Wage strikes in 1880s America: A test of the war of attrition model

By relating strike outcomes and durations to the value of the disputed wage change and to the cost to each side of continuing the strike, this paper tests the hypothesis that the war of attrition with asymmetric information model of strikes accurately describes the characteristics of strikes over wages in the United States in the early to middle part of the 1880s. That hypothesis is not rejected by linear, probit, or nonparametric kernel estimation. Specifically, variables that decrease a side's cost of striking or increase its opponent's cost are shown to increase its maximum holdout time, and vice versa, and strike duration increases with the value of the prize in dispute and with uncertainty about the outcome. Alternative game theoretic models of strikes-signaling and screening models, and models with ongoing negotiations-do not fit the data as well. We also explore why the strikes took the form of wars of attrition, and why later strikes did not. Our results have implications for modern union behavior in the face of globalization.