Transfer agents and paying agents sit deep in the back office of capital markets. One keeps the register of who owns what. The other distributes cash when an issuer needs to pay. They're easy to confuse, partly because the same banks often offer both services, partly because the line between them is thinner than the names suggest.
That line is now being redrawn onchain. Tokenized securities turn the transfer agent into a function of the token contract itself, and the paying agent into a programmable distribution layer. Knowing where one role ends and the other begins is table stakes for any issuer building on tokenization.
A transfer agent is the entity that maintains the official record of security ownership on behalf of an issuer. If you own shares in a US-listed company, your name (or your broker's nominee) sits in a registry kept by a transfer agent such as Computershare, Equiniti, or Continental Stock Transfer. When shares change hands, the agent updates the register. When a shareholder loses a certificate, the agent issues a replacement. When the issuer mails proxy materials or annual reports, the agent supplies the addresses.
In the US, transfer agents are registered with the SEC under Section 17A of the Exchange Act and follow detailed rules on recordkeeping, turnaround times for transfers, and safeguarding of unissued certificates. The role is conceptually simple but operationally heavy: tens of millions of holder records, constant corporate actions, name changes, deceased shareholders, escheatment, lost securities, fractional shares from dividend reinvestment plans. None of it is glamorous. All of it has to be right.
The transfer agent is also the issuer's source of truth on who is entitled to receive the next dividend, the next interest payment, or the next merger consideration. That information is then handed to the paying agent.
A paying agent distributes cash flows from an issuer to security holders. When a bond pays its semi-annual coupon, the issuer wires cash to the paying agent, which breaks it down across thousands of holders and pushes the payments out, net of withholding tax, often after FX conversion. The same mechanic applies to stock dividends, principal repayment at maturity, fund redemption proceeds, and cash consideration in M&A transactions.
The classic paying agents are large trust banks: BNY Mellon, Citi, Deutsche Bank, JP Morgan, Wilmington Trust. Their appointment is usually written into the bond indenture or the trust deed for a fund, and the role carries a fiduciary flavor. The agent sits between the issuer and the investor, contractually responsible for getting the right amount to the right account on the right date.
Paying agents are event-driven. They wake up on coupon dates, dividend record dates, merger closings, and fund distribution dates. In between, the cash isn't there. The infrastructure is.
| Transfer agent | Paying agent | |
| Core function | Maintains the holder registry | Distributes cash flows to holders |
| What they hold | Records and identity data | Cash, briefly |
| Operating cadence | Continuous | Event-driven |
| Regulatory regime | Registered with the SEC (US); prescriptive recordkeeping rules | Contractual appointment under indenture or trust deed |
| Typical examples | Computershare, Equiniti, Continental Stock Transfer | BNY Mellon, Citi, Wilmington Trust |
Three reasons:
A custodian bank that runs paying agency for a $5B corporate bond will frequently also act as registrar or share transfer agent for the same issuer. From the outside, it looks like one service. Internally, they're separate desks with separate books and separate fee schedules.
The paying agent can't pay anyone without a list of current holders. That list comes from the transfer agent. Every coupon date is a handoff between the two, and any error on one side shows up on the other.
Retail holders see their broker. Institutional holders see their custodian. The transfer agent and paying agent operate two or three layers of intermediation deeper, behind DTC in the US or Euroclear in Europe. The two roles only surface when something breaks: a missing dividend, a botched corporate action, an unclaimed coupon. Most market participants never need the vocabulary, so they never build the distinction.
Tokenization is no longer a pilot story. Bonds, treasuries, money market funds, and equity stakes are increasingly issued natively onchain, where settlement is instant, the holder registry updates in real time, and servicing the security costs a fraction of what legacy infrastructure demands.
That shift doesn't eliminate either role. It re-implements them.
The transfer agent function migrates into the token contract itself. An ERC-20 (or, for regulated securities, an ERC-3643 or similar permissioned standard) is mechanically a registry. Every transfer updates the holder list. Balances are queryable in real time. There's no batch end-of-day reconciliation between brokers and the issuer's registrar because the chain is the registrar.
That doesn't mean the regulated transfer agent disappears. For securities with KYC requirements, transfer restrictions, accredited investor checks, lock-ups, and reporting obligations, you still need a regulated entity to manage the whitelist, approve transfers, and maintain the official book of holders. Securitize is the most established example: an SEC-registered transfer agent operating natively on tokenized securities, plugging compliance logic into the token itself. Plume positions its chain as RWA-native infrastructure where transfer agent functionality is built into the protocol layer rather than bolted on.
The paying agent function migrates into a smart contract distribution layer. Once you know who holds the token at the snapshot block, paying a coupon, a dividend, or a yield distribution becomes a function call. No SWIFT messages, no batch files, no reconciliation between custodians, no T+2 settlement on the cash leg.
This is where Merkl operates. Merkl is the onchain programmable paying agent for tokenized finance: a distribution layer that takes a payment from an issuer and routes it to current holders, regardless of where the position sits. Tokens held directly in a wallet, deposited in a lending market, wrapped in a vault, used as LP collateral, the distribution still reaches the end user. Reward forwarders trace ownership through seven or more layers of nested contracts so the issuer doesn't pay the intermediary, the issuer pays the actual investor. That is what makes it programmable rather than simply automated. More than 250 companies have already used Merkl for onchain distributions, including PayPal, Circle, Coinbase, SG-Forge, and Kraken.
Two scenarios make the split concrete:
The issuer needs a transfer agent to maintain the holder registry under whatever regulatory regime applies (Reg D, Reg S, MiFID II, and so on) and a paying agent to distribute the semi-annual coupon and the principal at maturity. Onchain, this looks like Securitize as transfer agent and Merkl as paying agent, with the bond token enforcing transfer rules at the contract level.
A stablecoin issuer redistributing reserve yield to users, or an issuer rewarding liquidity providers, doesn't involve a security, so no transfer agent is needed. Just a paying agent function. Merkl alone covers it. See how it works for redistributing stablecoin yield in practice.
The pattern is consistent. If there's a regulated security, you need both. If it's an open distribution to non-securities holders, you need the paying agent layer only. The mistake to avoid is assuming one role can cover the other. A smart contract that distributes cash is not a transfer agent. A token that maintains a registry is not a paying agent. The two functions stay distinct even when they live on the same chain.
A transfer agent maintains the official register of who owns a security. A paying agent distributes cash flows from the issuer to those holders, such as dividends, bond coupons, or M&A proceeds. The transfer agent answers "who owns this?", the paying agent answers "how do they get paid?".
Yes. Large custodian banks like BNY Mellon and Citi often act as both for the same issuer, and the two services are commonly bundled in a single mandate. They remain legally and operationally distinct functions, with separate appointments and separate liability.
For publicly listed securities in the US, yes. SEC rules under Section 17A require issuers to maintain a registered transfer agent. Many private placements and tokenized offerings also use a regulated transfer agent to satisfy KYC, transfer restriction, and recordkeeping obligations.
For regulated tokenized securities, yes. The token contract acts as the technical registry, but a regulated transfer agent is still required to enforce KYC and transfer restrictions, manage the whitelist, and maintain the official book of holders for compliance purposes. Companies like Securitize provide this function natively onchain.
An onchain paying agent is a smart contract layer that distributes payments from an issuer to current token holders without intermediaries. It replaces the role traditional trust banks play for bond coupons, dividends, and fund distributions, with the difference that distributions can be programmable, real-time, and traceable through complex DeFi positions. Merkl is the leading example.
Merkl handles the distribution leg, the transfer agent handles the registry and compliance leg. The transfer agent provides Merkl with two inputs: a list of eligible holder addresses at the record date and the total amount to distribute. Merkl then executes the distribution onchain to those addresses. The transfer agent keeps full control over who is entitled to receive a payment, and Merkl handles only the technical execution.