Economics of the Bitcoin Transaction Fee Market
In 2008, the Bitcoin whitepaper was released, and the resulting fallout has been nothing less than earth shattering. Today, with cryptocurrency assets approaching a half a trillion dollar market cap, it is well past time to identify and model the market structures that have been created, the controls that have been instated over them, and the resulting effect that they have on the major players in the market.
In the freely available portion of this article, I would like to address Bitcoin itself and ignore alt-coins and Bitcoin forks such as Bitcoin Cash. While these additional cryptocurrencies are certainly an important part of the overall ecosystem, they are ultimately tangential to the core topic of this article, the Bitcoin Transaction Market. For those looking for more information, consider paying for the extended analysis beyond the pay-wall.
[NOTE FROM THE AUTHOR: I have now made all but a final note from me freely available. There is no more pay wall to the remaining significant content of this article. Also, while I have your attention, I'd like to note that some 7 months after writing this piece, I have learned a lot, and there is very little that I would like to change about this article. Regardless, aside from making the pay wall "disappear," I have changed nothing aside from a few edits on things like typos since the initial publishing of the article.]
Identifying the Market Players
In order to properly identify and create an accurate model of the market, it is first necessary to identify market players and their role in the ecosystem. Additionally, it is valuable to consider the economic structures under which each player operates to better understand the nature of the market as a whole. For this, we must examine at a high level what must occur for a transaction to be accepted on the network.
For a transaction to be accepted as valid and completed on the Bitcoin network, a user must transmit it to the network where it will be propagated to miners, who then validate the transaction according to the network protocol and place it into a block, which they secure through a proof of work mechanism. The proof of work mechanism consists of the use of mining computers, a form of capital good that must be run at a cost to the miner, to solve cryptographic hashing functions looking for an answer that satisfies conditions that the network deems valid. Additionally, the network protocol adjusts the conditions for validity such that the difficulty of finding a valid answer results in an average block find rate of one block every 10 minutes over the long run.
Considering this high level analysis, two major market players can be identified: The Miners, who serve the role of firms supplying mining power and transaction processing as well as to enforce network protocols, and the Users, who transact on the network, serving as the demand for transaction processing capacity.
While there are certainly additional technical roles within this market sector, they serve only minor economic purposes in facilitating the transaction market itself and can be ignored in this discussion without compromising the accuracy of the economic model.
Constructing the Market Models
To begin constructing the transaction market model, we will start with a simple supply and demand graph with the appropriate labels for each.
This graph should be familiar to most everyone, as it is a standard supply and demand graph. This is both the simplest and most natural state of the Bitcoin transaction fee market. What is traditionally labeled as the price is here labeled “Transaction Fee” and the quantity, “Transaction Quantity” with the intention of avoiding confusion with the overall price and supply of Bitcoin itself.
To fully understand the nature of the supply side of the transaction fee market, it is essential to explore the market structure of mining itself, and how that plays into the broader market of transaction fee pricing.
Those already well versed in basic economics will recognize this graph as that of a theoretical perfect competition prior to long run effects of competition. Over extremely short periods, this can be somewhat inaccurate, as in Bitcoin, the difficulty of mining blocks is fixed rather than fluid, such that additional mining will result in additional blocks found within an unadjusted difficulty period, therefor reducing fees, or vice versa. In a short term graph such as that, the Marginal Revenue curve would slope downward for large miners in response to the change in block find rates providing additional transaction capacity for the network.
However, over the medium and long terms, this effect is eliminated due to the difficulty adjustment algorithm regularly resetting block speeds to the intended 10 minute interval. Additionally over the long run, due to both competition and adjustments in difficulty, marginal revenue for the individual miner always trends towards marginal cost such that average total costs equal marginal revenue.
During difficulty adjustments, there is a sudden shift of expected revenue per hash performed by miners, inversely affecting marginal revenue for miners. In order to maximize profits, this leads miners to supply a quantity of hashing power, either turning on or turning off miners, that corresponds to the intersection between marginal cost and marginal revenue, as shown in the graphs.
Additionally, it is possible for marginal revenue to fall below a shutdown point where average total cost always exceeds marginal revenue at any quantity of mining. This is an effect that has seen to the demise of CPU, then GPU mining on the Bitcoin network as ASIC miners of increasing efficiency and hashing power have come online.
Other effects certainly exist, such as a lag in deployment of mining resources that makes it impossible for some miners to operate at equilibrium, as well as miner’s expectations of the future price of Bitcoin. However, in addition to being outside of the scope of this article, over the long run these effects trend strongly towards a perfect competition equilibrium of normal profit, as mining resources will eventually reach full deployment and the price of the underlying asset cannot continue rising indefinitely.
When aggregated across a competitive mining industry, the marginal cost of mining always trends such that marginal cost equals marginal revenue (the expected total block reward,) at the market equilibrium. Any underlying shifts in mining supply or user demand will always push marginal cost towards this same equilibrium over the medium to long run.
Accounting for the Block Reward Subsidy
The Block Reward, set at the time of this writing as 12.5 Bitcoins per block, serves as a subsidy for miners to operate on the network, both securing the network with hash power and accepting transactions into the blockchain as a byproduct of that process. Over time this effect will be removed, and the natural fee market as seen in the supply and demand graph without the subsidy will become the dominant source of miner revenue.
As can be seen, the block subsidy serves to increase the carrying capacity of the network (transaction quantity,) while reducing the cost to transact using the network. While typically considered an inefficient market manipulation, this has been essential in Bitcoin’s early days both to ensure sufficient mining power is applied to secure the network, and to encourage user adoption of the currency.
Accounting for Artificial Network Controls
The question then arises, how does the block size limit affect the market? For now let’s ignore the block reward so that we may analyze the block size limit in isolation.
The immediate effect of the block size limit is that it acts as a supply cap, enforced by network protocol that can only be changed with a hash rate of greater than 50%. It should be noted that this supply cap is not currently placed due to any economic factors, nor do technical limitations appear to be involved as simply allowing the network to find equilibrium would work if technical limitations were the defining factor.
In an economic sense, this means that the placement of the block size limit is arbitrary, meaning that any equilibrium to the left of the block size limit results in a natural fee structure, while any equilibrium to the right results in an artificially inflated fee as network users compete to have their transactions accepted into blocks.
Additionally, as shown in the previous analysis of the mining market, aggregate marginal cost always trends towards aggregate marginal revenue, which, in the near-perfect competition of the mining market, is the price, or in this case, the transaction fee paid by the user. The question then becomes, what happens to the market when the block reward subsidy is applied?
In this, the most accurate model of the current Bitcoin transaction fee market, it becomes clear that the transaction fees paid by users becomes uncoupled entirely from the effect of the block reward subsidy, the result of which is that user adoption is no longer encouraged. Additionally, transaction quantity does not increase due to the arbitrary cap on transaction acceptance, despite a rightward shift of market equilibrium.
But surely this is good for miners, as it greatly increases the reward for mining. In fact, the market structure applied here looks on first glance to be a perfect, enforced cartel designed to maximize miner profits, if only it were set in accordance with the right economic factors. However, there is one flaw in that analysis.
As is shown time after time, aggregate marginal cost always trends towards aggregate marginal revenue, with the result that miners do not actually benefit from this market structure. Instead, more resources are allocated to purchasing mining hardware due to a combined effect of two major distortions in the market - the block reward subsidy, and the block size limit. This raises costs of both honest miners and potential attackers equally, doing nothing to improve security of the chain, all while decreasing efficiency of the network as a whole. In fact, it can be argued that this makes it more likely that large interests will be able to out compete small miners in the short run at deploying mining resources, due to their greater capacity for absorbing short run fixed costs.
At best, this structure does ensure a higher miner profitability than on chains with the same hashing algorithm without such restrictions or with less adoption and use, thus artificially driving mining power towards this rather than other competing blockchains. But ultimately, this heightened profit is not entirely due to real market conditions, rather, it is due at least in part to arbitrary technical restrictions that do not otherwise benefit miners themselves, and that harm users directly by increasing both the cost to transact and the time that it takes for a transaction to be accepted on the blockchain.
While debate rages on over the scaling debate despite network forks that have occurred in an attempt to sidestep the issue as a whole, it is clear that the market structures discussed in this article are here to stay on the Bitcoin Core blockchain. Regardless of any technical concerns surrounding an unrestricted block size, or any tendency that may or may not exist for mining power to centralize into the hands of a few given a natural fee market, it is clear that the economic structures that have been created are not conducive to adoption by users, and attempts to increase adoption on the network beyond the block size limit will only result in greater costs of transacting on the network.
Some will challenge the analysis in this article, claiming that proposed scaling solutions aside from raising the block size limit have been ignored, and to those individuals I would suggest two things - first, that any scaling solution not currently in place must necessarily reduce either the number of size of on-chain transactions that are currently not crowded out of the market or they will do nothing to alleviate current on-chain scaling concerns, and second, to return to this author with such concerns when those scaling solutions are actually implemented and prove to be a substantial factor in alleviating pressure on the on-chain fee market. Until that time comes, there is no practical benefit derived from Layer 2 solutions, and as such, they can be ignored when considering current market models.
Otherwise the question arises, in light of knowledge that may be gleaned from this article, who stands to gain from installing a market structure that so clearly harms the cause of user adoption without benefiting the miners who actually run the network? That, it would seem, is the first step in discovering the true cause of the scaling debate as it has turned out - but that goes well beyond the scope of this article, and should therefor be left for now as an exercise for the reader.
For those curious to know more and to delve into some questions that cannot be answered quite as clearly as those covered in this free portion of the article, please explore the content beyond the pay-wall. Topics covered include an evaluation of the “store of value” argument, a short analysis on layer 2 solutions, additional explorations on the effects of the block size limit on transaction fees and what that means for the potential value of the network, and considerations on where potential users could go if and when they are priced out of using the Bitcoin network for their desired use cases.
Beyond the (Original) Pay-Wall
The question that ultimately drives the prices of underlying assets such as Bitcoin, Bitcoin Cash, Ethereum, Iota, oil, US Dollars, gold, bananas, etc. is, why are people buying the asset, and does that asset effectively fill the role desired by the individuals buying that asset?The analysis in this portion of the article is not meant to be predictive of future value, but to analyze specific effects and forces that play on the market. In no way does this amount to a complete analysis of all effects playing on this market, nor does it claim to effectively weigh the strength of one force over another in directing the price of any given asset.Please understand, this section does not merely serve as a disclaimer to avoid legal liability for actions taken in response to the analysis presented in this article, but to make the point that it is impossible for any one individual to understand the entire picture when it comes to market forces. Anyone claiming to know for certain what will happen are just asking to be blindsided by an effect that they did not previously account for.With that in mind, the following should be considered the opinion of the author, and is nothing more than a presentation of arguments that may, despite reasonable efforts by the author to be thorough and fair in his analysis, be incomplete. Additionally, it should be noted that there are no graphs in this section, not out of the author’s laziness, but because graphs tend to convey a sense of certainty that in this case is impossible to achieve.
On the "Store of Value" Argument
A lot of discussion has been had over whether or not the “store of value” argument holds weight in establishing Bitcoin’s price.It cannot be discounted that people’s personal predictions of future value of an asset do directly affect the market price for that asset over the short run, however, without an understanding of the fundamentals of the market, this argument alone can seemingly justify any price, if only the right emotional state can be teased out of the market.When applied on a practical level, this argument amounts to no more than one of three options: the price is going up/down because the price is going up/down; the price will go up/down because the price is currently going down/up too much; or, the price is staying the same because the price is staying the same.There is no equilibrium in this form of analysis, nor are there any fundamentals guiding the action of price movements, and therefor, no accurate price to be derived from it. If left as the only tool for establishing the price of an asset, the “store of value” argument will do nothing more than produce a price chart that is indistinguishable from random noise over any scale, long or short, until such a time that the market fails entirely.Even while this sort of argument plays on the psychology of the individual over the short run, the “store of value” argument cannot in and of itself account for the price of an asset being sustained over the long run as there is no predictive value in that argument that can survive the test of time.Additionally, the “store of value” argument implies the eventual desire to “unstore” that value by using it. How can something be stored if it is not easy to unstore it for immediate use? By analogy, should fresh water be considered stored if poured into the ocean, or is it more accurate to say it has been thrown away? Alternatively, what do you call a battery that no longer holds a charge?For these reasons, even those who believe that Bitcoin is a store of value cannot ignore market fundamentals, the realistic use cases for Bitcoin given its current or likely future construction, and the substitutes for Bitcoin if it is unable to service the needs and wants of potential users.
What is the Value of Bitcoin?
Bitcoins are a purely digital asset without any direct use value whatsoever. As such, Bitcoin is fundamentally nothing more than a currency and a unit of account.
Transactions as a Requirement of Use
Even the most elegant uses of the Bitcoin Blockchain, Layer 2 solutions included, require transactions to be placed on the network. The number of transactions required for a given use case depends on implementation of that use case. For a simple person to person exchange, a single transaction may be sufficient. Alternatively, operating a Lightning Network payment channel from start to finish requires at least two on-chain transactions regardless of the total number of transactions processed off-chain.Regardless of the use case being considered, bitcoins are the unit of account for transacting on the network because only bitcoin can be used to directly pay miners their fees. (It should be noted that it is possible to pay a miner off-chain to include, or even exclude, specific transactions in their blocks, but this is generally not a favorable approach over the long run due to various economic effects that run well beyond the scope of this article.)Even in cases where bitcoins themselves are not exchanged on the network, bitcoins are still the unit of account for use of the network. An example would be the storage of data on the blockchain, which could be placed in a transaction without a fee - however, in order to be included in a block, the miner must have leftover resources with which to process the free transaction, which implies that they have received enough value through other transactions, the block reward subsidy, or other means including a sense of altruism to serve this use case for free. More likely still is that such a use case would be coupled with a transaction fee that directly pays the miner for validating this transaction. In all cases, bitcoins are the unit of account.
Transaction Fees and the Unit of Account
To better understand this issue, let’s examine Iota and Ethereum. Iota’s Tangle could be used on an unimaginable scale while the price of Iota, the currency, is not effected, because Iota itself is not a factor in transacting on the Tangle. This is because the proof of work is done by the user, validating two transactions before their own, without receiving payment in Iota for having done so.There is no unit of account inherent to the Iota ecosystem, and unless individuals use it in real exchange at the risk of losing their entire purchase value, there is no unit of account anywhere in the system. Unless restricted from doing so by the Iota developers, any other currency could be loaded up as a resource on Iota and used as a pure replacement for Iota, perhaps with other perks included.On other chains, non-monetary uses of the network requires payment in the form of the underlying asset. For example, on Ethereum, execution of smart contracts requires “gas” payments, payable only in Ether. In the case of Ethereum, ether has at least some value as money because it is the unit of account in paying for gas to run smart contracts, which are themselves valuable.
Bitcoin's Use Cases
Returning to Bitcoin, given current limitations artificially placed on the network, namely, the block size limit, the Bitcoin network can only service, in theory, about 7 transactions per second. While short run hash power variations may alter this figure in the short run, at this time it is sufficient to suggest that 7 transactions per second is the limit.What is the highest asset price that can be supported by an economy that has a built in limit of roughly 7 transactions per second, or roughly 600k transactions per day? And once that limit is reached but the economy is still growing, what happens to the cost of transacting on that network, with that asset as the unit of account?The second question is easily answered using the models provided in the free section of this article, but the first question depends entirely on what people do with those transactions.While it would be satisfying to be able to put specific numbers down as answers to these questions, it is simply not possible to do so with any amount of accuracy. Others have speculated on the potential price by comparing Bitcoin to other markets, such as the international gold market, or the Visa payments network. Ultimately, these answers give a ballpark figure, but cannot possibly account for all the variables in human action that lead to one valuation over another.So what are people using Bitcoin for? Are they creating jewelry out of it, like gold? I certainly don’t see how, after all, bitcoins are a purely digital asset without any direct use value whatsoever. Are they buying coffee, or videogames with it? Due to the effects caused by the block size limit, the videogame publishing platform Steam no longer accepts Bitcoin, and coffee is the traditional example of supposed “spam” on the blockchain that is used all over to deride people for taking up precious space on the blockchain and in the hard drives of node operators.Many people believe that all that’s needed to handle these sorts of transactions is a Layer 2 scaling solution such as Lightning.
An Analysis of Layer 2 Solutions
Supporters of the Lightning Network and other Layer 2 scaling solutions believe that those 7 transactions per second on-chain can support a massive economy, by essentially compressing large volumes of off-chain transactions into a few on-chain transactions.From an economic perspective, there is nothing wrong with this analysis, and if Layer 2 solutions are proved out on a technical level, users adopt them, and their claims of the potential of Layer 2 solutions come to fruition, then the 7 transactions per second can support an economy of many more than 7 transactions per second. This, it would seem, would justify quite a high price of bitcoins, and could also justify high transaction fees on the network.Additionally, in accordance with their ability to take transactions off-chain, Layer 2 solutions would ease pressure on the on-chain fee market by shifting demand for on-chain transactions to the left as compared to total transactions conducted in the combined Layer 1 and Layer 2 economy.These promises, as mentioned in the free portion of the article, have yet to be proved out on a practical level. Additionally it should be noted that the potential ability to expand Bitcoin transaction capabilities off-chain does nothing from an economic perspective to justify a block size limit. Instead, if Layer 2 solutions do not lower demand for on-chain transactions to below the block size limit, the block size limit is just increasing the cost of doing anything on the chain, including opening and closing Layer 2 payment channels. Supporters must look elsewhere for their justifications of such a limit.With that in mind, let’s look at the extended effects of a block size limit on the Bitcoin transaction fee market regardless of the involvement of Layer 2 solutions. Specifically, what does a block size limit mean for adoption of Bitcoin, and how will potential users adjust given the effects of the block size limit on the transaction fee market.
Further Effects of the Block Size Limit on Bitcoin
As discussed in the free portion of this article, when the block size limit is reached, two things happen: fees decouple from the downward effect of the block reward subsidy due to an inability for the quantity of supplied transaction throughput to adjust to quantity demanded, and fees necessarily increase due to user competition over placement of their transaction in a block.This effect has been demonstrated at an increasing rate over the past months with increases in the mempool size, and resulting increases in transaction fees as the market adjusts to the new pricing model.What we have not yet addressed is what happens to users themselves whose transactions are priced out of acceptance into a block.On one hand, many of these individuals must simply wait for their transaction to eventually be accepted into a block as natural variances in transaction rates, block find rates, and mempool clearance rates move their transaction into the clearing batch for the next found block. This is a non-monetary cost to users, and can also be counted as an added, if hidden, cost of using the network. This has been the typical case up to this point, where the fee paid inversely corresponds to a wait time before acceptance into a block. However, this is not necessarily going to remain the usual state of affairs.As greater demand is placed on the network, the upward pressure on fees must come from transaction fee bidding wars that can only clear as transactions are cleared. With greater adoption comes greater upward pressure on fees, more transactions entering the mempool, and more bitcoins stuck awaiting confirmation into the blockchain.Not only is it the case that unconfirmed transactions are examples of value stuck in the mempool, but any transactions that users did not make due to the expected wait times or the expected transaction fees required for confirmation in a reasonable time represent lost opportunities for value exchange on the network.In essence, all value of demand to the right of the block size limit is value that is not being serviced by the Bitcoin network. This value is being priced out of Bitcoin.Is it possible for all of that value to be captured, if only the block size limit were removed? No, of course not. Instead, all value to the right of the supply and demand equilibrium would still be unserved. This is true of any and all mediums of exchange, as all mediums of exchange carry some form of cost in transacting with it. The question is, what is the cost of using Bitcoin as compared to other mediums of exchange, and how many users will switch from one medium of exchange to another in a given time frame?Asked differently, what are Bitcoin’s competitors, and what will users do in response to artificially high costs of using the Bitcoin network?
Where do Users go from Here?
This is where the fundamental value of Bitcoin comes into play. Will users stay in the Bitcoin ecosystem despite artificially high costs? Will new users continue to adopt the currency? Or will users jump ship, choose alt-coins and Bitcoin forks, or avoid cryptocurrencies altogether?Bitcoin isn’t a currency in a vacuum. Users who are priced out of Bitcoin will seek alternatives, and those alternatives include Bitcoin Cash, Ethereum, Litecoin, and yes, remaining in fiat currencies.Additionally, if Bitcoin costs more to use than alternatives, even users who can afford to use Bitcoin will choose to use alternatives, especially if those alternatives match their philosophical beliefs on how currency should be.Ultimately, businesses and everyday people are the ones who will choose the winners and losers in cryptocurrency. Must one currency win over all others? No, but there is undeniable value in weeding out the currencies that do not compete well. So what happens to Bitcoin if it doesn’t compete?It’s impossible to say what users will do for sure, but the conclusions to be drawn from economics are clear - if Bitcoin doesn’t compete, it will face downward price pressure from users who decide to go elsewhere. What, then, will justify the “store of value” argument? If the price starts going down, who is to say that it won’t continue to go down? The speculators and the get-rich-quick seekers will turn tail and flee just as quickly as they jumped on board this train. What value will the hodlers claim is stored, then?And if Bitcoin falls, what happens to the rest of the cryptocurrency ecosystem? Does another coin rise up and take its place? Does the Bitcoin of “Satoshi’s original vision,” Bitcoin Cash take the mantel? Or does cryptocurrency as we know it die, to be replaced by a shell of its former potential?
6 of 6 reviewers say it's worth paying for
0 of 6 reviewers say it's not worth paying for