I am not American and don't know much about the specifics of the welfare systems over there, and would like to understand more. I understand you have a food assistance system – SNAP, colloquially called "food stamps". I have also indirectly come to understand the system has been quite effective at alleviating hunger among very low income earners.
What I don't exactly understand is why this is the case, compared to simply giving cash. SNAP coupons are restricted to food items, but this doesn't seem to matter – I don't see why coupons couldn't be traded with other people for non-eligible goods anyway, either directly (historically – I understand it's all electronic now) or by converting to eligible goods first, making them essentially "cash with extra steps". And if the full SNAP benefit is spent on eligible food in a month anyway, it seems no different from a cash transfer at all.
But surely there is an administrative overhead. I don't know how big it is – it might be small, but I'd be very surprised if it was zero, and any money not spent on overhead can be spent as benefits instead. Furthermore, it seems generally true that people know their own utility functions best, and so we should expect people to make more optimal choices for themselves if simply given cash as opposed to coupons with politician-mandated eligibility criteria. So hence the question – why is SNAP better than simply handing out cash?
Is SNAP more efficient than simply giving cash, and if so, why?
byu/soldiercrabs inAskEconomics
Posted by soldiercrabs
1 Comment
It’s not more efficient—it’s strictly less efficient. So why have SNAP instead? The main reason is that [taxpayers have very paternalistic views about assistance for the poor](https://tobin.yale.edu/research/why-so-much-redistribution-kind-and-not-cash-evidence-survey-experiment), and so they prefer SNAP because they *want* to restrict the choices of recipients, they want to guarantee that recipients can’t spend tax money on alcohol and drugs. (Note that recipients strictly prefer cash!)
Even though SNAP should be cash-equivalent for consumers if they already spend enough on food (as you correctly point out), [in practice it is not](https://www.aeaweb.org/articles?id=10.1257/aer.20170866); consumers spend more on food when they receive SNAP than the cash equivalent, usually attributed to the bias known as “mental accounting.”
A [new paper](https://www.nber.org/papers/w34506) combines both of these ideas and says that the setup of SNAP is actually optimal *if* policymakers have paternalistic views *and* recipients engage in mental accounting, since SNAP money will actually be less likely to be spent on undesirable goods. (I should note that I think the empirics in this paper are kind of shady)