Productivity growth and structural transformation: Japan and the US

This paper examines the role of structural transformation in contrasting productivity growth performances in Japan and the US in the 1990s and 2000s, by decomposing aggregate productivity growth into intra-industry productivity growth and inter-industry resource reallocation. While a slowdown in intra-industry productivity growth is the main cause of the aggregate productivity slowdown, we find that the contribution of inter-industry reallocation of employment becomes almost zero in the 1990s and even significantly negative in the 2000s.

Discussing growth-reducing reallocation of labor under structural change, we show similar changes in the US and argue that the negative effect of the reallocation on productivity growth is unavoidable, but not very significant. We will argue that structural change or the lack of it may not be responsible for the lost decades in Japan, and that these contrasting outcomes come from common factors.

Turning to intra-industry productivity growth, we examine relative contributions of individual industries to aggregate productivity growth. We found some divergent contributions among industries over business cycles in Japan, i.e. cycle-robust industries such as manufacturing and transport and cycle-prone industries such as trade, finance and construction. While the former has remained to support aggregate productivity growth, the latter failed in the lost decades of Japan. In the recent US, both types of industries contributed to its aggregate productivity growth, generating the contrasting performance against Japan.

Thus, we conclude that long-term productivity growth inevitably accompanies structural change whose resource reallocation may or may not enhance growth, that business cycles affect sectors differently where structural change affects the pattern of cyclical impacts on aggregate productivity growth, and that, while we cannot deny the effect of inadequate structural change in Japan, its slower capital deepening and productivity growth suggest more importance of business cycle impacts as reasons for the lost decades.