The article to you link to is much more clear headed about the role of the model. It's a useful tool, not a glass ball.
There's a hell of a huge difference between a model predicting something, and people applying/adopting the model -- together with a LOT of other tools and insights -- to predict something.
It's as stark as the difference between executing a program and getting an exact output, and using a programming language, together with some bits of math, to write a program which computes a result for you.
Back to the point: since lots and lots others made similar predictions about the housing market (and beyond), I'm much more inclined to believe that it was a strong focus on data and observation -- not any particular model per say -- that allowed smart people to accurate predictions.