Megan McArdle defends a focus on the most disadvantaged:
But my point is also that chronically poor people are more likely to require extra government benefits because they don’t have any of the assets that the temporarily poor bring with them from the middle class: reliable cars, houses, savings accounts, credit cards, friends and family who have spare cash to help out. The chronically poor will need more help, for longer, than folks who are struggling through a temporary job loss or divorce. Which means that, at the very least, they take up a disproportionate share of resources. It seems entirely possible — perhaps even likely — that the chronically poor still account for the majority of spending in many programs.
So, mathematically, I think the argument being made by Bouie and Weissman fails; it obviously makes a lot of sense to focus on the group that generates a disproportionate share of our entitlement spending. At the very least, we should consider the strong possibility that those struggling with chronic poverty might need very different kinds of help than those dealing with a temporary income problem — rather than suggesting, as Bouie does, that we should obviously focus on doing whatever is best for people having an acute poverty episode because they’re the majority.
None of this necessarily means that Ryan’s anti-poverty plan is a good idea. You could still argue that it won’t work: that the chronically poor are basically beyond help, that we should just give them money without expecting that they will ever be able to wean themselves off substantial public assistance. Or that the particular sort of help that Ryan proposes won’t work and we should do something else instead.
I certainly hope that at least some of the chronically poor can, with help, haul themselves into the middle class. Empirically, I don’t know whether the Ryan plan will work — but this is exactly what makes its focus on experimentation and rigorous evaluation so valuable.