I'm doing some research on real estate valuation and came across Aswath Damodaran, professor at NYU Stern School of Business. He is a prolific author of books and youtube videos on valuation. Today I checked out his entertaining and informative "Introduction to Valuation":

He mentions that there are three, and only three, approaches to valuation: Intrinsic, Relative, and Options Pricing valuation.

The idea that I am toying with is that there might be another one called Causal Valuation.

Causal valuation is an approach that tries to take advantage of some recent mathematical and conceptual work on causal graphs by Judea Pearl. See his Engines of Evidence conversation for a taste. That research is starting to emerge into the mainstream but it is not trivial to master the tools or the ideas associated with causal graph theory.

To give you a flavor of what causal valuation might look like I'll show you what a simple causal graph looks like for a non-controversial case. Below is a causal graph depicting the relationship between smoking cigarettes, lung cancer, and yellow fingers.

We might contrast this causal graph with an associative graph between the Cigarette Smoking and Yellow Finger factors and the Lung Cancer outcome.

Note that there probably is an association between yellow fingers and lung cancer but it is not a causal one, it is an associative relationship.

Most of our statistical models are associative, not causal. They might tell you to clean your fingers more often if you want to avoid lung cancer. A causal model, on the other hand, explicitly tells you there is no causal relationship between yellow fingers and lung cancer because there is no connecting line between them. Causal models also provide more guidance on where you need to intervene in order to affect an outcome.

Real estate valuation often uses the "Relative" method of valuation, or what similar homes are selling for in the neighborhood. There may, however, be factors such as net migration, wage levels, and business growth or decline in the area that are actually driving real estate valuation levels. Instead of using associative metrics to arrive at a valuation, perhaps a causal approach would lead to better valuations. That is the big idea that I have started to explore.

If you want to play around with making causal graphs, a tool I would recommend is the free and powerful online drawing tool draw.io. I used it to make the graphs above.