So from that Pew survey, the more affluent, college-educated, younger and white demographics expect more from government than they're willing to pay for.
You've read it wrong. That graph isn't saying what that group thinks about spending/taxation, but rather what that group thinks
the American public thinks - it's clearer on the full report
here (see page 119). In other words, it measures pessimism about the rest of the American electorate - the younger, whiter, wealthier and more educated you are, the more pessimistic you are about what the rest of the electorate want.
However, I do actually think it's a very bad question because it isn't clear what type of pessimism is being tested. The question survey-takers got asked is "Does the public want more from government than it is willing to pay for?". Answering yes to that could imply that either a) the public asks for too much given acceptable taxation levels, or b) the public aren't willing to be taxed enough given acceptable spending levels. That means the question lumps pessimistic small-staters and large-staters together with no real way to discern between them, which is why I think that you get some otherwise strange cleavages. You can't really make any proposals on the basis of public opinion from that question, because it doesn't tell you what the public's opinion is (other than the fact they're generally pessimistic).
In general, being young and educated makes you more in favour of public taxation, being white and wealthy makes you less in favour of it, and vice versa on spending. It's very rare to see all those four things cleave together; the reason it is happening there is because of the above: the question is just not particularly good.
