SummaryThis paper reviews various mobility measures and establishes their asymptotic sampling distribution. The focus is on both transition matrix mobility measures and mobility measures which are based on the reduction in inequality occurring when the accounting period is extended. Statistical techniques are used to show the asymptotic normality of these measures and their variances. The empirical illustration examines the development of earnings mobility for both sexes in Germany between 1983 and 1992 using the Socio-Economic Panel data. It turns out that male earnings mobility fell during the eighties whereas the development of female earnings mobility is less clear. Comparing the levels of mobility females are more mobile than males, at least when inequality reduction mobility measures are employed. Considering the rather large number of observations the confidence intervals are often surprisingly wide. Therefore, confidence regions for mobility measures ought to be reported in empirical work whenever possible.
Application of confidence intervals and joint confidence regions to the estimation of psychometric functions
