Tokenomics: supply schedules, emissions and dilution

Almost every token has a document describing how many units will exist, who receives them and when. That document is the closest thing a token network has to a capital structure, and it is read far less carefully than a bond prospectus or a cap table. The reason is that the numbers look reassuringly precise. A fixed cap, a chart of monthly unlocks and a pie of allocations all suggest that the supply question has been settled.

It has not. A supply schedule tells you how the denominator of every per-token figure will change over time, and nothing at all about the numerator. The work is in connecting the two: understanding who is being diluted, by how much, in favour of whom, and whether the thing being bought with that dilution is worth the price.

Three supplies, and why the difference matters

Start with vocabulary, because the three common measures are routinely used interchangeably by people who should know better.

Circulating supply is the number of tokens that are transferable and not subject to a lock. It is the supply that can be sold today.

Total supply is every token that has been created, including locked treasury, team and investor allocations.

Maximum supply is the ceiling written into the issuance rules, if one exists. Bitcoin has one at 21 million. Ether does not.

From these come two headline numbers. Market capitalisation is price multiplied by circulating supply. Fully diluted valuation, or FDV, is price multiplied by maximum or total supply. The gap between them is where most of the trouble lives.

Take a token trading at CHF 2.00 with 100 million circulating and 1 billion total. Market capitalisation is CHF 200 million; FDV is CHF 2 billion. Ten per cent of the eventual supply is doing all the price discovery. That is a thin float in exactly the sense the term is used in equities, and thin floats price high for a mechanical reason: the marginal buyer competes against a small amount of available paper, while the eventual seller base is ten times larger.

Now suppose over the following year the circulating supply rises from 100 million to 130 million through scheduled unlocks, and the total amount of money willing to hold the token stays flat at CHF 200 million. The price becomes 200 ÷ 130 = CHF 1.538, a fall of 23.1 per cent from CHF 2.00. Nothing about the protocol changed. The denominator did.

Issuance rate is not dilution rate

These two are confused constantly and they are not the same number.

Continue with the example. Thirty million new tokens issued against an opening supply of 100 million is an issuance rate of 30 per cent. But a holder who owned 1 million tokens, or 1 per cent of the opening supply, now owns 1 million out of 130 million, which is 0.769 per cent. Their claim on the network has fallen by 23.1 per cent, not 30 per cent. Dilution measured against the closing supply is 30 ÷ 130 = 23.1 per cent; issuance measured against the opening supply is 30 per cent. Both are correct statements about the same event. Say which one you mean.

The same distinction settles a persistent confusion about staking rewards. Suppose total supply is 100, annual issuance is 8 tokens, all of it paid to stakers, and 60 tokens are staked. A staker holding all 60 receives all 8 and ends with 68 out of a closing supply of 108, which is 62.96 per cent, up from 60 per cent. In token terms the staker earned 8 ÷ 60 = 13.3 per cent. In ownership terms, the share that actually matters, the staker gained 62.96 ÷ 60 = 4.94 per cent. The non-staker holding 40 tokens now holds 40 out of 108, which is 37.04 per cent, a loss of 7.41 per cent of their claim.

So a headline "13.3 per cent staking yield" is, in this design, a 4.94 per cent real gain in ownership funded by a 7.41 per cent transfer away from those who did not stake. It is not income from an external source. It is a redistribution among holders, and the advertised rate is high mainly because participation is low. The lesson on where yield genuinely comes from takes this further; the point here is purely arithmetic.

Emission schedules: the shapes and what they are for

Bitcoin's schedule is the reference case. The block subsidy started at 50 BTC and halves every 210,000 blocks, roughly every four years: to 25 in 2012, 12.5 in 2016, 6.25 in 2020, 3.125 in 2024. The geometric decay converges on 21 million and the subsidy falls to nothing around 2140, when the halved amount drops below one satoshi. The design intent is that early issuance pays for bootstrapping security and adoption, and that transaction fees eventually replace it. Whether that handover works is the subject of the next lesson.

Ethereum shows the opposite architecture: no cap, but a variable net issuance. Since EIP-1559 took effect in August 2021, the base portion of each transaction fee is burnt rather than paid to the block producer. Since the move to proof of stake in September 2022, gross issuance to validators is a small fraction of what proof-of-work mining paid out. Net issuance is therefore gross issuance minus burn, and can be negative in periods of heavy on-chain activity. A cap is not the only way to constrain supply; a burn linked to usage is a different mechanism with different properties, because it makes supply endogenous to demand.

A third shape is the tail emission: a fixed number of new units per block forever. The absolute issuance is constant, so the percentage rate falls asymptotically toward zero as the base grows, while the security payment never disappears. This is a deliberate answer to the fee-only-future problem rather than an accident.

Cliffs, vesting and the unlock calendar

Allocations to teams and early investors are the part of the schedule that most often produces a visible price event, because the release is lumpy rather than smooth.

A standard structure is a one-year cliff followed by three years of monthly vesting. Take a 20 per cent team allocation on a 1 billion total supply, which is 200 million tokens, vesting over 48 months from launch with a 12-month cliff. On the cliff date, 12 ÷ 48 of the allocation vests at once: 50 million tokens. The remaining 150 million releases over 36 months at 150 ÷ 36 = 4.167 million per month.

Set that against a circulating supply of 100 million. The cliff alone is a potential 50 per cent increase in circulating supply on a single day. The monthly tranche thereafter, at a CHF 2.00 price, is 4.167 million × 2.00 = CHF 8.33 million of notional. If daily trading volume is CHF 5 million, the monthly unlock is worth about 1.67 days of volume.

That comparison is useful but it flatters the token, because volume is not depth. Volume counts every transaction over a day, including the same units changing hands repeatedly; depth is how much can be sold at one moment before the price moves materially, and it is typically a small fraction of daily volume. The honest reading is that an unlock of that size relative to the book is capable of setting the price for weeks, and that vested holders with a cost basis near zero have a very different reservation price from someone who bought at CHF 2.00.

Note also that vesting and lock-ups are protocol-level or contract-level promises about transferability. They do not stop an insider from hedging exposure through a derivative or a forward sale of the eventual delivery. A schedule constrains tokens, not economic positions.

Reading a schedule without being managed

A few habits separate a real reading from a marketing one.

Convert everything to a single unit before comparing. Percentages of "total supply", "circulating supply" and "supply at the end of the vesting period" are three different denominators, and allocation tables routinely mix them.

Extend the calendar past the horizon shown. Charts of unlocks often stop at 24 or 36 months. The interesting question is what the supply looks like at the end.

Ask what the emission buys. Issuance directed at validators purchases security. Issuance directed at liquidity mining purchases rented deposits that leave when the rate falls. Issuance directed at a treasury purchases optionality for the people who control the treasury. These are not the same expenditure.

Check whether burns are structural or discretionary. A burn funded by protocol fees is a real reduction in net issuance. A burn announced from the treasury is a transfer from one pocket to another that removes nothing the treasury could not re-issue.

Finally, treat the schedule as amendable unless there is a specific reason it is not. Supply parameters that can be changed by a governance vote are policy, not physics, and should be valued as policy.

The Swiss holder's angle: valuation dates and wealth tax

A supply schedule interacts with Swiss taxation in a way that has nothing to do with the protocol's design.

Switzerland levies an annual wealth tax at cantonal and communal level on worldwide net assets, valued at 31 December. Tokens are part of that base. The Federal Tax Administration publishes year-end tax values for a set of widely traded cryptocurrencies in its official price list, and its guidance for tokens without a published value directs the taxpayer to the year-end price on the platform where they trade, falling back to acquisition cost where no market price exists.

Two consequences follow for anyone holding an early-stage token. First, the tax bill is driven by a single date, so a position that spikes into year-end and collapses in January is taxed on the spike. Second, rates vary substantially across cantons and communes, which is why asset location within Switzerland is a live question for people with concentrated illiquid holdings in a way it is not in most neighbouring countries.

Locked or unvested allocations are the genuinely awkward case, because a claim that cannot be transferred still has value but no reliable market price. That is a matter for a tax adviser and, in practice, for a ruling with the cantonal authority rather than for a rule of thumb.

What to take away

• Issuance measured against opening supply and dilution measured against closing supply are different numbers for the same event: 30 million new tokens on a base of 100 million is 30 per cent issuance and 23.1 per cent dilution, and a document that does not say which it means is not telling you much.
• A high advertised staking rate is usually a redistribution among holders rather than external income, and the rate is high precisely because participation is low.
• FDV is not a valuation, it is a warning: it tells you the market capitalisation implied if the price survives every token that has yet to reach the float.
• A cliff unlock is the single most predictable supply shock in the asset class, and it should be measured against market depth rather than against reported trading volume.
• Swiss wealth tax fixes the value of a token position on 31 December, which makes a volatile holding's tax cost dependent on one day's price rather than on the year's average.

Sources

• Satoshi Nakamoto, Bitcoin: A Peer-to-Peer Electronic Cash System (2008), for the halving schedule and the 21 million cap.
• EIP-1559, Fee market change for ETH 1.0 chain, Buterin, Conner, Dudley, Slipper, Norden and Bakhta (Ethereum Improvement Proposals, 2019).
• Arvind Narayanan, Joseph Bonneau, Edward Felten, Andrew Miller and Steven Goldfeder, Bitcoin and Cryptocurrency Technologies (Princeton University Press, 2016).
• Andreas Antonopoulos and Gavin Wood, Mastering Ethereum (O'Reilly, 2018), on issuance and account mechanics.
• Swiss Federal Tax Administration, working paper on cryptocurrencies and initial coin offerings, and the annual official price list (Kursliste) of year-end tax values.
• Bank for International Settlements, Annual Economic Report 2022, chapter on the future monetary system, for the institutional framing of token issuance.