Why do PROC QLIM and PROC LIFEREG give different standard errors on parameters when estimating the same tobit model?


Differences in standard error estimates between the two procedures are due to differences in the algorithms they use for covariance matrix estimation. Specifically, when inverting the Hessian matrix PROC QLIM uses a G4 inverse which satisfies all four Moore-Penrose conditions while PROC LIFEREG uses a G2 inverse which satisfies only two Moore-Penrose conditions.