B1282
Title: Lagrange multipliers specification test for high dimensional one-way random-effects models
Authors: Nirian Martin - Complutense University of Madrid (Spain) [presenting]
J Miguel Marin - University Carlos III (Spain)
Abstract: Lagrange Multipliers (LM) tests are useful for assessing the intra-cluster covariance structure when having equicorrelated and homogenous responses within clusters, i.e. a one-way linear random-effects model. The framework encompasses the linear panel data model with unobserved random effects, consisting of $N$ cross-sectional units or panels; each observed over $T$ time periods. An intra-cluster specification test is proposed. Particularly, in the high-dimensional setting, it considers larger time periods than cross-sectional panels, where the classical likelihood ratio test is not applicable, and the LM tests remain robust to high-dimension.
