Thanks a lot for writing this up - it's nice to get some info on this literature. I didn't get though the relationship between the selfish option and "doing good ineffectively" - why do you think that rejecting the selfish option would be a response to the ineffective charity?
Remember that the core problem here is the tension that arises when people say that ineffective charity is impermissible while also saying that selfishness is permissible. This pair of views implies that a more beneficial option is more blameworthy, which seems paradoxical. But if we say that both of these options are impermissible, the problem goes away: the more beneficial option is never more blameworthy than the less beneficial option.
A review and critique of the third section of the volume The Ethics of Giving: Philosophers' Perspectives on Philanthropy, edited by Paul Woodruff, with an emphasis on issues relevant to the decision making of Effective Altruists.
Jeff McMahan writes about whether or not it is permissible to donate to a suboptimal charity. He attacks a basic tension in the views defended by several preceding philosophers: that it is permissible to be selfish, and nice to donate to the best charity, but *wrong* to donate to an ineffective charity. The prima facie contradiction should be obvious. How is it that my actions go from being permissible to being impermissible as they become more generous and more beneficial?
But in defense of that view, other authors have provided some reasonably compelling thought experiments. Suppose that I can do nothing or, at great cost to myself, I can save a man from losing either one or both of his arms. It seems reasonable to say that I can do nothing, as part of the general idea that morality doesn't demand us to sacrifice all the time. And it seems reasonable to say that saving both the man's arms at great personal cost would be praiseworthy and supererogatory. But to accept the same personal cost and yet only save one of the man's arms just seems perverse.
McMahan argues extensively about this and other thought experiments to show that the analogy fails to apply to charity. Once we revise such thought experiments to be similar in structure to donations, we no longer find them to be compelling. But McMahan never addresses a much more simple solution to these arguments, which is to deny that the selfish option is permissible in the first place. McMahan briefly mentions this as a potential response:
"A different strategy for effective altruists is to try to develop a robust defense of the demandingness of morality, according to which much less of doing good is supererogatory than we have hitherto supposed."
The way he presents it is odd: given the basic demanding arguments of consequentialism, Singer's demanding argument which formed the basis for the EA movement, the demanding donation practices of many EAs, and the fact there are good existing arguments against the demandingness objection, I would think that this should be treated as the default position. McMahan's argument can be taken to show that it's much more plausible than the mindset held by some EAs, which is that it's a free personal decision whether you want to give a lot of money to charity, but that that money must then be given to an effective one.
I don't think McMahan would find what you call a 'solution' very appealing: McMahan doesn't think that morality is demanding in the way e.g. Singer does. Further, what you suggest ought to be the default position - morality is really demanding - is something only a small percentage of philosophers (although many EAs) believe is correct.
He doesn't indicate that there is a reason to consider it unappealing. It's not a matter of whether he agrees, it's a matter of whether it's a substantive view that ought to be addressed.
The literature on moral demandingness is fairly split. The proposition that morality is demanding is different from the proposition that demandingness is not a moral problem. The majority of philosophers believe in theories which happen to not turn out to be very demanding in our current world; that does not necessarily mean that they did so because those theories are less demanding, so their views don't really diminish the viability of positing demanding moral principles in response to a new moral problem.