Abstract
Two aspects of the methodology for constructing index numbers of average earnings are investigated using data for the UK. Although trends in average earnings reflect changes in both industrial structure and the composition of the labour force, it is common to compute indices which ignore compositional changes other than those at the industrial level (as in the UK). First, this article examines the effect of standardising at both levels using UK data for the period 1971–92. Secondly, it considers the practical significance of using alternative index number formulae which possess certain desirable properties. It is concluded that the official series is subject to unnecessary error components which result in a non-negligible upward bias.
Get full access to this article
View all access options for this article.