15 March 2010

A "local" food program that actually increases food-miles?

James McWilliams has a piece on the Atlantic website describing his experience with a "community supported agriculture" company.  Everything is produced "locally."  I need to put that in quotes, because in his first batch the maximum distance traveled was 429 miles.  (Ignoring edge cases and density of population and arable land, about 15% of the country in within 429 miles.)

One of their major selling points is that they deliver a crate of food directly to your door.  This got me thinking -- does the farm-to-door delivery defeat the purpose of growing food locally?

If the food deliveries are enough to completely obviate the need of a consumer to go to the grocery store, and the efficiency of the delivery truck were not too much worse than the efficiency of whatever vehicle is used for grocery shopping, then having the delivery truck make a circuit of all participating households is definitely more efficient than having each of them make a round trip to the store.

But! if you need to go to the store anyway (and this is almost a definite, since you still need your TP, and your detergent, and your drinks, and...) then we have an empirical question on our hands. How many miles closer must your food originate in order to offset the added fuel consumption of the delivery truck? Keep in mind that the final delivery truck will consume more gallons of fuel per pound*mile of food (or per calorie*mile or whatever unit of food transportation) than will whatever means are used to get food to the grocery store.

My intuition is that the fuel you save from sourcing your carrots 200 miles away instead of 2000 is entirely offset by having those same carrots delivered to each house rather than to a central location.  Anybody feel like doing some agent-based modeling on a road network to testing my hypothesis?


  1. I haven't read it yet, but it looks as though this paper could be of interest:

  2. Thanks for the pointer. That looks promising.