Just adding on to Glenn's post, much opposition to the government actually doing anything decent for people comes from the idea that the government is going to take my tax money and give it to people who don't deserve it. The problem is that for decades the Dems have tried to get around this by making sure policies and programs were relatively small and incremental, everything targeted and means tested. But doing that effectively confirmed the critics' point. The big (giant) government programs which are most popular are the ones which are universal - Social Security and Medicare - and other less controversial government programs, like highway spending, are also perceived to benefit people across the board.
There's a couple of interesting things going on here that seem worth unpacking. The first is actually a legitimate point about how valid arguments against various kinds of redistribution tend, with unsettling ease, to shade into unsavory demonization of the folks on the receiving end of the transfer. Suppose someone suggests that the government should, either by regulation or direct subsidy, ensure that the indigent are provided with health care or that insolvent homeowners are protected from foreclosure. Now, there are a few types of objections people might raise. There's an argument from efficiency and incentives: To the extent that the risks associated with individual financial or lifestyle choices are borne by the public, there's a familiar problem of "moral hazard" reducing incentives for prudence. And there's an argument from property and autonomy, to the effect that even if people ought to help others in need, each person is entitled to decide whether and how to do so without compulsion. Neither of these implies any blanket judgment about the folks who find themselves in need of aid. The first argument does suggest that redistributive policy will make it rational for people to take more risks at the margin, but it does not follow from either that people who are having trouble meeting their mortgage payments, or people who get sick and cannot afford care, are bad or foolish or irresponsible or otherwise deserving of their fate. And it is a good thing for these arguments that no such conclusion follows, because it's clearly not true.
Yet in popular political rhetoric, it's disturbingly easy to find just such a leap being made. Think of Rick Santelli's jeremiad against "losers" under foreclosure getting bailed out by government. Is it just that people are inherently spiteful or unkind? In fact, the tendency to assume that people who are badly off must deserve it may be a result of what social psychologists call the Just World Hypothesis. In brief, faced with evidence that the world is often arbitrary and unfair, and that bad things often happen to good people, many of us prefer to preserve our faith in a basically fair and benevolent universe by assuming that the badly off must somehow deserve their fates—which is a stronger and (I think) rather morally uglier proposition than the more plausible notion that people are often significantly responsible for their fates.
There are at least three reasons to take some care to avoid this implication, given how easily human beings fall into it. The first is just that it's an ugly and callous attitude to have toward people who will often deserve our compassion whether or not they ought to receive government aid. The second is that people will readily—and sometimes intentionally—misconstrue an argument about incentives as an argument about the moral worthiness or personal virtue of the proposed recipients, which does not make for a particularly fruitful conversation. Finally, there's a paradoxically quite authoritarian implicit premise lurking behind this sort of argument—to wit, that it's the job of the government to determine who is or is not morally deserving of its largess, and that the central question is whether this or that particular class of prospective recipients qualifies. That's a frame people across the spectrum ought to be uncomfortable with.
As Atrios points out, strategic response to this on the part of progressives has been to embed what are essentially welfare programs within an elaborate—and functionally, if not politically, superfluous—superstructure of universal social insurance. My colleague Will Wilkinson has pressed this point cogently in the context of Social Security. The rationale for the program is ultimately that we hope it will prevent people from being mired in poverty in old age. There is no sane reason, on this rationale, for cutting Bill Gates a check when he reaches the age of eligibility—but we do it this way because progressives believe, perhaps correctly, that a means-tested aid program for the indigent elderly would be more politically vulnerable to cuts. Which, I think, underscores the perverse effect of thinking in terms of the desert of the recipients, since there's no actually-valid argument on which a universal need-blind benefit makes more sense than a narrow means-tested one. So one more reason to eschew desert-centered political discourse: It gives rise to policy that's less intelligent whether your underlying commitments are progressive or libertarian.