0.002
0.002
-0.072*
(0.042)
0.061
(0.062)
0.009***
(0.002)
0.002
(0.001)
-0.750*
(0.426)
-0.077
(0.143)
0.001
(0.016)
0.421
0.410
0.391
PEAK
TROUGH
IND(-6)
T
0.0107***
(0.002)
-0.008***
(0.001)
AR(1)
R-squared
0.351
0.011***
(0.004)
-0.008**
(0.003)
0.643***
(0.050)
0.613
0.015**
(0.006)
0.008
(0.021)
0.626***
(0.083)
0.644
0.002
(0.002)
0.001
(0.002)
0.453***
(0.074)
0.296
(6) OLS
Jan 30-Jul 36
@
Note: denotes significance at 11%, * at 10%, ** at 5% and *** at 1%
31
Table 2. Procyclicity of Strikes to Cost of Living Index Cycles. (% of average strike nr. over the cycle).
Trough
Phase 1
Phase 2
Expansion
Phase 3
Phase 4
Peak
Phase 5
Phase 6
Contraction
Phase 7
Phase 8
Trough
Phase 9
1st cycle
01-03/1914 04/14-05/16 06/16-07/18 08/18-09/20 10-12/1920 01/21-04/23 05/23-08/25 09/25-01/28 02-04/1928
73
31,0
146
269
73
130
47
3
0
2nd cycle
02-04/1928 05/28-08/29 09/29-01/31 02/31-06/32 07-09/1932 10/32-07/33 08/33-05/34 06/34-03/35 04-06/1935
0
0,0
115
330
244
73
73
24
0
Table 3. Ashenfelter and Johnson model with rivals.
Dep. Var.: Nr. Strikes
(1)
Nov 26 – Dec 35
(2)
Jan 30 – Dec 35
C
-2.231
1.918
0.145
0.798
-0.145
0.800
-0.004
0.803
1.112
0.826
2.815***
0.826
1.110
0.827
0.731
0.813
0.477
0.817
0.311
0.817
-0.148
0.800
-0.080
0.779
0.024
1.629
-0.213**
0.097
0.012***
0.003
1.737***
0.446
0.019***
0.006
11.820
(7.982)
-0.585
(1.036)
-1.055
(1.029)
-0.889
(1.034)
0.511
(1.125)
3.068***
(1.132)
0.602
(1.114)
0.284
(1.060)
0.138
(1.060)
0.026
(1.033)
-0.423
(1.023)
-0.522
(1.025)
-0.318
(2.750)
-0.144
(0.121)
0.002
(0.011)
@
1.007
(0.607)
-0.036
(0.022)
0.130
(2.794)
-0.397
(1.042)
-0.867
(1.035)
-0.650
(1.036)
0.927
(1.107)
3.510***
(1.110)
0.952
(1.105)
0.521
(1.062)
0.232
(1.072)
0.068
(1.045)
-0.486
(1.035)
-0.472
(1.032)
-0.360
(2.784)
-0.112
(0.121)
0.018***
(0.006)
1.394**
(0.561)
-0.388
(1.017)
-0.860
(1.015)
-0.643
(1.014)
0.932
(1.093)
3.516***
(1.095)
0.957
(1.092)
0.528
(1.043)
0.240
(1.052)
0.075
(1.027)
-0.480
(1.016)
-0.465
(1.018)
-0.356
(2.758)
-0.107**
(0.051)
0.018***
(0.004)
1.390**
(0.546)
0.326
0.439
0.415
0.415
D1
D2
D3
D4
D5
D6
D7
D8
D9
D10
D11
@MOVAV(DY,2)
UN
IND(-6)
Rivals
T
R-squared
(3)
(4)
Jan 30 – Dec 35 Jan 30 – Dec 35
Note: @ denotes significance at 11%, * at 10%, ** at 5% and *** at 1%
32
Table 3. Crime and dispair. Modernization theory under scrutiny.
Dep. Var.:
C
D1
D2
D3
D4
D5
D6
D7
D8
D9
D10
D11
@MOVAV(DY,2)
(1) Strikes
Jan 30 – Dec 35
(2) Strikes
Jan 30 – Dec 35
11.820
(7.982)
-0.585
(1.036)
-1.055
(1.029)
-0.889
(1.034)
0.511
(1.125)
3.068***
(1.132)
0.602
(1.114)
0.284
(1.060)
0.138
(1.060)
0.026
(1.033)
-0.423
(1.023)
-0.522
(1.025)
-0.318
(2.750)
25.601***
(8.523)
-0.438
(1.017)
-1.185
(1.007)
-1.087
(1.005)
0.054
(1.016)
2.570**
(1.018)
0.169
(1.011)
0.492
(1.049)
0.407
(1.051)
0.146
(1.016)
-0.082
(1.012)
-0.710
(0.999)
0.833
(2.652)
Y(-5)
UN
IND(-6)
Rivals
-0.144
(0.121)
0.002
(0.011)
@
1.007
(0.607)
Political violence
T
R-squared
-0.121
(0.116)
0.012
(-1.222)
-0.036
(0.022)
-0.115**
(0.051)
-0.079***
(0.024)
0.439
0.461
(3) Thefts
Jan 30 – Dec 35
115.877***
(19.493)
-6.551
(5.164)
-9.240*
(5.170)
-6.461
(5.167)
-6.906
(5.164)
1.946
(5.163)
-5.559
(5.166)
-13.458**
(5.163)
-10.075*
(5.165)
-5.923
(5.168)
-20.595***
(5.177)
-9.707*
(5.175)
-11.220***
(2.459)
-0.465
(0.466)
0.481
Note: @ denotes significance at 11%, * at 10%, ** at 5% and *** at 1%
33
Figure 1. Spanish and Biscayan strikes, 1905-1936 – a first look at similarities
1400
140
35
120
30
1000
100
25
800
80
20
600
60
15
400
40
10
200
20
5
0
0
Strikes declared Spain
1200
Average strike duration Spain
Strikes Biscay (rhs)
Average strike duration Biscay
0
1.050.000
31.500
Members UGT Spain (left hand)
Strikes
27.000
900.000
Profits AHV
20
1936
1934
1932
1930
1928
1926
1924
1922
1920
1918
150,0
1916
1914
1935
1933
1931
1929
1927
1925
1923
1921
1919
1917
1915
1913
1911
1909
1907
1905
25
Members UGT Biscay (right hand)
100,0
Stock exchange (rhs)
15
50,0
10
750.000
22.500
600.000
18.000
450.000
13.500
300.000
9.000
150.000
4.500
0,0
5
-50,0
0
-100,0
0
1935
1933
1931
1929
35.000
1927
Dec
1925
Nov
1923
Oct
1921
Sep
1919
Aug
1917
Jul
1915
Jun
1913
May
1911
Apr
1909
Mar
1907
Feb
1905
Jan
0
13
UGT Membership Biscay
30.000
12
Cost of Living
11
25.000
10
20.000
9
15.000
8
Source: Silvestre (2003: 75), Sanfeliciano(1990), Larrañaga (1977) and Olábarri (1978:
446 and 498)
7
10.000
6
5.000
5
0
4
1935
1934
1933
1932
1931
1930
1929
1928
1927
1926
1925
1924
1923
1922
1921
1920
1919
1918
1917
1916
1915
1914
34
Figure 2. Working class family, Bilbao 1914-36. Upper and lower bound cost of living versus daylabour earnings at Altos Hornos de Vizcaya and number of strikes throughout the period.
16
140
CLI max
Monthly blast furn earnings
Annual average earnings
124
117
14
CLI min
Monthly workers' average
Strikes (RH axis)
120
100
12
80
10
60
53
8
27
26
6
26
17
8
40
39
36
10
20
17
15
12
7
6
14
12
5
0
ene-14
jul-14
ene-15
jul-15
ene-16
jul-16
ene-17
jul-17
ene-18
jul-18
ene-19
jul-19
ene-20
jul-20
ene-21
jul-21
ene-22
jul-22
ene-23
jul-23
ene-24
jul-24
ene-25
jul-25
ene-26
jul-26
ene-27
jul-27
ene-28
jul-28
ene-29
jul-29
ene-30
jul-30
ene-31
jul-31
ene-32
jul-32
ene-33
jul-33
ene-34
jul-34
ene-35
jul-35
4
12
7
3
Figure 3. Welfare ratios for AHV steel worker’s families, 1914-1936.
1,5
1,4
1,3
1,2
1,1
1
0,9
0,8
0,7
oct-35
abr-34
ene-35
jul-33
oct-32
ene-32
abr-31
jul-30
oct-29
ene-29
jul-27
abr-28
oct-26
abr-25
ene-26
jul-24
oct-23
ene-23
abr-22
jul-21
oct-20
ene-20
jul-18
abr-19
oct-17
abr-16
ene-17
jul-15
oct-14
ene-14
0,6
35
Figure 4. Plot of residuals. OLS Ashenfelter and Johnson
12
10
8
6
4
2
0
1935
1934
1933
1932
1931
1930
1929
1928
1927
1926
1925
1924
1923
1922
1921
1920
1919
1918
1917
1916
1915
1914
-2
-4
36