Is the patent length an effective policy instrument in stimulating R&D? This paper develops a generalized variety-expanding growth model and then calibrates the model to the aggregate data of the US economy to analyze the effects of extending the patent length. The numerical exercise suggests that at the empirical range of patent-value depreciation rates, extending the patent length beyond 20 years leads to only a very small increase in R&D despite R&D underinvestment in the market economy. On the other hand, shortening the patent length can lead to a significant reduction in R&D and consumption. This paper also makes use of the dynamic general-equilibrium framework to examine the fraction of total factor productivity (TFP) growth that is driven by R&D, and the calibration exercise suggests that about 35% to 45% of the long-run TFP growth in the US is driven by R&D. Finally, this paper identifies and analytically derives a dynamic distortion of the patent length on saving and investment in physical capital that has been neglected by previous studies, which consequently underestimate the distortionary effects of patent protection.