Looking at the graph, 75% isn’t a great place to invest, historically, except at the start of the late bubble, RIP. The great place to invest looks more like 40%-50%. From a valuation point of view, the “right” number should be easier to find than that of any component asset. The inputs and outputs are much more predictable and stable than in a given firm: revenue (GNP), growth (1.8% per year per person), expenses (salaries, government, etc.) and risk (treasury rates.) The only harder thing to predict is the cost of commodities, although we can use futures prices for this to some extent. Run all this through a macro-DCF.
[The other variable that may move the graph is the amount of the domestic economy reflected in the stock market cap number. Not every company, or even every big company, is public. Much of the agricultural and professional services sectors are private, for instance. I tentatively extend as a hypothesis that the IPO boom of the late 90s is one reason for the run-up in the graph in that period, but I haven’t looked at a shred of data to back that up, so don’t quote me.]
It would be nice to have this model. It would not only be medium-term predictive, but it would also quantify the degree of uncertainty and/or irrationality in the markets It would also be a nice test of the model itself. Any aspiring academics out there?