The Regression Theorem Explains Bitcoin

1. A peculiar list

For a decade and a half, economists of my school declared Bitcoin impossible on the authority of Mises's regression theorem. The possibilities that it becomes money were pronounced null; its monetisation was a myth, a Ponzi scheme, an inflationist fraud (Shostak 2013; North 2013; LeRoux 2014). Several of these authors I know personally; I watched most of the verdicts take shape. By now there even exists a list of them (Kinsella 2025). This paper is not one more entry in that quarrel. I would rather tell how my knowledge of Austrian economics helped me explain, for many years, why Bitcoin would probably fail, and what I as an Austrian economist may have learned when it did not fail.

A confession first. I speak at about a dozen Bitcoin conferences every year about Austrian economics; the lecture behind this paper was the first I ever gave about Bitcoin at a conference of Austrian economists, and I was a little ashamed of the delay. The tables have turned in a way worth recording. A while ago one would have assumed that Bitcoin is a niche interest within Austrian economics, since we Austrians were of course open to private forms of money. By now Austrian economics survives rather as a niche interest within Bitcoin: most of the contemporary interest in the school, most of the young people eager to learn about it, comes from Bitcoiners. Something profound must have happened, and a school whose loudest representatives proved the phenomenon impossible owes its new students an account of what its central theorem actually says.

I discovered Bitcoin at the end of 2009, in a forum called anti-state.com, where someone was shilling the idea. Two things became apparent at once: it looked like a ledger-based form of money, and it looked like something libertarians would adopt, and both made me sceptical. A ledger is the mark of token money, which more or less necessitates the very trust that money should absolve us from; I had used e-gold, and I learned where the trust in a ledger ultimately resides when my balance ended up in the hands of an American agency. The libertarian token seemed confined to a niche besides: maybe five percent of libertarians take an interest in digital money, libertarians are perhaps five percent of the people who care about liberty, and those are five percent of the population, if one is optimistic. Nice to have, certainly nothing of economic significance. Still, in mid-2011, after the first bubble had burst, I told my readers that it could not hurt to acquire a few bitcoins, for the improbable case that they survive; an investment this was not, rather pure gambling (Taghizadegan 2011b). For years the scepticism remained reasonable. What may an economist learn when a thing his theory seemed to forbid refuses, cycle after cycle, to die?

2. What Mises claimed

The theorem was Mises's answer, in 1912, to a circle that embarrassed the subjective theory of value: money is demanded because it has purchasing power, and it has purchasing power because it is demanded. His way out ran through time. Why does the money we use today have value? We can only explain it by looking back. To explain the euro I must explain what gave the schilling its value; behind the schilling stands the krone, behind the krone the gulden and its silver, until the explanation reaches a good that people valued for its own sake, before any monetary employment (Mises 1912).

In my own textbook, the theorem appears in its strictest form: thought through to the end, it means that at the beginning of the discovery process of money the specific commodity must have been in industrial demand; money is therefore no invention of the state, it is discovered by market participants in the market process (Taghizadegan 2018, my translation). It is my own glossary, so the correction can begin at home. Two claims of very different rank sit in that entry. One is praxeological. Nobody can appraise a unit of anything as a medium of exchange without an anchor in exchange ratios he has already observed; the valuation of a medium is learned, and learned from remembered prices. The other claim says the chain of anchors must end in industrial use. That one is a historical conjecture about how the learning once began, plausible enough for cattle and silver. Mises wrote in the idiom of his time.

Bitcoin has not refuted the regression theorem. But the theorem must be understood as an economic explanation and no historical law: it asserts nothing about how things must unfold in every historical case; it explains the valuations of acting people. I have defended this reading since 2014, when a seminar participant first put to me the question whether Bitcoin had refuted Mises, and the debate among Austrians still counted as open. Only the praxeological claim carries apodictic force. Mises never wrote a licensing requirement for candidate monies. The prohibition was added by his readers.

A quieter accident deepened the confusion, and it was philological. Mises wrote Zeichengeld, literally sign money or token money. Batson's translation of 1934 made fiat money out of it. Words carry their associations along: Sachgeld sounds like physical stuff, fiat sounds like decree and moral verdict, use value sounds like objective usefulness. A subjectivist school should have distrusted that last sound most of all (Taghizadegan 2021). Instead a whole generation approached anything digital with tilted categories before any analysis had begun.

3. Two camps, one misreading

The sceptics deserve a fair reconstruction; their instinct was sound. Shostak (2013) insisted that a thing out of nothing cannot acquire the standing of money. North (2013), who had seen many manias, smelled the biggest Ponzi scheme in history. For Huerta de Soto the chances were simply null, as his students kept hearing (Kinsella 2025). LeRoux (2014) gave the orthodoxy its sharpest published form. Korda (2013) called the second act of the bubble; Dowd and Hutchinson (2015) trusted mining centralisation and competition to finish the job. Behind all this stands a caution that is genuinely Austrian. Monetary institutions grow, they are not decreed. Mises himself doubted that a pure token money had ever existed unmixed.

The defenders assembled early and argued better. Satoshi Nakamoto (2010) himself noted that a scarce thing with no initial value can bootstrap if people foresee its potential usefulness. Šurda (2012) placed Bitcoin carefully within Mises's own typology. Graf (2013) reconstructed the Mengerian stages of the early years, stage by stage. Murphy (2013) warned his fellow Misesians that the theorem explains the past of purchasing power and does not legislate its future. Davidson and Block (2015) went further: the theorem concerns the emergence of exchange media from barter, and it is simply silent about a good born inside a mature money economy, where dollar prices frame every valuation. St. Onge (2014) widened the admissible starting points to monetary properties themselves. Ammous (2018) carried the Mengerian saleability tradition to the widest audience the school has reached in a century. Then came the empirical question. Luther and Salter (2017) measured demand responding to bailout fear; Luther (2019) read the early mailing lists as a documented coordination process; Hazlett and Luther (2020) concluded that, by the ordinary definition, Bitcoin is money within a modest but real domain. Pickering (2019) replied that none of this embarrasses the theorem, whose point was always subjectivist explanation, never candidate selection. Selgin (2015) cut the knot taxonomically: synthetic commodity money, scarce like a commodity, useless like fiat.

My own position, held since the debates of the mid-2010s, sits in the second camp with one reservation. One cannot simply wave the thing away and say: Bitcoin is no money, because it violates the regression theorem. I find that ideological. But the reverse move is ideological too: it is money, so it must fulfil the theorem, and now let me construct the reasons why. Both treat a piece of explanatory theory as a party manifesto. And one must beware that the defence does not become a cost-theoretical argument. I have heard it often, also from Austrians, and it is completely wrong: because mining costs electricity, the thing is valuable; because it cost electricity from the beginning, it was valuable from the beginning. That is exactly the opposite of the insights of Menger and Mises. Costs are incurred because the thing is valued; the miners' electricity bill is a consequence of the money premium and of the security budget it finances. A school that spent a century correcting this inversion in the labour theory of value should recognise it in a data centre.

4. What stands at the end of the regression

If the theorem's praxeological core is the learning chain and its historical terminus an open conjecture, the natural question is what actually stands at the end of real regressions. Bernhard Laum, to whom my colleague Gregor Hochreiter once drew my attention, caused a brief stir with the thesis that most money goods have a mythical origin; that went somewhat against the rational spirit of the Viennese school, and its economists paid him little attention. Laum had worked through the earliest Greek money. At the end of the regression he kept finding less subjective use value than objective premises, objective here meaning extra-subjective, mythological ones. Cattle became the measure of value because cattle was what one owed the gods. The obolos was the iron spit of sacrificial meat (Laum 1924; Taghizadegan 2011a). Menger, to judge by his library, engaged remote tribes more thoroughly than European mythology. One need not follow Laum everywhere to accept the lesson. At the beginning of monetary regressions history shows us cult, status, play and ornament at least as often as industry; the strict industrial-demand clause was always the idealisation of a messier record, twelve years younger than Mises's treatise.

The regression path can also begin with a thing regarded as a collector's piece and lead to a thing regarded as an asset. That has happened with many goods; it happened with watches, it happened with artworks. Bitcoin's documented beginning looks exactly like this: for its first year and a half a provably scarce digital token circulated among cryptographers as a demonstration piece and a toy, valued by a handful of people whose preference structures we need not share to respect (Graf 2013; Luther 2019). It does not seem entirely alien to the human mind to attach some value to something provably scarce, with a fun history behind it and fun people around it. The first published exchange rates and the famous pizzas bought for ten thousand units gave it a price history, and the learning chain the theorem describes had its anchor.

The first time I understood that Bitcoin had a nonzero value was on Silk Road, which I studied, of course, for purely scientific reasons. There was something rare for monetary theory: an empirical comparison. One could compare the street prices of certain goods in Vienna with their darknet prices and observe the differential. A higher saleability of the digital good for a certain context of people, more exchange partners, better exchange relations; so it was rational to demand some bitcoin, though not yet to hold it. Years later, Chinese entrepreneurs told me how they used Bitcoin to move past capital controls, and again, without any value judgment: there is a demand, there is a use case, a higher saleability for a certain context of people (Taghizadegan 2019). This is Absatzfähigkeit in Menger's exact sense (Menger 1892), and it has an underrated dimension. The most crucial one, historically, is withdrawability: are you able to withdraw your money, the storage of what you worked for, from bad structures and from plundering? Economic historians tell of a great bullion famine in late medieval Europe and blame a shortage of money. The time was one of increased plundering, and of the plagues that follow it; it was no time to have one's money in the open. Money goes into hiding when it must, and it moves where it may. So the mobile metals carried European capital from falling Rome to Venice and Genoa, on to the Hanseatic towns, to Flanders, to Amsterdam, and along the route one keeps meeting the same families. Diamonds are clunky things, useless for daily transactions. In Amsterdam they became money-like all the same, for one context above all: one could sew them into a garment and pass borders carrying huge amounts of wealth, compared even to gold. A brain wallet of twelve remembered words beats the diamond on this margin by orders of magnitude. No large part of the adoption, for now. But it is the part that matters. The leaving minority has repeatedly founded the next centre of civilisation.

5. Commodity money without the commodity

The second refinement concerns the Sachgeld category itself. What is the economic essence of commodity money? A subjectivist cannot answer: its matter. The essence lies in the rising marginal costs of money production, which, unlike the costs of ordinary trade goods, do not fall with rising demand through economies of scale (Taghizadegan 2021). Hence the stock of the classical money commodities grows slowly and predictably; no producer can flood the holders. Bitcoin does something entirely new with this old property. A token money that algorithmically synthesises a commodity money had never existed; proof of work and the difficulty adjustment rebuild, in code, the supply behaviour that once made gold serviceable (Taghizadegan 2019). Meeting such a thing with scepticism was thoroughly reasonable, and nearly everyone did. Our own investor handbook of 2014 still filed Bitcoin as predominantly a token money, tension unresolved (Taghizadegan, Stöferle and Valek 2014). Selgin (2015) reached his synthetic commodity money from the opposite direction, extending the taxonomy of base money regimes while I was reinterpreting Mises's categories from within. Two routes, one object; that speaks for the category.

On this reading the regression theorem and the Sachgeld category, the two pieces of Austrian doctrine most often cited against Bitcoin, turn out to be the two most useful instruments for understanding it.

6. The theorem as an explanation of monetisation

The key lies in further theoretical subtleties. One is Hayek's insight that money should be an adjective rather than a noun (Hayek 1976): monetisation is a partial, evolutionary, uncertain process, no binary result. The textbook definition, a generally accepted medium of exchange, can only describe the end of such a process; demanded of a beginner, it defines all monetary emergence out of existence. One may deny competence and responsibility to corporations that shift their cash reserves from short-term government bonds into Bitcoin; one cannot deny the increased moneyness of Bitcoin that such acts constitute (Taghizadegan 2021).

And every monetisation means growing willingness to pay for the money premium. From history and theory we know that the use value of a good can become vanishingly small against its money premium; gold never had to apologise for that. From the outside a monetisation and a mania look alike. The difference shows only in the pattern over time. Here, unexpectedly, the regression theorem starts earning its keep. We remember the prices Bitcoin had in past cycles, and the memory changes the whole picture; at one hundred thousand dollars, Bitcoiners today lament a bear market. Corrections stop at levels once celebrated as summits. I know no better explanation for this pattern of ever higher lows than Mises's theorem itself: the valuation of a medium is learned from remembered prices, so a monetising asset carries its history with it. Rather than contradicting the story of Bitcoin, the theorem helps explain it. Beanie Babies and tulips each managed a single cycle; a monetising asset has a memory. Mises explained more than he knew.

From Austrian economics I also knew, all along, why Bitcoin might not fail. In the 1990s the biggest power on the planet had made clear that there was to be no civilian use of cryptography, and every reasonable person, myself included, would have considered something like Bitcoin impossible from the start. But software is just information, and information is hard to control. It passes from peer to peer. The early core version of Bitcoin would have fit into a book, and no border guard reads ciphers in every book. So the episode later called the crypto wars ended, unusually, with a win for the side of liberty. China banned mining in 2021; the majority of the hash rate went dark; the rigs went into containers, to Kazakhstan and elsewhere, and the hash rate stands higher today than before the ban. Unbanking waves followed. The hostility of most treasuries stands to this day. Böhm-Bawerk (1914) posed power against economic law and observed that power does not always win. Here it has not won yet.

7. What an economist may learn

Austrian economics does not recommend Bitcoin, and praxeology recommends nothing. I am not describing an investment; I am describing what hour of the adoption we have reached. Theory offers no shortcut past reality. It is the instrument that lets us make out shapes in the fog of uncertainty, and the regression theorem is such an instrument. It made the strangest monetary phenomenon of our time intelligible to me, slowly, over fifteen years, against my own expectations. Wielded as a prophecy instead, it embarrassed a school that of all schools should have known the difference between explaining valuations and forecasting them. The monetisation remains partial, reversible, hostage to real history. We observe it with humility. Where reality refuses our reading, the fault has so far never lain with reality.

References

• Ammous, S. (2018). The Bitcoin Standard: The Decentralized Alternative to Central Banking. Hoboken: Wiley.
• Böhm-Bawerk, E. von (1914). "Macht oder ökonomisches Gesetz?" Zeitschrift für Volkswirtschaft, Sozialpolitik und Verwaltung 23: 205–271.
• Davidson, L., and W. E. Block (2015). "Bitcoin, the Regression Theorem, and the Emergence of a New Medium of Exchange." Quarterly Journal of Austrian Economics 18 (3): 311–338.
• Dowd, K., and M. Hutchinson (2015). "Bitcoin Will Bite the Dust." Cato Journal 35 (2): 357–382.
• Graf, K. S. (2013). On the Origins of Bitcoin: Stages of Monetary Evolution. konradsgraf.com.
• Hayek, F. A. (1976). Denationalisation of Money. London: Institute of Economic Affairs.
• Hazlett, P. K., and W. J. Luther (2020). "Is Bitcoin Money? And What That Means." Quarterly Review of Economics and Finance 77: 144–149.
• Kinsella, S. (2025). "Austrians Who Say Bitcoin is Impossible Because of the Regression Theorem." stephankinsella.com, July 8, 2025.
• Korda, P. (2013). "Bitcoin: Money of the Future or Old-Fashioned Bubble?" Mises Daily, April 2013.
• Laum, B. (1924). Heiliges Geld: Eine historische Untersuchung über den sakralen Ursprung des Geldes. Tübingen: Mohr.
• LeRoux, C. (2014). "Why Bitcoin Can Never Be Money." Procesos de Mercado 11 (1): 249–271.
• Luther, W. J. (2019). "Getting Off the Ground: The Case of Bitcoin." Journal of Institutional Economics 15 (2): 189–205.
• Luther, W. J., and A. W. Salter (2017). "Bitcoin and the Bailout." Quarterly Review of Economics and Finance 66: 50–56.
• Menger, C. (1892). "On the Origin of Money." Economic Journal 2 (6): 239–255.
• Mises, L. von (1912). Theorie des Geldes und der Umlaufsmittel. Munich and Leipzig: Duncker & Humblot. English: The Theory of Money and Credit, trans. H. E. Batson, London: Jonathan Cape, 1934.
• Murphy, R. P. (2013). "Why Misesians Need to Tread Cautiously When Disparaging Bitcoin." consultingbyrpm.com, October 2013.
• Nakamoto, S. (2010). Reply in the thread "Bitcoin does NOT violate Mises' Regression Theorem." bitcointalk.org, August 27, 2010.
• North, G. (2013). "Bitcoins: The Second Biggest Ponzi Scheme in History." garynorth.com, November 2013.
• Pickering, G. D. R. (2019). "The Relevance of Bitcoin to the Regression Theorem: A Reply to Luther." Quarterly Journal of Austrian Economics 22 (4): 603–619.
• Selgin, G. (2015). "Synthetic Commodity Money." Journal of Financial Stability 17: 92–99.
• Shostak, F. (2013). "The Bitcoin Money Myth." Mises Daily, April 17, 2013.
• St. Onge, P. (2014). "Cryptocurrencies and a Wider Regression Theorem." Mises Daily, December 11, 2014.
• Šurda, P. (2012). Economics of Bitcoin: Is Bitcoin an Alternative to Fiat Currencies and Gold? Master's thesis, WU Vienna.
• Taghizadegan, R. (2011a). "Scholien 02/2011." Vienna: Institut für Wertewirtschaft.
• Taghizadegan, R. (2011b). "Scholien 03/2011." Vienna: Institut für Wertewirtschaft.
• Taghizadegan, R. (2018). Wirtschaft wirklich verstehen: Einführung in die Österreichische Schule der Ökonomie. Munich: FinanzBuch Verlag.
• Taghizadegan, R. (2019). Geld her oder es kracht! Graz: Leykam.
• Taghizadegan, R. (2021). "Die Bitcoin-Blase." scholarium.at, March 21, 2021. First published in eigentümlich frei.
• Taghizadegan, R., R. Stöferle and M. Valek (2014). Österreichische Schule für Anleger: Austrian Investing zwischen Inflation und Deflation. Munich: FinanzBuch Verlag.