Perhaps I can make it so that I am the good guy rather than the bad guy. But this makes no broader difference. Others will be helped and harmed (respectively) either way. I just get to choose which of 'me' and 'my counterpart' gets assigned to play which role. Put another way, I have various counterparts performing various actions, and I get to choose which one of these guys to locate myself within. This cannot be a morally significant decision. It makes no difference to the worlds. It's like choosing which movie to watch. We don't really influence the events therein. We merely choose which events to view, which pairs of eyes to see out of. The spectator's self-locating decision affects no-one but themselves.
David Lewis denies that his modal realism has such ethically repugnant implications. He claims that we should only care about our own world, and making this a better place. But apart from the crude tribalism, it's not as if we can change the world itself. Each world is what it is, and according to modal realism they are all equally real. So all we can do is change which world is ours. We can make "this world" a better place de dicto, by changing which world 'this' refers to. We don't thereby change the world itself (de re). As I wrote once before:
All we'd be doing is moving ourselves so that we were 'closer' to the well-off people rather than the suffering ones. And that hardly seems like a virtuous move (by my intuitions anyway).
Jeremy doesn't see it this way though:
If my action causes the bad, then I am to blame. If it causes the good, I'm to be evaluated positively. So if I'm the one who does the bad thing, and some duplicate of me in another part of the multiverse does the good thing, then I'm to blame and he is to be congratulated. So my not doing an action might logically (but not causally) entail someone else doing it, and it would lead to a result that's exactly similar to what happens if I do the other action. But that doesn't mean my action isn't bad if it's the bad one. In the world where I do the bad action, my action is bad. In the world where my duplicate does the bad action, his action is bad. The total amount of good or bad in the universe is irrelevant to whether my action right here is bad.
That means we should blame the one who does the bad thing, even if any choice leads to the same resulting future when you factor in the whole multiverse. This means that, even if this view that we have no evidence for is true, moral evaluation still makes sense.
By appealing to our standard moral practices of praise and blame, I think Jeremy fails to fully take on board the radical implications of modal realism here. Sure, we can still distinguish the good and bad actions. But an agent's choice between them has no significance, for the reasons explained above. The actions will all take place regardless of the agent's choice. So it makes no sense to say that he "should have done otherwise". He did! (At least, his counterpart did.) At worst you might criticize an agent for choosing to locate himself in the life he chosen. Perhaps he has poor aesthetic taste, prefering to experience the bad actions than the good ones. But again, it's not as if his self-locating/spectative decision impacts upon anyone else.
Worse, you might have suspicions about these "locative powers" according to which agents have the semi-magical ability to influence in which world their consciousness resides. Perhaps the most coherent interpretation of modal realism would simply deny that the "agents" in each world have any real choice at all. And then moral evaluation goes right out the window. (We can't even justify blame on pragmatic grounds, since nothing can influence multiversal consequences.)
Finally, Jeremy suggests that non-consequentialists are unaffected by the argument:
The argument is that the net result wouldn't be any different if I did an action usually considered better [or] worse, and therefore the action isn't really better or worse because the consequence of either action would be the same. If consequentialism is true, then the only morally relevant features of the action are its consequences, but anyone who denies consequentialism isn't going to buy this.
Again, this seems to underestimate what's going on here. The most plausible ethical views allow that actions have other morally relevant features besides consequences, but nevertheless recognize that these other features are ultimately grounded in consequential concerns. Ethics is important because (we assume) our choices can affect others and change the world. If this assumption is false, as modal realism would have it, then our decisions - and hence the norms governing them, i.e. ethics - are inconsequential, in the most derogatory sense. To care about ethics even when it makes no difference would be arbitrary and fetishistic.