Questions for Stock-to-flow Analysis
The Bitcoin halving is scheduled to occur in just a matter of days, on May 11th-12th, depending on respective time zones. In anticipation, many in the cryptoasset ecosystem have been focusing on the Stock-to-Flow (S2F) supply-focused valuation model for guidance as to Bitcoin’s future price action. S2F, in its application to Bitcoin, is the product of a pseudonymous European institutional investor, called ‘PlanB’ and attempts to project Bitcoin’s price based on its supply and issuance schedule.
Similar models have been used by some in the precious metals sector for several years before PlanB’s publication in March 2019 and the concept is based on the thesis that Bitcoin’s scarcity, a product of its finite total supply and resistance to counterfeit (both of which are supported by its decentralized governance and energy production costs) is the primary determinant of its market price. PlanB highlights the fact that Bitcoin currently has a relatively high Stock-to-Flow ratio, in that its annual issuance or (flow) is low relative to its circulating supply (stock) and that this ratio continues to increase in a predetermined manner with major accentuations at halving events, whereby block rewards are cut by 50%. By applying statistical modeling with the use of linear regression, he finds that there is a “statistically significant relationship” between Bitcoin’s S2F value and its price with a 95% R2 predictive value between the two variables, stating that there is close to a 0% chance of the relationship being purely coincidental.
Beyond this, PlanB points to the plotting of Silver and Gold, both precious metals arguably with similar monetary properties, on the same line as Bitcoin as further validation. In repudiating claims that S2F is invalid due to market efficiency pricing-in Bitcoin’s issuance schedule, PlanB explains that he believes the market is collectively over-estimating the risk of certain factors such as government censorship, software bugs, exchange hacks, 51% attacks, and hard forks and as such with the use of risk and return modeling, it is still possible to effectively project Bitcoin price with S2F. Since his initial article in March 2019, the author has published two auxiliary pieces, with his latest article (April 2020), projecting a Bitcoin price of $288,000 between 2020-2024.
Although PlanB’s application of S2F to Bitcoin has received praise and in some cases, validation from several other quantitative researchers, such as Nick Phraudsta, Marcel Burger, and Manuel Andersch, others. such as Orchid Research, have offered robust critiques of his methodology. While PlanB has undoubtedly made notable contributions to cryptoasset valuation and has arguably spawned an entirely new field of Bitcoin econometrics, there remains, in Smith + Crown’s judgment, several questions regarding the model’s implicit assumptions and underlying valuation approach that a more robust analysis would address. Readers should keep these questions in mind before accepting the S2F model’s ultimate conclusions.
Beyond questions of the statistical modeling’s validity, there are several theoretical concerns with the thesis and its conclusions. Most importantly, the sole focus on supply and issuance dynamics as the determinant of an asset’s price is controversial, to say the least. While an increasing S2F ratio may indeed increase Bitcoin’s attractiveness as a provably scarce monetary asset in a time of considerable expansions to global fiat currency supplies, this should be measured by way of market demand, rather than assumed. For instance, potential demand could be quantified by the aggregate demand for stores of value (SoV), with different projections as to what percentage of such demand may flow into Bitcoin and thereby determining annual SoV demand.
Beyond this, readers should be aware of the implicit assumptions made by the author. Of these, the most consequential is the notion that Bitcoin will be the sole or dominant cryptoasset to benefit from a high S2F ratio. In his first article, PlanB qualifies that for a cryptoasset to be properly considered for S2F modeling, it must have a supply cap, use Proof of Work (PoW), and a moderate or high hash rate while its monetary design cannot be affected by a small group of entities. While these conditions are in some cases vague, such as what constitutes a sufficient hash rate, other cryptoassets, such as Bitcoin Cash, Bitcoin SV, Litecoin and others, arguably meet these criteria, raising questions as to what extent this competition dilutes Bitcoin’s SoV demand and how much it will benefit from its high S2F ratio. In a world of programmable scarcity, as is the case with cryptoassets, where multiple such assets co-exist, it is reasonable to expect that SoV demand will be spread across a wide range of cryptoassets. Indeed, future research could quantify such an approach by dividing SoV demand by hash rate, adjusting it for different hashing algorithms as well as the upfront and ongoing capital costs for mining hardware.