from Discovering Urbanism
Randal O’Toole has proposed a thought experiment that he uses to “debunk the smart growth myth” of the ideal walkable neighborhood grocery store.
“For smart growth to work, then population densities must be high enough for businesses to have enough customers within walking distance to keep them going. Smart growth won’t work if businesses in pedestrian-friendly neighborhoods must attract hordes of auto drivers from other areas in order to survive. A modern large supermarket needs to draw patrons from a community of about 40,000 people. This is known as the trade population for this kind of store.
Joel Garreau says that, as a rule of thumb, ‘the farthest distance an American will willingly walk before getting into a car’ is 600 feet. However, ‘if you do everything you can to make casual use of the automobile inconvenient at the same time that you make walking pleasant and attractive, you maybe, just maybe, can up the distance an American will willingly walk to 1,500 feet’…
The population density required to place 40,000 people within 1,500 feet of a grocery store is almost 124,000 people per square mile. That’s about two-and-one-half times the density of Manhattan.”
Therefore, smart growth won’t work. QED.
O’Toole has asked a worthwhile question but plugged in the wrong numbers to answer it. Out of curiosity, I’d like to take a closer look at this hypothetical scenario to see how feasible the walkable grocery may really be.
Customer Base. O’Toole uses the concept of trade population, but this is begging the question. His trade population number is generated assuming an automobile-oriented environment, which is exactly what the proposed scenario is an alternative to. The more objective measure is the actual pool of customers a large grocery draws from. To get a feel for this I counted all of the food sources that serve the Charlottesville-Albemarle area. I came up with 14 large “modern grocery stores,” between 30,000 and 60,000 sq. feet. This does not include:
- medium-sized full-service grocers
- about a half dozen health food stores
- specialty ethnic food stores
- big box stores like Target that sell food
- dozens of small convenience stores
I’m only counting the kinds of places O’Toole considers an unreasonable sacrifice to do without.
Taking the combined service area’s population to be 134,086 from current ACS data (Charlottesville and Albemarle County), this breaks down to about one large grocery store per 9500 people. APA has determined the average customer base for a supermarket in the U.S. to be 8,412 , but I’ll just stick with my more conservative 9500.
Modal Split. O’Toole, as usual, characterizes the smart growth position as something far more extreme than anyone would actually propose: a 100% walking grocery store, as if bicycles, transit, and automobiles do not exist at all. The term walkable means able to be walked to, not only walked to. Every smart growth proponent I know would actually hope to see a multimodal balance to allow an array of transportation options. For the sake of this scenario, let’s suppose our store has 50% walkers (with some cyclists included in here), 25% transit users, and 25% drivers.
Walking Distance. If you ignore Garreau’s snarkiness, the 1,500 foot number is an alright estimate for typical behavior in a pedestrian-friendly environment. Walking 600 feet will seem intolerable if it’s between the Best Buy and the Bed, Bath, and Beyond through a parking lot and grass berm, but Charlottesville’s downtown pedestrian mall is 2,100 feet long and people will regularly walk its length for an errand because it is so enjoyable. A more objective way to go about this would be to fix the travel time. Apparently, the average drive to the grocery store currently takes about 20 minutes, including the walk from the parking lot. At a leisurely pace, the average walker could traverse about 3,000 feet in 20 minutes. I’ll stick with the 1,500 foot number, noting that this will cut the travel time in half even for the furthest walkers.
Needed Density. Considering all of these conditions, a back-of-the-envelope calculation shows that this grocer could be supported with a population density of 18,885 people per square mile surrounding it. This is about half the density of Brooklyn. If this still seems unreasonably high, it should be noted that this is only the density for one quarter of a square mile area. In theory, it could be surrounded by a greenbelt of parkland and have no effect on the calculation.