The effect of using administrative registers in economic short term statistics

In case of a single survey at one time point, it is well known that combining administrative registers with survey data often substantially improves the quality of statistic production. However, in short term statistics it is as important to measure changes over time as it is to measure the overall level. Using data from the Norwegian Labour Force Surveys and administrative registers, we demonstrate in this paper that the use of registers have little or no effect on the accuracy of estimates of change, both in terms of the sampling variance and the bias introduced by nonresponse. The survey data alone measures the changes over time. One consequence of interest for the rotation design is that a very high proportion of the sample can be retained from one survey to the next without seriously reducing the accuracy of the level estimates. We believe that these findings are relevant to short term statistics in general.