There’s no Such Thing as High Fees on Bitcoin

Veterans of the blocksize war that raged from 2015 to 2017 in Bitcoin were met with a curious sight this week: Bitcoiners complaining about high fees. If this doesn’t strike you as odd, it’s worth recalling some brief history.

Historically, “small-blockers” – the faction that retained the claim to Bitcoin in the blocksize wars – have supported the existence of fees during times of congestion as a necessary tradeoff to achieve decentralization. Big blockers – the faction that split off into BCH and BSV – were the ones aiming to keep fees low. But this trend has reversed for some Bitcoiners recently.

CoinDesk columnist Nic Carter is partner at Castle Island Ventures, a public blockchain-focused venture fund based in Cambridge, Mass. He is also the co-founder of Coin Metrics, a blockchain analytics startup.

Famously, big-block (and low fee enthusiast) Roger Ver lamented five years ago that “babies [were] dying” because Bitcoin core devs refused to raise the block limit (and alleviate fees). Central to the big blocker talking point was that more block space was required in order to accommodate more transactions. This would in theory drive down fees and make it easier to scale a blockchain to global usage for all manner of small payments.

The small blockers, by contrast, preached the importance of restraint, and sought to keep the blockchain as compact as possible, so that it might remain maximally decentralized and censor resistant. A larger blockchain, the thinking went, would only be maintained by industrial node operators, and could thus be trivially co-opted by state or corporate actors. In the scaling wars, instead of naively increasing block space to ease the fee pressure in the short term, Bitcoiners embraced instead a layered philosophy.

This is in fact how all payment systems scale: by deferring settlement. You don’t use wire transfers to pay for cigarettes – you use a network like Visa which uses batched, deferred settlement and ultimately settles to partner banks via ACH. You would typically reserve wire payments for larger transactions like down-payments on a house where finality is important.

Indeed, I’ve been sympathetic to the small blocker philosophy, and that’s why I supported Bitcoin over Bitcoin Cash, and that’s why I’m generally skeptical of the approach taken by blockchains like EOS or Solana. Accepting smaller blocks means tolerating larger fee spikes when the blockchain gets busy. Historically, large blockers used times of blockchain congestion as evidence in their favor, and Bitcoiners were forced to grit their teeth and resort to more philosophical arguments as to why the convenience wasn’t worth the tradeoff in terms of decentralization. In 2017, Gregory Maxwell, arguably the most influential developer on the small block side of the conflict, “popped the champaign (sic)” when Bitcoin fees topped the block reward for the first time.

In 2023, it’s 2017 again, except this time many hardcore Bitcoiners have flipped and are now embracing the “fees should be low” perspective – one that they (or their antecedents) explicitly fought against during the scaling wars.

The culprit is, ironically, Bitcoin’s own Taproot upgrade, which (perhaps unintentionally) opened up a new design space which permitted users to inscribe arbitrary content on the blockchain. Image NFTs (“Ordinals”) have stolen most of the attention, but the immediate catalyst for this spike was actually the creation of the BRC-20 standard, which relies on an exotic distribution method. BRC-20s are issued with a “proof of burned fee” mechanic, in which users must sacrifice transaction fees in order to create new tokens. This has driven fees in the short term to eye-watering levels, pricing out other sorts of conventional usage. Some Bitcoiners have even taken to calling the usage of the blockchain a deliberate “denial of service” attack or an attack on El Salvador’s Bitcoin mission.

The good news is that fee spikes generally don’t last. If history is any guide, the current mania for minting BRC-20 tokens will likely fade relatively quickly. It looks to me to be an event similar to the Otherside mint on Ethereum rather than anything enduring. However, it’s certainly the case that Ordinals and Inscriptions have unlocked a huge amount of latent demand for Bitcoin blockspace, and entered us into a new, structurally higher fee regime, even if this acute spike will fade in a matter of days.

This I consider to be an unequivocally good thing, as Bitcoin blockspace was a virtual wasteland from summer 2021 through early 2023. Miners need to be paid somehow, and as the miner subsidy further decays, fees will have to compensate for the lost revenue. The lack of sufficient miner revenue from fees has been the primary worry for myself and many other Bitcoiners who acknowledge the long term risks to the protocol. So I have been heartened to see Ordinals and Inscriptions trigger a new form demand for Bitcoin blockspace. I believe these kinds of more creative uses of otherwise-neglected blockspace could represent a path to sustainability for Bitcoin’s block reward.

Though I am sympathetic to the observation that high fees are pricing out individuals used to making smaller transactions on the Bitcoin base layer, especially folks in the global south like El Salvador or Africa, it’s simply mistaken to believe that Bitcoin owes anyone perpetually low fees. Bitcoiners made a very deliberate decision to cap block space to make validation cheap (and indeed – to make validation as inclusive as possible, especially for folks in the global south!). Mechanically, this means that fees must rise when congestion appears. This is an inescapable consequence of the Bitcoin design.

For a set of people who claim intellectual lineage from the Austrian tradition, the characterization of the price of blockspace as “too high” is incoherent. Fees are set by the market, and believing them to be too high is to endorse interventionism and central planning. To believe that the price of blockspace is “high” is to impose a normative view regarding which types of transactions Bitcoin should be reserved for, which is incompatible with a permissionless system.

The market price of a commodity can never be “wrong.” Prices simply reflect the aggregate attitudes of all market participants. This logic is akin to saying that oil prices ought to be $20/barrel so that even the poorest in society can afford gas. While that stance may have some populist appeal, someone always has to pay. In the case of Bitcoin, the price of low fees would be unbounded block space (a nonstarter from a decentralization standpoint), not to mention a permanent subsidy (as fees will be the sole driver of miner revenue in the long term). So Bitcoiners who claim that fees are “too high” are effectively demanding that their low fees be subsidized by node operators and, in the long term, by perpetual inflation in the Bitcoin protocol.

Those expressing dismay they can’t onboard newcomers to Bitcoin simply misunderstand the nature of the network. Base layer settlements are a finite commodity and cannot be expected to remain cheap in perpetuity. Instead, the settlement assurances sought should match the nature of the transaction. You don’t need a bank wire level of assurance to buy a coffee, and you don’t need a base layer transaction to settle a $5 test payment to a friend.

If Bitcoiners stymied by fees have any quarrel, it ought to be with the pace of development of L2s in the Bitcoin space. There’s no real excuse for this misconception either: a cursory look at the data shows that fees reached over $50 per transaction at certain points in both 2021 and 2017. This should have surprised nobody. Fees will always reflect the market’s demand for Bitcoin blockspace, and if your desired usage mode is priced out, that simply means that base layer Bitcoin isn’t a suitable network for you.

More puzzling is why former small blockers and self-appointed high priests Giacomo Zucco or Francis Pouliot have taken to complaining about the fee pressure (even if they enlisted on the small blocker side of the scaling wars). The answer is that for them, there are two types of Bitcoin usage: sacred and profane. Ordinals and Inscriptions are generally supported by the moderate “outgroup” and as of right now have primarily been used to issue and trade NFTs, or to speculate on new tokens.

To the hardcore fundamentalist crowd that absolutely disdains the ownership of anything other than Bitcoin, this is profane. Current Bitcoin maximalist doctrine stipulates that Bitcoin ought only be used for monetary transactions – but not those pertaining to non-Bitcoin asset types. Their perverse framework is one in which it’s morally acceptable for North Korea to use Bitcoin for sanctions evasion, but unacceptable for an artist to issue their NFT on the Bitcoin blockspace. If this strikes you as bizarre and inconsistent, it’s because the Bitcoin maximalist ideology has become highly reactive, technologically regressive, and more concerned with ideological purity than intellectual coherence.

For the few small blockers that have remained intellectually consistent from 2017 to today, the new fee pressure ought to be welcomed. It’s a catalyst for the further adoption of established L2 networks like Lightning, and where Lightning falls short, alternative L2 systems. As with all commodities, high prices are the cure for high prices. As a venture capitalist, I’m encouraged by the entrepreneurs I have been meeting lately who have begun designing novel L2s that explore alternative design spaces. In particular, rollups have clear product market fit on Ethereum, and I’m hopeful they can be added to Bitcoin in time.

Lightning is not a panacea, and it is best suited for high frequency, small granularity payments – which certainly doesn’t satisfy all types of Bitcoin transactional demand. There’s more than just one way to move dollars around – we have wires, ACH, credit and debit networks, Fednow, physical cash, remittances, hawala networks, money orders, and fintech apps, among others. More abstractly, these can be divided into “push” and “pull” approaches, real time versus deferred settlement, and gross versus net settlement models. Each has its own set of tradeoffs, and offers different transactional speeds and settlement assurances.

It’s naïve to think that one scaling network alone can satisfy the diverse transactional needs of Bitcoiners. The long term future for Bitcoin is a plurality of scaling approaches. The gist, however, is a move away from the one-baselayer-settlement = one-payment model, towards a more economically dense, one-baselayer-settlement = many-payment model. This is the right way to scale. This is the only way a broadcast system like a blockchain, where every node operator must be aware of every transaction, can scale to global usage.

The burst of activity around inscriptions and resulting high fees is an accelerant towards this more efficient future, and it should be celebrated. The BRC-20 mania, as perplexing as it is to many, is one of the best things that has happened to Bitcoin in some time.

DISCLOSURE