My friend Josh May suggests that the capital gains tax rate on stocks be set at time of purchase, and that it vary according to the price-earnings ratio of what you buy.
His idea is that if you buy a very cheap stock, you get a low tax rate. A 5 P/E would get you a 5% long-term capital gains rate. If, on the other hand, you buy a nosebleed IPO stock at a 70 P/E, 70% of your profits (if you make any, which is not likely) would eventually go to the IRS. We should encourage sound investment and discourage speculation, and this is Josh's way of doing it.
I like this idea! Though it is not exactly practical -- and Josh is proposing it with tongue pretty firmly in cheek. One problem is that some companies are losing money, and their stocks have no meaningful P/E.
He has a point, though. Could the tax rate vary based on the 10-year Shiller P/E of the whole U.S. market on the day of purchase?
I have one other version, more workable, of the May Plan. It may catch the spirit even as it abandons details: simply have the capital gains rate go down over time. Current law does this, but only with a single bump down (from the ordinary income rate to the lower capital gain rate, at the one-year mark). We could do more. The longer you hold, the lower should be your rate. Pay a 40% tax on your gains at one year, 39% at two, 38% at three, and so on. We could stop the discounting with a rate of, say, 21% when an investment is held for 20 years. (Keep going, if you think that would be better policy.)
This would help get people to think with a longer time horizon.
I guess this is a variant of a Tobin tax -- tax as a way of dampening too-frequent trading, i.e. speculation.