Should we force fees higher to set user expectations?
Last week in a (publicly logged) IRC channel, someone asked one of those most prominent Bitcoin Core developers the following question: “Are you arguing ‘fee pressure is good’ and therefore small blocks and zero growth are desirable?” The reply was:
“Fee pressure is an intentional part of the system design and to the best of the current understanding essential for the system’s long term survival. So, uh, yes. It’s good.”
In other words: higher fees will be necessary in the future; therefore, we should encourage them now.
Most arguments for smaller blocks are about a larger block size directly leading to more mining centralization, higher costs of running a full node, or mining revenue being reduced below the level that is necessary to secure the network. This argument is different in that it could be used even in the absence of those risks.
A deeper look at the argument for high fees
Why not just enjoy low fees until we are forced to give them up? Based on my conversations with Core developers on IRC, the argument is: if we artificially keep fees low now, then users will come to expect low fees. Business models will be built on low fees. These business models will be successful temporarily. However, when fees eventually have to rise to a long-term sustainable level, users and businesses will have their expectations painfully violated. These people are likely to clamor for technically unsound changes to Bitcoin which preserve the low fees that they’ve come to rely on, at the expense of security and decentralization. The more we pursue adoption via cheap fees now, the more people will be relying on cheap fees in the future, and the harder it will be to resist the changes they push for. The changes could essentially destroy Bitcoin, by turning it into something too similar to Visa or PayPal to be worthwhile.
Of course, there are costs to raising fees now. Higher fees directly reduce the usefulness of Bitcoin as a currency and payment network, making people less likely to use it. Because money has network effects, people being discouraged from using Bitcoin negatively impacts the entire Bitcoin ecosystem.
Do fees really need to be higher in the future?
The conclusion that we should encourage higher fees now relies on the assumption that users will need to pay higher fees in the future to secure the network. If you compare only fees for on-chain transactions this seems unclear — it depends on the future state of: Bitcoin’s price, the cost of bandwidth/storage/processing power, the number of Bitcoin users, and many other uncertain factors. We can sidestep the debate about future on-chain fee levels by realizing that a relevant fee comparison must account for changes in how transactions will be sent in the future.
Lightning is coming
It looks very likely that the Lightning network or something like it will allow users to send cheap transactions for as long as Bitcoin exists. Lightning can’t replace on-chain transactions for those using Bitcoin as digital gold (a use for which low fees are much less important). However, for those using Bitcoin as a currency or payment network, Lightning transactions will be an extremely close substitute for on-blockchain transactions. It won’t matter to users if on-chain transactions in the future are more expensive than on-chain transactions now, as long as they can accomplish the same thing for cheaper using Lightning.
The obvious question is: if users in the future will have a way to send what are essentially Bitcoin transactions for extremely low fees, why would we want to prepare them for high fee future that they will never experience?
If we want to encourage a particular fee policy not for its effect on decentralization or security, but in order to prepare users for what it will be like to use Bitcoin in future, we should instead aim for fees as low as we can safely provide.