Ever since the observation in 1995 that hunger and obesity co-existed among children, a flurry of research has sought to answer the question of whether those who are more “food insecure” are most likely to also become (ironically) obese. Many studies have now correlated participation in food stamp programs (since renamed the Supplemental Nutrition Assistance Program, or SNAP) with obesity. One in seven Americans are now on food stamps after the recession, so this finding may have important implications for the overall obesity epidemic. It may also seem like obvious selection bias to correlate food stamps to obesity; those who are poor are more likely to eat calorie-dense and cheap food, becoming both stuffed and starved. But a new study comparing food stamp users to eligible non-users found that the stamp users were disproportionately more likely to eat unhealthy foods. What’s going on here, and what should we do about it, if anything?
A series of studies over the past decade or more have found that food insecurity is associated–mostly among adult women in the cross-sectional National Health and Nutrition Examination Survey–with obesity and poor nutritional health (measured by standardized metrics like the Alternative Healthy Eating Index, which is predictive of cardiovascular mortality). The problems with these studies include the potential for a variety of selection biases (those who participate in stamps do not appear to be the same as non-participant eligible people, in terms of demographics) and endogeneity. But even fancier studies with instrumental variables and other statistical measures intended to isolate the effects of food stamps on obesity have found a striking relationship between the stamps and obesity outcomes. What might be going on?
One leading hypothesis is that the food stamp “cycle” is producing a feast-and-famine pattern among participants. That is, most people get SNAP benefits once monthly. They then buy their food in bulk (economically) and resolve their hunger, but run out of benefits (which are only about $3 to $4 per person a day on average) before the end of the month. The anticipation of hunger at the end of the month leads to more hoarding and calorie-dense eating at the beginning of each month, according to this theory. Some evidence from the Bureau of Labor Statistics’ consumer product expenditure survey seems to support this “food stamp cycle” theory:
The other possibilities are that those more likely to eat more food will avail of stamp benefits as compared to eligible people who don’t participate in SNAP. And yet another theory is still more selection bias: that people who live in those areas with the most “obesogenic environments” (no grocery stores, only cheap crappy food to eat, no fresh produce) are those with the most access to signing up for food stamps.
Whatever the cause, the solutions seem similar: in addition to engaging in the existing efforts to subsidize fresh fruits and vegetables and possibly de-incentivize cheap food consumption through taxes, one might change the nature of SNAP benefits. One option might be to increase the benefit to enable better purchasing. A USDA study argued against the idea that increased benefits would improve nutrition, but that study was fairly flawed: it just looked at whether income is correlated to diet quality, and found that it often wasn’t at very poor ends of the spectrum. That doesn’t mean that the nutrition of SNAP recipients won’t improve when they transition from a feast-and-famine state to a slightly more stable state. And the USDA analysis only casually mentioned that people do improve their diets when their incomes are above poverty levels; it’s just that the very poor who don’t, possibly because they’re stuck in the poverty trap of really bad, cheap food, and really transitioning out of that takes quite a bit of income, not just a marginal increase. Another strategy is to distribute SNAP benefits more evenly over the course of a month. In theory, this might encourage more grocery shopping and more consumption of produce. In reality, it may increase the time and transportation costs to get food among the very poor if they are forced to shop more often, potentially worsening food insecurity. It seems like it’s worth looking more closely at this “cycle” problem to figure out whether it truly exists and what might be the best way to remedy it…