Variations in Total Factor Productivity Analysed with Cointegration: Swedish Manufacturing Industry, 1950–94

This chapter aims to identify determinants for variations in total factor productivity (TFP). Not only have changes in productivity of Swedish manufacturing since the 1950s been sudden, but productivity on a disaggregate level has developed differently in various industries. Hence the causes of productivity variations may differ among industries with different methods of production. A comprehensive statistical analysis of three comparative data sets for labour-intensive, capital-intensive and knowledge-intensive industries in Sweden between 1950 and 1994 will explore this matter. The ‘real-business-cycle theory’ captures the leading explanation for short-term variations in TFP in which technological shocks constitute the key element (Black 1982; Kydland and Prescott 1982; Long and Plosser 1983; Prescott 1986). The most important macroeconomic evidence for such technological shocks is short-term variation in the Solow residual. In this study, TFP is calculated according to a Solow Cobb—Douglas production function in which the coefficients for labour and capital, both in logarithmic scale, are assumed to correspond to their marginal effect on TFP. This formula is used since, if certain conditions are met, it implies that TFP expresses shifts in technology (Solow 1957). These two conditions are perfect competition and constant return to scale. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.