Foundations of Market Structure as it applies to the Bitcoin Mining Industry
To date the most disruptive debate in Bitcoin has been, and in many ways, continues to be, the debate over block size, and, more fundamentally, the costs involved in running a node.
To summarize one side of the debate, many believe that high capital and operational expenses will "centralize" mining/node operation. They are concerned that barriers to entry will prevent home users from being able to interact with the network to defend against potential “evil miners.”
For these folks, this is one of the core elements in their reasoning for the block size limit, and other choices such as supporting the Lightning Network.
While this article is not intended to tackle all the concerns presented by those on the “small block” side, I believe it is important to accurately summarize their viewpoint.
In this article, I will focus primarily on the issue of market structure in the mining market. This topic covers the issue of “centralization,” and will at least clarify one of the more fundamental problems with the “small blocker” viewpoint.
On the surface, the claim that high barriers to entry will result in centralization of the mining industry seems to make a lot of sense, and a cursory look at economic literature will tell you that there’s a great line of reasoning behind this. After all, if it takes a lot of investment to break into a market, then current market players have an immediate advantage over potential newcomers to the market. It would seem to follow, then, that higher capital and operational expenses for running a node are necessarily a bad thing.
Great! Argument complete! Right?
Unfortunately, it is not so simple. To understand why not, we must first lay some groundwork for understanding the economics of market structure.
Like so many other examples of incomplete reasoning, (the likes of which bring us the “that’s not real socialism” trope, and the ever delightful Flat Earther movement,) the argument that high expenses will lead to centralized mining only factors in a portion of the involved forces: cost.
Cost, however, is not the only relevant factor. If it were, then Tesla and SpaceX could never have broken into their relevant markets (Musk is a master at finding the means of breaking into high barrier to entry markets,) and retail stores would never have to worry about the likes of Amazon, and governments – well, I suppose that’s an exception to this rule, at least for the short run.
What’s missing from the cost perspective is, of course, benefit. Value added. Production. Competition. It is a fact that markets will always, in the absence of external factors (such as government regulation, taxation, artificial supply limitations,) trend towards the supply/demand equilibrium that efficiently balances resource allocation against competing production opportunities. The exact mechanisms behind this trend depend on how competitive the firms in that market are at providing the good at low costs of production.
Say’s law states that “A product is no sooner created, than it, from that instant, affords a market for other products to the full extent of its own value.” This is often interpreted to say that supply creates its own demand, but this is not strictly true.
In fact, it is ultimately the values of consumers that lead to a market, as demand for a good leads to pressures to produce and sell that good at a profit, if a profit is possible. Though it is worth noting that many still attempt to produce valued goods and find they are unable to produce the goods at a profit. In short, without sufficient demand for a good (that is, enough demand to overcome the costs of production, allowing profitability to be achieved,) there is no producer for that good over the long run.
This is not to say that cost is irrelevant, (it is, after all, the real limiter on how much can be supplied at a given price over time,) but that it is not the sole factor that determines supply, let alone market structure.
Economies of Scale
Specialization is the means by which people increase their efficiency of production, lowering their costs of production for a given good. This is why it’s possible for mutually beneficial trade to exist.
If it were impossible to specialize, everyone would be just as well off producing everything for themselves as opposed to trading for goods they desire. In fact, they’d likely be better off, as they would not have to endure the costs of trading, such as seeking out another individual to trade with, negotiating a price, and physically exchanging the goods or services.
An important factor in an individual’s ability to specialize is their access to and adeptness at using capital goods. There is specialization in the use of capital goods, too, just as there is specialization in any other action.
Individuals with greater access to more efficient capital goods will be capable of augmenting their production capacity to produce even more than they could without that capital good.
As an aside, private property rights enable people to specialize in the use of capital goods by enabling them exclusive rights to the use of goods.
This is how economies of scale develop, as greater specialization begets greater specialization.
Over time, firms develop such that subtasks of production of a final good are divided among more people, who can specialize even further, under conditions agreed upon by the participants in the firm.
Those firms that are most specialized are best suited to produce a given good at the lowest cost.
That is to say, those firms that are most developed, with the most efficient division of labor, with the most efficient allocation of capital goods, and the most complete structure of production will be the most cost effective producers of that specific good which they are specialized in producing.
This necessarily implies that a firm that is specialized in producing more than one good, unless that good is a specific byproduct of the production of another good, is less cost effective than they could be at producing one specific good.
There are, however, costs to specialization. Investments in specialized equipment and skills are generally not easy to repurpose for other tasks. In some cases, the ability to change production priorities can be more valuable than specializing more completely in a given line of production.
This may be called agility, or flexibility, and may be more valuable than specialization in a rapidly changing economy where consumer demands are not fixed for long enough to justify lengthy investments in extreme specialization. Most firms seeking to break into a market utilize their relative ease of redistributing generalized resources to outmaneuver larger firms with greater inertia that results from their more mature specialization.
What is “a” Market?
McDonald’s has a monopoly on the production of Big Mac cheeseburgers.
Even if others were to duplicate their recipe, they would have to also source sufficiently substitutable ingredients, work within and sell from a building that is sufficiently substitutable for a McDonald’s building, and otherwise reproduce the entire experience of purchasing and eating a Big Mac for there to be another “Big Mac” firm in direct competition with McDonald’s in the market for, specifically, Big Macs.
But McDonald’s is not the only firm competing to sell the more generic good, cheeseburgers.
Nor are they the only firm competing to sell fast food.
Nor are they the only firm competing to sell food.
As the definition of the “good” being sold on the market changes, so does the number of firms selling in that market.
The question is, what goods are substitutes in the satisfaction of consumers’ demands? Are they perfect substitutes, or imperfect substitutes? Asked in more traditional economic parlance, are they homogenous goods, or non-homogenous goods?
McDonald’s is a fiercely competitive enterprise, operating in the “fast food” market, which satisfies ever changing demands of consumers, encompassing the desire for specific flavors, textures, smells, purchasing and dining experiences, speed of food delivery, health concerns, and innumerable other consumer demands.
Nobody cares that McDonald’s is the only seller of the Big Mac. That’s just one differentiation they can boast about to consumers in the fast food market. After all, they don’t sell Whoppers, or Six Dollar Cheeseburgers, or Dave’s Classic, or a Double Double, each of which serves as competitor’s differentiating factors.
And that is key. It is about market differentiation.
It is the differentiating factors between firms that create market segments, enabling them to sell to a consumer base separate from their competitors, giving them the ability to survive long term in a market for a given category of good.
McDonald’s isn’t a monopoly over fast food precisely because they can’t possibly specialize in selling the Big Mac, and the Whopper, and the Six Dollar Cheeseburger, or any of the other differentiated cheeseburgers on the market. If they were to try, they would lose their specialization, and competitors could undermine them at providing the same goods at a lower cost.
Instead, McDonald’s focuses on their particular competitive advantages, creating the conditions wherein nobody else can sell a Big Mac, so if a Big Mac is what a consumer wants in that moment, they must buy it from McDonald’s.
The Mining Market – The Good
In Bitcoin, the protocol design creates the mining market and the rules under which miners can produce goods for consumers. The protocol literally defines the category of good being sold to consumers. It can either encompass many opportunities for miners to differentiate their product, or few. It can be designed, in effect, to establish an equilibrium upon the number of firms that can survive over the long run in providing the goods being sold to consumers.
So, what is the good?
Transaction acceptance into the blockchain.
But it isn’t that simple. More accurately, that is the category of good being produced. There are qualities to transaction acceptance, such as the security of the blockchain, which is ensured via the requirement to find a hash of the block that satisfies difficulty requirements (an input to the production of blocks,) as well as the speed of transaction acceptance, the size and contents of transactions, acceptance and handling of smart contracts, etc.
On some chains, these qualities are artificially limited, usually for “technical reasons” of some sort or another. Perhaps the protocol doesn’t allow for certain qualities to be included in transactions, such as specific op codes. Or, in some cases, the supply of transaction acceptance is artificially limited by the protocol, thus creating restrictions on acceptance time, transaction size at a given price, etc.
Now I get to ask the question: have you been paying close attention?
If you have, you should have already noticed that those restrictions are artificially constricting the number of ways that miners can possibly differentiate their production of blocks. In short, those limitations are removing potential benefits over which miners can compete to provide to consumers.
This is the benefit side that so many are missing in their analysis.
And it alone, just like a cost-only analysis, is insufficient for understanding the mining market structure that will result in the long run.
Merging Cost and Benefit Arguments
What is required is a merger of the two sides of the market. Cost and benefit. Supply and demand. Production and consumption.
Analysis of only one side will never yield sensible conclusions about the final state of a market.
What’s required is an understanding that high barriers to entry (high costs,) lead to less firms capable of surviving in the market, and more competitive advantage opportunities lead to more firms capable of surviving in the market. Simultaneously, one could reverse those points. Lower barriers to entry lead to more firms capable of surviving, and less competitive advantages leads to less firms capable of surviving.
It should be recognized that “more” and “less” are vague, relative notions. It is not clear – and arguably impossible to clarify – the extent to which “more” and “less” apply in a given case. This sort of uncertainty should be familiar to anyone who has studied economics, as the same case applies in the analysis of shifts in supply and demand curves. It is, after all, the reason for the phrase “ceteris paribus,” which is to say, all other things being equal, which allows analysis of the effect of specific changes in the economy in isolation.
While we cannot determine an exact number of firms that can be carried in a market, nor their market share, under this framework we can start understanding what factors go into the end equilibrium firm carrying capacity of a market, and make comparative statements about what will happen under specific changes to a given market. We can also make educated guesses about how many firms may be carried under specific conditions over the long run based on how many sufficiently differentiated goods may be sold within the market being analyzed.
The Mining Market – Factors of Production
All production consists of the combination of multiple factors, such as labor, time, money, raw materials, and capital goods. The mining market is no different.
In mining, costs of production include labor, maintenance on equipment, hash power, node hardware, bandwidth, software development, warehousing, and various other business expenses.
For the purposes of this article, I will focus on hash power, software development, bandwidth, and node hardware, as these are the factors most closely related to protocol level market construction as discussed above. More specifically, these factors most affect the market differentiation opportunities, and the sources of capital and operational expenses.
To compete in the mining market, firms must invest in sufficient hash power to find blocks often enough that they have a steady revenue stream, thus crossing below their risk threshold. Additionally, they must be able to bring in revenue through block subsidies and transaction fees.
Over time, the block subsidy will be reduced in value until it is eventually eliminated, and is therefor not a long run revenue opportunity for miners. As such, the block subsidy will be ignored for the remainder of this article. If you wish to contest the conclusions based on this analysis because the subsidy is ignored, feel free to wait for another halvening and I’ll be happy to have that discussion then.
Market share in this model is not measured by percentage of blocks found, rather, by percentage of revenue derived from blocks. This is because miners don’t actually care about blocks, but about revenue and the costs to produce the blocks, ultimately resulting in their profit margin.
To maximize revenue, a miner must maximize blocks found and transaction fees derived.
To minimize cost, a miner must minimize resources (including work) required to mine blocks and accept transactions.
To maximize profit, a miner must optimize between their revenue and costs.
The Mining Market – Finding Competitive Advantages
As we have established, to remain sustainable over the long run, a miner must be able to differentiate themselves in the market, thus opening market segments that they can best serve, and manage costs effectively such that their profit margins are sufficient to avoid shutdown of the firm.
This combination, when achieved, is termed a competitive advantage. With a sufficiently differentiated market, multiple competitive advantages can be held in a single market, allowing each of the firms holding those advantages to survive in that market over the long run.
Firms competing over the same competitive advantage are essentially competing over specialization in providing that specific differentiated good, which reduces profit margins over all involved firms. Over the long run, only one firm can be sustained in providing that competitive advantage, though short run innovations can allow competing firms to remain relevant, potentially even displacing the leading firm.
The likelihood of more than two firms, one leader and one follower, surviving for very long in competition over the same competitive advantage is extremely low. More likely is that firms will seek differentiation, or shut down.
Fundamentally, there must be multiple competitive advantages available for multiple firms to survive in the long run.
Another factor worthy of noting is that competitive advantages tend to be mutually exclusive in that firms must invest in specializing to provide those specific differentiated factors that provide them a competitive advantage, and, while all funds are fungible, investment is not. Once resources are committed to a specific specialization, they are restricted by the generalization/specificity of the resources they have invested. For this reason, it is rare for one firm to be dominant over the long run in more than one market segment where multiple different competitive advantages are required.
An example: GPUs used for mining certain cryptocurrencies are more general purpose than ASICs, which are strictly specialized to hashing for a single algorithm. Purchases of ASICs are more specialized than purchases of GPUs, which may be more easily repurposed. Like all cases of specialization, ASICs are far better at performing their specific algorithm than GPUs are, so any miner repurposing GPUs to mine what the ASICs are specialized for will face a higher cost of mining. However, those investing in GPUs are far more agile in facing changing circumstances in the market, and may find that they can repurpose those GPUs into establishing new competitive advantages in the broader market.
In mining, market differentiation opportunities may include processing many small transactions, many chained transactions, or transactions with many inputs and outputs. It may also include processing large script or embedded data. It may also mean finding a balance between multiple different services, but with especially low costs of mining by focusing investment in efficient hashing machines.
Other potential specializations include operation as an archival node, or providing search functionality for address balance tabulation. These are not strictly competitive advantages in the mining market, but use more general purpose resources that allow the production of complimentary goods that may be dependent on specific protocol designs. For these firms, their competitive advantage lies in separate markets, and part of that advantage lies in having a foothold in the mining market as well.
The exact market share available for each differentiated market segment depends on the exact supply/demand for providing that particular service, and is expected to change as consumer demands change.
Each of these potential market differentiations requires specific, mutually exclusive investments. Purchasing hard drives to serve as an archival node means there are less funds available to purchase ASIC mining units. Investing in software to enable efficient search means there are less resources available to pay for bandwidth to receive unpruned (complete) transactions. Investing in ASICs requires a trade off in parallelization of transaction signature validation.
The Mining Market – Limiting Competitive Advantages
What happens when the protocol is designed in a way that artificially limits market differentiation opportunities?
Let’s compare a market with a blockchain that does not limit block size with one that has an artificial limit in place.
As a result of this protocol design choice, the limited blockchain cannot service all transactions in a single block. Accepting one transaction requires pushing out another from their block. Large transactions are particularly affected, as a single large transaction may fit with many small transactions, but multiple large transactions may not be able to fit in a single block.
The ability for miners to find a competitive advantage in processing large transactions has therefor been drastically limited.
Additionally, while small transactions aren’t affected as much as large transactions, it is conceivable that consumers will be forced to compete to get their small transactions into a block, too, thus raising prices and reducing consumer profits. In this case, the equilibrium for all transactions, large and small, is sub-optimal, and will result in dead weight loss in the market.
But there is one production input that remains unaffected entirely: that of the hash function.
Those miners specialized in providing cheap hash power will be benefited the most by this protocol limitation. Their competitive advantage never involved competing to include the most transactions of any type. Instead, they were always focused on finding the most blocks possible at the lowest cost to increase their market share.
Under this protocol limitation, some competitive advantages are eliminated, while a single competitive advantage is bolstered.
The result? Those miners focused only on providing cheap hash power increase their market share, and, because other miners must compete at a suboptimal equilibrium, other miners cannot afford to invest as much in hash power either, meaning hash power focused miners increase their percentage of blocks found, too.
Under this structure, and assuming a sufficiently low block size limit, there is only one optimal mining strategy – that is, to specialize in reducing hashing costs.
In a market with only one optimal strategy for profit maximization, the number of firms sustainable in that market is extremely limited. This is the market for Big Macs, not the market for Fast Food. Specialization is the name of the game, and vertical integration of producing blocks with the most hash power will lead to monopolization of the sector.
Monopolization, of course, is the enemy of “decentralization” in a Nakamoto Consensus model where 51% attacks exist. Under this model, only government protections of minority chains and miners can possibly prevent complete domination by a single entity in the long run.
In Bitcoin, there are only two possible protections from such a structure: ditch Proof of Work entirely, or open up the protocol to enable miners to differentiate themselves and establish a multi-miner equilibrium structure in the mining market.
The Mining Market – Achieving Maximum “Decentralization”
There is only one sound approach to maximizing the number of firms capable of being sustained in a market, and that is to reduce barriers to entry while increasing opportunities to differentiate in the market.
How might this be achieved?
Remove as many limits as possible on the protocol level while retaining integrity of the network itself.
Obviously, a broken network cannot support any service providers, and is therefore of no value to anyone. This is why it’s so important for a protocol to be “set in stone,” barring the discovery of bugs and significant flaws.
This allows firms to be established that compete over as many competitive advantages as possible.
The more competitive advantages that can be established, the less important it is for a firm to integrate all aspects of production (to specialize,) to be profitable, and the more important agility is in being able to find moment to moment optimums in the market.
Additionally, with less importance placed on integration of all factors of production, barriers to entry are actually reduced.
In the mining market, there may be hash power providers that nodes hire in order to provide needed specialization without giving up flexibility in the market, and without requiring up front investments to begin providing services to the market. It is, after all, much cheaper to buy a specific amount of hash power for a specific amount of time from someone who is specialized in providing it than it is to actually purchase all of the hardware, software, labor, and warehousing facilities required to efficiently provide your own hash power for your desired purposes. Roger Ver knows this fact well, considering his rental of hash power during the BCH/BSV hash war, though he also discovered that maintaining those costs over time has its own downsides.
This is also independent of the effect that general technological improvements will have on reducing barriers to entry over time. Reductions in the cost of production inputs such as bandwidth, computation, and storage will not only provide a constant disruptive force in the mining industry, putting large, inflexible firms at a disadvantage, but it will continually reduce barriers to entry for new firms. This aspect of the Red Queen Game trait of Bitcoin is another separate “decentralizing” force in the mining market that could easily take its own article to address in full.
First, if you’ve read this far – well done. I hope I’ve done a good enough job in explaining these subjects that it makes sense to you, and that I’ve made my case well enough that you at least see sense in my end conclusions.
To summarize a more than 4000 word piece that is itself a summary of rather deep concepts in economics, and to hopefully provide context for why I have written this novella, I’d like to clarify an important point:
The final segment of this piece, “The Mining Market – Achieving Maximum ‘Decentralization’,” is describing my ideal vision of Bitcoin. It is the version of Bitcoin that provides the most utility for the most people at the cheapest cost, and that just so happens to result in the most distribution of market share and block discovery in the mining market. In short, market power is most distributed, and consumers win.
This is why those who focus only on the cost side of the equation to advocate limitations on the protocol in the name of “decentralization” are wrong, and why Bitcoin should encompass everything it possibly can. The more ways to differentiate and compete in providing services in Bitcoin, the better.
This is the “why” behind my support of Bitcoin SV. It was the “why” behind my support of Bitcoin Cash before the BCH/BSV hash war. And, this piece as a whole makes the case for why BTC is not Bitcoin, and why it cannot actually succeed in achieving the goals its designers claim to be working towards.
I do not say what would be ideal because BSV happens to fit that description, rather, I support BSV because it does fit that description. I support Bitcoin SV only as long as it fits that description.
And if you find my arguments to be convincing, I hope you’ll begin working to build Bitcoin into what it should be, according to Satoshi’s original vision. The economics I describe in this article were clearly understood by Satoshi when he designed Bitcoin – they are, in my opinion, too complex and, oftentimes, counterintuitive, to be stumbled upon by accident.
Thank you for reading, Matthew Zietzke
P.S. If you're still not convinced, look at other markets in the world with high barriers to entry and high differentiation between products.
Two markets which are particularly interesting examples: US Domestic Airlines - https://www.statista.com/statistics/250577/domestic-market-share-of-leading-us-airlines/ Global automobiles - https://www.statista.com/statistics/316786/global-market-share-of-the-leading-automakers/
The airlines example is particularly interesting considering Airbus and Boeing are essentially operating in a duopoly for jetliner production, yet neither control the airline industry. This has interesting parallels to hashing machine production, as there is little differentiation possible over mining machines, but that does not imply that hashing machine producers will control the mining market, especially if competitive advantages are not limited to hash power, which would incentivize vertical integration in that market as opposed to flexibility by selling to multiple competing entities.