New Study: Low Bitcoin Transaction Fees Unsustainable

Joon Ian Wong
Jul 9, 2014 at 20:51 UTC
Updated Jul 10, 2014 at 02:25 UTC

Transaction fees have come to the fore again after core developer Gavin Andresen indicated that the next bitcoin core update would feature a new way to determine fees.

The issue of setting transaction fees has occupied core developers for some time. Andresen, for example, wrote about the issue last February, raising the possibility of a higher, fixed, fee for transactions.

Rising bitcoin transaction fees are the focus of a new paper that was published as part of the International Conference on Digital Security and Forensics held in the Czech Republic in June.

The paper’s author, Kerem Kaskaloglu is an instructor in cryptography at Ozyegin University in Istanbul. He warned that a new system for fee-setting was required because transaction fees are currently too low and are set to rise over the years.

“Right now, we have subsidies, so we don’t really need any transaction fees,” he said. “Because of subsidies, miners are willing to mine. But later on, as the subsidy goes away, we definitely need to figure out a balance between miners and users of the network.”

Satisfied miners

Kaskaloglu’s paper says that miners are currently satisfied with the value of block rewards, thus not requiring high transaction fees to motivate them to continue mining. Block rewards act as a form of subsidy for bitcoin transaction costs.

But the current model of transaction donations is unsustainable, the paper argues. As the supply of new bitcoins shrinks, this scarcity will outstrip even the accompanying price rise in bitcoin, thanks to the limited number of potential bitcoin users in the world, Kaskaloglu wrote.

As a result, the cost of mining will increase greatly while the rewards from each block will be insufficient to create enough profit after the increased costs are accounted for. Kaskaloglu noted this state of affairs is not imminent, but could take place anywhere between five and 20 years from now, explaining the difficulty in calculating this with certainty. He wrote:

“It is a daunting task to come up with an estimation about when bitcoin transactions should be charged by fees rather than donations […] the value of the block reward depends highly on the price of a bitcoin, which in turn is related with the adoption of bitcoin, which in turn is […] hard to foresee.”

Difficulties setting a fee

Kaskaloglu’s paper describes some of the difficulties in creating a new transaction fee system. One possible solution is to set a fixed fee for each transaction. In February, Andresen calculated that a miner should demand a fee of at least 0.0008 BTC, which at the time was worth $0.41, to include an average-sized transaction.

One problem with the fixed-fee system is that it severely hampers the utility of bitcoin for making micro-transactions, Kaskaloglu argued:

“Even that amount [Andresen’s $0.41 figure] is slightly frightening for small transactions, because bitcoin is highly promoted for little or no transaction fees for small transactions. But if you set a fixed fee, that wouldn’t allow that kind of application of bitcoin – for example, tipping.”

Another approach would involve limiting the number of transactions per block. This approach would make the number of included transactions more scarce, so that senders who pay a fee above a certain threshold would be more sure of inclusion.

But the problem with the limited-transaction approach lies in determining the threshold at which a transaction would be included. Since the volume of transactions fluctuates, the threshold constantly shifts. As a result, senders will face difficulty in calculating the correct threshold when they initiate a transaction. As Kaskaloglu wrote:

“This threshold will be a dynamic one depending on the instantaneous traffic of transactions so that the sender will not be able to figure out whether her transaction will be delayed or not in advance.”

New floating fees

Andresen’s proposed update to transaction fees is a step toward a dynamic fee system. The updated core software will ‘float’ fees, calculating the required fee for a quick confirmation, for each transaction. In some cases, there may be no fee for a quick confirmation – like a high-value transaction that re-spends old bitcoins – which is designated as a high-priority transaction.

Under the new proposal, fees could be as high as nearly 0.02 BTC for a quick confirmation, based on Andresen’s published test data. This is up from the current fee of 0.0001 BTC, Andresen noted in his blog post announcing the proposal.

Importance of a fee system

Kaskaloglu welcomed the planned improvements to the bitcoin core’s transaction fee system. The academic believes that maintaining low transaction fees is essential to bitcoin’s mass adoption.

“I think [low fees] are absolutely crucial,” he said.” There are two main charms of bitcoin: low transaction fees is the main one, and the other is that bitcoin reaches anywhere in the world with Internet.”

Members of Bitcoin Talk were distinctly nonplussed by Kaskaloglu’s paper. One poster, who goes by QuestionAuthority, responded to the paper thusly:

“OK, so what? Everything I read is already known.”

Kaskaloglu conceded that his paper makes no new technical contribution to the fee-setting system. But he was motivated to publish it as a way to underline the importance of improving on the way transaction fees are currently handled.

“I knew the hardcore users already knew about this stuff,” he said. “It’s not groundbreaking, but I wanted to alert people about the future of transaction fees. We need more studies about [this topic] and I would like to see more people be aware of the issue.”