A token buyback is a protocol or its treasury spending income to repurchase its own token on the open market. Unlike a burn, the purchase itself is a demand-side act: real money crosses the order book and someone sells into it. What happens to the tokens afterwards is a separate decision, and destroying them is only one option. Placeholder's Joel Monegro argued in 2020 that it is usually the wrong one.
The buyback does the work. The burn is a bookkeeping choice made afterwards. Monegro's line is that what affects the price is how many units participate rather than how many exist, and units sitting in a treasury already do not participate.
Scroll to see the full diagram
Two operations, usually discussed as one
Buyback and burn is a compound, and taking it apart is most of the analysis. The buyback spends income, usually held in another asset, to purchase the native token on an open market. The burn destroys what was purchased. Trade coverage describes the pair as one mechanism intended to grow token value by reducing supply as income grows.2
The two halves behave differently. A buyback is funded, contested and price-forming: someone has to sell to the protocol, and the protocol's bid competes with every other bid. A burn is unilateral and costs nothing at the moment it happens, because the protocol already owns the tokens.
So when a team tells us their buyback and burn creates buy pressure, only half the sentence is doing any work, and it is not the half in the headline.
Currency or capital, and why the answer changes the design
Monegro opens with a test that is worth adopting whole. An asset is currency when its value comes from exchange, from being spent to consume goods or services. An asset is capital when its value comes from governance or equity participation in a pool of resources. When a network generates income in one currency and redistributes that value to its token holders in any way, the token is a capital asset, because its fundamental value comes from those flows.1
The distinction matters because reducing supply does different things to the two. For money, reducing the supply can raise the unit value. For capital, issuance is how the system capitalises itself, and destroying units gets in the way of growing fundamental value.1
Most tokens running buyback programmes are capital assets by that test, since the buyback is itself a redistribution of income to holders. Which means the deflationary argument being made for the burn is borrowed from the wrong category.
What the equity analogy actually shows
In public equities, bought-back shares are not automatically destroyed. By default they sit with the company as treasury shares, and treasury shares cannot vote or participate in the economics of the organisation. Monegro's conclusion is that the buyback alone does the work, because what affects the price is how many shares participate rather than how many exist.1
Carry that across and the burn step looks thinner than its reputation. The tokens the protocol bought are already out of circulation in every sense that matters to a holder's proportional claim. Destroying them changes a number in a block explorer.
The governance argument for burning is that it prevents the tokens being reissued. Monegro answers that the people with the power to burn are usually the same people with the power to issue, so what actually controls supply is the governance protocol and the social contract around it. His example is Maker, which burns MKR as it earns income and issued new MKR during a solvency emergency in March 2020.1
Buyback and make, the alternative he proposed
The alternative keeps the purchase and drops the destruction. Monegro's proposal is a protocol-owned automated market maker, a Balancer smart pool controlled by the protocol's own contracts, acting at once as a buyback machine, a token issuance pool and a liquidity provider.1
His worked configuration is concrete. Take a pool indexed at 90% of the native token and 10% of the currency the network earns in, with a pre-minted supply deposited into it and only the protocol able to add or remove liquidity. Income is deposited into the pool. Whenever the currency side exceeds its 10% weight, the pool rebalances by selling the excess for the native token on the open market until the 90/10 index is restored, which is a buyback executed continuously by arbitrageurs rather than by a keeper bot the protocol has to fund.1
Run in reverse it is a fundraising line: withdrawing currency from the pool has the same effect as issuing the token and selling it. What the protocol accumulates stays usable for incentives, liquidity or collateral, which is the whole argument. A bought-back token is an asset. A burned one is not.
Uniswap is running the experiment
The largest live test of the opposite position started in late 2025, when Uniswap governance approved the UNIfication proposal and turned on the protocol fee switch. Coin Metrics describes the architecture as a pipes model: fees from v2, v3 and Unichain split between liquidity providers and the protocol, the protocol's share collects in a vault contract called TokenJar on each chain, and value can leave that vault only by burning UNI through a contract called Firepit.3
The numbers from the first read, published 13 January 2026 and covering twelve days, are these. Roughly 0.8 million USD of cumulative protocol fees, implying an illustrative annualised run rate of about 26 to 27 million USD if conditions held. About 100.17 million UNI burned in total, worth roughly 557 million USD, or about 10.1% of the original one billion supply, most of it a retroactive one-off. Ongoing burns running at roughly 4 to 5 million UNI a year.3 The authors present all of it as early and illustrative, and it should be read that way.
What makes this worth watching is that burning is not one option among several in that design. It is the only exit from the vault, by construction. Uniswap has taken the side of this argument that Monegro wrote against, and made it structural.
What we ask before signing off a buyback
Four things, and the first two decide the rest. Name the income that funds it, in the asset it arrives in, and show it is revenue rather than a treasury balance being spent down. Name the rule that triggers the purchase, and prefer one that executes without anyone deciding at the moment it fires, since a discretionary buyback is a trading desk and it will eventually be traded against. Then decide, separately and in writing, what happens to the tokens. And say plainly whether the programme continues in a quarter where revenue falls.
Uniswap's own governance forum established the principle years before any of this shipped: turning on fee retention does not create any expectation that the retained tokens will be paid out to holders.4 Retention, buyback and distribution are three separate decisions, and treating them as one is how a treasury policy gets read as a commitment.
The buyback is plumbing for a business that earns money. Where there is no income, a buyback is the treasury selling its runway to support its own price, and that is a decision about survival rather than a mechanism worth designing.
Common questions
What is a token buyback?
A token buyback is a protocol or treasury using income to repurchase its own token on the open market. The purchase is the demand-side event, since real money crosses the order book and someone sells into it. What happens to the tokens afterwards is a separate choice: they can be burned, held, redeployed as incentives, or used as liquidity or collateral.
Is buyback and burn good for a token?
The buyback and the burn have to be judged separately. Joel Monegro argued the buyback does the work, because what affects price is how many units participate rather than how many exist, and treasury units already do not participate.1 He also noted that burning does not guarantee protection against dilution, since whoever can burn can usually also issue, as Maker did in March 2020.1
What is buyback and make?
Buyback and make is Monegro's 2020 alternative to buyback and burn. A protocol-owned Balancer smart pool, in his worked example indexed at 90% native token and 10% of the currency the network earns in, absorbs income and automatically rebalances by buying the token from the market. The protocol keeps what it buys and redeploys it for incentives, liquidity or collateral instead of destroying it.1
Does a buyback create buy pressure?
The purchase does, for as long as it is funded. That is the honest scope of the claim. The size of the effect depends on the income behind it and on how it lands relative to liquidity, and the burn that usually follows adds nothing to it. A buyback funded from a shrinking treasury rather than from revenue is spending runway to support a price.
See Tokenomics Design for how this applies in practice.
Sources
- Stop Burning Tokens - Buyback and Make Instead
Joel Monegro, Placeholder, 2020
The currency-versus-capital test, the treasury-share analogy and the participation argument, the point that burning does not guarantee protection against dilution with Maker's March 2020 MKR issuance as the example, and the buyback-and-make design using a protocol-controlled Balancer smart pool indexed 90/10. - What is a token burn? How buyback-and-burn works
crypto.news
Trade explainer of buyback-and-burn mechanics, cited for how these programmes are commonly described rather than for mechanism analysis. - Uniswap Flips the Fee Switch: From Governance Token to Value Accrual, State of the Network #346
Tanay Ved, Coin Metrics, 2026
Published 13 January 2026. The TokenJar and Firepit architecture in which value can leave the vault only by burning UNI, roughly 0.8 million USD of protocol fees over the first 12 days, an illustrative 26 to 27 million USD annualised run rate, about 100.17 million UNI burned (roughly 557 million USD, about 10.1% of the original one billion supply), and ongoing burns of roughly 4 to 5 million UNI per year. Presented by the authors as early, illustrative data. - Fee Switch Design Space and Next Steps, thread 17132
Uniswap Governance Forum
States that turning on fee retention does not create any expectation that the retained tokens will be paid out to UNI token holders.
Last reviewed 2026-08
Know the terms but not sure how they apply to your project? That is what an engagement is for. We design, document, and stress-test the whole token economy inside the Tokenomics Data Room.
100+ projects advised. Complete tokenomics in 4 to 6 weeks.