Feeding America, the third largest non-profit in the United States, distributes billions of pounds of food every year. Most of the food comes from large firms like Kraft, ConAgra and Walmart that have a surplus of some item and scarce warehouse space. Feeding America coordinates the supply of surplus food with the demand from food banks across the U.S..
Allocating food is not an easy problem. How do you decide who gets what while taking into account local needs, local tastes, what foods the bank has already, what abilities the banks have to store food on a particular day, transportation costs and so forth. Alex Teytelboym writing at The Week points out:
…Before 2005, Feeding America allocated food centrally, and according to its rather subjective perception of what food banks needed. Headquarters would call up the food banks in a priority order and offer them a truckload of food. Bizarrely, all food was treated more or less equally, irrespective of its nutritional content. A pound of chicken was the same as a pound of french fries. If the food bank accepted the load, it paid the transportation costs and had the truck sent to them. If the food bank refused, Feeding America would judge this food bank as having lower need and push it down the priority list. Unsurprisingly, food banks went out of their way to avoid refusing food loads — even if they were already stocked with that particular food.
This Soviet-style system was hugely inefficient. Some urban food banks had great access to local food donations and often ended up with a surplus of food. A lot of food rotted in places where it was not needed, while many shelves in other food banks stood empty. Feeding America simply knew too little about what their food banks needed on a given day.
In 2005, however, a group of Chicago academics, including economists, worked with Feeding America to redesign the system using market principles. Today Feeding America no longer sends trucks of potatoes to food banks in Idaho and a pound of chicken is no longer treated the same as a pound of french fries. Instead food banks bid on food deliveries and the market discovers the internal market-prices that clear the system. The auction system even allows negative prices so that food banks can be “paid” to pick up food that is not highly desired–this helps Feeding America keep both its donors and donees happy.
Food banks are not bidding in dollars, however, but in a new, internal currency called shares.
Every day, each food bank is allocated a pot of fiat currency called “shares.” Food banks in areas with bigger populations and more poverty receive larger numbers of shares. Twice a day, they can use their shares to bid online on any of the 30 to 40 truckloads of food that were donated directly to Feeding America. The winners of the auction pay for the truckloads with their shares. Then, all the shares spent on a particular day are reallocated back to food banks at midnight. That means that food banks that did not spend their shares on a particular day would end up with more shares and thus a greater ability to bid the next day. In this way, the system has built-in fairness: If a large food bank could afford to spend a fortune on a truck of frozen chicken, its shares would show up on the balance of smaller food banks the next day. Moreover, neighboring food banks can now team up to bid jointly to reduce their transport costs.
Initially, there was plenty of resistance. As one food bank director told Canice Prendergast, an economist advising Feeding America, “I am a socialist. That’s why I run a food bank. I don’t believe in markets. I’m not saying I won’t listen, but I am against this.” But the Chicago economists managed to design a market that worked even for participants who did not believe in it. Within half a year of the auction system being introduced, 97 percent of food banks won at least one load, and the amount of food allocated from Feeding America’s headquarters rose by over 35 percent, to the delight of volunteers and donors.
Teytelboym’s very good, short account is working off a longer, more detailed paper by Canice Prendergast, The Allocation of Food to Food Banks.
Canice’s paper would be a great teaching tool in an intermediate or graduate micro economics class. Pair it with Hayek’s The Use of Knowledge in Society. Under the earlier centralized system, Feeding America didn’t know when a food bank was out of refrigerator space or which food banks had hot dogs but wanted hot dog buns and which the reverse–under the market system this information, which Hayek called “knowledge of the particular circumstances of time and place” is used and as a result less food is wasted and the food is used to satisfy more urgent needs.
The Feeding America auction system is also the best illustration that I know of the second fundamental theorem of welfare economics.
Even monetary economics comes into play. Feeding America created a new currency and thus had to deal with the problem of the aggregate money supply. How should the supply of shares be determined so that relative prices were free to change but the price level would remain relatively stable? How could the baby-sitting co-op problem be avoided? Scott Sumner will be disappointed to learn that they choose pound targeting rather than nominal-pound targeting but some of the key issues of monetary economics are present even in this simple economy.