Delivery versus payment (DvP) is a settlement method in which the transfer of securities from seller to buyer occurs simultaneously and conditionally with the transfer of cash from buyer to seller. Neither leg of the transaction settles independently: if either transfer cannot be completed, both are withheld. From the seller's perspective, the same mechanism is called receive versus payment (RvP) — the terminology reflects which side of the transaction is being described, not a different process.

DvP eliminates principal risk — the risk that one party delivers securities without receiving the corresponding cash, or pays cash without receiving the securities. This structural protection became the global standard following the October 1987 market crash, when the failure of several large settlements exposed the systemic danger of allowing securities and cash to move independently. The Bank for International Settlements (BIS) subsequently formalized DvP as the required settlement model for securities transactions in its 1992 report on delivery versus payment in securities settlement systems.

The move to T+1 settlement in the United States has reinforced DvP's importance by compressing the window between trade execution and settlement. At T+0, where atomic settlement eliminates that window entirely, DvP becomes the architectural foundation of the settlement layer rather than a procedural step within it.

In a DvP settlement, the buyer's cash account and the seller's securities account are debited at the same moment the corresponding credits are applied on the opposite side. The settlement agent — typically a central securities depository (CSD) or custodian bank — acts as the coordinating mechanism, holding both legs until both can be confirmed and released simultaneously.

Free of payment (FoP) settlement, where securities and cash move independently, exposes both counterparties to principal risk and liquidity risk during the interval between the two transfers. Cash or securities locked in transit cannot be deployed elsewhere, and if the second leg fails, the delivering party has lost both the asset and the use of the corresponding consideration. For institutional settlement at scale, this risk is structurally unacceptable. DvP removes it by design rather than relying on counterparty creditworthiness or timing discipline.

Three DvP models are recognized by the Bank for International Settlements (BIS):

• Model 1 — gross, trade-by-trade settlement of both legs simultaneously. This is the gold standard for high-value transactions and is the operational basis for many CSD-level real-time gross settlement (RTGS) systems.
• Model 2 — gross settlement of securities with net cash settlement at end of day. Securities transfer individually; cash settles as a single net position.
• Model 3 — net settlement of both securities and cash at end of day. Used in CSD environments where bilateral netting reduces settlement volume before final transfer.

Traditional DvP relies on the CSD or settlement agent as a trusted intermediary to enforce simultaneous transfer. On-chain DvP replaces the intermediary with smart contract logic: both legs are encoded in a single contract execution, and either both complete or neither does. This distinction is why a hybrid system must handle two trust models — the institutional intermediary model for traditional rails and the code-enforced model for on-chain settlement — under a single operational framework.