I don't think trolley problems are used to argue for policies. Rather, the point of trolley problems is to reveal that the way humans normally do moral reasoning is not shut-up-and-multiply utilitarianism.

Activist says, "We've got to take from this rich fat cat and give it to these poor people, or the poor people will starve and die. If you take the money, the fat cat will buy less cars and yachts, and the poor people will become much more successful and happy."

While activists may try to trot out utilitarian justifications for their political arguments, nothing about trolley problems can be seen as bolstering their claims (either that redistribution is utility-maximizing, or that utilitarianism is itself the "correct" moral theory.) Trolley problem research isn't normative, it's descriptive.