Ignoring Bad Real Estate Analysis for Fun and Profit: Postscript - Micro vs Macro

In Part 1 of my recent 3 part series on data measures you should mostly ignore, I wrote:

As long as the orthodoxy encourages this non-empirical intuition, it seems like there will be persistent profits from trading against it. Even the builders follow that intuition. And, it’s understandable, because quick changes in mortgage rates certainly affect trading activity among existing homes, foot traffic in sales offices, and some mortgage activity. And, it seems canonical that surely mortgage rates affect sales activity. And, it surely affects the distribution and shape of sales activity. But, gosh darn it, there just isn’t an important or reliable historical relationship between mortgage rates and sales.

Partly what makes reality deceptive is that there are probably various indirect ways in which higher mortgage rates are associated with more activity, which push against the direct and visible ways that they are associated with less activity.

One way I put it is that, for a developer, interest rates are axiomatically causal. They are a cell in a spreadsheet that is input by the user. And, when that rate rises, it hurts the viability of the project. What could be more obvious?

But, what changed the interest rate? Maybe the interest rate changed because various changing conditions made 3 other project viable, even at the higher rate. Yet, if you ask those 3 developers if higher interest rates lower construction activity, they will say, “Of course. Look at my spreadsheet.” They can see how interest rates affect them, but their effect on interest rates is communal and unseen.

At some point, the numbers say what the numbers say. Hardly anyone is willing to accept that, so profits are available. Mortgage rate causality is treated axiomatically rather than empirically.

I have an example, which I will share below the paywall.