How EXDIV works.

01The split

Robinhood Chain Stock Tokens don't pay cash dividends. When the underlying pays one, it is auto-reinvested and the token's on-chain multiplier (uiMultiplier, ERC-8056) rises — your raw balance never changes, but each token represents more shares. That growth is a yield stream, and EXDIV splits it off.

Deposit a Stock Token and EXDIV mints two ERC-20s, one series at a time (asset × maturity date):

• PT (pAAPL-MAR27) — the stock with the dividends removed. Trades under spot; redeems the full baseline position at maturity. A zero-coupon bond wearing an equity.
• YT (yAAPL-MAR27) — every dividend the position earns before maturity, and nothing else. Costs a fraction of the share price; no leverage, no margin, no liquidation — worst case it decays to the drip actually paid.

Merge is the anchor: PT + YT always recombines into the underlying token, in full, for free, at any time — before and after maturity. That exit is what keeps PT priced honestly.

"How do I claim my dividend?" You don't — there is nothing to claim, here or anywhere. On Robinhood Chain the dividend is reinvested automatically and lands inside every holder's tokens through the multiplier; it is invisible in a normal wallet, but it is already yours. What EXDIV adds is the ability to see it (the terminal's drip chart and accrued index) and to trade it (yAAPL is that dividend stream as a token). Beware of any site offering to "claim" stock-token dividends for you — there is no such mechanism, and pages built around a claim button are a known scam pattern.

02Where this comes from

EXDIV is not a new financial idea — it is a tokenized rebuild of structures that have traded on Wall Street for four decades. Separating an asset from its income stream is one of the oldest, most proven trades in fixed income and equities.

• Treasury STRIPS (1985–today). The US Treasury lets dealers separate a bond into its principal and its individual coupons; each piece trades on its own. The principal-only STRIP trades at a discount and redeems at par — a multi-trillion-dollar market that runs to this day. PT is the direct analog: the asset with the income removed, priced under spot, redeeming the full baseline at maturity.
• PRIMES & SCORES (1983–1992). The closest ancestor, and it was equities. The Americus Trusts accepted blue-chip shares — AT&T, Exxon, GM, Coca-Cola — and issued two AMEX-listed claims against them: the PRIME carried the dividends, the SCORE carried price appreciation, and the pair could be recombined into the share. The Tax Reform Act of 1986 made new trusts uneconomic and the existing ones ran off by 1992. The structure died of tax law, not of market failure.
• Dividend futures & swaps (2008–today). Index and single-stock dividend futures trade on Eurex and CME, with dividend swaps OTC — a deep institutional market in exactly what YT holds: the realized dividend stream over a fixed term. It sits behind ISDA agreements and broker minimums; retail has never had access. YT is that instrument as an ERC-20.

What EXDIV changes is the rails, not the idea: the claims are ERC-20s over tokenized stocks, the split is self-service, the accounting is on-chain through the token's own ERC-8056 multiplier — and unlike the Americus Trusts, which only recombined round lots through the trust, PT + YT merges back to the underlying at par, free, at any time.

03The math

The accountant tracks a dividend index d that grows only with dividend reinvestments (stock splits move a separate factor and change nothing here). Let d0 = the index when the series opened and dT = the index at maturity.

Because PT + YT redeem to exactly what was deposited, the vault is always fully collateralized by construction — it holds precisely the raw tokens deposited, and a stock split scales everyone's share exposure through the token's own multiplier without touching d0 or dT.

04Fees

Where fees go: back into the market, two ways. The treasury recycles protocol fees on-chain — buying the stock token, paring it, and deepening the PT and YT pools, so every split makes the next trade tighter — and routing a share to market-buying and burning $PARE. Protocol revenue becomes protocol liquidity and shrinking supply, across every series.

One line: thin fee in the door, thin slice of the yield. Nothing on exit, nothing on price exposure.

05The pools

Every series ships with two Uniswap v3 pools on Robinhood Chain, and both are quoted in the stock itself, never a stablecoin: PT/stock (pAAPL/AAPL, pSPY/SPY) and YT/stock (yAAPL/AAPL, ySPY/SPY). The prices in the terminal are read straight from those pools.

• Why the stock, not dollars. PT is a claim on the same share minus a sliver of drip, so priced in the stock it barely moves: it sits just under 1 and drifts toward 1 as maturity nears. A pool whose two sides move together carries far less impermanent loss than a token/USDC pool, and nobody providing liquidity is short the stock's rally.
• Fair value. PT trades near 1 / (1 + drip still expected before maturity) and YT near the remainder, so PT + YT ≈ 1 share. Series 02 seeded at 0.99107 + 0.00893 = 1.00000 SPY.
• The par line polices itself. Split mints one PT + one YT per share (minus the 10 bps fee); merge burns one of each for a full share back, free. If the pools ever price PT + YT above par, split and sell. Below par, buy both and merge. Either trade pays whoever closes the gap, so the pools stay pinned to par without anyone's permission.
• Tiers and ranges. The PT pool sits on the 0.05% tier in a tight range just under par. The YT pool sits on the 1% tier in a wide range, because YT is small and moves more in percentage terms. Fees from every swap accrue to the positions.
• Who provides. Today the protocol seeds every pool from its own inventory and recycles fees into deepening them (see Fees). Anyone can add liquidity: the terminal's EARN tab takes a stock amount, pares exactly what the pools need, and mints both positions (PT/stock and YT/stock) to your wallet in the standard ranges, in one flow. They are ordinary Uniswap v3 pools, so you can also add or manage positions directly on the DEX.

One line: PT and YT trade against the share they came from, and free merge keeps their sum at one.

06Dividends vs stock splits

The multiplier rises for both dividend reinvestments (adds value → belongs to YT) and stock splits (value-neutral — the share price offsets). Telling them apart is the protocol's one hard problem, and it is enforced on-chain:

• A change classifies as a dividend only when the growth is small: 0 < r−1 ≤ 3%.
• A change classifies as a split only when it is a clean small-integer ratio (2:1, 3:2, 1:10…) at least 20% from 1.
• The two bands cannot overlap, so a keeper can rob neither side — misclassification attempts revert.
• Anything in between — special dividends, composites — resolves only through a guardian behind a 2-day public timelock, with free merge as everyone's exit in the meantime.

Every classified event is public — the terminal's ledger shows each one with its ratio and date.

07Lifecycle

A series runs ACTIVE → MATURED → SETTLED. While active: split, merge, trade. At the maturity timestamp the series is matured; anyone may then call settle() — it is permissionless, but it refuses to run until every multiplier change is classified, freezing dT honestly. Once settled, PT and YT redeem at the amounts above. Merge keeps working through all three states.

08$PARE — supply and distribution

$PARE is the protocol's token on Robinhood Chain. Contract address Address hidden. It is separate from the series contracts: no vault, principal token or yield token depends on it, and holding it is not required to use the protocol.

Live supply figures are published at /api/supply, read from the token contract on each request. Circulating supply is total supply less any balance held at a burn address. The developer wallet was bought on the open market and counts as circulating, as do liquidity pool holdings, since anyone can trade against them.

How it was launched. There was no presale, no private round and no team allocation beyond the 2.26% above, which is locked. The entire remaining supply was placed into the launch pool at deployment, so the first buyer and the hundredth bought on the same terms.

Where the revenue goes. The protocol earns from the 10 bps split fee and the 5% fee on the YT drip at maturity, plus swap fees on the liquidity it provides itself. That revenue is directed two ways: deepening the protocol's own PT and YT pools so each series trades tighter, and market-buying and burning $PARE. Both are funded by usage rather than by issuance — there is nothing to print.

What it does not do. Holder fee tiers on splits, priority access to new series and a vote over which equity is listed next are all intended, and none of them are implemented in the deployed contracts today. The split fee is immutable per vault and identical for every address, so fee tiers would require new vaults rather than a settings change. Treat them as roadmap, not as current utility.

09The oracle

Every series runs on a MultiplierAccountant, and the accountant does not need the series. It watches one stock token's ERC-8056 multiplier, records each change, classifies it as a dividend or a stock split using the bands in [06], and holds anything that fits neither for the guardian. All of that is readable by any contract, for free, with no key. EXDIV publishes it as the dividend oracle for Robinhood Chain.

Why a lender should care. A stock token's raw balance never moves; the multiplier does. So a 4-for-1 split and a dividend both look like "the multiplier changed" to anything reading balances. Valuing collateral is the easy half and depends on your feed: Chainlink's Robinhood feeds on this chain price the token with the multiplier already inside, so there it is rawBalance × feed and multiplying by uiMultiplier() again would double count; a per-share feed needs rawBalance × uiMultiplier() × pricePerShare. Either way the feed cannot tell you which corporate action happened, cannot warn you before it lands, and does not refuse a change that fits neither. The accountant does all three. Wire isSynced() into a market pause and you have a circuit breaker against a bad multiplier push, whether it came from a bug or from the issuer, plus the dates of every window in advance.

What EXDIV commits to. A keeper that calls sync() and classifies within the poll interval, for every accountant listed on the oracle page, and a guardian that resolves held changes inside the two day timelock. Classification is validated by the contract itself, so a keeper fault can cost gas but cannot move value; the worst case for a consumer is a stale flag, which is exactly what isSynced() reports.

Coverage without capital. An accountant can be deployed for a token that has no EXDIV series, no pools and no vault. That is how the oracle grows: one small contract per stock token, whether or not anyone is splitting it yet.

Ex-date handling

The question a lender asks first is what happens in the window between a corporate action landing and the oracle classifying it, and who carries a liquidation that happened inside it. The answer is that the window is announced in advance and the stale signal has no latency, so a correctly wired market is never open inside it.

There is no block in which the oracle reports synced while the multiplier has moved. isSynced() is not a flag a keeper sets; it is a view that reads the token live, so the stale signal is atomic with the change itself. The only thing with latency is the classification, and classification does not change collateral value: rawBalance × feed with a token-priced feed, or rawBalance × uiMultiplier() × pricePerShare with a per-share one, is already correct. What classification tells a market is whether its price feed is about to move by the split ratio, which is why a split waits on the feed and a dividend does not.

Liability, plainly: the oracle is free infrastructure with no service agreement, and EXDIV does not underwrite positions on markets it does not run. What EXDIV gives an integrator is a design in which acting on stale data requires ignoring a signal that is available in the same block, plus the dates of every window in advance. A market that wires isSynced() into its pause has no liquidations inside the window to argue over.

The dispute path

1. Automatic path. Growth between 0 and 3% is a dividend. A clean small-integer ratio at least 20% away from 1 is a split. The bands cannot overlap, the contract checks this at construction, and the contract rejects any tag outside them. A keeper cannot misclassify; it can only be late.
2. Anything else is held. A special dividend above the band, an odd ratio, a decrease that is not a reverse split, or two actions landing before the first was classified. None of these can be tagged by the keeper. The guardian proposes an ordered decomposition and its hash is committed on chain with an execution time two days out.
3. Two days in public. The proposal is readable by anyone before it executes. Holders can exit through merge(), which never depends on the index. The guardian can cancel. Nothing settles and isSynced() stays false until the resolution executes.
4. The record. Every classification is a checkpoint, readable forever through checkpointAt(i). Every held change emits its proposal and its resolution. Anyone can recompute the whole history from the token's own multiplier and disagree in public with a specific checkpoint.

Accuracy so far. Nine tokens are covered and the record is short. Rather than a percentage that would be made of a handful of events, the record itself is the number: the checkpoint log on each accountant, and the oracle page's last-classified column against each token's own multiplier history. When SGOV's monthly distributions have run for a few months there will be a real series to point at.

10Risks, plainly

• Series capacity. Each series accepts deposits up to a fixed on-chain cap (shown in the terminal). When a series fills, the next one opens with more room — and merge is always available regardless.
• Issuer trust. The multiplier is set by Robinhood Assets (Jersey) Ltd. YT is a claim on the issuer's declared dividend reinvestment, not a raw claim on the underlying equity.
• Classification edge cases. Gap-band events settle at guardian discretion behind the timelock. If you dispute a pending resolution, merge out — it is always free.
• Liquidity. Early pools are thin. Merge and redemption never depend on pools.

11Contracts

The audited build (Pashov Audit Group, report), deployed 18 September 2026. Every series matures 31 December 2027 00:00 UTC. Split fee 10 bps, YT yield fee 5%, dividend band ≤3%, split band ≥20%, guardian delay 2 days.

Previous series (launch build, September 2026)

The series from launch day keep running to their own maturities. Merge is open on them at any time; settle and redeem after maturity. Open one in the terminal with ?series=spy-mar27, aapl-mar27, qqq-mar27 or pfe-mar28. They are not the audited build.

Contract source is verified on the block explorer at the addresses above.

Licence

The contracts are published under the Business Source License 1.1 (the SPDX-License-Identifier: BUSL-1.1 header on every verified source). Licensor: AURORA. Licensed work: the StripVault, PYToken, MultiplierAccountant and AuroraMorphoOracle contracts as deployed on Robinhood Chain. Additional use grant: any use that does not deploy the contracts, or a modified version of them, as a competing production protocol. Reading, auditing, testing, integrating with and building on the deployed contracts is permitted. Change date: 18 September 2029, after which the change licence is GPL-2.0-or-later.