Validity of LSE for trimmed and differenced panel data due to absorbing attrition

Under attrition, usually panel data are trimmed to include only the individuals with complete records; then, to allow for a unit-specific effect related to regressors, LSE is applied to the first-differenced model. This paper provides conditions for the LSE to be valid.