Abstract

The Labor Force Survey (LFS) was originally conducted once every quarter to estimate labor statistics in the Philippines. Beginning 2021, however, the LFS is conducted monthly in response to the volatile nature of employment due to CoVID-19. To simplify the survey operations, it uses dependent replicates between regular (i.e., quarterly) and additional monthly surveys. However, the trade-off is that the computation of variance in pooled surveys is complicated as the regular-to-additional-month covariance is not equal to zero. As a result, additional steps are required to account for non-zero covariances. Hence, this paper aims to demonstrate a method to compute the variance of labor statistics that captures replicate dependence.