Economies of scale in British coalmining between the wars

Abstract

In the course of a wide-ranging attempt to question the standard historiography of interwar British coalmining, Fine considers the failure of the industry to reap substantial scale economies.' To demonstrate the existence of these economies, he estimates a production function, and reports substantially increasing returns to scale.2 Indeed the estimate for returns to scale, as Fine admits, appears surprisingly high. It seems that a doubling of scale would have led to an increase of more than four times in coal output between the wars. Rather than being sceptical, Fine believes the result to be qualitatively correct. Since the finding of substantially positive returns undermines earlier views, Fine's claim warrants further investigation. This comment has a threefold purpose. First, it draws attention to other production function studies which suggest that scale economies in interwar coalmining were roughly constant. Secondly, it questions Fine's own estimate, primarily by casting doubt on the reliability of his capital stock series. Thirdly, it offers revised production function estimates which support the notion of constant returns. I In addition to the work of Rhodes mentioned by Fine, at least three other attempts-by Lomax, Leser, and myself-have been made to estimate coalmining production functions.3 This research denies that significant scale economies were available to the industry. The pioneering work of Rhodes receives short shrift from Fine since constant returns are initially assumed. Yet Rhodes deploys seasonal data to test the constant returns assumption, and reports scale economies estimates for a variety of mining regions ranging from o.9i9 to I.i39 (unity indicates constant returns). Rhodes stressed the nearness to unity of all the figures, arguing that the evidence favoured constancy of returns.4 Moreover, Rhodes simply estimated a coalface production function, and his findings therefore exclude an important source of diminishing returns: longer below-ground haulage requirements, which affect coalmining more generally. Lomax estimated coalmining production functions with cross-sectional as 1 Fine, 'Economies of scale'. 2 Fine also uses multiple regression to examine the connection between productivity, mechanization, and mine size. Since coal output appears both as the dependent and (twice) as an independent variable in this equation, little credence can be attached to the result. 3Rhodes, 'Output, labour, and machines'; Lomax, 'Coal production functions'; Leser, 'Production