David, on your first point: I don’t know if I’d say that the Democratic share of the low-income vote has been “steadily” increasing. Clinton, by my calculations, got about 59 percent of the two-party vote among people making less than $50,000 in 1996.
On the second point: This is just a restatement of the Democrats’ longstanding advantage among lower-income voters. It is equally true that if only people making more than $75,000 a year had voted, Bob Dole would have won in 1996.
The change in Democratic support over time does not appear to be concentrated disproportionately among low-income voters. To reiterate: There’s a theory that Obama ought to be doing very poorly by the historical standards of American politics, but is being boosted in the polls because the welfare state has expanded and an increased number of beneficiaries are supporting him. That theory does not appear to be true: The expansion of benefits and the expansion of the Democratic base vote over time don’t line up demographically.