Measuring Income Mobility with Dirty Data (published in Ethnic and Racial Studies, 22(3), May 1999)

We examine the performance of measures of mobility when allowance is made for the possibility of data contamination. We find that 'single-stage' indices - those that are applied directly to a sample from a multivariate income distribution - usually prove to be non-robust in the face of contamination. However, 'two-stage' models of mobility - where the distribution is first 'discretised' into income intervals and then a transition matrix or other tool is applied - may be robust if the first stage is appropriately specified.