Dietary Shapes

A new essay and some new stories about diet optimization.

Last week in Zócalo Public Square, I wrote a piece on optimizing diets, adapted from the second chapter of The Irrational Decision.¹ It’s one of my favorite stories in the book. Though it predates computers by several years, it’s a microcosm of the computer age and fits in seamlessly with today’s oddly dominant online culture of wellness optimizers. In a spat with USDA nutritionist Hazel Stiebeling about what sorts of recommendations are acceptable for the government to publish, the prickly economist George Stigler solved a complicated tableau by hand to find a rather unpalatable “minimum cost subsistence diet” of wheat flour and navy beans. Go read the essay, and then come back here and read a few fun addenda.

I tell the story of the diet problem in most of my book talks, and I always receive fun feedback. Jeff Linderoth sent me a hilarious reflection by George Dantzig, the inventor of linear programming, on his apparently futile attempt to find his own optimal diet to lose weight. Though Dantzig knew Stigler’s optimization problem had nothing but absurd and disgusting solutions, he figured he was adept enough at building linear programming models to patch Stigler’s simplistic assumptions with appropriate shaping of objectives and constraints. He used spare cycles of the IBM 701 at the RAND Corporation to churn out ever more innovative meal plans. But he kept getting bizarre recommendations, like drinking gallons of vinegar or consuming mass quantities of bouillon. He diligently refined the diet over the course of a week before his wife, Anne, got fed up:

Speaking firmly so that I would know who was boss, she said, “I have been studying the various menus the computer has been generating. There are some good ideas there that I can use. I’ll put you on MY diet. She did and I lost 22 pounds.

Steve Stigler—not only an amazing statistician and historian but also George’s son—attended my talk at the University of Chicago. Steve told me how this paper made its way out of academia and into national newspapers, ruffling feathers from coast to coast. George would receive angry letters scolding him about how “this is no way to feed growing boys.” I thought it was pretty clear from reading the original paper that George didn’t think anyone should try to eat his diet. He was trying to prove a point about the impossibility of optimal diets and the paternalistic nature of government recommendations. But people ended up taking him literally. Papers that start as sardonic jokes can surprisingly take on a life of their own.

Stigler’s paper is part of a broader conversation about the scope of government policy. The idea of a computable government was central to economic discussions during the Great Depression and throughout the Second World War. Experts and government officials argued about what is optimal, what can be planned centrally, what individuals should be allowed to navigate for themselves, and what sort of information is beneficial and which is coercive. These debates strongly influenced von Neumann, as you can see in his and Morgenstern’s engagement with contemporary economic debate in the introduction to their revolutionary book on game theory. For another fun example of the people building computers closely interacting with the people designing policy, here’s a 1958 photograph of a sharply dressed Claude Shannon at the Center for Advanced Study in the Behavioral Sciences at Stanford, taken by George Stigler on Steve Stigler’s camera.²

The diet debate also raises the uncomfortable central theme in Elizabeth Popp Berman’s book, Thinking Like an Economist. Everyone across the political spectrum is arguing about efficiency, as if that’s the only thing the government should think about. Left-wing technocrats (aka the Democrats) apply this sort of economic thinking to the utility of the population. Right-wing policymakers (aka Republicans) apply economic thinking to the utility of the individuals in that population. No matter their politics, everyone is thinking like an economist. The valence of the arguments and the parties making those arguments remain uncannily similar today.

Given the grand scale and ambition of the federal government, USDA dietary guidelines should be a fourth-order concern. But there’s something about worrying about what we should eat that galvanizes the popular imagination. It’s fun to walk through the original arguments about what should be in the food pyramid, especially given the weird steak-centric geometry being pushed by RFK’s cuckoo version of HHS. A steak every day sure sounds more appealing than a bean pie.

1. If you haven’t grabbed your copy yet, you should! The book tells a fun history of how we computerized everything and remains a solid snapshot of the argmin mindset. Rob Nelson tells me that I should periodically remind people that it’s out and you can buy it.

2. Sent to me in an email from Steve!