Staff Report, Federal Reserve Bank of New York 336
We study two designs for a liquidity-saving mechanism (LSM), a queuing arrangement used with an interbank settlement system. We consider an environment where banks are subjected to liquidity shocks. Banks must make the decision to send, queue, or delay their payments after observing a noisy signal of the shock. With a balance-reactive LSM, banks can set a balance threshold below which payments are not released from the queue. Banks can choose their threshold such that the release of a payment from the queue is conditional on the liquidity shock. With a receipt-reactive LSM, a payment is released from the queue if an offsetting payment is received, regardless of the liquidity shock. We find that these two designs have opposite effects on different types of payments. Payments that are costly to delay will be settled at least as early, or earlier, with a receipt-reactive LSM. Payments that are not costly to delay will always be delayed with a receipt-reactive LSM, while some of them will be queued and settled early with a balance-reactive LSM. We also show that parameter values will determine which system provides higher welfare.
Liquidity-saving mechanisms real-time gross settlement large-value payment systems