The paper compares the way economies with exogenous and endogenous innovation respond to capital income taxes. If innovation is exogenous, tax cuts increase saving. If innovation is endogenous, tax cuts increase innovation as well. Faster innovation raises capital productivity and calls forth still more saving. A larger capital stock lowers the discount rate, increases the present value of monopoly profit and calls for faster innovation. How large a difference endogenous innovation might make is an open question. We calculate numerical solutions of a model including features of the U.S. tax code that affect incentives to innovate. The results suggest that models with exogenous innovation substantially underestimate long–run effects of capital income taxes.