[Chief Executive of Oxfam GB] Goldring says it would be wrong to apply the EA philosophy to all of Oxfam's programmes because it could mean excluding people who most need the charity's help. For a certain cost, the charity might enable only a few children to go to school in a country such as South Sudan, where the barriers to school attendance are high, he says; but that does not mean it should work only in countries where the cost of schooling is cheaper, such as Bangladesh, because that would abandon the South Sudanese children.
Fuzzy group-level thinking allows one to neglect real tradeoffs, and pretend that one is somehow helping everyone if you help each group a little bit. But this is obviously not true. If there are more Bangladeshi children in need of education than your current budget can provide for, then by spending the rest of your budget on educating a few kids in South Sudan, you are abandoning a greater number of Bangladeshi children.
If we don't have the resources to help everyone, then (inevitably) some people will not be helped. To put it more emotively, you could say that we are "abandoning" them. That's a good reason to try to increase our philanthropic budgets. It is not any sort of reason at all to spend one's budget inefficiently, leading to the philanthropic "abandonment" of even more children.
Goldring's reasoning is 'collectivist' in the bad sense: treating groups rather than individual persons as the basic unit of moral consideration. That you have already helped some Bangladeshi children is cold comfort to the other Bangladeshi children you have spurned for the sake of instead helping a smaller number of South Sudanese. They would seem to have a legitimate complaint against you: "Why do you discount my interests just because I share a nation with other individuals that you have helped? You have not helped me, and I need help just as much as those you chose to prioritize in South Sudan. Since you chose to help a smaller number of South Sudanese children when you could have instead helped a greater number of children in my community, you are effectively counting my interests for less. That is disrespectful and morally wrong."
By contrast, if you focus your philanthropic resources on providing as much good as possible, no-one has any legitimate complaint. You may imagine a South Sudanese child asking, "Why did you not help me? Just because it's more expensive to provide schooling in my country, does not make my educational needs any less morally important than anyone else's!" To which the obvious answer is, "Indeed, I give equal weight to the interests of all, including yourself; but that is precisely why I must prioritize the educating of a larger group of individuals over a smaller group, if my resources only allow for one or the other. If we could fund your education without thereby taking away the funding for multiple other people's education then of course we would! But it would not be fair on those others to deprive several of them of education in order to educate just you. I count your education as being equally as important as the education of any other one individual. If I can then educate a second individual as well for the same amount of resources, then that is what treating each person's interests equally requires me to choose."
We may vividly demonstrate the irrationality of the collectivist's thinking by mentally subdividing the group of Bangladeshi children into two groups. Call 'B' the group that is helped by the current budget, and 'C' the group of additional children who will be aided if and only if funding is redirected to Bangladesh from South Sudan. Call 'S' the group of children currently aided in South Sudan, who stand to lose funding if we instead help C.
By stipulation, C is a larger group than S. And we may suppose that "all else is equal": the individuals across each group are equally deserving, and stand to gain just as much from being educated. I submit that it is plainly irrational to prefer to aid B+S rather than B+C. The argument that the latter choice "abandons the children" in S is more than counterbalanced by the fact that choosing B+S abandons the (even greater number of) children in C. It's wrong, and it's a completely irrational and unmotivated form of wrongdoing that could be easily avoided if only people were to think more clearly about how to justly adjudicate moral tradeoffs.
We absolutely should focus our philanthropic efforts towards the most efficient opportunities. It's true, but inevitable, that when under-resourced we will be forced to "abandon" some in need. But we should seek to minimize the burden of that abandonment: to abandon fewer individuals (or else those with less urgent needs, or less to gain) where possible. One does not achieve this by fuzzy group-level thinking that obscures the number and needs of individuals. Helping some members of a group is not the same as helping all members of the group; it is not an intellectually honest way to avoid dealing with the burdens of abandonment.