Most of the discussion by economists of the appropriate capital gains tax rate is about a very narrow criterion: the effect of capital gains tax rates on capital gains tax revenues. But in a 2009 study done for the Institute for Research on the Economics of Taxation (IRET), Ohio State University economist Paul D. Evans considers a broader criterion: the effect of capital gains tax rates on overall federal tax revenues.
What's the difference? Because capital gains taxes discourage capital formation, they also cause other tax revenues to be lower. If there's less capital formation, workers have less capital to work with and, therefore, are less productive. If they're less productive, the government collects less tax revenue from them.
Professor Evans looks at data from the overall economy from 1976 to 2004, a period in which there was a lot of variation in the marginal tax rate on capital gains. He concludes that in 2004, the tax rate on capital gains that would have maximized overall federal government revenues was 9.69 percent. But if the government taxes to maximize revenues, the ratio of the delta in deadweight loss to the delta in tax revenue is infinity. [I rewrote this previous sentence in response to a correction from Robert Murphy in the Comments section.] Therefore, if the revenue-maximizing capital gains tax rate was 9.69 percent, the optimal tax rate was even lower. So a greedy, grasping government that wants to maximize tax revenues should cut the marginal tax rate on capital gains and if that government cares at all about taxpayers, it should cut the rate even further.