Wage Rates and Work Intensity in the South Wales Coalfield, 1874-1914

Abstract

In a recent article in this Journal, Hirsch and Hausman (1983) attempt to identify and measure the influences determining labour productivity in the British and especially the South Wales coalfield during the 40 years before the First World War. In particular, attention is focused on the idea, commonly expressed by contemporary coal owners, that the upturn in wage rates (both nominal and real) during the 1880s exerted a major depressing influence on labour The findings of Hirsch and Hausman are largely consistent with the views of traditional economic historians, notably Taylor (1961) and Walters (1975). While less emphasis is placed by Hirsch and Hausman upon the effects of diminishing returns, they, find strong support for the contention that rising wage rates played a major role in explaining decreased productivity. Essentially, Hirsch and Hausman claim to offer a firm empirical foundation for the persuasive ... [but] ... far from conclusive arguments of Taylor and Walters by using regression analysis to isolate the separate influences determining labour If the estimates presented by Hirsch and Hausman are considered legitimate, their results are striking. For South Wales, elasticities of labour productivity with respect to both nominal and real wages of approximately -0 27 are reported. Juxtaposition of this parameter with the nominal wage index employed by Walters, which rises around 55 per cent between 1888 and 1914, suggests that over 60 per cent of the 24 per cent fall in labour productivity during the same period was accounted for by rising wage rates. A consistent, though moderately less strong, interpretation is produced if the real-wage index, which increases some 40 per cent in the same period, is preferred. The contention of this paper is that the inverse association between wages and productivity postulated by Hirsch and Hausman for the South Wales coalfield between 1874 and 1914 is without statistical foundation. The basis for this assertion is that Hirsch and Hausman replicate unacceptable elements in the framework, albeit without the qualifications, offered by Walters. Indeed, this is done by intent: We employ a somewhat ad hoc model reflecting the discussion by economic historians. It will be argued that this desire to test the literature leads Hirsch and Hausman to estimate a model for which the results are difficult to interpret, and to reach conclusions that are implausible. Specification problems are compounded by the choice of data, particularly by the adoption of the wage rate index employed by Walters, but also by the use of proximate indicators to measure the balance between technical progress and diminishing returns.