Rates of return and alternative measures of capital input: 14 countries and 10 branches, 1971-2005

Abstract

We employ the EU KLEMS database to estimate the real rate of return to capital in 14 countries (11 in the EU, three outside the EU) in 10 branches of the market economy plus the market economy as a whole. Our measure of capital is an aggregate over seven types of asset: three ICT assets (computers, communications equipment, and software) and four non-ICT assets (machinery and equipment, nonresidential structures, transport equipment, and other). The real rate of return in the market economy does not vary very much across countries, with the exception of Spain where it is exceptionally high and in Italy where it is exceptionally low. The real rate appears to be trendless in most countries. Within each country however, the rate varies widely across the 10 branches, often being implausibly high or low. We also estimate the growth of capital services by two different methods: ex-post and exante, and the contribution of capital to output growth by three methods: ex-post, ex-ante and hybrid. Our implementation of the ex-ante method uses an estimate of the required rate of return for each country instead of the actual, average rate of return to calculate user costs and also employs the expected growth of asset prices rather than the actual growth. These estimates are derived from exactly the same data as for the ex-post method, ie without any extraneous data being employed. For estimating the contribution of capital to output growth, the ex-ante method uses ex-ante profit as the weight, while both the ex-post and the hybrid method use ex-post profit. We find that the three methods produce very similar results at the market economy level. But differences are much larger at the branch level, particularly between the ex-post and ex-ante methods.