Capital Stocks and Productivity in Industrial Nations
O'Mahony, Mary O., National Institute Economic Review
Differences in physical capital intensity are frequently thought to be important determinants of differences in both productivity levels and growth rates across nations. But quantifying the impact of physical capital on productivity requires estimates of capital stocks which are measured on a consistent basis across countries. Official estimates of the stocks of physical capital are available for most industrialised countries but assumptions underlying their measurement by the statistical offices are in many ways not internationally comparable. This article uses a standardised method to measure capital stocks and employs these estimates to quantify the impact of differences in capital stocks on both levels and growth rates of labour productivity. The analysis is carried out for five industrial nations: the United States, the United Kingdom, Germany, France and Japan.
The article first presents relative levels of output per worker hour, both for the whole economy and manufacturing, for the five countries. It draws on estimates of labour productivity levels in manufacturing which were mostly published in the National Institute Economic Review, i.e. Van Ark (1990), O'Mahony (1992a), Van Ark (1992), and on Pilat and Van Ark (1991) and updates these estimates to 1989. Estimates for the total economy are derived using data primarily from the OECD.
Section 3 discusses the issues involved in consistently measuring capital stocks across countries. Estimates of gross physical capital stocks are presented together with an indication of the sensitivity of the results to changes in the underlying measurement method. Section 4 looks at the contribution of physical capital to countries' relative productivity levels in 1989, the latest year for which it was possible to obtain data for all five countries. Section 5 then examines time series on relative productivity levels and the contribution of physical capital to productivity growth.
2. Relative labour productivity levels
Table 1 shows labour productivity levels relative to the UK for 1989 in both the total economy and manufacturing. The economy-wide estimates were derived by deflating gross domestic product by GDP purchasing power parities for 1985. The manufacturing estimates used price ratios for producer goods to convert output to a common currency; details of the methodology used is outlined in Van Ark (1990) and O'Mahony (1992a). The manufacturing estimates for France, Germany and the United States were binary comparisons with the UK for 1984, 1987 and 1987, respectively, whereas the Japanese estimate was derived by linking two binary comparisons for Japan and the US and the US and the UK, both for 1987. The estimates were updated to 1989 using growth rates of real output and labour input (annual hours worked) for each country. The resulting estimates, together with a description of data sources, are shown in Table 1.
Table 1 shows that, in general, the UK labour productivity performance in 1989 was poor relative to other major industrial nations. For the aggregate economy only Japan had lower levels of GDP per worker hour than the UK in 1989. In manufacturing all four countries had higher productivity levels than the UK and the productivity gaps were larger than those for the aggregate economy. In Japan, the large difference between manufacturing and aggregate economy relative productivity levels reflects low productivity levels in that country in agriculture, mining and some service sectors (see Pilat (1992)).
Table 1. Relative output per worker hour, 1989 (UK=100) Total Economy Manufacturing United States 134.1 158.1 Germany 113.1 116.3 France 115.2 124.2 Japan 82.0 124.3 Sources: Total Economy: GDP in 1985 prices and employment primarily from OECD 'National Accounts', converted to US$ using 1985 GDP purchasing power parities from OECD (1985), and from Summers and Heston (1991). …