Introduction

Nine of the papers presented here originate from a conference on the measurement of capital held at the University of Southampton in the summer of 1976. These papers were subsequently revised for publication in this volume. Two further papers, one by Alan Armstrong and the other by William Peterson (on total factor productivity), which are germane to the conference theme but were not available at the time, are also included. The conference was convened and organised by David Heathfield whose chief concern was to promote discussion between theorists, statisticians and applied economists. This aim seems particularly relevant in the case of capital measurement. Here we have all theorists agreed that an aggregate measure of the stock of physical assets is impossible except under extremely stringent conditions1. In addition the theorists have reached a consensus that the particular aggregation conditions required for the specification of capital within an aggregate production function will not be met in reality; any parameter estimates based on such an exercise may be misleading except under certain circumstances. Furthermore, the measure of capital usually used in this function is not strictly independent of prices and distribution2. Meanwhile, back at the Central Statistical Office, and elsewhere, statisticians provide measures of something they call capital stock and applied economists continue to use these aggregate measures in econometric models based on either strict neoclassical parables about aggregate production functions or stories of another ilk.