Matt Zwolinksi thinks there's something odd about philanthropists like Warren Buffett calling for higher taxes when their own philanthropic efforts are instead directed towards funding non-governmental organizations. But there's nothing remotely odd about this. It is exactly what we should expect if what's politically feasible is more constrained than what's individually feasible, such that:
(i) It is individually, but not politically, feasible to direct funds to one's favoured NGO. That is, you can do it with your own money, but you don't expect to be able to redistribute (much) taxpayers' money to this end.
(ii) Increasing government revenue is believed to be the best of the politically feasible options for large-scale redistribution -- in particular, better than the alternative of a marginal increase in consumer spending that would arise in the absence of additional taxation/redistribution.
I'm not here going to argue that (ii) is true. My point is just that it is obviously believed by those who advocate for higher taxes. And there's nothing particularly surprising that it should be believed by them concurrently with (i), i.e. that there are NGOs that they would prefer to be funded, whenever feasible.
This obvious point seems to have been largely neglected in the broader blogospheric discussion, which has focused more on standard "coordination problem"-type justifications for non-voluntary taxation (see, e.g., Will Wilkinson, and Matt's reply). That works fine for explaining why people might reasonably advocate for higher taxation while being unwilling to engage in any individual acts of philanthropy themselves. But it doesn't address Matt's question of why even philanthropists don't voluntarily donate money to the government.
The answer to that is simply: there are better philanthropic options available to them as an individual to donate to. But, contra Matt, this really doesn't tell us anything about whether we should force others to pay into the public treasury, because those better options may not be politically available. (I'd certainly rather that tax increases instead went to fund GiveWell, were that an option!)
We have less control over others' money than we do over our own, and the public treasury may, in practice, be our only option for redirecting it. And the mere fact that we have better options for our own philanthropic efforts does nothing at all to suggest that the public treasury is a worse place for the marginal dollar than consumers' wallets -- assuming that most consumers are not otherwise going to direct those marginal dollars to superior NGOs. (Of course, there might be other reasons for thinking that taxation has bad consequences on net, but that would be a different argument.)