Underground, Overground : Labour Productivity in British Coalmining, Co. Durham, 1890-93

It is widely believed that one of the main causes of productivity decline in British coal-mining in the late 19th century was a backward-bending effort supply curve: as wage rates rose, miners reduced their effort either by adjustments of output per shift, or by changes in their attendance behaviour. The paper shows that this view is mistaken. The negative correlation between measured wage rates and productivity found in aggregate time series data does not accurately reflect the response of worker effort to changes in wage rates, because it conflates that response with the fixing of wage rates to compensate for geologically induced productivity differentials. Individual panel data from a single coal-mine in County Durham are used to analyse the identification problem. When it is properly taken into account, estimated responses of worker effort and attendance are changed dramatically. While it is true that the data used here are limited both geographically and in timespan, they are ideally suited to illustrate the main point. The data are used to replicate the techniques and results of previous studies as closely as possible, and then to show that the same data yield strikingly different estimates when the identification problem is correctly addressed. The results show that the short run response of worker effort is positive, rather than negative as previously believed. The long run response is also positive but small. Responses of attendance are less well-determined, but there is no support for the idea that absenteeism increased when wage rates rose