Induced innovation, the distributive cycle, and the changing pattern of labour productivity cyclicality: a SVAR analysis for the US economy

The empirical literature on induced technical change has explored the long-run relationship between real wages and labour productivity but still lacks an explicit treatment of the implications of the wage-productivity nexus for the business cycle. The present paper aims to fill this gap. By employing a four-dimensional structural vector autoregressive (SVAR) model for the US economy (1948-2019), we test an extended version of the Goodwin model that includes aggregate demand and decomposes the labour share into real wages and labour productivity. This paper adds to the existing literature in some respects. First, it contributes to the induced innovation literature, by showing that wage shocks have positive and persistent effects on labour productivity at business cycle frequencies. Second, it adds to the debate and empirical evidence on the distributive cycle. Impulse response functions show that, even when decomposing the labour share, empirical evidence supports the Goodwin pattern, although the profit-led regime turns out to be driven more by technology than distributive shocks. Finally, we address two relevant cyclical stylized facts of the US economy: since the mid-1980s, the procyclical pattern of labour productivity has vanished, and real wages have no longer been correlated with employment over the business cycle. We explore the hypothesis that the two changes are linked. In light of the theory of induced innovation, we argue that the decline in the cyclical correlation between output and labour productivity can be explained by a lessened incentive to invest in labour-saving innovations due to missing wage growth in the upturn of the business cycle. Impulse response functions qualitatively support this intuition.