I just skimmed @ole_b_peters’ Ergodicity Economics textbook, and it already feels like a much clearer way of understanding economic rationality. The core idea: model people as maximizing the average growth rate of their wealth over time, rather than a utility function of wealth.

Their headline example: a game where your wealth increases by 50% if you flip heads and decreases by 40% if you flip tails. Each coin flip increases your expected wealth. But if you keep flipping your wealth will almost certainly approach 0. (Intuitively, W * 0.6 * 1.5 = 0.9 W.)

How can both be possible? When growth is multiplicative (as it typically is in the long run), optimizing for the “ensemble average” concentrates wealth in very few possible worlds, with the others going to 0. (Related: the St Petersburg paradox, which the book also discusses.)

That is: the utility function that leads to the most consistent growth is different in different environments. Ergodicity economics then says: so let’s work with the (empirically validated) assumption that people are trying to get consistent growth not utility. Very elegant!

Having said that I’m not an economist and don’t have a great sense of how confused the field of economics previously was about these ideas. If you have a strong econ background and are interested in reviewing the book (even briefly) DM me your address and I’ll order you a copy.

I’ll also reiterate that I’ve only skimmed the book, so any mistakes in the exposition are mine. Lastly, Scott Garrabrant wrote a sequence on related ideas under the heading “geometric rationality”: I’m curious what ergodicity economists think of that!