This paper develops several comparative dynamics results for Ramsey-ty
pe multidimensional dynamic optimization problems in economics. We pro
vide sufficient conditions for the value function to be monotone and s
upermodular, and for the extremal optimal policies to be monotone in t
he state and in other parameters. These results are natural extensions
of the one-dimensional case when the varying parameters are endogenou
s (state variables), but are rather restrictive for exogenous paramete
rs (discount rate and length of horizon). The analysis relies mostly o
n the fundamental results in lattice programming.