Food stamps, time of the month, and outcomes

This is Joseph

A recent set of studies had some interesting things to say regarding food stamps, and what happens in the last week of the month of benefits.  Emily Badger concludes:

None of these studies can say for certain that the food stamps shortfall causes hypoglycemia, or poorer test scores, or student behavioral problems. And the relationship is likely complex, a product of fewer calories, rising stress, financial tradeoffs, or lost sleep. But these studies suggest that families struggle in multiple ways when the food assistance runs out, and in ways that have to do with more than hunger.

These are good experiments, as the participants are serving as their own controls, greatly mitigating potential confounders of this association.  Instead, week of the month seems to be acting like an instrument (acting only through benefit and wage cycles), which is a very good property for an exposure to have.

I think that there are two pieces here.  One, it is pretty clear that we are probably spending more public money to keep food stamp benefits low -- as being admitted to a hospital is expensive.  So it is a odd kind of frugality that insists we need to pay more to reduce benefits to people with low levels of resources.

Two, the test scores piece makes it clear that these decisions also affect children, who are not really moral agents in regards to family financing.  Instead, it is more of a "luck of the draw" as to what income strata one is born into.

I will also note that this could have interesting effects on high stakes testing in areas with a lot of low income families.  Just the variance as to which week of the month a class takes their test could influence teacher evaluations -- a clear example of an exogenous factor we really don't want driving our results.