A Good Time to Stay Out? Strikes and the Business Cycle
| Published date | 01 June 2011 |
| DOI | http://doi.org/10.1111/j.1467-8543.2010.00800.x |
| Date | 01 June 2011 |
A Good Time to Stay Out? Strikes and
the Business Cyclebjir_80070..92
Paul J. Devereux and Robert A. Hart
Abstract
In this article, we compile a unique historical dataset that records strike activity
in the British engineering industry from 1920 to 1970. These data have the
advantage of containing a fairly homogeneous set of companies and blue-collar
workers. They cover a long period with varying labour market conditions,
include information that enables the addition of union and company fixed
effects, and provide geographical detail that allows a district-level analysis that
controls for year and seasonal effects. We study the cyclicality of strike dura-
tions, strike incidence and strike outcomes, and distinguish between pay and
non-pay strikes. We find evidence that strikes over pay have countercyclical
durations. However, in the post-war period, the magnitude of this effect for
pay-related strikes is much reduced when union and firm fixed effects are
included. These findings suggest that it is important when studying strike dura-
tions to take account of differences in the types of companies and unions that are
involved in strikes at different points of the business cycle. We also find that
strike outcomes tend to be more favourable to unions when the national unem-
ployment rate is lower.
1. Introduction
Economists have long been interested in how strike durations and strike
incidence change with business conditions. Evidence has been decidedly
mixed. Work with US data by Kennan (1985) and with Canadian data by
Harrison and Stewart (1989) suggests that strike durations are strongly coun-
tercyclical; these contrast with the US findings of McConnell (1990) of little
cyclical variation in strike duration. North American studies tend to support
procyclical strike incidence (Harrison and Stewart 1994; Kennan 1986;
Vroman 1989). However, work with British data by Ingram et al. (1993)
produces strong support for countercyclical strike incidence. In this article,
Paul J. Devereux is at the School of Economics and Geary Institute UCD, CEPR and IZA.
Robert A. Hart is at the Division of Economics, University of Stirling and IZA.
British Journal of Industrial Relations doi: 10.1111/j.1467-8543.2010.00800.x
49:S1 June 2011 0007–1080 pp. s70–s92
© Blackwell Publishing Ltd/London School of Economics 2010. Published by Blackwell Publishing Ltd,
9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.
we use a new dataset that we have assembled from the strikes records of the
member companies of the Engineering Employers Federation (EEF)1in
Great Britain. The data cover 10,870 company-level strike incidents within
the engineering, metal goods and automotive industries over the period
1920–1970. The companies are unionized, and virtually all strikes involve
employer–union bargaining and/or consultation. These unique data allow us
to study strikes in a relatively homogeneous industry and so differ from the
cross-industry studies that dominate earlier research. They also provide evi-
dence on the cyclicality of strike behaviour for a European country; to date,
the related literature has been dominated by North American findings.
We have five main objectives. First, and most broadly, we wish to explore
the main characteristics of the EEF data. This should provide useful back-
ground to researchers in this area, since they can access the full electronic
dataset at the UK Data Archive.2Second, we investigate whether the data
cast new light on one of the existing labour market research themes in the
study of strike behaviour — the connection between strike activity and the
business cycle. Predominantly, we proxy the cycle by rates of unemployment
because, for an important section of our analysis, we are able to disaggregate
unemployment by monthly rates and at travel-to-work engineering district
levels. This allows us to match unemployment rates to both the exact location
and the timing recorded in the EEF strikes records. Third, in line with
Harrison and Stewart (1989, 1994), we separately investigate the cyclical
characteristics of strikes undertaken for pay and for non-pay reasons. We
argue, and offer supporting empirical evidence, that this distinction matters
in the analysis of strike cyclicality. Fourth, for the 1960 to 1970 period, our
data allow us to control for company, union, time (month and year) and local
labour market fixed effects. This increases the likelihood that we are captur-
ing true cyclical effects rather than selection effects. Fifth, we take advantage
of the fact that the EEF data allow us to study strike outcomes. In particular,
we examine for the first time in a study of this type whether unions are more
likely to obtain successful outcomes towards the peak of the business cycle.
2. Strikes, cycles and types of company disputes
Guidance from Theory
The theoretical literature does not provide strong guidance about cyclical
effects. An important strand of theory involves bargaining models of strike
behaviour that emphasize private information and shared rents (e.g.
Cramton and Tracy 2003; Kennan and Wilson 1993). In a management–
union context, for example, strikes can be seen as a means of one (or both)
party’s willingness to incur costs in order to elicit more information from the
other side of the dispute. The predominant modelling assumption is that
information is asymmetric. Thus, union members are unable to observe fully
aspects of company variables that are important in determining employers’
ability to pay higher wages. Company profits have featured most commonly
Strikes and the Business Cycle s71
© Blackwell Publishing Ltd/London School of Economics 2010.
To continue reading
Request your trial