Where Staking Yield Actually Comes From

Staking yield comes from three places: newly issued tokens, transaction fees paid by users, and value extracted from transaction ordering. They are usually quoted as a single percentage, which hides the fact that they behave completely differently. Issuance is paid by every holder through dilution. Fees are paid by users who wanted something done. Ordering revenue is paid by traders who did not know they were paying. Knowing which one dominates a given return tells you more about its durability than any historical APR chart.

Key takeaways

• Issuance is not income. It is a transfer from holders who do not stake to those who do, and it leaves total value unchanged.
• Transaction fees are the only component genuinely paid in from outside the token supply.
• Ordering revenue is real but volatile, concentrated in a few periods, and unevenly distributed between validators.
• A protocol that burns fees hands its stakers a smaller headline number and a stronger claim on real economic activity.
• The question to ask about any advertised rate is not how large it is, but which of the three it mostly consists of.

Issuance: the component that is not income

Proof-of-stake networks pay validators in newly created tokens. The rate is set by protocol rule rather than by market demand, and it exists to make honest validation worth the capital it locks up. It is the largest component of most advertised staking rates and the least meaningful.

The reason is arithmetic. If a network issues new tokens and distributes them to stakers, the stakers' share of total supply rises and everyone else's falls. Nothing has entered the system. A staker earning the issuance rate exactly is not richer in proportional terms; they have merely avoided the dilution imposed on holders who did nothing. That is a defensible reason to stake, but it is a defensive one, and it is not what the word yield normally implies.

The practical consequence is that issuance-heavy returns are close to a floor rather than a return. If a network issues at a given rate and you are earning roughly that rate, your position relative to other holders is flat. To be genuinely ahead you need one of the other two components, or you need the network's total supply to be doing something other than growing.

Transaction fees: the only external payment

Fees are different in kind. Somebody wanted a transaction included, valued that inclusion, and paid for it out of tokens they already held. No new supply is created. This is the component that behaves like revenue in the ordinary sense: it scales with demand for the network's blockspace and disappears when nobody wants to use it.

Fee revenue is therefore the component worth watching if you are trying to judge whether a staking return survives a bear market. Issuance continues regardless — it is a protocol constant. Fees collapse when activity collapses. A network whose staking return is almost entirely issuance will publish a stable-looking rate through a downturn while the economic activity underneath it evaporates, which is exactly the moment the rate is least informative.

Some networks burn a portion of fees rather than paying them to validators. This lowers the headline staking rate and raises the quality of what remains: burned fees reduce supply, which benefits every holder including the staker, rather than transferring value between them. A protocol that burns aggressively will look worse on a yield comparison table and may well be paying its stakers better in real terms.

Ordering revenue: real, volatile, unevenly shared

The third component comes from the right to decide what order transactions are processed in. That right has value, because certain orderings are profitable — placing a trade immediately before a large known trade, for instance, and closing it immediately after. Validators can capture some of this value, either directly or by selling the ordering right to specialists who bid for it.

Three things make this component awkward to reason about. First, it is lumpy: it concentrates in periods of high volatility and is close to nothing in quiet markets, so an annualised figure derived from a busy month is misleading. Second, it is not evenly distributed — validators connected to a competitive market for ordering rights capture far more than those that are not, which means two validators on the same network can deliver materially different returns from identical stake. Third, the money is ultimately coming from other users, generally traders who received worse prices than they otherwise would have.

That last point is worth sitting with. Ordering revenue is genuinely external to the token supply, like fees, and it is not obviously a payment for a service rendered. If a large share of a staking return comes from this source, the return is real, and it is being funded by a transfer from people trading on the network.

Why the split changes what you should compare

Once you separate the components, several common comparisons stop making sense.

Comparing headline rates across networks compares three different things bundled in three different proportions. A high-issuance network with little activity can advertise a larger number than a low-issuance network processing enormous volume, and the second is the healthier position.

Comparing a staking rate to a savings rate is a category error for the issuance component specifically. Interest on a deposit is paid by a borrower. Issuance is paid by other holders of the same asset, which is not the same relationship and does not carry the same expectation of continuation.

Comparing providers on the same network is where the split becomes genuinely useful. Issuance is identical for every validator that stays online, so any difference between providers on the same network comes from fee capture, ordering revenue, and commission. That makes it a fair comparison — and it is the comparison most people skip in favour of the cross-network one that is not.

What to check before you commit capital

• Find the protocol's issuance schedule. It is published, it is a rule rather than a forecast, and it sets the floor of any return on that network.
• Find fee revenue over a full cycle rather than a quarter. A rate calculated across a busy period will not repeat.
• Ask the provider whether ordering revenue is passed through or retained. Some pass a share back and say so; retention is not disclosed as often as it should be.
• Check whether the network burns fees. A burn lowers the number you are quoted and raises what it is worth.
• Compare providers on one network before comparing networks. The first comparison is like-for-like; the second usually is not.

None of this tells you whether a given rate is good. It tells you what the rate is made of, which is the prior question, and the one an advertised percentage is designed not to raise.