You seem to compare different models of supply in demand, each based on different assumptions, and each expected to work on different time scales, and declare one of them "incorrect". Presumably what you mean by a more "correct" model is that it explains existing data and predicts future events better than the alternatives. If so, one should take your outline, make it into a proper simulation, plug in the available data, tweak the free parameters to match, then check if its predictions match the data not used to calibrate the model. Only after that you will have an argument for correctness. And only then, say, make a recommendation to the EA community re ways to decrease farm animal production.

To be fair, this bias towards theorizing instead of modeling and testing is a common pitfall of this community. I find it pretty disappointing, but maybe it's just me.