Small area estimation for domain discontinuities

Changes in the survey process usually have a systematic effect on the results. This is sometimes referred to as a discontinuity. In the case of samples that are carried out repeatedly, this disrupts the continuity of a time series. To prevent this, the effect of the changes in a survey process can be investigated by means of an experiment. In this paper, a bivariate Fay-Herriot model is developed to estimate such discontinuities as accurately as possible.