Quantifying tax effects under policy foresight$

Studies of tax effects make the conventional information assumption that changes in period-t taxes become known at t. Legislative lags, however, imply that news arrives before tax changes take place. Under policy foreknowledge, the conventional information structure is therefore misspecified. Simulations of a standard neoclassical growth model suggest that foresight of only one quarter can distort substantially the estimates of tax effects obtained under the no-foresight assumption. Also, it is crucial to model capital and labor taxes separately: anticipated changes in these two tax policies have opposite effects on consumption, investment, labor, and output before policy realization. r 2005 Elsevier B.V. All rights reserved.