Pond Street Ledger

What a Stock Split Does to a Stock Token, and Who Is Supposed to Press the Button

Dividends, splits, mergers and ticker changes are routine in equity markets and awkward onchain. The mechanism that handles them is the least glamorous and most load-bearing part of a tokenized equity.

✓ 1593.efrogs.eth2026-11-228 min

A tokenized share is usually sold on the strength of one claim: it tracks a real share, one for one. That claim holds easily on a quiet Tuesday. It is tested on the day the issuer does something to the share itself. Companies split their stock, pay dividends in cash and sometimes in more stock, spin off subsidiaries, get acquired for cash or for paper, change their ticker, consolidate shares in a reverse split, and offer rights to existing holders on a deadline. Each of those is a corporate action, and each one changes either the number of shares you hold, the identity of what you hold, or the cash that arrives alongside it.

In traditional markets this is an entire industry, with its own vendors, message formats and error rates. Onchain, it becomes a question about who has the authority to change a token supply, a token price basis or a token's reference asset, and how fast they can do it relative to a market that does not close. Most of the hard problems in tokenized equities that look like market structure problems are actually corporate action problems wearing a different coat.

The three things a corporate action can change

Sort them by what they do to the token and the work becomes clearer. The first group changes the share count without changing economic value: forward splits, reverse splits, stock dividends, bonus issues. A two-for-one split doubles the shares and halves the price. The second group pays something out: cash dividends, special dividends, return of capital, interest on preferred. The third group changes the identity of the asset: mergers, acquisitions, spin-offs, ticker and name changes, delistings. A token programme has to have an answer for all three, and the answers are different.

Splits, and the uncomfortable fact of a mutable supply

The cleanest way to handle a two-for-one split is to double every holder's token balance, because that mirrors what happens in the custody account holding the underlying shares. Doing that requires either a rebasing token, where balances change under holders without a transfer, or an administrative mint distributed to a snapshot of holders. Both have costs. A rebasing balance breaks a surprising amount of downstream plumbing: automated market maker pools, lending collateral accounting, wrapped versions of the token, and any contract that cached a balance rather than reading it. An administrative mint requires a snapshot, which means picking a block, which means deciding what happens to tokens sitting in a liquidity pool or posted as collateral at that block.

The alternative is to not change supply at all and instead change the ratio: one token now represents two shares, or half a share. That keeps every integration intact and moves the complexity into the disclosure and the redemption maths. It also means the token stops being a one-for-one representation, which was the selling point. This is a real tradeoff and there is no free version of it. You are trading integration safety against the simplicity of the claim.

Dividends, and the problem of paying someone you cannot identify

Cash dividends are harder onchain than splits, for a reason that has nothing to do with blockchains being slow. The share is held in a custody account and the dividend arrives there in fiat. Getting it to token holders requires knowing who the token holders are at the record date, converting the cash to something payable onchain, usually a stablecoin, and distributing it. Each step has friction. Withholding tax applies at the custody level and depends on the beneficial owner's jurisdiction, which an issuer serving an anonymous onchain holder base may not know. The pragmatic answer most programmes reach for is to withhold at the highest applicable rate, or to not pay the dividend out at all.

Not paying it out is more common than you might expect, and it takes two forms. The token can accrue the dividend into its reference value, so the token quietly becomes a total-return instrument rather than a price tracker. Or the dividend can be reinvested in more of the underlying share and the token ratio adjusted upward. Both are defensible. Both mean the token's price will diverge over time from the screen price of the stock it is named after, and a holder who does not read the documentation will experience that divergence as a tracking error rather than as the dividend they were owed.

Mergers and delistings, where the token can stop referencing anything

The sharpest edge is an acquisition. When a company is bought for cash, the custody account's shares become cash on the closing date and the token has nothing left to track. The programme has to force a redemption, pay holders out in stablecoin at the deal price, and burn the supply. That is a mandatory event with a date attached, and anyone holding the token in a lending protocol as collateral, or in a liquidity pool, needs to be out by then or needs the protocol to handle it for them. A stock-for-stock deal is slightly better, because the token can be migrated to reference the acquirer's shares, but it still means every integration points at a new asset.

Delisting is the worst case, because it has no clean closing price. A share that stops trading on an exchange still exists, but the mechanism that produced the price feed does not. A token referencing it is left with a stale oracle on a venue that trades continuously, which is precisely the combination that produces liquidations at prices nobody would transact at. Programmes handle this by freezing transfers and opening a redemption window, which is the correct answer and also an admission that the token's always-on property is conditional on the underlying behaving normally.

What this tells you about a programme's quality

Read a tokenized equity's corporate action policy before you read anything else about it. Four questions get you most of the way. Does the document name a specific record-date mechanism, or does it say actions will be handled at the issuer's discretion. Are dividends paid out, accrued or reinvested, and in what asset. What happens to tokens held in third-party protocols at a snapshot. And who bears the cost when the handling is late, which it will sometimes be, because the custodian's notification arrives on the custodian's schedule and not the chain's. A programme that has written these down has thought about the boring half of the product. One that has not is selling you the price feed and hoping nothing happens to the company.

Where it breaks

The structural weakness is a timing mismatch. Corporate actions are announced, go ex-date and settle on a calendar run by the issuer, the exchange and the central securities depository, all of which operate on business days. A token trades continuously. Between the ex-date and the moment the token programme implements the adjustment, the token and the share are referencing different things, and arbitrageurs will price that gap correctly while retail holders will read it as the token being broken. Nothing about putting equities on a ledger shortens the corporate action calendar. It only makes the gap visible in public, tick by tick, to an audience that can trade it.