Stackelberg differential games are useful settings in which optimal go
vernment policies can be studied. This paper argues that the analysis
of these games involves a key technical issue. In particular, we quest
ion the necessity for optimality of one boundary condition invoked in
existing literature. The issue is of key interest because the boundary
condition is largely responsible for the time inconsistency results p
reviously obtained. We show that the boundary condition is not necessa
ry in some cases. As a result, our finding undermines the credibility
of the existing conclusions. (C) 1997 Academic Press.