Nassim Taleb is one of the most influential authors of our time, so I think it’s important to take a close look at his recently published and purportedly damning “Bitcoin Blackpaper.”
Let’s start with his two main conclusions: Bitcoin is neither a sound currency (it’s worthless) nor a great technology. As the second conclusion is based on arguments about the first, let us first consider what constitutes a sound currency.
Economists generally agree that a currency should serve three functions. It should work as a “store of value,” meaning that it more or less retains its purchasing power into the future. It should work as a “unit of account,” or a reasonable measure of value to allow for meaningful interpretation of costs, prices, and profits. And it should work as a “medium of exchange,” meaning that it can be readily swapped for goods and services.
Taleb uses quantitative finance methods and economic arguments to argue that Bitcoin fails at all three of these functions.
To provide a summary, I list what I see as his line of argument below (my words in bold, Taleb’s italicized). I organize these arguments into two groups — primary arguments that are the key assumptions of his argument that Bitcoin is worth zero and secondary arguments that are unnecessary for that conclusion but support that Bitcoin doesn’t work well as a medium of exchange and unit of account.
- Bitcoin fails at being a store of value because of absorbing barriers including miner extinction, technological obsolescence, and user rejection (“…if we expect that, at any point in the future, the value will be zero when miners are extinct, the technology becomes obsolete, future generations get into other such ‘assets’ and Bitcoin loses its appeal to them, then the value must be zero now.”)
- Bitcoin also fails at being a store of value because it pays zero dividends (“If any non-dividend yielding asset has the tiniest probability of hitting an absorbing barrier, then its present value must be 0.”)
- Bitcoin fails at being a medium of exchange because of volatility in purchasing power (“To be able to regularly buy goods denominated in Bitcoin … it requires a parity Bitcoin-USD of low enough volatility to be tolerable and for variations to remain inconsequential.”)
- Bitcoin fails at being a medium of exchange because transactions are expensive and slow (“Transactions in Bitcoin are considerably more expensive than wire transfers or other modes, or ones in other cryptocurrencies, and order of magnitudes slower than standard commercial systems used by credit card companies.”)
- Bitcoin fails at being a unit of account as almost nothing is priced in it (“A true numeraire must be one of minimum variance with respect to an arbitrary basket of goods and services.”)
- Bitcoin is a poor inflation hedge because it is a failed unit of account (“… we can look at the inflation hedge as the analog of a minimum variance numeraire.”).
In my response, I focus on the primary arguments as I see them as more potentially damning to Bitcoin’s future than the others. The secondary arguments are diminished by the fact that Bitcoin is an asset that is rapidly capitalizing so volatility should be expected and early in its adoption life cycle so we shouldn’t expect much to be priced in it yet.
Taleb’s supports his claim that Bitcoin is worth zero with a mathematical proof. But, that proof rests on two key assumptions: (a) absorbing barriers and (b) zero dividends. These assumptions should be scrutinized because if either are invalid, so are any inferences based on them.
I’ll start with “zero dividends” because it’s a more straightforward argument with fewer degrees of freedom.
My immediate response was, what about Bitcoin lending? You can earn interest by lending Bitcoin for margin trading as many trading pairs are denominated in Bitcoin. You can also earn interest by lending Bitcoin in DeFi applications that also benefit from liquidity providers.
My second response has to do with a statement Taleb makes himself when defining non-dividend yielding assets:
“We exclude collectibles from that category as they have an aesthetic utility, as if one were, in a way, renting them for an expense that maps to a dividend — no different from perishable consumer goods.”
In my eyes, Bitcoin was a collectible from its birth until now. When Bitcoin was introduced in 2009, it was worth zero. It had no value beyond being a collectible for cypherpunks and geeks who saw promise in its technology. A year later, it was still worth less than a penny. Now, we have the five-figure Bitcoin, but I don’t see how it has lost its status as a collectible. There is a certain allure of owning one or more of the 21 million Bitcoin that will ever exist. Bitcoin could even be considered art.
But unlike Taleb, I won’t imply that my opinion is fact. You may agree that Bitcoin bears zero dividends, but then you are then assuming that (a) cryptodividends are not dividends and (b) Bitcoin is not a collectible or work of art, thus joining Taleb in the red quadrant below:
If Taleb is making some sort of legal argument about what constitutes “explicit yield,” that would be one thing. But if we’re interested in decision making under risk, which would make more sense, then subjectivity should be taken into account.
To this end, if there’s a non-zero probability of individuals outside of the red quadrant, then the there’s also a non-zero probability that Bitcoin pays dividends.
We should also consider that there’s actually a fifth “I don’t know” group as well. Given Bitcoin is still in its early adoption phase, it’s likely that the majority of people in the world fall into this fifth group. Then, even if we were to put the odds in Taleb’s favor with regard to the other four groups, we could still conclude that the expected dividends of Bitcoin are > 0. Then, Taleb’s thesis that Bitcoin is worthless would be invalidated, even if his argument about absorbing barriers is valid. Let’s turn to that next.
Absorbing Barriers, Ergodicity, Cumulative Ruin
Taleb believes that there are safer ergodic assets (time average = expectation value) like gold and riskier “non-ergodic” assets (time average ≠ expectation value) like Bitcoin. He asserts that Bitcoin is non-ergodic because it will be rendered permanently worthless (ruined) by some extinction event that is certain to happen sometime in the future (absorbing barriers). Gold, on the other hand, doesn’t have absorbing barriers, according to Taleb:
“Gold and other precious metals are largely maintenance-free, do not degrade over a historical horizon, and do not require maintenance to refresh their physical properties over time.”
The argument about absorbing barriers itself rests on ergodicity economics, a nascent field of economics that challenges mainstream economic models but still has many kinks to work out. Let’s discuss that in brief.
The Ergodicity Question
Ergodicity first developed within nineteenth-century physics when Ludwig Boltzmann suggested using ensemble averages (averaged value of many identical systems at a single time) instead of time averages (averaged value of a single system over many times) to model macroscopic features of fluids such as pressure and temperature. Where it is valid, dynamical descriptions can be replaced with simpler probabilistic ones that don’t require modeling time.
To illustrate, imagine flipping a coin 10,000 times. Because a coin toss is ergodic, the distribution of heads and tails would be the more or less the same as if 10,000 coins were flipped at the same time. Therefore, time can be ignored in modeling the time average of coin tosses.
Modern economic theory extends the ergodicity assumption to models of wealth. But, wealth isn’t like a coin toss because it evolves over time — wealth grows. Some argue that this presents a flaw in the foundations of mainstream economics. Ole Peters, a prominent scholar of ergodicity economics, points out in his 2019 Nature Physics paper:
“To make economic decisions, I often want to know how fast my personal fortune grows under different scenarios. This requires determining what happens over time in some model of wealth. But by wrongly assuming ergodicity, wealth is often replaced with its expectation value before growth is computed. Because wealth is not ergodic, nonsensical predictions arise.”
He goes on to criticize mainstream economics as a whole, but perhaps that’s unwarranted. The authors of this rebuttal make several comments we should consider before overextending the reach of where ergodicity economics is useful. For example, on modeling actual human preferences:
“… some ergodic predictions seem to be faithful to preference, others do not. Would a person ever prefer a process that, after three rounds, diminishes wealth from US$10,000 to 0.5 cents over one that yields a 99.9% chance of US$10,000,000 and otherwise US$0?”
And on whether time is always the most central aspect in actual human decision making:
“Although it is true that our consumption of economic goods develops over time, time is not the most central aspect of all our decisions. For many of our decisions, other equally ubiquitous aspects such as risks, strategy, and the balancing of pros and cons are more central.”
So, is it really nonsensical to value Bitcoin above zero when it suits a multifaceted investment strategy or to conclude it’s worth zero based on time alone?
Unlike Taleb, I don’t see how assets can be objectively categorized as having or not having absorbing barriers. All assets are destructible human constructs and nothing has intrinsic value. Objectively speaking, they only have some energetic form. It seems more reasonable to consider a continuum of risk of an asset’s value infrastructure irrevocably collapsing (“absorption”). For gold, the chance of it mutating or degrading under normal conditions is low, but there’s a reasonable chance that, at some point in the future, people lose interest in storing value in it or it is replaced by a superior material in its industrial uses.¹ For Bitcoin, what are the odds of its value infrastructure collapsing? Let’s see what Taleb has to say and then consider some alternative perspectives.
“A central attribute is that Bitcoin depends on the existence of such miners for perpetuity.”
First, there appear to be some issues with this statement on the technical front. Bert Slagter, the founder of Procurius, objects that Taleb confuses the information for the medium:
“It’s not about the book, it’s about the text. Not the CD, but the music. Not the blockchain, but the UTXO [unspent transaction output] set.² Not mining, but the protocol … Bitcoin-the-asset (or UTXO set) does not depend on any specific technical implementation.”
The conflation of information and medium presents a challenge to the ergodicity argument. Slagter points out that risks such as insufficient security budget for miners, hacks, and bugs are not actually absorbing when you consider the adaptability of the blockchain and protocol to support the UTXO (as)set:
“The ecosystem can collectively decide to switch data structure, algorithm, or consensus mechanism. And fully recover.”
So, technically, miner extinction is neither a sufficient nor necessary condition for absorption.
As evidence of such technological adaptability, the second-largest cryptocurrency by market cap, Ethereum, is already in the process of switching its protocol from proof of work to proof of stake, effectively eliminating mining altogether and thus precluding miner extinction while preserving miners’ function: ordering transactions and creating new blocks.
Second, Daniel Brain, a lead engineer at PayPal, suggests that Taleb muddles the cause and effect of Bitcoin mining:
“For Bitcoin to go extinct, the actual cause must be for transactions to tail off. So long as there are transactions and fees, there will be miners; a certainty which is guaranteed by Bitcoin’s block rules and difficulty adjustment mechanism.”
But can Bitcoin’s transactions ever tail off? Brain replies:
“Only if the network stalls in gaining adoption. This is a given; nobody argues that Bitcoin will be a success with or without adoption.”
It’s hard to make a case that adoption will fail anytime soon in light of recent developments like the explosive rise of DeFi, multibillion-dollar fintech companies including PayPal and Robinhood opening Bitcoin exchange and custodial services, and waves of institutional adoption and facilitation. Also, despite being only 12 years old, it’s estimated that 70M+ people owned Bitcoin as of January 2021, with over a hundred thousand joining the network every day.
Third, Bradley Rettler, philosophy professor at the University of Wyoming and founding member of Resistance Money, a Bitcoin research collective, argues that Taleb’s conclusions are diminished by omissions and skimpy arguments:
“There’s no added ‘there will be no miners’ or ‘the value will go to 0 when there are no miners,’ nor is there any argument for either of those. So rather than an argument against Bitcoin as it seems purported to be, it’s rather a general point about expected value.”
I also found it strange that Taleb claims Bitcoin will be worthless “when miners are extinct” without any discussion of when or how likely that might be. Practically speaking, it’s less important for many Bitcoin owners that Bitcoin holds value indefinitely than that it can help them preserve or grow their wealth during their lifetime. Also, the final Bitcoin isn’t expected to be mined until 2140. By that time, Bitcoin would be more than 80 years older than fiat USD is now.³
For those seeking a longer-term store of value, it’s not absurd to think that Bitcoin will last longer than a few generations, especially given the built-in infrastructure adaptability. For example, Bitcoin’s protocol uses a difficulty adjustment mechanism based on block time that keeps block time constant, inflation predictable, and miners incentivized. A predefined schedule is already in place for gradually transitioning to using transaction fees as a sufficient security budget for miners. And, as Slagter also points out, Bitcoin doesn’t technically depend on the existence of miners for perpetuity anyway. Rather, Bitcoin depends on a trustless consensus mechanism that can adapt to future conditions.
“Technologies tend to be supplanted by other technologies with a vulnerability in proportion to their past survival duration (>99% of the new is replaced by something newer), whereas items such as gold and silver have proved resistant to extinction.”
Bitcoin is not that new in information age time. If the internet had a child at 26 years old, Bitcoin would be its 12-year-old adolescent now.⁴ Also, I don’t know how Taleb got his “>99%” replacement figure nor how relevant it actually is to Bitcoin, but I’d bet Bitcoin has outlasted a large proportion of the technologies that were factored into it.
With that aside, there are at least two relevant topics that Taleb doesn’t address here: (a) technological adaptation and (b) technological layering.
Taleb reveals his disregard of technological adaptability by drawing a weak analogy between Bitcoin and the gettoni, the Italian national telephone tokens that were accepted as tender during the 1970s but became obsolete as communications got cheaper and telephone calls got replaced by superior technologies. It just doesn’t make sense to compare Bitcoin, a “living” multilayered information network to a pressed chunk of bronze.
There are better parallels to be drawn between the internet and Bitcoin. Like the IP/TCP network and transport layers of the internet, the network and transaction layers of Bitcoin adaptively address limitations and vulnerabilities. Most of us don’t worry about absorbing barriers of the IP/TCP layer. When we store anything of value on the internet or invest in internet technologies and platforms, we are fully trusting in its infrastructure and adaptability. Why wouldn’t we do the same for Bitcoin?
Or when should we trust Bitcoin’s infrastructure and adaptability? I’m not the one to say, but neither is Taleb. It depends on individual risk preferences.
As for technological layering, Brain points out:
“Taleb also ignores that successful technologies usually build upon other technologies. The internet, which Bitcoin is built atop, is the culmination of a century or more of computer science. And said computer science is built atop centuries more of prior science and innovation. So, let’s assume there is a successor to Bitcoin a century from today. It would not surprise me if that successor continues to use Bitcoin and its existing store of value as a base layer.”
In some sense, then, Bitcoin may only be 12 years old but its technological infrastructure is arguably much older. In this case, the Lindy effect that Taleb is referring to — or the use of time in existence as a judge of a technology’s robustness — would predict a pretty long shelf life.
“Transactions in Bitcoin are considerably more expensive than wire transfers or other modes, or ones in other cryptocurrencies, and order of magnitudes slower than standard commercial systems used by credit card companies.”
Taleb seems to imply that Bitcoin’s transaction cost and speed are a reason for technological obsolescence. But a comparison between a third-layer credit card network to a first-layer final settlement network doesn’t make much sense. Furthermore, as Brain points out:
“This argument completely ignores the Lightning Network, which sits atop Bitcoin as a transactional layer, and is both instantaneous and virtually feeless. Bitcoin is designed to act as the store-of-value and settlement layer beneath. As such, we can completely ignore this argument.”
Taleb rebuts with alleged certainty that the Lightning Network doesn’t exist. But the thing is, it does. It just hasn’t been widely adopted yet. As an experimental second-layer solution, nothing about this is surprising. Plus, traction seems more likely to increase than decrease with recent developments such as Lightning adoption by Bitfinex (currently 4th largest CEX by 24-hr volume) in 2020 and Kraken (8th largest CEX) this year. Strike also already deployed a Bitcoin Lightning payment app in El Salvador, where Bitcoin was adopted as legal tender last month.
Also, is it not ironic for Taleb to quickly dismiss “what doesn’t exist now⁵” when his thesis rests on future possibilities?
“We cannot expect a book entry on a ledger that requires active maintenance by interested and incentivized people to keep its physical presence, a condition for monetary value, for any such period of time — and of course we are not sure of the interests, mindsets, and preferences of future generations.”
Let’s dig into what exactly Taleb is postulating here. Bitcoin “exists” in binary data and code flowing through the internet. To really destroy it, you would need to wipe the data of all of the 80,000+ Bitcoin full nodes in the decentralized network or destroy the internet. But like the internet, the Bitcoin network is designed with no single point of failure where all the data flows through, and the Bitcoin client is prepared for possible segmentation or outages.
Then, what Taleb is suggesting is that user interest will not just wane, but there will certainly be a global and unanimous rejection of Bitcoin, resulting in permanent data loss. When stated that way, it does make me wonder whether Taleb is abusing ergodicity economics to induce fear (FUD) when he could use his intellect for a more useful discussion about separating probable risks from those that are merely possible.
If Taleb were to instead argue that there’s a probable risk that interest in Bitcoin will wax and wane, no one would debate this. But it also wouldn’t constitute an absorbing barrier. The fact is that even if the great majority of Bitcoin users lost interest in it, Bitcoin would retain value and people would mine and transact it.
But for Taleb’s sake, let’s assume the hypothetical is true. Bitcoin is entirely forgotten about. Would its value go to zero? Yes, value is a human construct. Could its value recover from zero? Yes, as long as the UTXO set is not irrevocably destroyed. Is it probable? I don’t think so.
To understand why, we should consider the strength of Bitcoin’s technological and social network layers.
While the technological aspects of Bitcoin, such as its blockchain, are marveled at, its human side receives much less attention, even when it arguably adds a robustness that could parallel that provided by the blockchain.
As a social species, humans create emergent organizations beyond the individual — structures that range from friends and family to cities and civilizations. Our very biology, from our genes to our brains and bodies, has evolved to support these emergent structures because the social behavior they facilitate helps us survive and reproduce. History suggests that this evolutionary progression toward cooperative organizations of greater scale and evolvability is unstoppable.
In the case of Bitcoin, its decentralized network layer has engendered a global monetary alliance with over US$600 billion of skin in the game, united by distributed trust in the blockchain. As Andres Antonopoulos, author of Mastering Bitcoin, explains:
“Bitcoin is a highly sophisticated decentralized trust network that can support myriad financial processes … The interaction between many nodes is what leads to the emergence of the sophisticated behavior, not any inherent complexity or trust in any single node.”
Like self-interest at the level of genes, self-interest at the level of Bitcoin’s node operators will not stop cooperation and evolvability from increasing. It is by Bitcoin’s very design that self-interest strengthens the network. Running a full node provides you with the full ledger so you don’t need to query a third party about it, protects your Bitcoin investment by making the network more robust, increases your blockchain query speeds, and maximizes your privacy. There’s also an optional signaling value, which should not be overlooked when it drives the growth of incentive economies in online communities like Wikipedia. It is such features of the blockchain that makes me doubt that centralized regimes have the upper hand against Bitcoin or some evolution of it.
Although Antonopoulos leans toward the technical aspects of Bitcoin, it’s implied that Bitcoin is brought to life by the merging of machine and human, a decentralized technological structure that is invigorated by a supporting decentralized social network. He writes:
“The highly intelligent and sophisticated behavior exhibited by a multimillion-member ant colony is an emergent property from the interaction of the individuals in a social network. Nature demonstrates that decentralized systems can be resilient and can produce emergent complexity and incredible sophistication without the need for a central authority, hierarchy, or complex parts.”
As the Bitcoin community grows, so does its adaptive resilience according to evolutionary theory and its value for individual members and intrigue for non-members according to network effects. In other words, network-driven self-reinforcement is likely to go hand in hand with increasing not decreasing returns — much less, complete devaluation.
All in all, what I see holding back Taleb’s argument is that it rests too heavily on a snapshot of what Bitcoin is now to predict what it will become, as if information technologies are as immutable as gold. It neglects where Bitcoin came from and what it is becoming. Don’t be fooled by what’s in the frame, be wise about what’s left out.
A better valuation would take into account the mutual evolution of the people involved in Bitcoin and the technological infrastructure of Bitcoin itself — both of which are crucial to Bitcoin’s value infrastructure. If we neglect either, we risk distorting our view of the future by assuming minute or distant possibilities are certainties. We overestimate risks and underestimate the timescale at which they might happen. Or worse yet, we throw the baby out with the bathwater.
¹ Fiat currencies can be rendered useless if people lose faith in the governments that issued them. Property like land and houses can be destroyed by natural disaster or in warfare. Jewelry and collectibles can lose their aesthetic utility.
² The UTXO set acts as a global database of all the spendable outputs available when constructing a Bitcoin transaction. The entire UTXO set is tracked by Bitcoin full nodes to ensure that a user cannot spend Bitcoins that have already been spent, preventing double spending.
³ The USD switched from representative currency to fiat currency when President Nixon decoupled USD from gold in 1971, making fiat USD about 50 years old. Bitcoin will be about half the age of fiat USD in 20 years.
⁴ The Internet was born when ARPANET and the Defense Data Network officially adopted the TCP/IP standard on January 1, 1983, allowing all predecessor networks to be interconnected as the Internet.
⁵ Even if Taleb is obstinate about this requirement, he shouldn’t ignore that there are already massive exchanges acting as de facto secondary layers processing billions of dollars’ worth of Bitcoin daily at web speed.
Inquiries: firstname.lastname@example.org (open to work)
Searching for a Black Swan: A Brief Analysis of Nassim Taleb’s “Bitcoin Blackpaper” was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.