A Half-Baked Thesis on Stablecoins

What follows began as snippets of conversations I have been having for years, on and off, here and there. It will likely eventually be collated into a piece I have been meaning to write on “payments” as a whole. I foolishly started writing this piece years ago, not realizing that the topic is gargantuan and for every week I spend writing it I have to add two weeks to my plan. That may or may not ever come to fruition, but in the meantime, Tether announced it was issuing on Taproot Assets and suddenly everybody is interested again. This is as good a catalyst as any to carve out my “stablecoin thesis”, such as it exists, from “payments”, and put it out there for comment and feedback.

In contrast to the “Bitcoiner take” I will shortly revert to, I invite the reader to keep the following potential counterargument in mind, which might variously be termed the “shitcoiner”, “realist”, or “cynical” take, depending on your perspective: that stablecoins have clear product-market-fit. Now, as a venture capitalist and professional thinkboi focusing on companies building on Bitcoin, I obviously think that not only is Bitcoin the best money ever invented and its monetization is pretty much inevitable, but that, furthermore, there is enormous, era-defining long-term potential for a range of industries in which Bitcoin is emerging as superior technology, even aside from its role as money. But in the interest not just of steelmanning but frankly just of honesty, I would grudgingly agree with the following assessment as of the time of writing: the applications of crypto (inclusive of Bitcoin but deliberately wider) that have found product-market-fit today, and that are not speculative bets on future development and adoption, are: Bitcoin as savings technology, mining as a means of monetizing energy production, and stablecoins.

I think there are two typical Bitcoiner objections to stablecoins of significantly greater importance than all others: that you shouldn’t be supporting dollar hegemony, and that you don’t need a blockchain. I will elaborate on each of these, and for the remainder of the post will aim to produce a synthesis of three superficially contrasting (or at least not obviously related) sources of inspiration: these objections, the realisation above that stablecoins just are useful, and some commentary on technical developments in Bitcoin and the broader space that I think inform where things are likely to go. As will become clear as the argument progresses, I actually think the outcome to which I am building up is where things have to go. I think the technical and economic incentives at play make this an inevitability rather than a “choice”, per se. Given my conclusion, which I will hold back for the time being, this is a fantastically good thing, hence I am motivated to write this post at all!

Objection 1: Dollar Hegemony

I list this objection first because there isn’t a huge amount to say about it. It is clearly a normative position, and while I more or less support it personally, I don’t think that it is material to the argument I am going on to make, so I don’t want to force it on the reader. While the case for this objection is probably obvious to this audience (isn’t the point of Bitcoin to destroy central banks, not further empower them?) I should at least offer the steelman that there is a link between this and the realist observation that stablecoins are useful. The reason they are useful is because people prefer the dollar to even shitter local fiat currencies. I don’t think it is particularly fruitful to say that they shouldn’t. They do. Facts don’t care about your feelings. There is a softer bridging argument to be made here too, to the effect that stablecoins warm up their users to the concept of digital bearer (ish) assets, even though these particular assets are significantly scammier than Bitcoin. Again, I am just floating this, not telling the reader they should or shouldn’t buy into it.

All that said, there is one argument I do want to put my own weight behind, rather than just float: stablecoin issuance is a speculative attack on the institution of fractional reserve banking. A “dollar” Alice moves from JPMorgan to Tether embodies two trade-offs from Alice’s perspective: i) a somewhat opaque profile on the credit risk of the asset: the likelihood of JPMorgan ever really defaulting on deposits vs the operator risk of Tether losing full backing and/or being wrench attacked by the Federal Government and rugging its users. These risks are real but are almost entirely political. I’m skeptical it is meaningful to quantify them, but even if it is, I am not the person to try to do it. Also, more transparently to Alice, ii) far superior payment rails (for now, more on this to follow).

However, from the perspective of the fiat banking cartel, fractional reserve leverage has been squeezed. There are just as many notional dollars in circulation, but there the backing has been shifted from levered to unlevered issuers. There are gradations of relevant objections to this: while one might say, Tether’s backing comes from Treasuries, so you are directly funding US debt issuance!, this is a bit silly in the context of what other dollars one might hold. It’s not like JPMorgan is really competing with the Treasury to sell credit into the open market. Optically they are, but this is the core of the fiat scam. Via the guarantees of the Federal Reserve System, JPMorgan can sell as much unbacked credit as it wants knowing full well the difference will be printed whenever this blows up. Short-term Treasuries are also JPMorgan’s most pristine asset safeguarding its equity, so the only real difference is that Tether only holds Treasuries without wishing more leverage into existence. The realization this all builds up to is that, by necessity,

Tether is a fully reserved bank issuing fiduciary media against the only dollar-denominated asset in existence whose value (in dollar terms) can be guaranteed. Furthermore, this media arguably has superior “moneyness” to the obvious competition in the form of US commercial bank deposits by virtue of its payment rails.

That sounds pretty great when you put it that way! Of course, the second sentence immediately leads to the second objection, and lets the argument start to pick up steam …

Objection 2: You Don’t Need a Blockchain

I don’t need to explain this to this audience but to recap as briefly as I can manage: Bitcoin’s value is entirely endogenous. Every aspect of “a blockchain” that, out of context, would be an insanely inefficient or redundant modification of a “database”, in context is geared towards the sole end of enabling the stability of this endogenous value. Historically, there have been two variations of stupidity that follow a failure to grok this: i) “utility tokens”, or blockchains with native tokens for something other than money. I would recommend anybody wanting a deeper dive on the inherent nonsense of a utility token to read Only The Strong Survive, in particular Chapter 2, Crypto Is Not Decentralized, and the subsection, Everything Fights For Liquidity, and/or Green Eggs And Ham, in particular Part II, Decentralized Finance, Technically. ii) “real world assets” or, creating tokens within a blockchain’s data structure that are not intended to have endogenous value but to act as digital quasi-bearer certificates to some or other asset of value exogenous to this system. Stablecoins are in this second category.

RWA tokens definitionally have to have issuers, meaning some entity that, in the real world, custodies or physically manages both the asset and the record-keeping scheme for the asset. “The blockchain” is at best a secondary ledger to outsource ledger updates to public infrastructure such that the issuer itself doesn’t need to bother and can just “check the ledger” whenever operationally relevant. But clearly ownership cannot be enforced in an analogous way to Bitcoin, under both technical and social considerations. Technically, Bitcoin’s endogenous value means that whoever holds the keys to some or other UTXOs functionally is the owner. Somebody else claiming to be the owner is yelling at clouds. Whereas, socially, RWA issuers enter a contract with holders (whether legally or just in terms of a common-sense interpretation of the transaction) such that ownership of the asset issued against is entirely open to dispute. That somebody can point to “ownership” of the token may or may not mean anything substantive with respect to the physical reality of control of the asset, and how the issuer feels about it all.

And so, one wonders, why use a blockchain at all? Why doesn’t the issuer just run its own database (for the sake of argument with some or other signature scheme for verifying and auditing transactions) given it has the final say over issuance and redemption anyway? I hinted at an answer above: issuing on a blockchain outsources this task to public infrastructure. This is where things get interesting. While it is technically true, given the above few paragraphs, that, you don’t need a blockchain for that, you also don’t need to not use a blockchain for that. If you want to, you can.

This is clearly the case given stablecoins exist at all and have gone this route. If one gets too angry about not needing a blockchain for that, one equally risks yelling at clouds! And, in fact, one can make an even stronger argument, more so from the end users’ perspective. These products do not exist in a vacuum but rather compete with alternatives. In the case of stablecoins, the alternative is traditional fiat money, which, as stupid as RWAs on a blockchain are, is even dumber. It actually is just a database, except it’s a database that is extremely annoying to use, basically for political reasons because the industry managing these private databases form a cartel that never needs to innovate or really give a shit about its customers at all. In many, many cases, stablecoins on blockchains are dumb in the abstract, but superior to the alternative methods of holding and transacting in dollars existing in other forms. And note, this is only from Alice’s perspective of wanting to send and receive, not a rehashing of the fractional reserve argument given above. This is the essence of their product-market-fit. Yell at clouds all you like: they just are useful given the alternative usually is not Bitcoin, it’s JPMorgan’s KYC’d-up-the-wazoo 90s-era website, more than likely from an even less solvent bank.

So where does this get us? It might seem like we are back to “product-market-fit, sorry about that” with Bitcoiners yelling about feelings while everybody else makes do with their facts. However, I think we have introduced enough material to move the argument forward by incrementally incorporating the following observations, all of which I will shortly go into in more detail: i) as a consequence of making no technical sense with respect to what blockchains are for, today’s approach won’t scale; ii) as a consequence of short-termist tradeoffs around socializing costs, today’s approach creates an extremely unhealthy and arguably unnatural market dynamic in the issuer space; iii) Taproot Assets now exist and handily address both points i) and ii), and; iv) eCash is making strides that I believe will eventually replace even Taproot Assets.

To tease where all this is going, and to get the reader excited before we dive into much more detail: just as Bitcoin will eat all monetary premia, Lightning will likely eat all settlement, meaning all payments will gravitate towards routing over Lightning regardless of the denomination of the currency at the edges. Fiat payments will gravitate to stablecoins to take advantage of this; stablecoins will gravitate to TA and then to eCash, and all of this will accelerate hyperbitcoinization by “bitcoinizing” payment rails such that an eventual full transition becomes as simple as flicking a switch as to what denomination you want to receive.

I will make two important caveats before diving in that are more easily understood in light of having laid this groundwork: I am open to the idea that it won’t be just Lightning or just Taproot Assets playing the above roles. Without veering into forecasting the entire future development of Bitcoin tech, I will highlight that all that really matters here are, respectively: a true layer 2 with native hashlocks, and a token issuance scheme that enables atomic routing over such a layer 2 (or combination of such). For the sake of argument, the reader is welcome to swap in “Ark” and “RGB” for “Lightning” and “TA” both above and in all that follows. As far as I can tell, this makes no difference to the argument and is even exciting in its own right. However, for the sake of simplicity in presentation, I will stick to “Lightning” and “TA” hereafter.

1) Today’s Approach to Stablecoins Won’t Scale

This is the easiest to tick off and again doesn’t require much explanation to this audience. Blockchains fundamentally don’t scale, which is why Bitcoin’s UTXO scheme is a far better design than ex-Bitcoin Crypto’s’ account-based models, even entirely out of context of all the above criticisms. This is because Bitcoin transactions can be batched across time and across users with combinations of modes of spending restrictions that provide strong economic guarantees of correct eventual net settlement, if not perpetual deferral. One could argue this is a decent (if abstrusely technical) definition of “scaling” that is almost entirely lacking in Crypto.

What we see in ex-Bitcoin crypto is so-called “layer 2s” that are nothing of the sort, forcing stablecoin schemes in these environments into one of two equally poor design choices if usage is ever to increase: fees go higher and higher, to the point of economic unviability (and well past it) as blocks fill up, or move to much more centralized environments that increasingly are just databases, and hence which lose the benefits of openness thought to be gleaned by outsourcing settlement to public infrastructure. This could be in the form of punting issuance to a bullshit “layer 2” that is a really a multisig “backing” a private execution environment (to be decentralized any daw now) or an entirely different blockchain that is just pretending even less not to be a database to begin with. In a nutshell, this is a decent bottom-up explanation as to why Tron has the highest settlement of Tether.

This also gives rise to the weirdness of “gas tokens” - assets whose utility as money is and only is in the form of a transaction fee to transact a different kind of money. These are not quite as stupid as a “utility token,” given at least they are clearly fulfilling a monetary role and hence their artificial scarcity can be justified. But they are frustrating from Bitcoiners’ and users’ perspectives alike: users would prefer to pay transaction fees on dollars in dollars, but they can’t because the value of Ether, Sol, Tron, or whatever, is the string and bubblegum that hold their boondoggles together. And Bitcoiners wish this stuff would just go away and stop distracting people, whereas this string and bubblegum is proving transiently useful.

All in all, today’s approach is fine so long as it isn’t being used much. It has product-market fit, sure, but in the unenviable circumstance that, if it really starts to take off, it will break, and even the original users will find it unusable.

2) Today’s Approach to Stablecoins Creates an Untenable Market Dynamic

Reviving the ethos of you don’t need a blockchain for that, notice the following subtlety: while the tokens representing stablecoins have value to users, that value is not native to the blockchain on which they are issued. Tether can (and routinely does) burn tokens on Ethereum and mint them on Tron, then burn on Tron and mint on Solana, and so on. So-called blockchains “go down” and nobody really cares. This makes no difference whatsoever to Tether’s own accounting, and arguably a positive difference to users given these actions track market demand. But it is detrimental to the blockchain being switched away from by stripping it of “TVL” that, it turns out, was only using it as rails: entirely exogenous value that leaves as quickly as it arrived.

One underdiscussed and underappreciated implication of the fact that no value is natively running through the blockchain itself is that, in the current scheme, both the sender and receiver of a stablecoin have to trust the same issuer. This creates an extremely powerful network effect that, in theory, makes the first-to-market likely to dominate and in practice has played out exactly as this theory would suggest: Tether has roughly 80% of the issuance, while roughly 19% goes to the political carve-out of USDC that wouldn’t exist at all were it not for government interference. Everybody else combined makes up the final 1%.

So, Tether is a full reserve bank but also has to be everybody’s bank. This is the source of a lot of the discomfort with Tether, and which feeds into the original objection around dollar hegemony, that there is an ill-defined but nonetheless uneasy feeling that Tether is slowly morphing into a CBDC. I would argue this really has nothing to do with Tether’s own behavior but rather is a consequence of the market dynamic inevitably created by the current stablecoin scheme. There is no reason to trust any other bank because nobody really wants a bank, they just want the rails. They want something that will retain a nominal dollar value long enough to spend it again. They don’t care what tech it runs on and they don’t even really care about the issuer except insofar as having some sense they won’t get rugged.

Notice this is not how fiat works. Banks can, of course, settle between each other, thus enabling their users to send money to customers of other banks. This settlement function is actually the entire point of central banks, less the money printing and general corruption enabled (we might say, this was the historical point of central banks, which have since become irredeemably corrupted by this power). This process is clunkier than stablecoins, as covered above, but the very possibility of settlement means there is no gigantic network effect to being the first commercial issuer of dollar balances. If it isn’t too triggering to this audience, one might suggest that the money printer also removes the residual concern that your balances might get rugged! (or, we might again say, you guarantee you don’t get rugged in the short term by guaranteeing you do get rugged in the long term).

This is a good point at which to introduce the unsettling observation that broader fintech is catching on to the benefits of stablecoins without any awareness whatsoever of all the limitations I am outlining here. With the likes of Stripe, Wise, Robinhood, and, post-Trump, even many US megabanks supposedly contemplating issuing stablecoins (obviously within the current scheme, not the scheme I am building up to proposing), we are forced to boggle our minds considering how on earth settlement is going to work. Are they going to settle through Ether? Well, no, because i) Ether isn’t money, it’s … to be honest, I don’t think anybody really knows what it is supposed to be, or if they once did they aren’t pretending anymore, but anyway, Stripe certainly hasn’t figured that out yet so, ii) it won’t be possible to issue them on layer 1s as soon as there is any meaningful volume, meaning they will have to route through “bullshit layer 2 wrapped Ether token that is really already a kind of stablecoin for Ether.”

The way they are going to try to fix this (anybody wanna bet?) is routing through DEXes, which is so painfully dumb you should be laughing and, if you aren’t, I would humbly suggest you don’t get just how dumb it is. What this amounts to is plugging the gap of Ether’s lack of moneyness (and wrapped Ether’s hilarious lack of moneyness) with … drum roll … unknowable technical and counterparty risk and unpredictable cost on top of reverting to just being a database. So, in other words, all of the costs of using a blockchain when you don’t strictly need to, and none of the benefits. Stripe is going to waste billions of dollars getting sandwich attacked out of some utterly vanilla FX settlement it is facilitating for clients who have even less of an idea what is going on and why North Korea now has all their money, and will eventually realize they should have skipped their shitcoin phase and gone straight to understanding Bitcoin instead …

3) Bitcoin (and Taproot Assets) Fixes This

To tie together a few loose ends, I only threw in the hilariously stupid suggestion of settling through wrapped Ether on Ether on Ether in order to tee up the entirely sensible suggestion of settling through Lightning. Again, not that this will be new to this audience, but while issuance schemes have been around on Bitcoin for a long time, the breakthrough of Taproot Assets is essentially the ability to atomically route through Lightning.

I will admit upfront that this presents a massive bootstrapping challenge relative to the ex-Bitcoin Crypto approach, and it’s not obvious to me if or how this will be overcome. I include this caveat to make it clear I am not suggesting this is a given. It may not be, it’s just beyond the scope of this post (or frankly my ability) to predict. This is a problem for Lightning Labs, Tether, and whoever else decides to step up to issue. But even highlighting this as an obvious and major concern invites us to consider an intriguing contrast: scaling TA stablecoins is hardest at the start and gets easier and easier thereafter. The more edge liquidity there is in TA stables, the less of a risk it is for incremental issuance; the more TA activity, the more attractive deploying liquidity is into Lightning proper, and vice versa. With apologies if this metaphor is even more confusing than it is helpful, one might conceive of the situation as being that there is massive inertia to bootstrap, but equally there could be positive feedback in driving the inertia to scale. Again, I have no idea, and it hasn’t happened yet in practice, but in theory it’s fun.

More importantly to this conversation, however, this is almost exactly the opposite dynamic to the current scheme on other blockchains, which is basically free to start, but gets more and more expensive the more people try to use it. One might say it antiscales (I don’t think that’s a real word, but if Taleb can do it, then I can do it too!).

Furthermore, the entire concept of “settling in Bitcoin” makes perfect sense both economically and technically: economically because Bitcoin is money, and technically because it can be locked in an HTLC and hence can enable atomic routing (i.e. because Lightning is a thing). This is clearly better than wrapped Eth on Eth on Eth or whatever, but, tantalisingly, is better than fiat too! The core message of the payments tome I may or may not one day write is (or will be) that fiat payments, while superficially efficient on the basis of centralized and hence costless ledger amendments, actually have a hidden cost in the form of interbank credit. Many readers will likely have heard me say this multiple times and in multiple settings but, contrary to popular belief, there is no such thing as a fiat debit. Even if styled as a debit, all fiat payments are credits and all have credit risk baked into their cost, even if that is obscured and pushed to the absolute foundational level of money printing to keep banks solvent and hence keep payment channels open.

Furthermore! this enables us to strip away the untenable market dynamic from the point above. The underappreciated and underdiscussed flip side of the drawback of the current dynamic that is effectively fixed by Taproot Assets is that there is no longer a mammoth network effect to a single issuer. Senders and receivers can trust different issuers (i.e. their own banks) because those banks can atomically settle a single payment over Lightning. This does not involve credit. It is arguably the only true debit in the world across both the relevant economic and technical criteria: it routes through money with no innate credit risk, and it does so atomically due to that money’s native properties.

Savvy readers may have picked up on a seed I planted a while back and which can now delightfully blossom:

This is what Visa was supposed to be!

Crucially, this is not what Visa is now. Visa today is pretty much the bank that is everybody’s counterparty, takes a small credit risk for the privilege, and oozes free cash flow bottlenecking global consumer payments.

But if you read both One From Many by Dee Hock (for a first person but pretty wild and extravagant take) and Electronic Value Exchange by David Stearns (for a third person, drier, but more analytical and historically contextualized take) or if you are just intimately familiar with the modern history of payments for whatever other reason, you will see that the role I just described for Lightning in an environment of unboundedly many banks issuing fiduciary media in the form of stablecoins is exactly what Dee Hock wanted to create when he envisioned Visa:

A neutral and open layer of value settlement enabling banks to create digital, interbank payment schemes for their customers at very low cost.

As it turns out, his vision was technically impossible with fiat, hence Visa, which started as a cooperative amongst member banks, was corrupted into a duopolistic for-profit rent seeker in curious parallel to the historical path of central banks …

4) eCash

To now push the argument to what I think is its inevitable conclusion, it’s worth being even more vigilant on the front of you don’t need a blockchain for that. I have argued that there is a role for a blockchain in providing a neutral settlement layer to enable true debits of stablecoins. But note this is just a fancy and/or stupid way of saying that Bitcoin is both the best money and is programmable, which we all knew anyway. The final step is realizing that, while TA is nice in terms of providing a kind of “on ramp” for global payments infrastructure as a whole to reorient around Lightning, there is some path dependence here in assuming (almost certainly correctly) that the familiarity of stablecoins as “RWA tokens on a blockchain” will be an important part of the lure.

But once that transition is complete, or is well on its way to being irreversible, we may as well come full circle and cut out tokens altogether. Again, you really don’t need a blockchain for that, and the residual appeal of better rails has been taken care of with the above massive detour through what I deem to be the inevitability of Lightning as a settlement layer. Just as USDT on Tron arguably has better moneyness than a JPMorgan balance, so a “stablecoin” as eCash has better moneyness than as a TA given it is cheaper, more private, and has more relevantly bearer properties (in other words, because it is cash). The technical detail that it can be hashlocked is really all you need to tie this all together. That means it can be atomically locked into a Lightning routed debit to the recipient of a different issuer (or “mint” in eCash lingo, but note this means the same thing as what we have been calling fully reserved banks). And the economic incentive is pretty compelling too because, for all their benefits, there is still a cost to TAs given they are issued onchain and they require asset-specific liquidity to route on Lightning. Once the rest of the tech is in place, why bother? Keep your Lightning connectivity and just become a mint.

What you get at that point is dramatically superior private database to JPMorgan with the dramatically superior public rails of Lightning. There is nothing left to desire from “a blockchain” besides what Bitcoin is fundamentally for in the first place: counterparty-risk-free value settlement.

And as a final point with a curious and pleasing echo to Dee Hock at Visa, Calle has made the point repeatedly that David Chaum’s vision for eCash, while deeply philosophical besides the technical details, was actually pretty much impossible to operate on fiat. From an eCash perspective, fiat stablecoins within the above infrastructure setup are a dramatic improvement on anything previously possible. But, of course, they are a slippery slope to Bitcoin regardless …

Objections Revisited

As a cherry on top, I think the objections I highlighted at the outset are now readily addressed – to the extent the reader believes what I am suggesting is more or less a technical and economic inevitability, that is. While, sure, I’m not particularly keen on giving the Treasury more avenues to sell its welfare-warfare shitcoin, on balance the likely development I’ve outlined is an enormous net positive: it’s going to sell these anyway so I prefer a strong economic incentive to steadily transition not only to Lightning as payment rails but eCash as fiduciary media, and to use “fintech” as a carrot to induce a slow motion bank run.

As alluded to above, once all this is in place, the final step to a Bitcoin standard becomes as simple as an individual’s decision to want Bitcoin instead of fiat. On reflection, this is arguably the easiest part! It's setting up all the tech that puts people off, so trojan-horsing them with “faster, cheaper payment rails” seems like a genius long-term strategy.

And as to “needing a blockchain” (or not), I hope that is entirely wrapped up at this point. The only blockchain you need is Bitcoin, but to the extent people are still confused by this (which I think will take decades more to fully unwind), we may as well lean into dazzling them with whatever innovation buzzwords and decentralization theatre they were going to fall for anyway before realizing they wanted Bitcoin all along.

Conclusion

Stablecoins are useful whether you like it or not. They are stupid in the abstract but it turns out fiat is even stupider, on inspection. But you don’t need a blockchain, and using one as decentralization theatre creates technical debt that is insurmountable in the long run. Blockchain-based stablecoins are doomed to a utility inversely proportional to their usage, and just to rub it in, their ill-conceived design practically creates a commercial dynamic that mandates there only ever be a single issuer.

Given they are useful, it seems natural that this tension is going to blow up at some point. It also seems worthwhile observing that Taproot Asset stablecoins have almost the inverse problem and opposite commercial dynamic: they will be most expensive to use at the outset but get cheaper and cheaper as their usage grows. Also, there is no incentive towards a monopoly issuer but rather towards as many as are willing to try to operate well and provide value to their users.

As such, we can expect any sizable growth in stablecoins to migrate to TA out of technical and economic necessity. Once this has happened - or possibly while it is happening but is clearly not going to stop - we may as well strip out the TA component and just use eCash because you really don’t need a blockchain for that at all. And once all the money is on eCash, deciding you want to denominate it in Bitcoin is the simplest on-ramp to hyperbitcoinization you can possibly imagine, given we’ve spent the previous decade or two rebuilding all payments tech around Lightning.

Or: Bitcoin fixes this. The End.

• Allen, #892,125

thanks to Marco Argentieri, Lyn Alden, and Calle for comments and feedback