The benefit in the Marketplace of defining a maximum supply for Pocket Network has been expressed by many, such as in the following comment from @thedoc in the SER proposal discussions:
The original SER draft proposed an emission strategy going forward that would asymptote to a maximum supply of 3.0B POKT in the absence of app burn or other sinks of POKT. IT was decided by the authors to abandon that approach in favor of defining a more aggressive emission reduction strategy over the next twelve months that gets to single-digit inflation by the end of the year and forces the DAO to revisit emission strategy for V1.
Were the roughly 4% monthly reduction outlined in the final draft of SER be continued indefinitely, it would correspond to a maximum supply of just over 2.0B $POKT. Other monthly reduction levels correspond to different maximum supplies, as seen in Table 1.
Table 1 – Max POKT Supply as a natural consequence of small, stable monthly emission reductions.
|monthly emission||Corresponding||Corresponding Period||Other projects w/|
|reduction||max POKT||to reduce emissions||similar long-term|
|(percent)||supply (Billion)||by half (Years)||emission reductions|
|1.4%||3.00||4.1||BTC, ADA, LTC, BEAM|
Using the SER proposal as a mechanism to define a Maximum Supply proved problematic due to the inability to reconcile two incompatible design principles within the confines of a single proposal, namely the goal of not unreasonably constraining emissions in the future if the network encounters the need to rapidly grow, and the goal of bringing inflation to single-digit levels by the end of this year. If monthly reductions were set too small, such as the original 1.4% pre-proposal, the reductions were seen as too small as to be meaningful and resulted in taking too long to reach single-digit inflation. If monthly reduction were set to a higher value such as 4% it would not leave enough headroom to accommodate growing emission during a future growth phase.
The proposed solution to this conundrum is to separate actual from allowed emission.
An actual emission schedule does exactly what it’s name implies - dictates actual emissions. Examples include WAGMI, FREN and SER.
An allowed emission would sit above the actual and would trace what in the following I call an emissions “bound” or emissions “envelope” that is higher than and imposes constraints on the actual emission schedule. It represents what users and investors of the token can reasonably expect the supply never to exceed, even if actual emissions are not precisely known into the far future and even in the face of follow-on DAO governance actions.
Were the DAO to pass a proposal that constrains the allowed emissions (e.g., by defining a 3.0B maximum supply), it would be free to pursue any actual emission schedule it desired within that bound, but not above the bound. (Caveat: the DAO always retains the right to vote to revoke or modify an allowed bound but the goal would be to right-size the bound in the first place such that the DAO would never need to revoke or modify it in the future).
The logical method by which allowed governance actions are constrained is via the Pocket Network Foundation Constitution. For example, article 4.25 of the constitution constrains how validators may be rewarded and was invoked several times over the course of 2022 to test if one or another proposed governance action was allowable or not (see, e.g., PIP-22, PIP-25 discussions):
The course of action being investigated is to submit a proposal to amend the PNF constitution by adding a clause just after the above quoted clause. The wording is still being worked on, and is subject to community imput, but a draft laungage may be as follows:
4.26. As investors are the first building block upon which Pocket is built, care must be taken not to dilute investors’ token holdings via unlimited expansion of the token supply. Toward that end, the total supply of $POKT shall not exceed 3.0 billion $POKT. The Council must never approve Protocol Upgrades that would cause the token supply to grow in any one month by more than ten percent of the difference between the then current supply and the maximum supply of 3.0 billion $POKT.
Note that it is important that the language of this clause reference supply growth and not emissions. In the future when Pocket Network handles hundreds of billions of paid RPC requests. emissions is likely to grow to be very large, but when balanced by app burn, the supply growth will be small, zero or even negative (i.e., POKT become deflationary).
The proposed value of 3.0B as a maximum supply is strategic in that it allows for the claim that over 50% of possible tokens have already been minted while giving so much space for growth of net emissions in the future, if such growth is ever needed, as to cause the constrained to never be felt in practice.
Current supply as of March 1 is expected to be ~1540MM tokens. This is 51% of the Maximum Supply if Maximum Supply is set to 3.0B. This would position Pocket to be in good company with several other tokens such as LINK and NEAR, as can be seen in Table 2 that was complied by @caesar for the SER proposal and copied over here. This provides a much cleaner narrative for reporting services such as Messari, etc. than reporting that the token supply is unlimited.
At the same time, the constraints placed by the above clause are never likely to get tested. Every way that I have sliced and diced the numbers leads to the conclusion that the constraint is nominal only and will never get tested in practice, as illustrated below.
TEST 1: COMPARISON OF CURRENT EMISSIONS TO EMISSIONS CONSTRAINT AT END OF SER
Supply as of March 1 of this year is approximately 1540MM. IF SER passes, anticipated supply at the end of the one year is 1740MM. Emissions at the end of the SER period would be 0.42MM per day or 12.8MM per month. Now compare this to the constraint imposed by the above proposed clause. The headroom between the then current supply at the end of SER and the max supply is anticipated to be 1260MM POKT (3000MM – 1740MM). 10% of this value is 126MM POKT. Thus, the constraint imposed by the above clause on monthly net emissions is anticipated to be ten times greater than the actual emissions at the end of SER. While fluctuation in emissions may be desired to implement Vitaly’s ideas on demand-centric emissions and to accommodate growth, it is not anticipated that Pocket will ever need to go back to 10x the current level of emissions.
TEST 2: DURATION ALLOWED AT 420K/DAY EMISSIONS BY THE EMISSIONS CONSTRAINT
If 420k daily emissions were to continue indefinitely into the future following the 12 months of SER, there would come a point in the future, as POKT supply got close to the maximum supply, that the 420k/day emission would itself start to exceed the constraint imposed by the above clause. This happens when the supply reaches 2872MM POKT (3000MM – 2872MM = 128MM and 10% of this is 12.8MM). This would not happen for over 7 years. In other words, Pocket Network could continue until summer of 2031 at 420K/day emissions in the absence of app burn before being forced to further reduce net emissions by the above constraint. Even in the very worst-case scenario, app burn will turn on long before this and render this to be a moot point.
TEST 3: DURATION ALLOWED AT CONSTRAINT ENVELOPE
Suppose, on the other hand, that following the conclusion of SER, the decision was made to turn on emissions to the maximum extent allowed by the above clause – that is, to cause actual supply growth to trace the envelope defined above (10% per month of the headroom between current supply and max supply). While this is somewhat of an artificial construct, it helps inform the discussion to know how long the network could be run at “full throttle” in the absence of app burn before emissions came back down to the 420k/day level. As can be seen in the hypothetical emission schedule shown in Table 3 - where emissions are ramped up almost 10x at the conclusion of SER - the network could be run at the full throttle allowed by the proposed clause until January of 2026 before emissions reduced back down to end-of-SER levels. All while never endangering breaching the 3.0B defined maximum.
Actual POKT supply will, in practice, most likely never approach anywhere close to 3.0B POKT.
SER, if continued indefinitely traces to just over 2.0B. Ending reduction measures at the end of SER will cause supply to grow linearly rather than asymptote to any value, but at a rate of only 0.15B added POKT per year. App burn will turn on long before 3.0B is even on the horizon. It is likely that POKT supply will never exceed 2.2 to 2.3B POKT, no matter what tokenomics policy is employed going forward. Imposing a max Supply of 3.0B is safe from the perspective of not unduly tying DAO hands in governance actions it takes in the future. Enforcing the integrity of the Maximum by not allowing any governance action that would cause supply to grow in any month by more than 10% of the remaining headroom between current supply and max supply is a stable and often used methodology and is also free from imposing any practical constraint on likely desired DAO action.
The only downside to defining a Maximum Supply of 3.0B is that it plants a number in peoples head that is likely larger than it needs to be, potentially giving an impression that POKT is more inflationary that it actually is. Defining a Maximum of 2.5B rather than 3.0B is better in this regard, but is more constrictive in the case the DAO finds the need to rapidly expand emission for a period of time in order to grow the network. Further, the narrative of 2.5B is only slightly better than 3.0B (current supply sits at 62% of max supply for Max of 2.5B vs 51% of max supply for a max of 3.0B). The goal is to find the right balance between controlling the narrative of POKT in the marketplace while being the least constrictive of potential future DAO actions.
Feedback from the community is solicited on this approach to defining a Max Supply for POKT (i.,e the use of constitutional amendment to constrain allowed governance action so as to ensure a max supply) as well as on the actual values being suggested.