Decomposing sources of growth in a growth accounting framework relies heavily on price measures as proxies for input productivity. We show how different price measures affect the contribution of investment to growth in labor productivity and determine the share of this contribution that results from improved quality of ICT capital. The price measures we consider are the rate of return to capital and the rate of decline in ICT asset prices. If they represent productivity in an undistorted way, their effects correspond to the effects of marginal productivity of capital and embodied technological progress. The analysis is based on data from the EU KLEMS database for seven countries in the period of 1990−2007. Our aim is to disentangle the effect of the overall level of the rate of return and the overall decline in ICT prices from the effect of differences in these measures across countries, sectors and time. While it is plausible to consider that the overall level reflects economic and technological conditions, we presume that the differences are at least more prone to measurement error.

Alternatively to the measures from the EU KLEMS database, we introduce a constant real rate of return to capital of 4% and a rate of decline in ICT asset prices equal to the US average between 1990 and 2007. The sensitivity analysis reveals that most growth accounting results with data from the EU KLEMS database are comparatively robust. Solely for the period from 1995 − 2000 we obtain a substantial decline in capital contributions in both the UK and the US using the constant rate of return to capital. This in turn points to a greater contribution of multi-factor productivity. The decrease in the contribution of non-ICT capital is more pronounced than the decrease in the contribution of ICT capital. As the constant rate of return, alternative ICT deflators somewhat downplay the role investment played relative to growth in multi-factor productivity in the UK and the US during 1995 − 2000.

The second main element of our paper is the decomposition of ICT capital contributions into contributions of quantity, quality and the change in the composition of ICT capital. The latter describes variations in the shares of hardware, software and telecommunications equipment in total ICT capital. We show that more than half of the ICT contribution to labor productivity growth results from changes in quality and the composition of capital. Between 2000 and 2007, the contribution of ICT quantity is close to zero in some countries, while replacement of depreciated capital goods leads to a positive contribution of quality.

Keywords

ICT Capital, Asset Prices, Rates of Return, Growth Accounting