[lordnacho]

> I'd disagree, you must have come across the wrong models. Theoretical economics has many beautiful models precisely laid out. For example, the Arrow-Debreu equilibrium model [1] utilises "the Kakutani fixed-point theorem on the fixed points of a continuous function from a compact, convex set into itself." I doubt that any engineer can find fault with the specification of that model. Similarly, the Heckscher–Ohlin model of international trade is well specified and yields five informative theorems [2].

There's a lot of theorems in economics that I basically think of as math theorems. Arrow's Impossibility Theorem for instance. Various things in game theory as well, just about all of Tirole's book (IO? Can't remember).

But I think of them as math, with the very particular term "theorem" precisely because they are defined like math problems, with very specific assumptions.

They are really math theorems that are dressed in economics words like "demand function" in the same way that you can have a theorem in physics, eg the Bubble Theorem about what angles arise. Or that theorem that says you can't balance a magnet statically.

> To which extent those models reflect, or can be applied to, the real world is an entirely different question of course.

That's what engineers tend to care about though.

Now about averages, I think the point really does matter that the person who spends the average basket isn't representative. The guy in the 1% just doesn't care much at all about his beans and rice getting expensive. The family in the bottom 1% may well end up not eating for a day. Sweeping everything into one number causes some real problems with our decision making.