Wednesday, October 08, 2008

Systematic thinking about food deserts

It is surprisingly difficult to design an objective way of measuring "food deserts" -- neighborhoods with inadequate food retail.

I was thinking about this recently, because USDA's Economic Research Service is hosting a conference on the topic this week, which unfortunately I cannot attend.

I read a lot of advocacy research on food deserts, but I also try to "ground truth" the formal research. A few months ago, I did a post on a walk through the South End and Roxbury in Boston. Below, I discuss a walk in June through Central Harlem with a friend who lives there. I have much longer experience living in mixed-income but high-poverty neighborhoods of downtown Washington, DC.

A key challenge is defining retail conditions that characterize a food desert, but not a typical wealthy neighborhood. The walk through Central Harlem, for example, passed Citarella, an upscale small storefront market with fabulous produce. If one excludes such stores, then many wealthy neighborhoods would also be classified as food deserts, which is clearly not the intended meaning of the term.

Does the presence of a Citarella make a low-income neighborhood adequate for fruit and vegetable retail? Many readers may say "no." Does the same store make a high-income neighborhood adequate? Many readers may say "yes." If these are the answers, has the classification become tautological -- having nothing to do with the retail, but really based on income alone?

To make things more complicated, suppose one decides that low-income neighborhoods must have fresh fruits and vegetables in retail settings targeted to low-income consumers, in order not to be classified as a desert, but a Citarella is adequate for a wealthier neighborhood. In that case, what does one make of fruit stands like this one from the walk in Central Harlem?

Is a neighborhood with fruit and vegetable stands no longer a desert? Do we assume that a brand-name corporate supermarket, typically oriented toward automobile traffic, is essential for adequate food retail in a densely populated low-income neighborhood? If we say neither the fruit stand nor the Citarella counts, then are we "fixing" the definition out of determination to classify this neighborhood as a desert?

While we are on the topic of driving, what is our expectation for driving distances? If you plot supermarkets on Google Maps, richer suburban neighborhoods generally seem to have less retail access than low-income urban neighborhoods. Does that make the rich neighborhoods supermarket deserts? In the United States, even many low-income people travel to the grocery store by car. Clearly, a single neighborhood may have adequate retail from the perspective of some neighbors and not others. Do we assume that nobody in poor neighborhoods shops by car, while assuming that everybody in rich neighborhoods shops by car?

The research on food deserts sometimes turns not on adequacy but on the ratio of "good" stuff like fruits and vegetables to "other" stuff like fast food restaurants or corner stores without fresh produce. For example, see this 2007 study of the same Harlem neighborhoods discussed above. Does the presence of liquor stores or fast food restaurants drown out the value of the fresh produce that exists in a neighborhood, such that the ratio is what matters? I have doubts.

I don't ask these questions out of heartlessness. They are important because some of the implicit remedies, such as tax breaks to bring in supermarkets, have seriously expensive tradeoffs for poor neighborhoods that are already suffering from an inadequate commercial property tax base. I am just not sure that retail location decisions are the sort of decisions that governments make better than markets do.

What do you think? How would you define a "desert" objectively, so that the definition is both reasonable and captures low-income neighborhoods you have in mind, without applying equally well to richer neighborhoods?


Rebecca Blood said...

The fruit stands and small markets in San Francisco usually also stock some staples like rice and tortillas and beans. Personally, I would not consider a neighborhood that had such a fruit stand to be a "food desert".

You're right to be thinking this through in this way. I guess it would come down to your goals. If your goal is to provide "ready access to affordable fresh food", a fruit stand might disqualify a neighborhood as a food desert - but a high-end market instead would not.

On the other hand, if your goal is to provide "ready access to a wide variety of foods", a grocery store might be the only thing that would satisfy the criterion.

Then there are cultural differences. A supermarket that provides dairy products and flour, but no tofu or rice might be inadequate for a neighborhood with a high Asian population. One with a wide variety of breads but no tortillas might be inadequate for a neighborhood with a high Latino population. Fortunately, most businesses cater to their customers, so I'd think that kind of situation would be unlikely to occur - except in the case of the upscale market that is designed to cater to a small segment of the local population.

For me, the criteria would be similar to the first one I named: "ready access to affordable fresh food and staples".

But I'm making the huge assumption there that people in every neighborhood are willing and able to cook from scratch every night. If your goal is "ready access to easy-to-prepare food", a place that sells only frozen pizzas might trump the fruit stand.

Jack said...

I think a great way to differentiate between rich and poor neighborhoods in terms of access to food (and other resources) would be to map the area around the neighborhood that residents typically travel. This would help to capture the culture differences between the neighborhoods. An indicator of actual time spent within the boundaries of the neighborhood would also be useful. This would help to demonstrate that even though richer, suburban neighborhoods often do not have any access to food within the boundaries of the neighborhood there is no access issue because most of the residents spend a large portion of their time outside of the neighborhood. Conversely a large number of residents in many lower-income urban neighborhoods spend much of their time in and in the area immediately surrounding the neighborhood. I’ve gotten a bit ramblie here but I hope you understand what I’m saying about using aggregate statistics showing the distance an average resident of a particular neighborhood travels outside of the neighborhood on a daily basis and the average amount of time the average resident spends outside of the boundaries of his/her neighborhood. I’m thinking in my head: a map of the neighborhood with a circle drawn around it to show how far residents travel and the circle shaded in, darker indicating more time spend outside of the neighborhood. I would suspect that in lower-income areas the circle would small and shaded lightly and in upper-income areas the circle would be large and dark. Then look at the availability of food within the circle to determine the ‘food-desertieness’ of the neighborhood. These statistics would prove to create a difficult surveying project indeed; however, I think could be used as a useful tool to show the difference between low and high income areas.