This paper presents two tests for strict exogeneity of the covariates in a correlated random effects panel data Tobit model. The tests are applied in an analysis of hours of work of US women. Estimation procedures when a model does not pass a test for strict exogeneity are discussed.