Current Issues2 November 26, 2010
Monitoring banking sector risk
Empirical evidence demonstrates thatbanking sector risk arises mainly from twosources:(i) a common exposure of banks to(domestic or foreign) macro risks,
(ii) contagious effects between banks,markets and countries.In systemic crises, usually both kinds of riskadd to and reinforce each other. Thus, bothaspects need to be considered whenmonitoring banking sector risk. Although anumber of tools have been developed toassess banking sector risk, there is noconsensus yet as to how to measuresystemic risk.
Complementing the approach presented in thispaper, Weistroffer and Vallés (2009) developed amonitoring tool to assess the risk of macro shocks tothe banking sector.
A more elaborate discussion about the challengesof measuring systemic risk can be found in Borio andDrehman (2009).
AU AustraliaAT AustriaBY BelarusBE BelgiumBG BulgariaCA CanadaHR CroatiaCY CyprusCZ Czech RepublicDK DenmarkEE EstoniaFI FinlandFR FranceDE GermanyGR GreeceHU HungaryIE IrelandIT ItalyJP JapanNL NetherlandsPL PolandPT PortugalRO RomaniaSP SerbiaSK SlovakiaSI SloveniaES SpainSE SwedenCH SwitzerlandTR TurkeyUK United KingdomUS United States
The Bank for International Settlements (BIS) consolidated bankingstatistics provide a rich data set of aggregate cross-borderexposures. The data are used primarily by the BIS, central banksand supervisory authorities to monitor vulnerabilities of and possiblespill-over effects for national banking sectors. It allows quantifying towhich extent banks are exposed to foreign credit risk
therebycomplementing the national view on bank credit exposures. The BISstatistics provide a valuable data source not only to the officialbodies, but also to institutional investors and internationally activebanks. Using the data on cross-border exposure can greatly benefitthe assessment and understanding of bank systemic risk
especially for the developed countries, which form the bulk of theBIS reporting countries.This article provides a primer on how the BIS consolidated bankingstatistics can be used to monitor banking sector risk that stems fromcross-country
exposure. The data can also be used to lookat the transmission of shocks through banks
which we plan to address in a follow-up paper. Inaddition to providing a brief introduction to the BIS statistics, thispaper demonstrates how network analysis can be deployed to
produce a bird’s eye view on interlinkages and structural changes in
The analytical framework presented in thispaper puts the data into perspective and helps to uncover
sometimes not so obvious
cross-country dependencies.We start by describing the scope and limitations of BIS data. Next,we calculate simple ratios at the country level that help assess thevulnerability of lenders to cross-border exposures. We then proceedby establishing a network of 19 BIS reporting countries and assesstheir mutual dependencies.
Responding to the heightened interestin the EU peripheral countries, we finally show how problems inthese countries translate into exposures of the international bankingsystem.
New interest in cross-country exposures
The financial crisis has demonstrated
that significantrisks to national banking sectors can stem not only from domesticasset and credit markets but also from cross-border exposures.Germany provides a case in point. Prior to the financial crisis,country risk indicators at the national level typically issued no alerts.However, German banks held a significant portion of claims on USborrowers (although to a good deal off-balance sheet), which leftthem highly vulnerable to the international credit crisis. Likewise,Belgium, the Netherlands and Switzerland were adversely affectedthrough their
Our approach is closely related to work by McGuire and Tarashev (2007), Hattoriand Suda (2007), Espinosa-Vega and Solé (2010) as well as to a recent reportpublished by Fitch Ratings (2010), who all apply network analysis to the BIS
banking statistics. By contrast, Von Peter (2010) looks at the BIS
banking statistics to identify important banking centres using networkmethods. Castrén and Kavonius (2009), in turn, use euro area flow of funds datato identify sectors and channels through which local shocks may propagatethrough the financial system.
Australia, Austria, Belgium, Canada, Denmark, France, Germany, Greece, Ireland,Italy, Japan, the Netherlands, Portugal, Spain, Sweden, Switzerland, Turkey, theUK and the US.