Journal of Financial Market Infrastructures

In this issue of The Journal of Financial Market Infrastructures I would like to draw your attention to an alternative to financial market infrastructures (FMIs): correspondent banking. Pictured as a network, corresponding banking relationships usually look like a random irregular graph, while the direct participants of an FMI form the familiar star-like graph. It is important that we study correspondent banking alongside FMIs, because a lot of payment traffic travels via that network. For the euro area, a total of about €1100 billion in value is transferred on a daily basis. The ratio of correspondent banking euro payments to large-value payment systems settling in euro is approximately 1W3. Correspondent banking is an alternative to the central processing of payments in an infrastructure, but it does not always cover the whole chain between the debtor and the end beneficiary. Roughly half of all correspondent banking payments pass through an FMI at some stage in the chain.

There is a reason why the literature on this topic is rather thin: it is not easy to collect relevant data due to its decentralized nature and, moreover, the data is highly sensitivity (the corresponding banking network provides information on where banks hold accounts with other banks). After all, correspondent banking is defined as an arrangement whereby one bank (the service-providing bank) makes or receives payments on behalf of another bank (the service-user bank). And yet, from the standpoint of risk control it does not matter whether the systemic risk is caused by a failure in correspondent banking or by the failure of an FMI.

You should be able to guess the topic of the first paper in the issue by now. It is entitled "Correspondent Banking in Euro: bank clustering via self-organizing maps" and it is by Fabio Franch. The paper takes its data from the survey on correspondent banking that is conducted regularly by the European Central Bank. The author analyzes the network structure using a data-mining technique called self-organizing maps. The outcome of the analysis confirms that the correspondent banking network in euro contains the two groups that are distinguished a priori in the survey: retail banks and wholesale banks (the line of division between the two groups is an average transaction size of €10 000). However, in addition, the author finds two additional groups: banks that operate domestically and banks that predominantly service customer banks outside their country of residence.

The issue's second paper, "Competition in bank-provided payment services" by Wilko Bolt and David Humphrey, aims to estimate the degree of competition in the field of payments services provided by banks. The authors use a frontier-based approach to overcome the data problem for assessing relative competition. They find that the largest banks in the United States come out in the middle of the range when it comes to their competition efficiency measure. Location within the United States also matters: banks in states with a relatively high income per capita and a high population density are usually in the most competitive quartile.

Our third paper, "A bill of goods: central counterparties and systemic risk" by Craig Pirrong, provides a critical assessment of the impact of the G20 central clearing obligation for standard over-the-counter derivatives. In this forum contribution the author argues that a truly systemic view of risk is missing. That is, as well as taking into account the impact of central clearing on central counterparties, the effects on its clearing members and the system-wide consequences of asking for more collateral (potentially in a procyclical manner) should be assessed. Pirrong then discusses some regulatory issues that need to be addressed.

I hope you enjoy reading this issue of The Journal of Financial Market Infrastructures.

Ron Berndsen
De Nederlandsche Bank and Tilburg University