In theory, tax rates are lower on capital gains and dividends because this raises the stock of capital, which is good for the economy. But as Bruce dryly points out, "The empirical question of whether sharply lower taxes on capital, and hence the wealthy, has actually raised saving, investment and productivity is one I will revisit. Suffice it to say that since 2003, when the current tax rates on capital gains and dividends were instituted, the economy offers little, if any, evidence on this score."
I want to resist the temptation to say that this is a circumstance in which the theory says one thing and the empirical evidence says something else. One point is that even insofar as people have specific retirement income targets they're trying to hit, lower taxes on investment income could easily reduce savings. The other point is that offsets matter. If you paid for a reduction in investment taxes with new taxes on consumption or with cuts in Social Security benefits, there'd be a strong theoretical argument for a net shift out of consumption and into investment. But what we did in 2003 is finance a tax cut with increased borrowing. In that case for the tax cut to increase the capital stock the shift in incentives has to be so large as to outweigh the direct crowding out that's involved with deficit finance in a non-depressed economy.