This paper extends the dynamic theory of optimal fiscal policy with a representative

agent in several environments by using a generalized version of recursive preferences. I

allow markets to be complete or incomplete and study optimal policy under commitment

or discretion. The resulting theories are interpreted through the excess burden of taxation,

a multiplier, whose evolution gives rise to different notions of “tax-smoothing.” Variants

of a law of motion in terms of the inverse excess burden emerge when we allow for richer

asset pricing implications through recursive preferences. I highlight a common unifying

principle of taxation and debt issuance in all environments that revolves around interest

rate manipulation: issue new debt and tax more in the future if this can lead to lower

interest rates today.