Measurement Error and Misclassification : A Comparison of Survey and Register Data

The authors provide both a theoretical and empirical analysis of the relation between register and survey data. By distinguishing between different sources of deviations between survey and register data the authors are able to reproduce several stylized facts in the literature. In doing so, they deviate from the almost universal assumption that the register data represent the truth. They illustrate the implications of different error sources for estimation in (simple) econometric models. The analysis is applied to Swedish data that have been collected for a validation study as part of a larger European health and retirement study (SHARE: Survey of Health, Ageing, and Retirement in Europe). Thus this paper makes two contributions: (1) it adds to the limited number of empirical validation studies of earnings measurement in surveys and (2) it shows the sensitivity of some findings in the literature for the assumption that register data represent the truth. They find in particular that the common finding of substantial mean reversion in survey data largely goes away once we allow for a richer error structure