Reviewing “Modelling Bitcoin’s Value with Scarcity” — Part III: The Fall Of Cointegration
Why S2F and bitcoin price can’t be cointegrated
Introduction
We are a little more than a year further down the road since PlanB wrote his initial piece on the S2F model. Both phraudsta and I have put in quite some work to run model validations and quite recently our findings were criticised by Sebastian Kripfganz (Assistant Professor in Econometrics at the University of Exeter), who pointed out that the deterministic elements in the S2F timeseries should be accounted for.
Now that the halving is behind us, the majority is probably less interested in these model validations. Most likely I’m writing this piece as a rectification of my earlier work and to bring some clarity to myself. Clarity w.r.t. the question whether or not it would be possible that the two main variables within PlanB’s model are cointegrated or not. Why did cointegration matter? If the two variables are cointegrated, the key take away is that there is a long term relation between the two non stationary timeseries.
This piece is actually the missing piece in my series on the stock to flow model. (If you paid close attention, there never was a part III, while I did publish a part IV). In part IV I already wrote what the theoretical framework should look like and how it should lead us in terms of model selection. This piece will mainly readdress the concepts of cointegration and…