Review of Taleb's ‘Bitcoin, Currencies, and Bubbles’

Nassim Taleb recently wrote a paper talking about how bitcoin doesn't have value.

Taleb, N. N. (2021) ‘Bitcoin, Currencies, and Bubbles’. Available at: https://www.academia.edu/49313911/Bitcoin_Currencies_and_Bubbles (Accessed: 21 June 2021).

There is a fair bit of boisterous tweeting on the topic on either side which is not that entertaining.  

I will focus on my thoughts on the paper.  I will follow the headings in the paper:

  • The Blockchain
  • Vulnerability of revenue-free bubbles (or why BTC is worth exactly 0)
  • Success in wrong places
  • Principles for a currency
  • The difficulty with inflation hedges
  • Some additional fallacies
  • And the conclusion

I'm generally enjoy Taleb's work and I'm long bitcoin so there is that too.

The Blockchain (or should it be time chain?)

I think this section describes Bitcoin and how it operates at a broad level though some of the details are a little off:

  1. Nodes validate the transactions, but miners validate and also order the transactions in a distributed fashion by placing them into blocks.  This is the solution to the "Byzantine generals' problem". How to agree on what happened when is the big problem.  Not the actual order as much as how to come to consensus on the order.  The order of transactions is what controls which coins were spent when and hence prevents the same coins from being spent twice. This is what Nakamoto consensus solves.  Nodes also maintain and store the "database" or ledger of transactions.  Miners decide the ordering of updates to the ledger. Miners are nodes (broadly speaking) but nodes are not miners.  The article confuses these roles a bit throughout.
  2. There are no direct incentives to run a node, but it's relatively cheap (and kept so).  So the state of bitcoin is easy to track with a node and indeed this allows everyone with a node to verify that all is as it should be with bitcoin the currency and Bitcoin the network.  So as a receiver or spender of bitcoin I can do this through a node and verify that all is as expected.  That is the indirect benefit of running a node.
  3. Miners gain value from bitcoin rewards as well as transaction fees.  The balance of which will switch gradually to transaction fees over time according a predefined schedule finishing in the far future.
  4. The degree of mining difficulty is adjusted based on the speed of the blocks over certain time frames (not the speed of transactions as is stated in the paper).  It gets easier to mine the less blocks are mined and vice versa.

Vulnerability of revenue-free bubbles

In this section Taleb states the following:

The implication is that, owing to the absence of any explicit yield benefiting the holder of bitcoin, 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.

He then postulates if bitcoin drops below "its level" it hits an absorbing barrier and stays 0.  This state is left undefined though.  The implicit definition is price or is it hash rate. The thing is, as far as I can tell, bitcoin escaped an "absorbing state" of zero price and zero mining at least once when there was a single miner on the network (the inventor of Bitcoin) and bitcoin the currency was without value.  So what is an absorbing state in this context?  Clearly if someone values it and difficulty drops low enough they can and will mine it.  

In the paper the "absorptive state" leads to the principle of cumulative ruin:

If any non-dividend yielding asset has the tiniest probability of hitting an absorbing barrier, then its present value must be 0.

But clearly bitcoin has already escaped this absorptive state at least once.  The long tail of crypto currencies losing value but seemingly never going to zero is also testament to this.  

One could postulate as long as the Bitcoin chain data is maintained by one node there is a non-zero chance that the Bitcoin network and hence currency can reappear as long as the owner of that node starts it up and starts mining.  So the absorptive state is not absorptive even if all activity ceases.  Only when the chain data is lost does the bitcoin finally disappear, so there may be some absorptive states.  But other stuff has this property perhaps?

Taleb excludes collectibles from his cumulative ruin  principal but he fails to exclude traditional fiat currency.  Does fiat not have exactly the same properties as bitcoin?  A currency note in my pocket generates no revenue.  And in fact the note has a non-zero chance of zero value, if for example I forget it in pocket when washing my clothes, but also if, for example, the government backing that currency fails.  Then, according to Taleb, fiat currencies too must have zero value?

A currency note in my pocket is the closest equivalent to a private key to bitcoin in a wallet in my phone, PC or preferably a secure hardware wallet.  Now you may argue that most currency are deposits in banks and those yield interest (or not in some place, but let's leave that conversation).  But is that really currency or is that not just a debtor where you earn income?  If I lend my bitcoin would I not also expect income?  In fact I can earn money by lending out bitcoin already so either we only count notes in our pockets and bitcoin in non-custodial wallets (i.e. wallets where the owner has the keys) or bitcoin can earn revenue and fiat too.

So either bitcoin is not a revenue yielding asset, but then neither is fiat currency, or both are revenue yielding assets.

Fiat is not mined you say, but who are the keeper of the ledgers?  For notes there are no ledgers, but to transact I have to move my note myself which costs energy.  Not a problem when getting milk around the corner but a problem when I'm sending money around the world.  Deposits are kept in ledgers of bank accounts and those banks will keep track of those as long as they have value (to deduct fees from or to lend out).  So in a fiat world the banks are closest to nodes and miners.  Banking is unlikely to exist for long if the value of the transacted currency goes too low.  Similarly I'm not going to use money if I need a wheelbarrow to use it.  So we have a similar absorptive state?

At the end of the day, in my mind, there is no real reason why currency should have any value.  It's just boils down to whether it is accepted as such by enough people.  This applies to bitcoin, but also applies to USD or ZAR (South African Rand) or EUR etc.  

So Taleb theory is either right, and all currencies should, in theory have no value, or the theory doesn't work.  I actually prefer the former to be honest as I think that explains some of the oddness of currency.  It's a shared delusion that something has value when it should not, but it is practical that we have something like currency, because let's face it I'm not going to harvest potatoes and send them to pay for my website hosting.  That's my lay understanding of currency.  A collective delusion.  

Success in the right places?

There are a few flaws here in the argument postulated which itself is not very clear.

  1. Miners do not have to make their money from increases in the price of bitcoin. Clearly they do but, they could also make their money under a stable bitcoin price in say USD terms.  This applies to all system participants.  Bitcoin can be valuable without price increases, but the volatility probably comes from the perceived value.
  2. Transactions in bitcoin base are not more expensive than all other modes. International wire transfers are more expensive generally than bitcoin transactions (and slower and also prone to being reversed or never executed).
  3. Bitcoin transactions are also more final than most transactions available to individuals.  Credit card transactions are not final at all and can be reversed.  O
  4. Bitcoin also has second layer technologies such as lightning that allows near instantaneous transactions at fraction of the cost of any similar technologies in the traditional financial systems.

I do agree that for it to be considered on par with the USD or EUR that eventually we would need native bitcoin prices, but to expect that in 12 years is a bit hasty.  As bitcoin is used more I expect this will follow, but cannot prove this.

All currencies can't have this stable state in all circumstances.  ZAR prices are valid in South Africa, but at large enough amount and for purchases across borders price negotiations are likely to switch to EUR or USD terms.  I'm unlikely to be able to purchase a Lamborghini in ZAR terms.  Similarly the old Zimbabwe Dollar (ZWD) was probably valid for prices for everyday locally produced food items in Zimbabwe (at least at some point), but prices for higher value items (say for a car, TV or PC) were set in ZAR or USD terms given the volatility of ZWD.  Everything is on a spectrum.  Taleb acknowledges this himself:

Simply, there is a free market of fiat, with the most reliable at the time used by third parties. Before the Euro, there were plenty of currencies in Europe. But real contracts were drawn in deutschmark or Swiss franc, sometimes the U.S. dollar; drachmas, liras, and pesetas were there for petty expenditures.

This also implies that bitcoin can have value even if it's more volatile than USD or EUR.  It also implies that if bitcoin is less volatile than some existing currency there is value?

Difficulty of inflation hedges

So clearly it will take time to displace fiat as "it is not easy".    The problem is tying to a basket of goods is impossible in a decentralised way.  Stable coins have managed to tie cryptocurrency to fiat currency but that requires trust in the link between the two or the entity providing the backing for these stable coins. This does not guarantee a stable value to a price of goods.

Bitcoin chose to address this inflation problem by capping the issuance of bitcoin ultimately to 21m bitcoin.  The guarantee being that any inflation or deflation of bitcoin will not be driven by human intervention in issuance.  Clearly this inflation hedge could be stronger in the short term by more connections between bitcoin and the real economy and hence a stable price.

But as a holder of ZAR I can also buy USD to escape ZAR inflation but will also then be subject to the volatility of ZAR vs. USD and hence potentially lose out on inflation protection in ZAR terms.  

So inflation is difficult to hedge anyway.

Additional fallacies?

Libertarianism

I'm not sure what is meant about this issue.  The "politics" or "law" of libertarianism is irrelevant.

Financial Tail Risk

Clearly bitcoin is not designed as a financial tail risk element and simply is not seen by the majority as that.  It cannot however be argued that it won't become that.

Protection from tyranny

I think there are some problems here with Taleb's arguments.  Bitcoin is traceable but some of the examples are also that of poor security.  The ransomware payments in bitcoin were not hacked but the systems storing the private keys were subject to subpoena.  It's clear that the attackers could easily have gotten away with the money if they had implemented basic key security measures such as putting keys on third party computers.

Will bitcoin provide 100% protection to tyranny?  No.  Is it infinitely better than the traditional financial system?  YES!

Agency

In the case of agency between government and financial industry it results in financial benefit that is target explicitly and unfairly to those able to influence the process.

Taleb indicates bitcoin has an agency issue due to the distribution of bitcoin balances. However the distribution of bitcoin was as fair as could be attempted.  Bitcoin was announced in advance and issued fairly and provably so (see the quote in the first genesis block).  This is much fairer than most cryptocurrencies that followed.  Not clear what is fair in fiat terms?  Essentially it started with who could find gold the best and then those were grandfathered into the new system.  Is that fair?

Furthermore one could still take part and buy bitcoin or run a miner.  The only requirement is trading with others or using electricity.  This is not perfect agency but it's not been claimed to be perfect.  

One would also add in the traditional financial system those with the most money can influence the money system (to big too fail for example?), whereas in Bitcoin those with most money can still not influence the transaction or money process.  So the direct impact on money is zeroised under bitcoin.

Taleb's Conclusion

It's bitcoin not blockhain.  I think blockchain's best (only?) use is for a cryptocurrency that's not tied to the real world (other than through mining and difficulty adjustments linking it to raw energy used to run the computing hardware).  

Any use of blockchain technology to track shipping containers, or land ownership will be centralised to some central party (or parties) that will need to resolve issues when the blockchain does not tie up with the real word.  Only when a blockchain does not track real world items can it be truly decentralised and hence  valuable.

Conclusion

I found the paper to be based on somewhat theoretical arguments about absorbing states that are not clearly shown to apply.  My theory and argument using the same level of details is that the same arguments apply to fiat currency which, in of itself, does not generate revenue unless given to others.  Bitcoin given to others should also be able to generate revenue and does.  So either the theory doesn't apply or equally applies to the fiat currencies.  Either way we have a conundrum.  

My way to solve this conundrum is to say that bitcoin should have no value in the same way that any currency should not have value.  Currencies only have the value we conveniently assign to them and hence remain a collective delusion.  The edge of Bitcoin is to ensure that this delusion is not devalued or controlled by some.  

The strongest other point in the article is that bitcoin is volatile and hence tricky to use as a currency and clearly not convenient to set prices.  I agree but I also think that this is a problem that exists on a scale of currencies from USD and EUR all the way down to the Zim Dollar (ZWD).  Is bitcoin not just on this scale?  In that volatility will limit some use cases until it reduces.  

I think Bitcoin network also gives access to a decentralised currency across the world.  The true value will first be felt in corners of the world where the alternative is far worse, but at some point bitcoin will be better than all other options.  This I cannot prove in the same way Taleb hasn't proven to me why bitcoin has no value.  I can just point to the stability of bitcoin.  1 bitcoin = 1 bitcoin and there will be only 21m.  That's a promise Bitcoin can keep which the issuers of currencies have failed to do since currencies began.

The final point of the paper I can agree with.  The technology solves real world problems in remittances to El Salvador, in Africa and many other places.  And that is where I agree with the paper.  Bitcoin should be judged on the problems it solves and the problems it solves are everywhere.  Some are possibly quite large.  Fiat economies seem to be driven by the need to consume and produce more.  What if bitcoin can solve that?