Settlement bank behaviour and throughput rules in an RTGS payment system with collateralised intraday credit

A simple two-period, two-bank model of an RTGS system with collateralised intraday credit is presented. It is shown that two types of outcome are possible - inefficient or efficient - depending on whether banks care about payments imbalances between them in the first period. If they do, banks delay payments to each other, increasing their aggregate liquidity requirements. It is argued that efficiency is not guaranteed even when banks face repeated interaction in a real payment system, largely because of imperfect information and the competitive dynamics of the payment industry. An efficient outcome can be achieved by the imposition of throughput rules on the value of payments banks must make by a certain deadline. These can both reduce aggregate liquidity requirements and increase the contestability of the payments market, encouraging a higher degree of direct access to payment systems. Throughput rules could therefore also have risk-reduction benefits if they help to reduce the level of tiering in the financial system. The detailed characteristics of these rules are shown to be important, and a number of design issues are addressed, such as how frequently requirements should be set, and whether throughput rules should apply on an aggregate or bilateral basis.