Testing for spatial lag dependence and homoskedasticity in a random effects panel data model
推导了随机效应空间自回归面板模型中空间相关性与同方差性的联合拉格朗日乘子检验，并进行了蒙特卡洛实验评估小样本表现。
This paper derives a joint Lagrange Multiplier (LM) test for spatial correlation and homoskedasticity for a spatial autoregressive (SAR) panel model with random effects. The heteroskedasticity in the random effects term is an unknown function of known strictly exogenous variables as in Holly and Gardiol (2000). The latter paper deals with a panel random effects model with heteroskedasticity that ignores the spatial correlation. Conditional LM tests for homoskedasticity given spatial lag correlation, as well as zero spatial lag correlation given heteroskedasticity are also derived. Monte Carlo experiments are performed to study the small sample performance of these tests.