Estimates from the Current Employment Statistics (CES) Survey are produced based on the data collected each month from the sample of businesses that is updated once a year. In some estimation cells, where the sample is not large enough, the Fay-Herriot model is used to improve the estimates. Under the current approach, the model combines information from a set of areas and is estimated independently every month. Given the design of the survey, it may be beneficial to borrow information not only cross-sectionally but also over time. This paper explores the feasibility of applying such a model. The results are evaluated based on historical "true" employment data available on a lagged basis.