It may seem strange to people both inside and outside the cryptocurrency space that micropayments have not taken off as a major use case. Turning money into digital packets, with all the flexibility and extensibility that entails, should have brought a swift death to the “Pay by Credit Card” form we are met with at every online purchase. We could have replaced the 3-digit security code of our credit cards – littered across the databases of Amazon, Shopify, and others – with cryptographic invoices. Yet here we are, in AD 2023, with very few retailers even offering crypto payments at all. How did cryptocurrency fail at its most direct use case?
The design space of products where sending (or streaming) values under one cent remains wide and unexplored, and cryptocurrencies differ widely in how they work and what they promise. We’re going to look at a few major cryptocurrencies to find the lowest micropayment we can send today, from one person directly to another, and see how they stack up: Bitcoin, Ether, Stablecoins like USDC and Tether.
The King: Bitcoin
Bitcoin is the most established cryptocurrency, and today secures north of USD $400 billion worth of assets. On any given day, it will facilitate the movement of somewhere between $15 to $40 billion across a global, peer-to-peer decentralized network, without having to trust a central authority for final settlement of transactions. A criticism that has been levied against Bitcoin is its inability (or lack of desire) to accommodate micro-transactions, say sub $1. Let’s look at what the limits are for small payments on the base Bitcoin network, as well as the constraints of its second layer, the Lightning Network.
Bitcoin implements a fee market for block space, meaning that when demand for inclusion of transactions in a block is high, fees rise, and vice versa. However, in order to prevent spamming of the network with miniscule value payments (sometimes called “dust”), the Bitcoin network also implements a minimum value that any transactions on Bitcoin must abide by; as of January 2023, denominated as 0.00000546 BTC, or 546 satoshis, or approximately $0.13 as of 03/01/2023. Any transaction below that will be rejected as invalid.
In practice, there is always a fee, even if it is very low. Crucially, the transaction fee doesn’t depend on the amount being transferred, but on the amount of data being transferred. At the lowest fee possible (1 sat/vbyte) and an approximate average transaction size of 250 bytes, we get a $0.11 fee. Adding the fee to our smallest transaction (0.11 + 0.13), we get $0.24 for a 45% fee for the transaction–not workable in practice. Keeping the transaction byte size equal, and for a $10 transaction, you’re paying a 1% fee, and so on: the higher your transaction amount, the smaller the relative fee. Along with its other assurances, this has made Bitcoin very attractive for sending larger amounts, especially in a global context, but remains problematic for small transactions.
Ride the Lightning
It’s clear that Bitcoin, even in its most favorable market conditions, will be unable to support a market for microtransactions or even just casual transactions. (Friendly reminder that $2.50 USD can get you pretty far in some countries, and Bitcoin is by default a global network.) In order to extend Bitcoin’s network to this use case, a second layer has emerged called the Lightning Network. It works similar to a bar tab: you and a counterparty both contribute to a Bitcoin transaction, which creates a “channel” (i.e., a tab), where you can send as many transactions back and forth of any denomination (buy as many drinks as you want without running your card), and at the end, when you decide to close the channel, each side gets their last accounting state (your final tab) and receives it on the Bitcoin main chain. Lightning is not a blockchain, nor is it a distributed state machine. It simply allows two or more parties to create arbitrary, collaborative payment channels that settle to Bitcoin. These channels can be linked into a network of channels, which define the “payments layer” for Bitcoin–the Lightning Network.
There are several costs to consider when using the Lightning network:
- Channel open and close cost, which are native BTC transactions
- Base routing fees through the Lightning network (what individual channels charge for routing your payment)
- Liquidity fees which scale with the amount you’re trying to send
It makes sense to open a channel with someone if you intend to do multiple payments with them, and the base Bitcoin fee will be a negligible percentage of the amount sent back and forth. But you don’t need to open a channel to make a payment; you will however incur fees for asking others to route the payment for you. CashApp, for example, lets you pay Lightning invoices in-app, and you rely on their infrastructure to route the payment successfully.
So let’s imagine you want to send a single payment to a friend, without setting up a channel. Each node along the route sets its own fee rate, and these may be difficult for your wallet to estimate. However, a quick look at the largest nodes gives us fees between 0.00005 and 0.001 satoshi per satoshi sent. For example, if your payment is $1 (approximately 2500 sats), your fee might be (0.0005 * 2500) or 0.125 satoshis–less than a thousandth of a cent. In practice, every node includes a base fee for routing anything, usually just 1 sat–again, negligible fees. Even as these fees are minuscule, we can expect competition for routing fees to intensify as more capacity and demand comes online.
Into the Ether
Ethereum emerged in 2015, framing itself as both a competitor and an evolution of Bitcoin’s ideas. In the years since, Ethereum has enabled a bevy of new experiments in financial engineering–AMMs, liquidity pools, yield farming, NFTs, etc. While Ethereum’s network is in practice far less “peer-to-peer” than Bitcoin, we can still ask how it fares regarding person-to-person payments.
Once upon a time, Vitalik Buterin called any cryptocurrency with $50 fees per transaction a “de-facto liveness failure”. Unfortunately, the average transaction fee today on Ethereum’s mainnet hovers between $20 and $80 depending on demand and estimated costs of running the transaction (i.e. “gas cost” or the estimated cost of computing the output of a transaction). Ethereum is unlikely to compete for payments.
Two Sides, Same Coin
At two opposite ends of the regulatory spectrum, we have USDC (USD Coin, issued by Centre, a joint venture by Circle, Coinbase, and Bitmain) and USDT (aka Tether, a subsidiary of Bitfinex, one of the largest cryptocurrency exchanges by volume). USDC and Tether are stablecoins, meaning digital tokens which trade at par with the US dollar, such that 1 USDC and 1 Tether always equal $1 USD.
Before we dive into the transaction costs, consider why these might be useful. If you want to send USD from one jurisdiction to another, you must go through traditional payment methods to do so. These systems can be slow, especially for international settlement, and crucially, they cannot enable the programmability and digital nature of cryptocurrencies. A stablecoin is a US dollar IOU which behaves like a cryptocurrency, greatly enhancing a financial institution’s ability to move dollars globally.
Both USDC and USDT are implemented as Ethereum smart contracts (i.e. ERC-20 tokens), and claim to be fully backed by USD reserves. When you buy either one with USD, a token is created out of thin air, because you send USD which “backs it”; and vice versa: if you redeem a token for USD, that token is permanently destroyed, since the backing entity is “paying out” the USD. Centre is the company behind USDC, and remains a fully centralized, regulated financial institution even as it operates on a decentralized platform. Tether is not regulated.
This brings us to the transaction costs: even as both are centralized in issuance, they are fully exposed to Ethereum’s fee market as described above. So we see exchanges like Kraken charge a high minimum transfer amount as well as a high minimum fee. In a way, casual users get the worst of both worlds: highly centralized entities which are unable to handle micropayments at all.
Whatever the technical implementation, micropayments remain somewhat elusive in practice. While cryptocurrencies’ digital nature makes these use cases approachable, implementing a system where micropayments make sense financially has proved much more difficult, hence the current emphasis on “layer 2” scaling in both Bitcoin and Ethereum. Every payment has a cost and it will never be free.
But there’s a bigger issue at play here: the quest for payments media is more than just a technical question – it’s also a question of the success of money as a network. I could spin up a new crypto-token and run it on AWS with almost free payments. Whom would this benefit? If only 10% of the population accepts your currency, it won’t matter much whether micropayments are possible on its network or not. The network that gets the largest reach is the one people will accept payment in, which helps cement its reach as a network, in an example of a positive feedback loop. Therefore, it’s unlikely that a new up-and-coming cryptocurrency will win payments market share in any meaningful way.
Given Bitcoin’s head start as money, I suspect the Lightning Network will end up spurring the most activity in bringing micropayments into reality. Ethereum, and by proxy stablecoins such as USDC, will have to figure out how to solve its base layer fees before it can hope to offer micro-payments to the world. The ability to support micro-payments sheds light on an important, and often overlooked aspect of cryptocurrency: it is as much about tech as it is about monetary network effects, and no amount of one can help you solve the other.