Small Area Estimation Under A Pseudo-Unit Level Model with First-Order Autoregressive Time Effects and Measurement Error

Statistics Indonesia (BPS) provides employment data through the National Labor Force Survey (SAKERNAS). SAKERNAS data collection is carried out quarterly under a rotating panel survey with a sample unit of the rotation group. Our study focuses on the Rao-Yu unit level model and its modifications. The modified Rao-Yu unit level model is required to handle pseudo panel units i.e., the units are replaced by the summary of units in the rotation group. In this context, these units are referred to as pseudo units. A modification was made by adding a measurement error pseudo unit into the Rao-Yu unit level model. The simulation study showed that the greater variance between areas, the better the capability of Rao-Yu pseudo unit level model compared with the Rao-Yu unit level model in estimating the parameters of interest. Our method proposal was applied to data on the quarterly unemployment rate at the district/city level in one of Indonesia's provinces according to the 2011-2014 SAKERNAS panel rotation data.