Search results

It has long been recognised that cohort and cross‐sectional age‐earnings profiles differ. A standard procedure, which is quite reasonable in the absence of more information, is to obtain a cohort profile by simply adding the general rate of growth of real earnings to the growth of earnings associated with age, as shown by cross‐sectional data. Indeed, this would seem to be supported by the observation that cross‐sectional earnings profiles for a number of different years show a great deal of stability in their general shape. The main question considered here is whether cohort profiles can in fact be estimated in this simple way. A basic statistical model of age‐earnings profiles is described in the next section. The model is then applied to several groups of professional scientists, chemists and physicists in Britain and Australia, in the third section. The data were obtained from special surveys of career histories and are described briefly in the Appendix. A feature of the surveys is that sufficient data were collected to enable separate analyses of male and female scientists to be carried out.