Megan McArdle offers some reasons why we shouldn't place too heavy a tax burden on "the rich."
High taxes on a narrow base are about the opposite of optimal tax theory. This is not because economists are mean, cruel people who are primarily interested in serving their corporate overlords, but rather because the narrower the base, and the higher the rates, the more sharply the marginal returns to rate increases diminish.
Take an extreme example. The top 1% of households, about 1 million in all, have about 20% of national income. They've also experienced most of the income gains in the last twenty years. So let's say we want to fund federal operations entirely out of their pockets. Well, to do so, we'd need an income tax rate of 100%. Even ardent liberals will surely concede that at these levels, the supply-siders are right, and we'll soon end up with no tax base.
Even a less extreme example--make them pay half the tax burden--ends up with a 50% effective rate on high earners. And to get a 50% effective rate, you need an even higher marginal rate. The problem for people who want to load tax increases on these people while cutting taxes for everyone else is that if you actually succeed in shifting the tax burden this way, you'll rapidly end up on the wrong side of the Laffer Curve.
There is another reason too - "the rich" are highly susceptible to changes in the economy. They have done reasonable well for the last several years because the economy, in general, has done very well. But when the economy starts to do poorly the rich tend to see some pretty dramatic drops in their income. Its very much a high risk/reward system.
Do we really want the tax base so highly dependent on these wild swings? It might be entertaining to watch for a bit, but probably not very good for the overall health of the country.