> For the complete documentation index, see [llms.txt](https://docs.stakingrewards.com/staking-data/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.stakingrewards.com/staking-data/methodologies/ethereum-srb.md).

# Ethereum SRB

Learn how Ethereum staking rewards are calculated. ETHSRB formula covers consensus and execution layer rewards, validator performance, tips, and MEV yield.

ETHSRB is a benchmark representing the mean, annualized staking rate across all active Ethereum validators.

ETHSRB is calculated and published by Staking Rewards via the [Ethereum Profile](https://www.stakingrewards.com/asset/ethereum-2-0) and [Data API](https://www.stakingrewards.com/data-api).

> **TL;DR:** Ethereum staking rewards come from two sources: consensus layer earnings (block proposals + attestations) and execution layer earnings (priority tips + MEV). The reward rate equals the sum of both layers divided by total staked ETH. Validators can be penalized through slashing or inactivity leaks.

$$
ETHSRB = (cle + ele) / st
$$

<table data-full-width="false"><thead><tr><th width="141">Formula key</th><th width="158">Metric name</th><th>Description</th></tr></thead><tbody><tr><td>cle</td><td>Consensus Layer Earnings</td><td>Validators earn rewards for proposing and attesting to blocks in this layer.</td></tr><tr><td>ele</td><td>Execution Layer Earnings</td><td>Transaction fees (tips/priority fees) earned by validators post-EIP-1559 implementation</td></tr><tr><td>st</td><td>Staked Tokens</td><td>The total number of ETH tokens staked with active validators.</td></tr></tbody></table>

{% hint style="info" %}
**Learn more about the Consensus and Execution Layer Rewards in the section below**
{% endhint %}

* ETHSRB takes into account all accumulated rewards, whether they have been claimed or remain unclaimed.
* ETHSRB operates on a non-compounded basis, meaning it excludes returns generated from the reinvestment of rewards.
* ETHSRB is based on an average validator effectiveness.

***

### Consensus Layer Earnings

The consensus layer earnings per validator can be broken down again into the following components.

$$
cle = bpr + cr - sp - op
$$

<table data-full-width="false"><thead><tr><th width="147">Formula key</th><th width="143">Metric name</th><th>Description</th></tr></thead><tbody><tr><td>bpr</td><td>Block Proposal Rewards</td><td>Block Proposal Rewards are the specific earnings validators receive for successfully proposing a new block, serving as an incentive for active participation in maintaining the Ethereum network's integrity.</td></tr><tr><td>cr</td><td>Attestant Rewards</td><td>Attestant Rewards are the earnings validators receive for attesting to the validity and correctness of blocks proposed by others, supporting the Ethereum network's security and consensus mechanism.</td></tr><tr><td>sp</td><td>Slashing Penalties</td><td>Slashing Penalties are severe fines imposed on Ethereum validators for serious violations such as double signing or similar malicious activities, serving as a deterrent to preserve the network's security and reliability.</td></tr><tr><td>op</td><td>Other Penalties</td><td>Other Penalties include Inactivity Leak for validators offline in consensus, missed Proposer Rewards due to failure in block proposal, reduced earnings from Whistleblower Reward Adjustments, and Surround Vote Penalties for violating specific voting rules in Ethereum's consensus mechanism.</td></tr></tbody></table>

***

### Execution Layer Earnings

The execution layer earnings per validator can be broken down again into the following components.

$$
ele = te + me
$$

<table data-full-width="false"><thead><tr><th width="147">Formula key</th><th width="139">Metric name</th><th>Description</th></tr></thead><tbody><tr><td>te</td><td>Tips Earned</td><td>This metric reflects the optional fees paid by users to Ethereum validators for prioritizing transaction processing.</td></tr><tr><td>me</td><td>MEV Earned</td><td>This metric measures the additional earnings Ethereum validators gain from strategically including, excluding, or ordering transactions within a block, over and above standard block rewards and gas fees.</td></tr></tbody></table>

***

#### Observation Period:

1. The ETHSRB is determined based on the latest 24h at the point of evaluation.
2. This data is then annualized for a comprehensive year-long projection.

#### Calculation

The calculation happens every 2 hours and is immediately published via [Data API](https://www.stakingrewards.com/data-api).

{% hint style="info" %}

#### Limitations

* ETHSRB does not guarantee the specified reward rate for individual validators, as it is subject to various factors such as network conditions, validator performance, and adherence to protocol rules.
* ETHSRB's annual projections are estimates and may fluctuate due to changes in network participation, Ethereum usage patterns, and evolving network dynamics, making long-term predictions inherently uncertain.
  {% endhint %}

***

### ETH Real Reward Rate (ETHSRB^R)

The real reward rate calculates the ETHSRB adjusted for inflation in the network.

$$
ETHSRB^R = 1+ETHSRB /(1+i)-1
$$

<table data-full-width="false"><thead><tr><th width="150.33333333333331">Formula key</th><th width="187">Metric name</th><th>Description</th></tr></thead><tbody><tr><td>i</td><td>Inflation Rate</td><td>Inflation refers to the increase in the total supply of Ethereum due to block rewards distributed to validators for block production and attestations, adjusted by the burning of transaction fees as per EIP-1559, which can impact the net inflation rate.</td></tr></tbody></table>

The inflation rate is the annualized net change in ETH supply: consensus-layer issuance minus the EIP-1559 burn, over circulating supply.

$$
i = \frac{apr\_{cl} \cdot st - t\_y \cdot b\_d}{cs}
$$

<table data-full-width="false"><thead><tr><th width="150">Formula key</th><th width="187">Metric name</th><th>Description</th></tr></thead><tbody><tr><td><span class="math">apr_{cl}</span></td><td>Consensus APR</td><td>Mean consensus-layer (attestation, proposal and sync committee) reward rate across validators over the latest 24h. <span class="math">apr_{cl} \cdot st</span> is the annual consensus-layer issuance in ETH.</td></tr><tr><td><span class="math">st</span></td><td>Staked Tokens</td><td>Active stake in ETH.</td></tr><tr><td><span class="math">b_d</span></td><td>ETH Burned 24h</td><td>Average daily EIP-1559 base-fee burn in ETH over the last 30 days.</td></tr><tr><td><span class="math">t_y</span></td><td>Year Constant</td><td>365 days</td></tr><tr><td><span class="math">cs</span></td><td>Circulating Supply</td><td>Circulating ETH supply.</td></tr></tbody></table>

* Execution-layer rewards (priority fees and MEV) are **not** part of the inflation rate. They are transfers from users and searchers to block proposers, not newly issued ETH, so they raise ETHSRB without changing supply.
* The result can be negative. A negative inflation rate is real net deflation (burn exceeding issuance) and is published as is.
* When any input is unavailable, the inflation rate (and with it ETHSRB^R) is omitted for that run rather than published as zero.

{% hint style="info" %}

#### Inflation Rate Behaviour

Ethereum's inflation rate is influenced by its block reward system and the burning of a portion of transaction fees as per **EIP-1559**. This [EIP](https://consensys.io/blog/what-is-eip-1559-how-will-it-change-ethereum) introduced a base fee for transactions, which is algorithmically adjusted and burned, potentially reducing the overall circulating supply of ETH. However, the inflation rate is not solely dictated by EIP-1559; it also depends on network congestion and block size. Block size is elastic: each block may use up to the gas limit, which validators adjust over time (60 million gas as of November 2026, raised from 30 million in 2025), and the base fee rises or falls to steer usage toward half of that limit. This flexible block size mechanism is a crucial factor in determining the base fee and, consequently, affects the inflation dynamics. While EIP-1559 introduces elements that could theoretically lead to deflationary pressure on Ethereum's supply, the overall inflation rate is still subject to a range of factors including network demand and the amount of ETH staked.
{% endhint %}

***

### Other Metrics

<table data-full-width="false"><thead><tr><th width="200">Metric</th><th>Description</th></tr></thead><tbody><tr><td>Staking Wallets</td><td>Number of distinct on-chain owners of active validator stake: the distinct withdrawal credentials of all active validators (active and exiting) in the finalized beacon state. For 0x01 and 0x02 credentials the owner is the execution-layer withdrawal address, so one address running several validators, of either credential type, counts once; a 0x00 (BLS) credential counts as its own owner. Ethereum has no delegation, so this counts stake owners as the protocol sees them, not individual people: a pooled staking contract such as a liquid staking vault is one owner however many holders it serves, and per-position contracts (for example minipools or restaking pods) each count separately. Computed once a day.</td></tr><tr><td>Pending Staked Tokens</td><td>ETH deposited and waiting in the activation queue (beacon pending deposits).</td></tr><tr><td>Active Validators</td><td>Validators currently attesting, excluding those still activating or exiting. Since EIP-7251 a validator can hold up to 2,048 ETH, so the validator count no longer tracks staked ETH or the number of stakers.</td></tr></tbody></table>

***

### ETH LST Staking Rewards Benchmark ETHSRB(lst)

ETHSRB(lst) is a benchmark representing the annualized staking rate for each Ethereum Liquid Staking Reward Option.

A liquid staking token pays its holder through the rising redemption value of the token against ETH, not through new tokens arriving in the holder's wallet. ETHSRB(lst) therefore measures the growth of that redemption value across an observation window and annualizes it.

$$
ETHSRB(lst) = (er\_t / er\_0)^{t\_y/\Delta t} - 1
$$

<table data-full-width="false"><thead><tr><th width="141">Formula key</th><th width="158">Metric name</th><th>Description</th></tr></thead><tbody><tr><td><span class="math">er_t</span></td><td>Exchange Ratio</td><td>The current redemption value of one LST unit in ETH, read from the token's own on-chain rate function.</td></tr><tr><td><span class="math">er_0</span></td><td>Past Exchange Ratio</td><td>The same exchange ratio observed at the start of the window, read at a historical block or from the recorded exchange ratio history.</td></tr><tr><td><span class="math">\Delta t</span></td><td>Elapsed Time</td><td>The measured time in seconds between the two observations, taken from the timestamps at which each ratio was recorded.</td></tr><tr><td><span class="math">t_y</span></td><td>Year Constant</td><td>365d (31,536,000s)</td></tr></tbody></table>

* The growth factor is **relative** — the ratio of the two exchange ratios, not the difference between them. An absolute difference is not a rate: it overstates the result by a factor equal to the exchange ratio itself, which for a mature LST sits meaningfully above 1.0.
* $$\Delta t$$ is **measured** from the timestamps of the observations themselves rather than inferred from an assumed block interval, so the rate stays correct when block production drifts or a source returns a ratio older than requested.
* The exponent compounds the observed growth over the year, which is the behaviour a holder actually experiences: the exchange ratio grows on its own accrued value, with no claim or restake step.
* **Rebasing tokens.** A rebasing LST keeps its published Exchange Ratio at 1 by construction (see Staked Tokens below), so the ratio itself cannot carry the growth. For Lido stETH, $$er\_t$$ and $$er\_0$$ are the protocol's share rate — pooled ETH per share, `getPooledEthByShares(1e18)` — read on-chain at both block heights. There is no recorded-history fallback for it: a window whose on-chain read fails is not published. eETH follows the same rule through weETH's `getRate()`, which is eETH's share rate.

{% hint style="warning" %}

#### Reprice-anchored windows and non-reward exchange-ratio steps (swETH)

**Elapsed time is anchored to reprices.** swETH's exchange ratio changes only when Swell's bot calls `reprice()`, which posts all accrual since the previous reprice in one step and records that reprice's time on-chain (`lastRepriceUNIX()`). Reprices are irregular: about two a day, but with gaps of up to 19 days. For swETH, $$er\_t$$ and $$er\_0$$ are therefore read together with the time of the reprice that posted each of them, at the same two blocks, and $$\Delta t$$ is the time between those two reprices rather than between the two blocks. Measured block to block, a window that happens to contain a multi-day catch-up step would annualize accrual earned before the window started, and the next window would understate by the same amount; anchored, the 30-day rate stays within about ±0.3pp of the long-run yield (standard deviation 0.16pp against 0.44pp block to block, over 1,477 replayed two-hourly runs). Between reprices the published rate holds steady at the rate as of the last reprice. There is no recorded-history fallback for swETH, because recorded exchange ratios carry no reprice time: a window whose on-chain read fails is not published.

**Non-reward steps.** The formula assumes every change in the exchange ratio is staking reward. A one-off step that is not — an unbacked reprice, an oracle error or a correction — would be annualized as if it recurred. For swETH, a +6.15% step on 21–22 October 2026 (six hourly reprices at the 1% cap, with no fees taken and no new beacon-chain deposits) would otherwise have published a 30-day rate of about 110% against about 2.2% of underlying yield.

For swETH, a window rate is therefore **not published** (never published as 0) when:

* Swell's access-control contract reports bot, core or withdrawal methods paused, or the pause state cannot be read — the exchange ratio is then neither updated by the protocol nor redeemable;
* the last reprice is more than 21 days old (the longest gap observed is 19 days), which means the repricer has stopped even though the protocol is not paused;
* no reprice happened inside the window, or the two reprices are less than 12 hours apart, so no reward, or only an hour-scale step, was observed;
* the annualized window rate falls outside ±15%, a range ordinary ETH staking yield cannot leave, so the window contains a non-reward step.

**The October 2026 incident.** The six unbacked steps were reversed on 31 October 2026, but the reset level carried accrual only up to 21 October, and a partial reprice on 1 November added only about 1.7 days of it; the catch-up reprice of 9 November 2026 paid the rest. A window whose starting block falls between the first unbacked step and that catch-up therefore takes its starting observation from the last honest reprice before the incident (21 October 2026) instead. The window is then longer than nominal, which the anchored $$\Delta t$$ measures exactly, so no window is lost to the incident.

The headline rate uses the longest surviving window (30, 14, 7, then 1 day), and is withheld when no window survives.
{% endhint %}

{% hint style="info" %}

#### Provider fees are already reflected — ETHSRB(lst) is a net rate

ETHSRB(lst) carries **no separate provider fee term**.

An LST's exchange ratio rises only by the rewards that actually reach the token holder. The provider's commission is taken out of staking rewards before they reach the pool backing the token — whether deducted at distribution or realized by minting shares to the provider — so exchange ratio growth is a post-fee quantity by construction. Subtracting the fee a second time would double-count it.

This is measurable. For wbETH, whose provider charges a 10% commission, independent validator-level measurement of the underlying validator set shows **2.62%** gross, while the on-chain exchange ratio grows at **2.37884%** annualized over 90 days. The ratio of the two is **0.908**, matching the published commission to within measurement noise. Applying (1 − fee) again would publish **2.141%** — understating the rate the holder receives by roughly 0.24 percentage points, and falling below the fee-adjusted gross rate.

A second reward option confirms the mechanism at a different fee level and by a different route: osETH's rate contract deducts accrued treasury assets from total assets *before* dividing to produce the rate, so its realized on-chain growth divides out against its gross reward rate at exactly (1 − 5%), its published commission.

The same rule applies to the reward options that still source their rate from a provider-published APR instead of the exchange ratio: published LST APRs are quoted net of the provider's commission and are used as reported. Lido stETH, Lido wstETH and Rocket Pool rETH are not among them; they use the exchange-ratio formula above.

The commission is not discarded. It is published separately as Commission, and it is what Fee Revenue is derived from — by grossing the net rate back up before applying the fee:

$$
fr = st \* (ETHSRB(lst) / (1 - lpf)) \* lpf
$$

where $$fr$$ is Fee Revenue, $$st$$ is Staked Tokens and $$lpf$$ is the LST Provider Fee.

For ether.fi weETH and eETH, $$lpf$$ in the Fee Revenue formula is not the nominal 10% fee. It is the measured effective fee share over the trailing 30 days: the protocol fee divided by gross rewards, both read from the ether.fi LiquidityPool's on-chain rebase events.

$$
lpf\_{ether.fi} = \frac{\sum fees}{\sum fees + \sum S\_i (r\_i - r\_{i-1})}
$$

where $$fees$$ are the `ProtocolFeePaid` amounts and $$S\_i (r\_i - r\_{i-1})$$ is the holders' gain on each `Rebase` (eETH shares times the change in the pooled-ETH-per-share rate). The protocol fee is ether.fi's whole take; node operators are paid out of it. The nominal 10% is still published as Commission. If the effective fee share cannot be measured, Fee Revenue is omitted for that run.
{% endhint %}

{% hint style="info" %}

#### Staked Tokens and AuM are denominated in ETH, not in LST units

A value-accruing LST holds its share count still and lets each share become redeemable for a growing amount of ETH. Its total supply is therefore a count of **shares**, not a quantity of ETH, and the two diverge by exactly the exchange ratio. Staked Tokens for a liquid staking reward option is the ETH-denominated figure:

$$
st = ts \* er\_t
$$

<table data-full-width="false"><thead><tr><th width="141">Formula key</th><th width="158">Metric name</th><th>Description</th></tr></thead><tbody><tr><td><span class="math">st</span></td><td>Staked Tokens</td><td>The ETH backing the reward option, i.e. what the outstanding LST supply is collectively redeemable for.</td></tr><tr><td><span class="math">ts</span></td><td>Total Supply</td><td>The LST's own outstanding share count, denominated in LST units and published unconverted under its own key.</td></tr><tr><td><span class="math">er_t</span></td><td>Exchange Ratio</td><td>The current redemption value of one LST unit in ETH — the same quantity the rate formula above uses.</td></tr></tbody></table>

Assets Under Management then follows from Staked Tokens and the ETH spot price:

$$
aum = st \* p
$$

* **The exchange ratio is applied exactly once, and only in** $$st$$**.** AuM multiplies the already-converted Staked Tokens by price — never the raw share count, and never the share count and the ratio again. Deriving AuM independently from total supply is the error this rule exists to prevent: applying the ratio in both places squares it, and applying it in neither understates both figures by the whole accrued premium. For a mature LST that premium is material — an exchange ratio of 1.0855 makes the difference between the two mistakes a factor of roughly 1.18.
* **Total Supply is published unconverted**, in LST units, under its own key. It is the share count and is correct as denominated; the conversion belongs to Staked Tokens alone.
* **Staking Market Cap prices the share count at the LST's own price**: $$smc = ts \* p\_{lst}$$, where $$p\_{lst}$$ is the market price of one LST unit. This equals $$st \* p$$ up to the token's deviation from its peg. Multiplying the ETH-denominated Staked Tokens by the per-share LST price would apply the exchange ratio twice. Rebasing LSTs, whose supply already equals Staked Tokens, are unaffected; an LST that publishes no Total Supply falls back to $$st \* p\_{lst}$$.
* The ratio is read live from the token's own on-chain rate function. Where that read fails, Staked Tokens and AuM are **left absent rather than published against an assumed ratio of 1** — a pinned ratio silently understates every ETH-denominated figure and inverts Peg Deviation, which divides by it.
  {% endhint %}

{% hint style="warning" %}

#### Non-reward exchange-ratio steps (rswETH)

The rate formula assumes every change in the exchange ratio is staking or restaking reward. A one-off step that is not — an unbacked reprice, an oracle error or a correction — would be annualized as if it recurred. For rswETH, a +5.10% step on 21–22 October 2026 (five hourly reprices at the protocol's 1% cap, with no fees taken and no matching deposits, after which Swell paused the protocol) would otherwise have published a 30-day rate of about 92.6% against about 2.4% of sustained underlying yield.

For rswETH, a window rate is therefore **not published** (never published as 0) when:

* rswETH's own access-control contract reports bot, core or withdrawal methods paused, or the pause state cannot be read — the exchange ratio is then neither updated by the protocol nor redeemable;
* the annualized window rate falls outside ±15%, a range ordinary ETH staking and restaking yield cannot leave, so the window contains a non-reward step;
* the exchange ratio did not move across the window, so no reward was observed.

The headline rate uses the longest surviving window (30, 14, 7, then 1 day). A step therefore stops affecting the headline once it is older than the window in use, and the rate — with the Fee Revenue and Real Reward Rate derived from it — is withheld when no window survives.
{% endhint %}

***

#### Observation Period:

1. ETHSRB(lst) is evaluated over trailing windows drawn from 1, 7, 14, 30 and 90 days, each published in its own right as Reward Rate 1d / 7d / 14d / 30d / 90d. Which windows a given reward option publishes depends on how far back its exchange ratio can be resolved and how often the provider updates it.
2. The headline ETHSRB(lst) is the **longest window up to 30 days** that resolved for that reward option. Windows longer than 30 days are published alongside as reference series and do not feed the headline rate, so that a single anomalous reading cannot dominate the published rate for a full quarter.
3. Shorter windows are used only when longer ones cannot be resolved, which keeps the headline rate from being driven by the update cadence of providers that reprice their exchange ratio once a day or less often. For swETH, whose reprices are irregular, the elapsed time of every window is additionally measured between the reprices that posted its two observations (see the swETH note above), so the headline does not swing with where a catch-up reprice falls.

#### Calculation

The calculation happens every 2 hours and is immediately published via [Data API](https://www.stakingrewards.com/data-api).

{% hint style="info" %}

#### Limitations

* ETHSRB(lst) reflects the growth of the exchange ratio over a trailing window and is not a forward guarantee; it moves with validator performance, network conditions and the provider's fee policy.
* ETHSRB(lst) does not take into account the inflation rate of the asset. Please refer to ETHSRB^R for this.
* A window only resolves if a past exchange ratio is available for it. Newly listed reward options publish short windows first and longer ones as history accumulates.
* A subset of Ethereum liquid staking reward options is still being migrated to the compounding relative form above, and currently publishes a non-compounded rate derived from an absolute difference between two exchange ratios. Because an absolute difference is not normalized by the starting ratio, those rates read high roughly in proportion to the token's exchange ratio — about 18% high at an exchange ratio of 1.19 — until the migration completes.
  {% endhint %}

***

#### Data Sources:

* [Staking Rewards](https://www.stakingrewards.com/) from indexed Ethereum Blockchain RPC Endpoints
* Liquid staking token contracts — exchange ratio (or share rate, for rebasing tokens) and total supply, read on-chain at current and historical blocks. The same two reads feed both the ETHSRB(lst) window calculation and the ETH-denominated Staked Tokens and AuM figures.
* Liquid staking provider APR endpoints — for the reward options that still publish their own net rate
* CoinGecko — price and supply data
* Validator-level consensus-layer reward rates and active stake indexed from the Ethereum beacon chain — the issuance term of the inflation rate
* Ethereum beacon chain validator registry (finalized state) — withdrawal credentials of active validators, for Staking Wallets
* Ethereum execution-layer EIP-1559 base-fee burn totals (30-day sum) — the burn term of the inflation rate

***

#### References:

* <https://ethereum.org/en/developers/docs/consensus-mechanisms/>
* <https://www.stakingrewards.com/asset/ethereum-2-0>

***

:books: [Learn more about ETH Staking](https://www.stakingrewards.com/asset/ethereum-2-0)

:bar\_chart: [Calculate your ETH Staking Rewards](https://www.stakingrewards.com/calculator?asset=ethereum-2-0)

:gear: [Access historic Ethereum Staking Data via API](https://www.stakingrewards.com/data-api)

```
Last updated on 11/17/2026
```
