Economist Robert Frank argues for progressive taxation in today's NY Times by noting that market pay scales are progressive anyway. The most productive workers on any team earn slightly less than the value of their marginal productivity and the least productive earn slightly more than their productivity.
In his discussion, Frank invokes the anti-progressivity libertarian straw man. But one can cheer for a more progressive distribution of incomes without signing on to existing federal programs. One could suggest that real school reforms (for example, not killing programs like Washington DC's voucher program) would go a lot further in that direction than the current tax code.
But that begs the question of the role of the state in impacting the distribution of well-offness. There are many levels of government, many taxes and many expenditure programs. What is their net effect in terms of redistribution? To my knowledge, the big question has never been rigorously answered.
What do we know? In the U.S., the role of the state has grown steadily over the last 80 years, but the chorus of complaints is that "inequity" is now greater than ever. As with all failed programs, the chorus responds by insisting we have not really bought enough of it. (Unhappy with how program X performs? Increase its budget.) This is not how the real world works. We typically (still) downsize or discard the failures.