One popular argument compares the Bitcoin block size limit to the coin production schedule that sets up a terminal maximum of 21 million bitcoins that can ever be created. Raising the block size limit, this argument continues, could set a precedent for changing the coin production schedule, and then what? Changing the block size limit opens up a slippery slope that could threaten to lead to the end of cryptocurrency standards and boundaries. Just as the coin limit is an essential value proposition of Bitcoin, so other types limits must be conservatively protected as well.
How can this type of argument be considered?
First, note that this represents an approach opposite to the one I have taken. I have identified and discussed the block size limit as something uniquely and importantly different within Bitcoin from an economic standpoint. The above argument, in contrast, presents these different “limits” as quite similar to one another for this purpose and therefore ripe for analogizing.
Next, one might note how Bitcoin started with its production schedule already in place, whereas the block size limit was added about 20 months later and at just under 1,200 times larger than the average block size of the time. The limit’s original proponents defended it from critics as a merely temporary measure and thus of no real concern.
A common retort to such observations is, in effect, “that was then, this is now.” The project is at a more advanced stage. The current developers have more experience and a more mature view than the early pioneers. The system now carries far more value and the stakes are higher. Today, we can no longer afford to be so cavalier as to just put a supposedly temporary limit right into the protocol code where it could prove difficult to change later…
That is…we can no longer be so cavalier as to just remove such a previously cavalierly added temporary limit...That is…it is time to move on from reciting old founder tales and look to the present concerns.
And indeed, such matters of historical and technical interpretation are subject to many differing assessments. However, there is an altogether different and more enduring level on which to consider this matter. There are substantive economic distinctions between a block size limit and a coin production schedule that render the two remarkably different in kind and thus weaker objects for analogy than they could at first appear.
When “any number will do” and when it will not
This is because raising the total quantity of a monetary unit by changing its production schedule has completely different types of effects from changing the total quantity of a given service that can be provided. Producing an increased quantity of a given cryptocurrency is entirely unlike producing an increased quantity of transaction-inclusion services. This follows from a unique feature of monetary units as contrasted with all other economic goods and services. An arbitrary initial setting for the production of new coins (which operates to define an all-time maximum possible production quantity) works quite well for a cryptocurrency, but does so only for unique and distinctive reasons.
With money, barring certain divisibility issues of mainly historical interest, any given total quantity of money units across a society of users facilitates the same activities as any other such total quantity. This includes mediating indirect exchange (facilitating buying and selling), addressing uncertainty through keeping cash balances (saving; the yield from money held), and facilitating lending and legitimate commercial credit (not to be confused with “credit expansion”). The particular total number of money units across a society of money users is practically irrelevant to these functions. What is critical to a money unit’s value is users’ confidence that whatever this total number (or production schedule) is, money producers cannot arbitrarily alter it, especially upward, so as to rob money holders through devaluation.
A hypothetical model of physical commodity money production on a free market differs in certain important respects from both cryptocurrency and fiat money and bank-credit models. We should therefore closely consider the meaning of arbitrary with regard to these distinct cases.
With precious metal coins produced by ordinary businesses on a free market, the number of units cannot be increased arbitrarily for reasons rooted directly in physical constraints. Each additional precious metal coin to be produced requires specific scarce materials and energy combined with various manufacturing and other business costs, from mining to minting. Each such coin is much like any other good produced and exchanged on the market in that it is a product to be used in the market as money as opposed to a product to be used in the kitchen as dinner. Material scarcity itself protects money users from rouge money producers by preventing arbitrary changes to the quantity of money units. Changes in quantity supplied reflect supply and demand for such coins, including marginal production costs, as with other products.
In sharp contrast to this, a state-run system of fiat money and bank credit supports “flexible” increases in the “money supply.” These are arbitrary in that, unlike hypothetical commercial precious metal coin makers, these legally privileged money producers can generate additional money units at little to no cost to themselves. Notes can be printed and differing numbers of zeroes can be designed into printing plates as the denomination at no difference in printing cost. Likewise, cartel-member bankers can issue “loans” of nothing, filling customer accounts with what has been aptly described as “fountain pen money,” limited to a degree by the current policies and practices of those managing the banking cartel (“regulators,” etc.). Legal frameworks provide some protection for users of such money, most of the time (except when they do not), but such protections are far weaker and less reliable than those from the harder constraints of mineral reality.
Against this backdrop, some cryptocurrencies, led by Bitcoin, feature a novel and innovative third way to protect money users from arbitrary increases in new add-on supply. A production schedule can be specified within the effective definition of what a given cryptocurrency is.
Now in considering the exact number of possible units of a given cryptocurrency, consider two almost identical parallel universes, A and B, which differ in only one respect. Assuming sufficient divisibility in both cases (plentiful unit sub-division is possible), 30 widgetcoins out of a 300-trillion widgetcoin supply across a given society in Universe A carry the same purchasing power as 60 halfwidgetcoins out of a 600-trillion halfwidgetcoin supply across a given society in Universe B.
In each universe, one can buy the same kilogram of roast beef, in one case with 30 units, in the other with 60. Since the 300-trillion versus 600-trillion total money supply is the only difference between these two universes, it makes no difference whether the roast beef is bought with 30 units in Universe A or with 60 units in Universe B. Since the people in the two universes are wholly accustomed to their own respective numerical pricing conditions, their psychological and felt interpretations of the value associated with “30” in the one case and “60” in the other, are likewise indistinguishable.
Naturally, many individuals and organizations in any universe dream of having “more money.” For example, considering that 20 units of a good is worth more than 10, it is easy to equate having more units with having more wealth. Twenty good apples represent an amount of wealth (ordinally) greater than 10 such apples do. This is also the case with holding quantities of the same monetary unit. Twenty krone represents more wealth than 10.
But the crucial point now arrives: the foregoing “more is better” with regard to money applies to the number of units in a given party’s possession, but does not apply—as it does with ordinary non-money goods and services—to the wealth of the society of money users as a whole. Viewed across an entire society, intuitive associations from personal and business experience between larger numbers and greater wealth do not translate into a way to raise overall wealth. Political funny-money schemes with names such as “monetary policy” and “credit expansion” instead produce only sub-zero-sum transfers of wealth from some monetary system participants to others. Such transfers produce win/lose results in which some gain at the expense of others, not to mention the additional net losses from the transfer process itself (thus sub-zero-sum).
With Bitcoin, when the initial design was set—but not afterwards—42 million units, or other possible numbers, would have been as serviceable as 21 million. After the system launched, however, no general benefits could follow from increasing the quantity of possible bitcoins beyond their initially defined schedule. Such a later increase would instead tend to 1) reduce the purchasing power of each unit below what it would have otherwise been, 2) transfer wealth to recipients of new add-on units away from all other holders of existing units, 3) raise uncertainty about the coin’s reliability, likely depressing its market value with an uncertainty discount, 4) create demand for an analog of a “Fed watching industry” that speculates on what might happen next with the malleable production schedule, and 5) give rise to an industry of lobbyists, academics, and other experts dedicated to influencing such decisions.
While the block reward framework does indeed also “transfer wealth” in a sense to miners from existing bitcoin holders as in item (2) above, it crucially does so only in a predefined way, knowable to all participants in advance. The block reward schedule, defined before launch, provides a form of compensation for mining services in the system’s early days. This has enabled the system to evolve and succeed from its launch to the present. This follows not from any arbitrary change to the production schedule, but merely from the ongoing operation of the production schedule initially set.
One free pass only
In sum, a peculiar characteristic of money units when viewed across an entire society of money users provided a one-time and unique economic free pass for setting an arbitrary number of possible bitcoins at 21 million. This free pass could only be valid before initial launch (prior to 2009, or at the very latest, prior to the evolution of any tradable unit value). Changing the schedule later, especially in such a way as to increase unit creation, would have completely different and wholly negative effects from a systemic perspective.
Now returning to non-money goods and services the case is quite different again. The foregoing unique monetary free pass is entirely absent, whether after launch or before it. When non-money goods and services are likewise viewed at the level of a given society as a whole, “almost any number will do” does not apply. An increased total quantity of a non-monetary good or service supplied can be in the general interest, not only in special interests. It can be win/win and not win/lose. If there are more apples or cattle to go around in a given society (as opposed to just more pesos), this does tend to lower the costs of acquiring those goods in a meaningful way. This does enhance wealth in society, not just transfer it around. It represents a real increase in production, not just a “flexible” money fraud as in the case of arbitrary inflation on the part of money producers.
Miners provide one such ordinary “non-money” service when including a given transaction in a candidate block. This is a scarce service provided (or not) to a specific end user by specific miners. It does not fall under the unique category of the total number of monetary units in a society of money users. The total possible number of bitcoins, however, does fall under this unique category. The two numbers differ in kind and for that reason make poor objects for analogy. Both may, indeed, be viewed as “limits,” but it is important to recognize the contrasting economic roles and natures of these two types of limits.