Skip to content

The Economics of Strikes

Literature › Works

John Kennan (1986). The Economics of Strikes. Handbook of Labor Economics, Volume II (eds. O. Ashenfelter and R. Layard), Chapter 19. Elsevier Science Publishers BV.
Type book chapter
Scope Comparative and methods
Method synthesis review
Period studied 1870–1980
Regions United States, Britain, Canada, New Zealand, Australia, France, Germany, comparative multi-country
Themes unions strikes, comparative theory, methods data sources, labour markets wages
Citation key kennan1986

Argument

Kennan surveys economic theories and evidence on strikes, starting from the 'Hicks Paradox': if parties are rational they should avoid the costs of a strike by agreeing in advance. He reviews and criticises the Nash and Cross bargaining models and the Ashenfelter-Johnson model, then presents the joint cost theory and private information theories (a cooperative incentive-efficient mechanism model) as ways to escape the paradox. He then assesses empirical work on cyclical, seasonal and duration patterns of strikes, finding strike frequency procyclical and contract-strike duration countercyclical, and argues the empirical literature is largely measurement without theory. He closes by calling for developed private-information theory, micro-data on contracts, and experiments.

Evidence and method

Secondary literature review of North American, British and other strike statistics (BLS, Canadian contract data, Farber, Tracy and Gramm matched contract-strike datasets), Burns-Mitchell cycle analyses (Jurkat and Jurkat, Rees, Weintraub), regression studies (Ashenfelter-Johnson, Pencavel, Paldam-Pedersen), hazard-function studies of duration, stock-market event studies (Neumann), and bargaining experiments (Malouf-Roth, Coursey). Includes Kennan's own hazard estimates of US contract strikes 1968-76.

Findings

  • The central theoretical puzzle is that predictable strikes should be bargained away in advance, since strikes shrink the pie. (p. 1)

    if one has a theory which predicts when a strike will occur and what the outcome will be, the parties can agree to this outcome in advance, and so avoid the costs of a strike."}

  • The Cross bargaining model is judged fatally flawed because bargainers behave irrationally and the model lacks empirical content. (p. 10)

    The Cross model has two fatal flaws. First, the bargainers do not behave rationally."}

  • Under private information, strikes can be incentive efficient only if the union's expected payoff from the good state exceeds the bad-state pie; otherwise efficiency requires no strikes. (p. 19)

    If so, efficiency requires 3' = 1, which means that there will be no strikes. If not, any point in the smaller triangle in Figure 19.4 represents an efficient mechanism"}

  • If bargaining occurs before the firm learns private information, strikes will not occur. (p. 20)

    if only one side receives private information, and if bargaining takes place before this information is available, strikes will not occur."}

  • Unions are much less likely to win long strikes than short ones across several countries and periods. (p. 23)

    These studies uniformly agreed that unions are substantially less likely to win long strikes than short strikes."}

  • Averaged over 14 US business cycles 1915-80 the Burns-Mitchell analysis shows strikes tracing the cycle perfectly, though the fit is weaker in later subperiods. (p. 29)

    The last row of the table shows, however, that when all 14 cycles between 1915 and 1980 are averaged, the fit is perfect."}

  • Cross-country regression surveys find strikes negatively related to unemployment and recent real wage increases, and positively related to inflation. (p. 32)

    a uniform pattern emerged, indicating that strikes are negatively related to the unemployment rate, and to recent real wage increases, and positively related to the inflation rate."}

  • Contract-strike incidence in North American data is roughly 11-23 percent of negotiations. (p. 34)

    The incidence of strikes here was 23 percent (34 strikes in 149 contracts)."}

  • Incidence and duration of contract strikes move in opposite directions over the cycle: strike duration is shorter when industrial production is high. (p. 38)

    the mean duration of strikes when industrial production was high was 37.2 days, as compared with a mean duration of 52.3 days when industrial production was low."}

  • Strikes are strongly seasonal, and seasonality is not explained by the bargaining calendar. (p. 41)

    the average incidence was 21.3 percent (131 strikes in 614 negotiations) in March, as compared with only 5.7 percent (15 strikes in 263 negotiations) in December."}

  • Stock-market evidence indicates strikes are costly but imperfectly predictable. (p. 41)

    Neumann (1980) estimated the cost of an average strike as about 0.9 percent of the value of the firm."}