Setting up an internal market for food-banks, and “what should I donate to food banks”?
Here’s an interesting prospect; suppose you run a charity that distributes food collected from grocery stores or other major contributors? How would you do it?
This was a question asked in a recent JEP(1) article, and, what surprised me more than the mechanism design (which was fairly straight forward) was some of the results that came to light after a market was implemented. If you’re more interested in what food banks value on the margin (ie. what should I donate to food banks?), skip to the end for the tl;dr.
The specifics of the market are as follows. A charity called Feeding America exists to coordinate donations from supermarkets. These usually involve fairly large amounts of food that have to be allocated to nearby food banks (50 daily truckloads at 30,000 lbs./truck). What they had been doing was to run a queue system where a certain lot would be available and the food bank at the front of the line could pick it up or refuse it after which they would be moved to the back of the line. What was implemented was a new currency and an online bidding system that turned out to be more effective in allocating food, but also provided a mechanism for allocating undesirable food (more on that later)
The individual food banks were given a type of money (I’ll call them FA bucks), based on how many they served, and submit a sealed bid online (like the grocery store auctions for paintings, but with no minimum) for a load of food described in detail (food is denoted in pounds and categorized quite finely). Additionally, small banks have access to credit, and can jointly bid on loads. Winners pay their FA bucks into the pot, and the day’s total take is redistributed by the same formula as the initial distribution.
2 things stand out here – 1 is that the most ardent opposition at the beginning of the project from an individual: [and I quote] “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.’’ To later becoming one of the most ardent supporters. This tells me that markets are in general poorly understood, likely due in part to apathy on the part of people at large, the complexity with which they tend to present themselves in the real world, and poor performance by economists at really explaining the first and second fundamental welfare theorems.
Here’s my new attempt: Markets do not inherently give advantage to some parties over others. When well constructed, they lead to the efficient outcome (meaning that once the market has settled, nobody can be made better off without harming someone else). If John has six marbles and Stacey has 9 I can take Stacey’s marble to give to John, but I’ve hurt Stacey. The market is there, not to assure that everyone has the same, but that there are no marbles left on the ground.
A market works even if you do redistribution beforehand. I don’t lose my efficiency result just because I take some resources from one person to give to another. Where this gets a bit messy in the real world is that when you’re born the only resource you have is time. The rest has to be earned/bequeathed and taking these resources actually does distort the market and reduce efficiency (leaving marbles on the ground). This is a central difference between left and right wing thinking on inequality – with the left generally believing that distortions are small, and the right believing that these distortions are large.
The second standout is that there are not a ton of real world examples of negative prices in action. In theory, they are sound and will present in cases of undesirable goods. In this case, loads of food that need to be picked up, but that no food banks value (maybe they are of very low quality and are additionally somewhat far away for pick-up purposes.) Since the market imposed that every load get picked up, negative prices emerged in a number of cases (about 5% of auctions, the reason that every load has to be picked up is for donor relationship management.) When food banks can get FA bucks for picking up an undesirable load, the donors and food banks (on net) are all happier.
Analysis of food bank outcomes after the fact was very positive. In addition to having a greater degree of freedom over what food they receive, Feeding America has been able to keep donors happy by allowing negative prices, and the socialist was won over.
If you’re interested in learning more, the paper is:
Prendergast, Candice. 2017. "How Food Banks Use Markets to Feed the Poor." Journal of Economic Perspectives, 31(4): 145-62. DOI: 10.1257/jep.31.4.145
It is available free of charge.
Now for the bit you’ve all been waiting for:
Part of the analysis involved finding the average price paid by the food banks for food. This is what you should be donating:
The far and away winner in terms of food bank preference is cereal; think of it, it’s durable, would handle the major part of 1/3 of the meals and is often pretty healthy. Next up: diapers; diapers are expensive and nobody thinks to donate them - also non-perishable. Other notables, cleaning supplies, condiments, pasta. Note that these numbers are likely different in Canada, and the donation season likely sees the relative supply of some goods fall compared to the data used in this paper, but it’s likely a fair approximation.
What's the best food/stuff to donate?
Prices in this market are denominated in shares (FA bucks) / pound. Your prices are denominated in $/unit. So to get the best bang for your donation buck you need to maximize FA bucks/dollar you spend which means we need a conversion factor. Some things can easily be written off. Fruit snacks are light and expensive, and have a midling value to the charity per pound, they will not be a good idea for the individual. Diapers have a high value to the charity, but they are very light (I make it to be 1oz. /diaper 174oz. per box of Costco diapers @ $37.99) which makes for 0.28 pounds per dollar and 1 FA buck per dollar you spend.
Short comparison chart
Product Price/Qty I used Charity Equivalent Spend
Cereal 755g @ $2.67 2.36FAbucks/dollar
Diapers 174 diapers @ $37.99 1 FAbuck/dollar
Rice 8kg @ $8.975. 5.11FAbucks/dollar
Ketchup 1L @ $2 1 FAbuck/dollar
Baby food 100mL @ $1.14 0.11 FAbucks/dollar
Pasta 900g @ $1 3.1 FAbucks/dollar
So despite the fact that diapers are one of the most sought after items by charities, it doesn’t make much sense to donate it from an individual perspective, no better than condiments which are much further down the list, and far worse than the grand champion, rice. What’s more interesting, is that these shadow price data could feasibly be used to inform the public what the optimal donation to food banks is on an ongoing basis (if I had a sufficiently large readership, the shadow price of rice might fall).
Of course the best thing you can donate is cash or time... but that's a post for another time.