Working Paper, University of Massachusetts, Department of Economics 2014-03
This paper uses panel cointegration and error correction models to unveil the direction of long-run causality between the real product wage and labor productivity at the industry level. I use two datasets of manufacturing industries: the EU-Klems dataset covering 11 industries in 19 developed economies, and the Unido Industrial Statistics Database covering 22 industries in 30 developed and developing economies. In both datasets, I find evidence of cointegration between the two variables, as well as evidence of two-way, long-run Granger causality. These findings are consistent with theories of directed technical change, which claim that a rise in labor costs sparks the adoption of labor-saving innovations. They are also consistent with distributive theories whereby real wages keep apace of labor productivity growth, giving rise to long-run stability in functional distribution.