Whenever you construct toy models, you reach back to the first- and second-year textbooks. Because that's where you learned to build toy models. So if you teach people nonsensical toy models in the first- and second-year textbooks, they will come away with a nonsensical conventional wisdom.
You can teach more sophisticated models later to your heart's content, but the qualitative, intuitive understanding - their bullshit detector, if you will - will still be based on their conventional wisdom.
- Jake If you only spend 20 minutes of the rest of your life on economics, go spend them here.
Those advanced models "replace" the simple first-year models, "in the same sense that quantum theory replaced classical mechanics." And since the undergraduate-level stuff has been supplanted, it doesn't need to be revisited. After all, you don't revisit the classical physics textbooks much either.
The difference that this physics envy conceals is that the relative importance of the simplifications that go into physics models does not change over time. The three assumptions that were most constraining on a model of the motion of the gyroscope two hundred years ago (air resistance, contact point friction and rigidity) are still the three assumptions that are most constraining today.
The same cannot, alas, be said for economic models. So you have a toy model that stipulates liquid, competitive markets as the three most important assumptions, and doesn't even bother to mention capital immobility. Which was not a bad way to prioritise the assumptions two hundred years ago. Unfortunately, today competitive markets is an afterthought compared to the enormity of assuming that capital is immobile. But toy-model trade theory is Settled Science, so there's no need to revisit those models. Right?
