I haven't been following the moral economics posts that closely, but I wanted to add a musing of my own. This isn't important to read. Summary: You can effectively pay an EA without giving them anything by donating on their behalf. This transfer has economic properties that tingle my brain when I think about them.
Think when someone gives you a birthday present. Sometimes you like it and it has value to you. Great! Sometimes you don't like it and it doesn't have that much value. Bummer!
The birthday present cost a certain amount of money for the buyer to purchase and it's worth a certain amount to you. Often times, the present is worth less to you than the buyer spent to purchase it, or you would have gotten it yourself. This is why economists often bemoan the loss of value inherent in gift giving. (Of course, this is just one type of value, and there are many non-monetary benefits to gift giving that economists might not pay much attention to!)
Perfect gift giving (from an economics perspective) is when you get a present where the amount of money it cost the person to buy the present is equal to (or less than) the value you get from the present.
One such example of perfect giving is if the present-giver could have also given you cash equal to the present's value and you could have bought a present you liked better yourself. Indeed, this is the entire idea behind gift cards -- provide more value to the person by letting them buy what they want, but not lose the social institution of gift giving by making it just a straight cash transaction. (Though many aunts and uncles often give a card with cash inside.)
However, one could also imagine a perfect gift-giver that always gives you a gift that you would have purchased on your own (except you haven't yet). Since you would have definitely spent the money yourself, this gift is basically as good as cash. For example, while a gift card to Arby's isn't of much value to me (I'm a vegetarian but I do like their mozzarella sticks), given the amount I spend on Amazon, a gift card to Amazon is basically as good as cash.
The point is here that I can pay someone without paying them money, provided I give them something that is as good as money. For EAs who donate a substantial portion of their income to charity, this gets a bit more interesting. Since they would donate a certain amount of money to a certain charity, I can then pay them using this principle by (a) donating to their chosen charity myself and (b) communicating that amount to the person so they can reduce their intended donation by that amount.
This is cool because the money transferred is acausal -- it never is sent to them, yet it ends up paying them an amount. This means you can transfer them money at any time, without their permission, without needing much coordination, without needing to set up a way to make payments between each other, in any country where a donation can be made, regardless of what country they are in. It's also a power only really available to EAs. ...And that's kinda cool to think about.
This is a neat capability, and it seems like it might sometimes be useful. I'm not entirely sure how recipients are likely to reason about this -- will they tend to take it at face value and reduce their donations (if they wouldn't normally donate windfall cash)? Or will they want to follow a chain of reasoning about what the money you've given would otherwise have been used for?
Note on terminology: it seems to be a kind of indirect transfer, but it's definitely a causal one: your actions cause them to have more money. I think it's good to reserve the term "acausal" to situations where causality is not involved.