financial stability report 2005 - swiss national bank...financial strength ratings* (moody’s)...

Financial Stability ReportBericht zur FinanzstabilitätRapport sur la stabilité financière

2005

Swiss National BankFinancial Stability Report

Bericht zur FinanzstabilitätRapport sur la stabilité financière

Table of Contents

4 Bericht zur Finanzstabilität 2005 (Übersicht)4 Vorwort4 Bankensektor7 Finanzmarktinfrastruktur

9 Rapport sur la stabilité financière 2005 (Synthèse)9 Avant-propos9 Secteur bancaire

13 Infrastructure des marchés financiers

14 2005 Financial Stability Report 15 Introduction15 Overall assessment17 Data and data sources

18 Part I: Banking sector19 1 General conditions26 2 Profitability30 3 Risks34 4 Capital base38 5 Market assessment41 6 Stress index for the banking sector

44 Part II: Financial market infrastructure45 1 Introduction46 2 Oversight of systemically important infrastructures53 3 Business continuity planning in the Swiss financial sector55 4 Developments in large-value payment systems

SNB 4 2005 Financial Stability Report

Bericht zur Finanzstabilität 2005(Übersicht)

VorwortDer vorliegende Bericht behandelt die unter

dem Aspekt der Stabilität massgebenden Tendenzendes schweizerischen Finanzsektors. Es ist das dritteMal, dass die Schweizerische Nationalbank (SNB)einen Jahresbericht zur Stabilität veröffentlicht.1 DieSNB bezweckt damit, die Öffentlichkeit über denZustand des Finanzsystems zu informieren. Sie über-mittelt auf diese Weise ihre Einschätzung der Stabili-tät des Finanzsystems und stellt der Öffentlichkeiteine Übersicht an Informationen und Indikatoren zurVerfügung. Der Stabilitätsbericht gibt der SNB dieMöglichkeit, auf Spannungen oder Ungleichgewichtehinzuweisen, die ein Risiko für die Stabilität darstel-len könnten. Der Bericht ist Teil der Beurteilung derStabilität des Finanzsystems, zu der die SNB gemässNationalbankgesetz Art.5, Abs.2, Bst. e. beizutragenhat. Der Bericht dient nicht dazu, die Solvenz einzel-ner Finanzinstitute zu beurteilen. Einzelne Bankenwerden nur betrachtet, wenn dies für das Gesamtbildwesentlich ist.

Ein stabiles Finanzsystem zeichnet sich dadurchaus, dass dessen Komponenten ihre Funktion erfüllenund sich gegenüber Schocks als widerstandsfähigerweisen. Dieser Bericht beschränkt sich auf zweiwesentliche Komponenten des Finanzsystems: denBankensektor und die Finanzmarktinfrastruktur.

BankensektorDie Einschätzung der Stabilität des Bankensek-

tors erfolgt in zwei Schritten. Zuerst wird die Ent-wicklung der Risikofaktoren analysiert, die im makro-ökonomischen Umfeld und auf den Finanzmärkten fürdie Stabilität des schweizerischen Bankensystemsrelevant sind. Anschliessend erfolgt die Beurteilungder Widerstandsfähigkeit des Bankensystems in Bezugauf diese Risikofaktoren. Dazu werden die Rentabili-tät, die eingegangenen Risiken sowie die Eigenmittel-ausstattung im Bankensektor gemessen. Ergänzendwerden Modelle berücksichtigt, die den auf den Ban-kensektor ausgeübten Stress und dessen Bezug zummakroökonomischen Umfeld quantifizieren.

Im Jahr 2004 war das Umfeld des schweizeri-schen Bankensektors mehrheitlich durch positiveEntwicklungen geprägt. In den Vereinigten Staatenund Japan gewann die Konjunktur weiter an Fahrt.Zudem verbesserte sich die Wirtschaftslage auch in

der Europäischen Union und in der Schweiz, die inden Vorjahren unter einer schwachen Konjunkturgelitten hatten. Gleichzeitig blieb das Zinsniveau inder Schweiz wie auch auf den meisten anderenFinanzplätzen sehr tief. Parallel dazu verzeichnetendie Börsen eine leichte Aufwärtstendenz bei starksinkender Volatilität. Aus den verfügbaren Indikato-ren geht hervor, dass diese Situation die Bonität ins-besondere bei den grossen in- und ausländischenUnternehmen positiv beeinflusst hat. Die Ratingshaben sich allgemein verbessert, und die Risikoprä-mien auf den Schulden sind weiter zurückgegangen.Der Anstieg der Konkursrate in der Schweiz kontras-tiert jedoch mit dieser Entwicklung und deutet daraufhin, dass sich die Bonität bei den kleineren und mitt-leren Firmen leicht verschlechtert haben könnte.

In diesem mehrheitlich positiven Umfeld konn-ten im schweizerischen Bankensektor die hohenGewinne des Vorjahres nicht nur gehalten, sondernüberwiegend sogar gesteigert werden. Da dieserGewinnanstieg mit einer proportionalen Erhöhungder Bilanzsumme einherging, veränderte sich dieRentabilität (Return on Assets) in Vergleich zu 2003jedoch kaum (vgl. Grafik 1). Während die gutenErgebnisse des Jahres 2003 primär auf Kostensen-kungen zurückzuführen waren, sind die positivenResultate des letzten Jahres gesteigerten operativenErträgen – insbesondere aus dem Kommissions- undDienstleistungsgeschäft – zu verdanken. Zudemkonnten die guten Vorjahresergebnisse im Zinsge-schäft mehrheitlich wiederholt werden. Ausserdemgelang es, das im Vorjahr deutlich verbesserte Cost-Income-Verhältnis konstant zu halten. Schliesslichwurden die im historischen Vergleich sehr niedrigenRückstellungen des Vorjahres weiter reduziert. Diesepositive Entwicklung kontrastiert zwar mit demAnstieg der Konkurse in der Schweiz, steht aber imEinklang mit der relativ günstigen Wirtschaftslageund der positiven Entwicklung der übrigen Indikato-ren der Schuldnerqualität. Die niedrigen Rückstel-lungen sind wohl zudem auf die relativ vorsichtigeKreditvergabepolitik der Banken während der letztenJahre zurückzuführen.

Die hohen Erträge führten zu einer Erhöhungder Eigenmittel im Bankensektor. Ende 2004 lag der Eigenmittelüberschuss deutlich über demhistorischen Durchschnitt (vgl. Grafik 2). Die ver-stärkte Eigenmitteldecke der Banken hat die Fähig-keit des Bankensektors verbessert Schocks abzufe-dern. Eine wichtige Ausnahme stellen allerdings dieGrossbanken dar, die ihre Eigenmittel trotz hoherGewinne kaum erhöhten. Die Lage der Grossbanken

1 Frühere Ausgaben des Stabilitätsberichts sind auf der Internetseiteder SNB (www.snb.ch) abrufbar.


bezüglich Eigenmittelausstattung hat sich somit imVergleich zum Vorjahr nur unwesentlich geändert.Aufgrund der für die Regulierung massgeblichenrisikogewichteten Eigenmittelquote liegen sie iminternationalen Vergleich weiterhin auf den vorderenRängen. Wird dagegen die ungewichtete Eigenmittel-quote als Massstab herangezogen – sie unterliegt inder Schweiz keiner formellen Regulierung –, ist dieEigenmittelausstattung der Grossbanken im interna-tionalen Vergleich nach wie vor niedrig.

Grafiken 1 und 2:Quellen: Eidgenössische Bankenkommission (EBK), SNB

Grafik 2*Einen wesentlichen Teil der Eigenmittel der Raiffeisenbanken bildet

die Nachschusspflicht der Genossenschafter. Seit 1995 kann diese nurnoch teilweise angerechnet werden, was zu einem Einbruch bei denEigenmitteln führte.

1999 2000 2001 2002 2003 2004

0.0

0.2

0.4

0.6

0.8

1.0

Return on Assets (nach Bankengruppen) Grafik 1

%Grossbanken Kantonalbanken Regionalbanken Raiffeisenbanken Bankensektor

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

0

10

20

30

40

50

60

70

80

Eigenmittelüberschuss in Prozent der erforderlichen Eigenmittel Grafik 2

%Grossbanken Kantonalbanken Regionalbanken Raiffeisenbanken* Bankensektor


Die aus den Bewertungen durch die Finanzmärk-te hervorgehenden Indikatoren bekräftigen in derRegel den allgemeinen Eindruck der Robustheit, dendie hohen Gewinne und die gute Eigenmittelausstat-tung der Banken vermitteln. Der schweizerische Ban-kensektor erscheint aufgrund der Spreads der Obliga-tionenrenditen und der Credit-Default-Swap-Preisesowohl im historischen wie auch im internationalenVergleich als robust (vgl. Grafik 3). Aufgrund derAktienkurse und der Ratings der einschlägigen Agen-turen lässt sich jedoch nicht ableiten dass der Marktdie Schweizer Banken im internationalen Vergleichsystematisch als überdurchschnittlich sicher ein-schätzt (vgl. Grafiken 4 und 5).

Der allgemeine Eindruck, dass der SchweizerBankensektor momentan in einer soliden Verfassungist, wird durch den SNB-Stressindex bestätigt.2 DieserIndikator fasst eine Vielzahl von Informationen übermögliche Stresssymptome im Bankensektor zusam-men. Er berücksichtigt insbesondere die Entwicklungder Gewinne, der Aktienkurse und der Risikoprämienauf den Obligationen des schweizerischen Bankensek-tors. Der Index deutet darauf hin, dass der im Jahr2004 vom Schweizer Bankensektor erfahrene Stress inVergleich zum historischen Durchschnitt sehr tief war.(vgl. Grafik 6). Zudem zeigten sich in keinem Wirt-schaftsbereich grössere strukturelle Ungleichgewich-te, die sich als Stressfaktoren auf den Bankensektorauswirken könnten.

2002 2003 2004 2005

0

50

100

150

200

250

Renditedifferenzen (spreads) Grafik 3Zwischen Banken- und Schweizer Bundesobligationen*

BasispunkteZürcher KB BC Vaudoise BC de Genève Raiffeisenbanken Bankensektor

<C- C- C C+ B- B B+ A- A

0

5

10

15

20

25CS UBS

CSFB

Financial strength ratings* (Moody’s) Grafik 4

Anzahl BankenMoody’s

<C/D C/D C B/C B A/B A

0

5

10

15

20

25CSG UBS

Individual bank ratings* (Fitch) Grafik 5

Number of banksFitch

Grafik 3: Quellen: SNB, Thomson Datastream

Grafik 3:*Mittelwert der Spreads auf allen Obligationen, welche die folgenden

Kriterien erfüllen: fixe Couponauszahlungen, keine Optionen, in CHF,minimale Restlaufzeit von zwei Jahren. Stand am Ende des Monats.

2 Für eine detaillierte Beschreibung dieses Indicators siehe Kapitel 6.

Grafik 4: Quelle: Moody’s, Mai 2005Grafik 5: Quelle: Bankscope, Mai 2005

Grafiken 4 und 5:*Auswahl der gemäss „The Banker“ (Juli 2004) weltweit grössten

Banken aus Nordamerika, Japan und Europa, sofern sie über ein Ratingvon Moody’s, Standard & Poor’s und Fitch verfügten. Ein fehlendesFinancial Strength Rating auf Gruppenebene wurde durch das Ratingder grössten Tochtergesellschaft ersetzt.


Bezüglich der Zukunftsaussichten sind wir derAuffassung, dass das makroökonomische Umfeld unddie Finanzmärkte für die Stabilität des schweizeri-schen Bankensystems grundsätzlich keine grossenRisiken bergen. Die Perspektiven für 2005 deutenzwar auf eine Verlangsamung des Wirtschaftswachs-tums in den meisten Regionen hin, allerdings ineinem moderaten Ausmass. Zudem geht aus den ver-fügbaren Indikatoren hervor, dass die Gefahr einerdurch Ansteckung übertragenen Krise für den Schwei-zer Bankensektor zurzeit klein ist. Zum einen schei-nen die weltweit wichtigsten Banken- und Versiche-rungssektoren ebenfalls ziemlich robust zu sein. Zumanderen ist der Schweizer Bankensektor gegenüberden Hedge Funds (siehe Box 1, S. 23) nicht wesent-lich exponiert. Schliesslich deutet die Analyse desHypothekarmarktes darauf hin, dass sich der Wettbe-werb in diesem Geschäft verstärkt haben könnte, dadie Margen aus dem Zinsgeschäft tendenziell zurück-gegangen sind. Wir verfügen jedoch über keine Indi-katoren, die Hinweise darauf geben, dass dieser Wett-bewerbsdruck destabilisierend gewirkt hat. So bliebder Preisanstieg auf dem schweizerischen Immobi-lienmarkt im historischen und im internationalenVergleich moderat; eine abrupte Anpassung diesesMarkts nach unten erscheint daher unwahrscheinlich.Zudem sind in den letzten Jahren weder die Hypothe-karforderungen insgesamt noch der Anteil der risiko-reicheren Hypotheken im zweiten und dritten Rang imHypothekarportfolio der Banken stark gewachsen. Esgilt jedoch zu betonen, dass dies eine Lockerung derVergabekriterien der Banken im Hypothekargeschäft– zum Beispiel durch eine von der SNB nicht beob-achtbare Erhöhung der Belehnungswerte innerhalbder Hypothekenränge – nicht ausschliesst.

Zwei potenzielle Gefahrenquellen bestehenjedoch. Erstens könnte eine starke konjunkturelleAbschwächung negative Auswirkungen auf die Quali-tät der Kreditportfolios und auf die Börsenkursehaben. Zweitens könnte, im Fall einer nachhaltigenWirtschaftserholung, eine unerwartet starke Erhö-hung der nach wie vor sehr niedrigen Zinssätze dieQualität der Kreditportfolios ebenfalls beeinträchti-gen, insofern die verschuldeten Haushalte und Unter-nehmen durch die steigende Zinslast unerwartetstark belastet würden. Zudem würde eine starke Zins-erhöhung die Banken auch direkt treffen, da ihreZinsbindung bei den Aktiven tendenziell länger ist alsbei den Passiven. Angesichts der niedrigen Rückstel-lungen sowie der eingegangenen Zinsrisiken imschweizerischen Bankensektor könnten die negativenKonsequenzen einer unerwartet starken Zinserhö-hung relativ gross sein. Aus unserer Szenario-Analysegeht allerdings hervor, dass der schweizerische Ban-kensektor über genügend Eigenmittel verfügt, umeine deutliche Verschlechterung der Konjunktur undder Börse sowie einen erheblichen Anstieg der Zins-sätze zu verkraften (siehe Box 3, S. 29).

FinanzmarktinfrastrukturIm Bereich der Abrechnung und Abwicklung von

Zahlungen und Geschäften mit Wertschriften undanderen Finanzinstrumenten verfügt der FinanzplatzSchweiz über eine reibungslos funktionierendeFinanzmarktinfrastruktur, die hinsichtlich Sicherheitund Effizienz auch im internationalen Vergleich sehrgut abschneidet. Aus Sicht der Stabilität des schwei-zerischen Finanzsystems besonders bedeutsam sinddie innerhalb der «Swiss value chain» miteinander

Grafik 6: Quellen: EBK, SNB, Thomson Datastream

*Je höher das Niveau des Index, desto grösser ist das Stressniveau desschweizerischen Bankensektors. Der Index ist in Standardabweichun-gen von seinem historischen Durchschnitt (1987–2004) bemessen. Einpositiver (negativer) Wert bedeutet, dass der Stress grösser (kleiner)ist als der historische Durchschnitt.

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

-3

-2

-1

0

1

2

3

Stressindex* Grafik 6In Standardabweichungen

Effektiv Prognostiziert 90%-Vertrauensintervall


verbundenen Zahlungs- und Effektenabwicklungs-systeme. Dazu zählen das Zahlungssystem Swiss Inter-bank Clearing (SIC), das Wertschriftenabwicklungs-system SECOM und die zentrale Gegenpartei x-clear.Die ersten beiden dieser Infrastrukturen haben sichseit Jahren bewährt. x-clear wurde 2003 eingeführtund hat sich bisher ebenso bewährt. Alle diese Infra-strukturen vermindern aufgrund ihrer Architektur dieAbwicklungsrisiken wesentlich. Durch die Verbindungvon SIC und SECOM ist namentlich die real-time-Abwicklung von Effektengeschäften nach dem PrinzipLieferung gegen Zahlung möglich. Die zentrale Gegen-partei x-clear bringt für die Teilnehmer der elektroni-schen Handelsplattform virt-x zudem den Vorteil derEliminierung der einzelnen Gegenparteirisiken.

Aufgrund ihrer Bedeutung für die Stabilität desschweizerischen Finanzsystems werden SIC, SECOMund x-clear durch die SNB überwacht. Seit demInkrafttreten des revidierten Nationalbankgesetzesam 1. Mai 2004 hat die SNB die Überwachung dieserSysteme vorbereitet und vollzieht zurzeit die erstenpraktischen Schritte um die Einhaltung der in derNationalbankverordnung festgelegten Mindestanfor-derungen durch die Betreiber dieser Systeme fortlau-fend zu überprüfen. Ebenfalls von systemischerBedeutung für die Stabilität des schweizerischenFinanzsystems ist das MehrwährungszahlungssystemContinuous Linked Settlement (CLS), welches dieAbwicklung von Devisengeschäften in 15 Währungengemäss dem Prinzip Zahlung gegen Zahlung ermög-

licht. CLS wurde von der SNB von der Einhaltung derMindestanforderungen befreit, da der in New Yorkansässige Betreiber, CLS Bank International, durchdas amerikanische Federal Reserve angemessen über-wacht wird. Das Federal Reserve konsultiert regel-mässig die übrigen Zentralbanken, deren Währungenin CLS integriert sind. Dies gilt auch für die SNB.

Das von operationellen Risiken ausgehendeSchadenspotenzial im Finanzsektor ist beträchtlich.Die Stärkung der Widerstandsfähigkeit des Finanz-systems in Krisensituationen durch eine verbesserteKrisenvorsorge ist deshalb ein Thema von grosserBedeutung, sowohl für die einzelnen Finanzinstituteals auch für die Finanzmarktinfrastrukturen. DerFinanzplatz Schweiz hat sich zum Ziel gesetzt, fürGeschäftsprozesse, die aus Sicht der Systemstabilitätkritisch sind, die Sicherheit bei Auftreten von grossenStörfällen zu erhöhen und damit Instabilitäten inner-halb des Finanzsystems als Folge solcher Störfälle zuvermeiden. Eine zu diesem Zweck durchgeführte sek-torweite Analyse der bestehenden Vorkehren bei denzentralen Finanzmarktinfrastrukturen und den kriti-schen Systemteilnehmern vermittelt ein insgesamtpositives Bild: Der Finanzplatz Schweiz ist auch auf grössere Störfälle gut vorbereitet. Dennochbestehen Verbesserungsmöglichkeiten; entsprechen-de Empfehlungen werden zur Zeit vorbereitet undwerden nach ihrer Verabschiedung von den einzelnenFinanzinstituten bzw. Infrastrukturbetreibern umzu-setzen sein.


Rapport sur la stabilité financière 2005 (Synthèse)

Avant-proposLe présent rapport met en évidence les grandes

tendances, sous l’angle de la stabilité, dans le sec-teur financier suisse. Il s’agit du troisième rapport(annuel) sur la stabilité financière publié par laBanque nationale suisse (BNS).3 En publiant un telrapport, la BNS a pour objectif de fournir au publicdes informations sur l’état du système financier. Ellefait part de son évaluation de la stabilité de ce sys-tème, met à la disposition du public une synthèsed’informations et d’indicateurs et signale, le caséchéant, des tensions ou des déséquilibres suscep-tibles de constituer un risque en matière de stabilité.Ce rapport fait partie de l’appréciation de la situationde la BNS sur le plan de la stabilité du système finan-cier, à laquelle elle a pour tâche de contribuer (art. 5,al. 2, let. e, LBN). Il n’a pas pour objet d’évaluer lasolvabilité d’établissements financiers pris indivi-duellement. Des établissements ne sont considéréssur une base individuelle que lorsque cela joue unrôle déterminant pour la vue d’ensemble.

Un système financier stable est un systèmedans lequel les diverses composantes remplissent leurfonction et sont en mesure de résister à d’éventuelschocs. Le présent rapport se concentre sur deux com-posantes essentielles du système financier: le secteurbancaire et les infrastructures des marchés finan-ciers.

Secteur bancaireNotre évaluation de la stabilité du secteur ban-

caire se fait en deux étapes. Nous analysons d’abordl’évolution des facteurs de risque qui, dans l’environ-nement macroéconomique et financier, sont perti-nents pour la stabilité du système bancaire suisse.Nous évaluons ensuite la capacité de résistance dusystème bancaire face à ces facteurs de risque. Cetteévaluation repose sur une mesure de la rentabilité,des risques encourus ainsi que de la dotation enfonds propres du secteur bancaire. Pour compléternotre analyse, nous utilisons des modèles qui per-mettent de quantifier le stress subi par le secteurbancaire et son lien avec l’environnement macroéco-nomique.

En 2004, le secteur bancaire suisse a opéré dansun environnement dans l’ensemble favorable. LesEtats-Unis et le Japon ont connu une croissance éco-

nomique soutenue. Par ailleurs la situation écono-mique s’est améliorée dans l’Union Européenne et enSuisse, après plusieurs années marquées par unecroissance faible. Parallèlement, le niveau des tauxd’intérêt est resté très bas, à la fois en Suisse et sur laplupart des autres places financières. En outre, lesmarchés boursiers ont amorcé un léger redressementaccompagné d’une baisse substantielle de la volati-lité. Cette situation favorable semble avoir eu uneinfluence positive sur la solvabilité des grandesentreprises suisses et étrangères. Leurs ratings ontgénéralement augmenté et la prime de risque sur leurdette a diminué. L’augmentation du taux de faillitesen Suisse contraste cependant avec cette évolution etsuggère que, pour les petites et moyennes entre-prises, la situation pourrait s’être légèrement dété-riorée.

Dans cet environnement dans l’ensemble favo-rable, le secteur bancaire suisse a généré des béné-fices élevés, en augmentation par rapport à l’année2003. La rentabilité du secteur bancaire, mesurée parle biais de la rentabilité des actifs (Return on Assets),est cependant restée pratiquement inchangée, lahausse des bénéfices s’étant accompagnée d’uneaugmentation proportionnelle de la somme desbilans (cf. graphique 1). Alors que les bons résultatsenregistrés en 2003 avaient avant tout découlé d’uneréduction des coûts, ceux de 2004 sont dus à unaccroissement des produits opérationnels, notam-ment par le biais d’une hausse des revenus des com-missions et des prestations de services. Parallèle-ment, les bons résultats des opérations d’intérêtsatteints en 2003 ont généralement pu être mainte-nus. Par ailleurs, le ratio coût/revenu, qui avait puêtre nettement réduit l’année précédente, est restépratiquement inchangé. Finalement, les provisionsont encore été réduites. Cette baisse des provisions,dont le niveau à fin 2003 était déjà particulièrementbas en comparaison historique, contraste avec l’aug-mentation du nombre de faillites en Suisse. Elleconcorde cependant avec la situation économiquefavorable et avec l’amélioration de la qualité desdébiteurs suggérée par d’autres indicateurs. Le basniveau des provisions pourrait en outre refléter lapolitique relativement prudente suivie par les banquesen matière de crédits durant ces dernières années.

Les bénéfices élevés ont conduit à un renforce-ment de la dotation en fonds propres du secteur ban-caire. A fin 2004, l’excédent de fonds propres sesituait sensiblement au-dessus de la moyenne histo-rique (cf. graphique 2). Cette couverture accrue aamélioré la capacité du secteur bancaire à absorber

3 Les précédentes éditions du rapport sur la stabilité financière sontdisponibles sous www.snb.ch.


Graphiques 1 et 2:Sources: BNS, Commission fédérale des banques (CFB)

Graphique 2:*Une part importante des fonds propres des banques Raiffeisen est

constituée des versements supplémentaires auxquels se sont engagésles sociétaires. Depuis 1995, ces versements supplémentaires nepeuvent être comptés qu’en partie comme fonds propres, ce quiexplique la forte diminution observée cette année-là.

1999 2000 2001 2002 2003 2004

0.0

0.2

0.4

0.6

0.8

1.0

Rentabilité des actifs (par catégories de banques) Graphique 1

%Grandes banques Banques cantonales Banques régionales Banques Raiffeisen Secteur bancaire

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

0

10

20

30

40

50

60

70

80

Excédent de fonds propres, en pour cent du montant exigé Graphique 2

%Grandes banques Banques cantonales Banques régionales Banques Raiffeisen* Secteur bancaire

des chocs. Les grandes banques, dont les fondspropres sont restés quasiment inchangés malgré desbénéfices élevés, constituent toutefois une exceptionnotable. En matière de dotation de fonds propres, lasituation des grandes banques n’a donc que peu évo-lué. Sur la base des ratios de fonds propres pondérésen fonction des risques, déterminants sur le planréglementaire, les grandes banques occupent tou-jours les premiers rangs en comparaison internatio-nale. Par contre, mesurée à l’aune du ratio de fondspropres non pondéré – ratio qui ne fait pas l’objetd’une réglementation formelle en Suisse –, leur dota-tion en fonds propres reste faible en comparaisoninternationale.

L’impression générale de robustesse qui ressortdes bénéfices et de la dotation en fonds propres desbanques est dans l’ensemble corroborée par les indi-cateurs reflétant l’évaluation faite par les marchésfinanciers. Sur la base des écarts dans les rendementsdes obligations ainsi que des prix des swaps surdéfaillance (credit default swaps), le secteur bancairesuisse apparaît solide en comparaison historique etinternationale (cf. graphique 3). Les indicateurs baséssur le cours des actions ainsi que les ratings desagences spécialisées ne permettent cependant pas deconclure que le marché considère la solidité du secteurbancaire suisse comme systématiquement supérieure àla moyenne internationale (cf. graphiques 4 et 5).


L’indicateur de stress de la BNS confirme l’im-pression de robustesse du secteur bancaire suisse.4

Cet indicateur présente la synthèse d’un ensemble devariables constituant chacune un symptôme possiblede stress dans le secteur bancaire. Il tient comptenotamment de l’évolution des bénéfices, des coursdes actions et des primes de risque sur les obligationsdu secteur bancaire suisse. L’indicateur montre que leniveau de stress auquel le secteur bancaire suisse aété soumis en 2004 était très bas en comparaison his-torique (cf. Graphique 6). En outre, nous n’avonsconstaté aucun déséquilibre structurel, dont larésorption constituerait un facteur de stress im-portant pour le secteur bancaire.

En ce qui concerne les perspectives, nous consi-dérons que, dans l’ensemble, l’environnementmacroéconomique et les marchés financiers ne pré-sentent pas de menace majeure pour la stabilité dusystème bancaire suisse. Les prévisions pour 2005indiquent certes un ralentissement de la croissanceéconomique dans la plupart des régions, mais celui-cidevrait être modéré. De plus, il ressort des indica-teurs disponibles que le danger d’une crise par conta-gion est actuellement faible pour le secteur bancairesuisse. D’une part, sur le plan mondial, les principauxsecteurs bancaires et secteurs d’assurances semblenteux aussi relativement robustes. D’autre part, l’expo-sition du secteur bancaire suisse aux fonds d’arbi-

2002 2003 2004 2005

0

50

100

150

200

250

Ecarts de rendements Graphique 3Entre obligations de banques et de la Confédération suisse*

Points de baseZürcher KB BC Vaudoise BC de Genève Banques Raiffeisen Secteur bancaire

<C- C- C C+ B- B B+ A- A

0

5

10

15

20

25CS UBS

CSFB

Financial strength ratings* (Moody’s) Graphique 4

Nombre de banquesMoody’s


0

5

10

15

20

25CSG UBS

Individual bank ratings* (Fitch) Graphique 5

Nombre de banquesFitch

Graphique 4: Source: Moody’s, mai 2005Graphique 5: Source: Bankscope, mai 2005

Graphiques 4 et 5:*L’échantillon se compose des plus grandes banques au niveau mondial

de l’Amérique du Nord, du Japon et de l’Europe d’après "The Banker"(juillet 2004) qui sont notées à la fois par Moody’s, Standard&Poor’set Fitch. La notation d’un holding bancaire ne disposant pas de ratingest remplacée par celle de sa plus grande filiale.

Graphique 3: Sources: BNS, Thomson Datastream

Graphique 3:*Ecarts (spreads) moyens pour l’ensemble des obligations satisfaisant

aux conditions suivantes: coupon fixe, absence d’options, libellés enCHF, durées résiduelles égales ou supérieures à 2 ans. Données de finde mois.

4 Voir la section 6 du rapport pour une description détaillée de cet in-dicateur.


trages ou «hedge funds» (voir encadré 1, p. 23) estmodeste. Finalement, l’analyse du marché hypothé-caire suggère, sur la base du repli des marges dans lesopérations d’intérêts, que l’intensité de la concur-rence pourrait s’y être renforcée. Nous ne disposonscependant d’aucun indicateur montrant que la pres-sion concurrentielle aurait entraîné des consé-quences déstabilisantes. La hausse des prix sur lemarché immobilier en Suisse est en effet restéemodérée en comparaison historique et internatio-nale; un ajustement brusque à la baisse sur ce marchéapparaît donc improbable. Par ailleurs, ni lescréances hypothécaires ni, parmi celles-ci, la partrelativement risquée des prêts hypothécaires en 2e et3e rangs n’ont augmenté fortement ces dernièresannées. Il convient cependant de signaler que cettesituation n’exclut pas un relâchement des critèresd’octrois de prêts hypothécaires par les banques; untel relâchement peut, par exemple, revêtir la formed’un relèvement de la limite de crédit accordée ausein d’un rang hypothécaire, relèvement que la BNSn’est pas en mesure d’observer.

Deux sources de dangers potentiels doiventnéanmoins être signalées. Premièrement, un ralentis-sement sensible de la conjoncture pourrait avoir desrépercussions négatives sur la qualité des porte-feuilles de crédits ainsi que sur les cours boursiers.Deuxièmement, dans le cas d’une reprise économiquedurable, une hausse plus forte qu’anticipée des tauxd’intérêt – dont le niveau actuel reste particulière-ment bas – pourrait elle aussi entraîner une dégrada-tion de la qualité des portefeuilles de crédits, dans la

mesure où elle se traduirait par une augmentationbrusque et marquée du fardeau de la dette desménages et des entreprises. En outre, une haussesensible des taux d’intérêt affecterait directement lesbanques, du fait que les taux d’intérêt sur leurs actifssont généralement fixés pour des durées plus longuesque sur leurs passifs. Considérant le bas niveau desprovisions ainsi que l’exposition au risque de tauxd’intérêt dans le secteur bancaire, une hausse destaux d’intérêts plus forte qu’anticipée pourraitentraîner des répercussions négatives relativementimportantes. L’analyse de scénarios que nous avonsconduite indique cependant que la dotation en fondspropres du secteur bancaire suisse est suffisante pourrésister à une dégradation notable de la conjonctureet de la bourse ainsi qu’à une hausse sensible destaux d’intérêt (voir encadré 3, p. 29).

Sources: BNS, CFB, Thomson Datastream *Une valeur élevée de l’indice correspond à un niveau de stress élevédans le secteur bancaire suisse. Une valeur positive (négative) signifieque le stress est supérieur (inférieur) à sa moyenne observée entre1987 et 2004. La déviation par rapport à la moyenne est exprimée entermes d’écarts-type.

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

-3

-2

-1

0

1

2

3

Indice de stress* Graphique 6Déviation par rapport à la moyenne (écarts-type)

Final Prévision Intervalle de confiance à 90%


Infrastructure des marchés financiersDans le domaine de la compensation et du

règlement des paiements et des opérations sur titreset autres instruments financiers, la place financièresuisse dispose d’une infrastructure qui fonctionnebien et qui, sous l’angle de la sécurité et de l’effica-cité, occupe une position de choix en comparaisoninternationale. Les systèmes de paiement et de rè-glement des opérations sur titres qui sont reliés entreeux dans la «swiss value chain» revêtent une impor-tance particulière pour ce qui a trait à la stabilité dusystème financier suisse. Il s’agit surtout du SwissInterbank Clearing (SIC), pour les paiements, duSECOM, pour le règlement des opérations sur titres, etde x-clear, la contrepartie centrale. Les deux pre-miers systèmes ont fait leurs preuves depuis desannées; quant à x-clear, qui a été introduit en 2003,il a lui aussi bien fonctionné jusqu’à présent. Toutesces infrastructures contribuent, par leur architecture,à réduire sensiblement les risques de règlement.Etant donné que le SIC et le SECOM sont reliés entreeux, il est possible notamment d’assurer en tempsréel le règlement des opérations sur titres selon leprincipe livraison contre paiement. La contrepartiecentrale x-clear offre en outre aux participants de laplateforme de négoce électronique virt-x l’avantaged’éliminer les risques vis-à-vis de chacune des contre-parties avec lesquelles des opérations ont étéconclues.

Du fait de leur importance pour la stabilité dusystème financier suisse, le SIC, le SECOM et x-clearfont l’objet d’une surveillance de la part de la BNS.Depuis l’entrée en vigueur, le 1er mai 2004, de la loirévisée sur la Banque nationale, cette dernière a pro-cédé aux préparatifs en vue de la surveillance de cessystèmes. Actuellement, elle met en place les pre-mières mesures concrètes pour pouvoir vérifier enpermanence le respect des exigences minimales parles exploitants de ces systèmes. Le système de paie-

ment multidevises Continuous Linked Settlement(CLS), qui permet un règlement des opérations dechange dans 15 monnaies selon le principe paiementcontre paiement, est lui aussi d’importance systé-mique pour la stabilité du système financier suisse.La BNS a exempté CLS de l’obligation de satisfaire auxexigences minimales suisses, car l’exploitant CLSBank International, qui a son siège à New York, estdéjà soumis à une surveillance appropriée de laRéserve fédérale américaine. Celle-ci entretient descontacts réguliers avec les autres banques centralesdont les monnaies sont intégrées dans le systèmeCLS. La BNS compte également au nombre de ces ins-tituts d’émission.

Dans le secteur financier, le potentiel de dom-mages associé aux risques opérationnels est considé-rable. Renforcer la capacité de résistance du systèmefinancier en améliorant la prévention des crises revêtdonc une grande importance, aussi bien pour les éta-blissements financiers pris individuellement que pourles infrastructures des marchés financiers. La placefinancière suisse s’est fixé comme objectif d’accroîtrela sécurité, lors de l’apparition de perturbationsmajeures, des processus opérationnels qui sont sen-sibles sur le plan de la stabilité systémique et, par-tant, de juguler toute instabilité qui pourrait endécouler au sein du système financier. Menée à cettefin dans tout le secteur concerné, une analyse por-tant sur le dispositif de sécurité en vigueur actuelle-ment dans les infrastructures financières centrales etchez les participants d’importance sensible donneune image positive de la situation d’ensemble: laplace financière suisse est bien préparée pour faireégalement face à des perturbations majeures. Maisdes améliorations sont toujours possibles. Desrecommandations appropriées sont actuellement enpréparation et, une fois adoptées, devront être appli-quées par les établissements financiers et les exploi-tants d’infrastructures.

2005 Financial Stability Report



IntroductionThis report highlights the main trends in the

Swiss financial system with a view to their impact onstability. It is the third annual Financial StabilityReport published by the Swiss National Bank (SNB).5

The report aims to inform the public of the state ofthe financial system. Through this report, the SNBcommunicates its evaluation of the stability of thesystem, provides a set of information and indicatorsand highlights potential tensions or imbalances thatcould jeopardise system stability. The report formspart of the assessment of the financial system stabil-ity, to which the SNB is required to contribute accord-ing to the National Bank Act (art. 5 para. 2 (e) NBA).It is not the purpose of this report to analyse the sol-vency of individual financial institutions. Individualbanks are only considered if it is deemed relevant forthe overall picture.

A stable financial system can be defined as asystem where the various components fulfil theirfunctions and are able to withstand the shocks towhich they are exposed. This report focuses on twovital elements in the system: the banking sector andthe financial market infrastructure.

Overall assessmentBanking sectorThe analysis of the stability of the banking sec-

tor is divided into two steps. Firstly, we analyse thedevelopment of risk factors in the macroeconomicenvironment and in the financial markets which arerelevant to the stability of the Swiss banking sector.Secondly, the resilience of the banking system withregard to these risk factors is assessed. This involvesmeasuring the profitability, the risks taken and thecapital adequacy in the banking sector. To completethe analysis, we use indicators provided by modelsquantifying the level of stress experienced by thebanking sector and its connection to the macroeco-nomic environment.

In 2004, the Swiss banking sector operated in apredominantly positive environment. Economicgrowth in the US and Japan was robust. The econom-ic situation in the European Union and Switzerlandalso saw a recovery after having been dogged by weakgrowth in the previous years. Furthermore, the inter-est rate level remained very low in Switzerland, as tooin most other financial centres. Moreover, the stockmarkets reported a slight upswing accompanied by a

substantial drop in volatility. The available indicatorssuggest that this situation led to an improvement inthe financial standing of the large domestic and for-eign companies. Their credit ratings improved, whilethe risk premiums on their debts declined further.However, the increase in the number of bankruptciesin Switzerland contrasts with this trend. This sug-gests that in spite of the positive conditions, the situa-tion of the small and medium-sized companies mighthave deteriorated slightly.

In this predominantly positive environment, the Swiss banking sector not only succeeded in main-taining the high profit levels of the previous year, itactually managed to exceed them in most cases. How-ever, because the rise in profits was accompanied bya proportional increase in the balance sheet total,there was hardly any change in profitability (returnon assets). While the good results in 2003 were pri-marily attributable to cost-cutting measures, the posi-tive figures for 2004 can be ascribed to the increasein operating earnings, particularly from the commis-sion business and services. At the same time, thefavourable results achieved in 2003 in the interest-earning business could for the most part be repeated.It was also possible to maintain the cost-income ratioat last year’s improved level. Finally, the level of pro-visions – which was already very low by historicalstandards – was reduced yet further. On the onehand, this trend contrasts with the rise in the numberof bankruptcies in Switzerland, while on the otherhand, it is consistent with the relatively favourableeconomic environment and the positive developmentof other credit standing indicators. Moreover, the lowlevel of provisions might reflect banks’ relatively cau-tious lending policies in recent years.

The high earnings enabled the banking sector toincrease its capital base, leading to levels of excesscapital that significantly exceed the historical aver-age. This increase improved the banking sector’sresilience. One major exception, however, are the bigbanks, which – despite higher profits – hardlyincreased their capital base at all. The situation ofthe big banks with regard to capitalisation thusremained almost unchanged compared with the pre-vious year. Given the risk-weighted capital ratio rel-evant for regulation, they still rank among the bestwhen compared with their foreign competitors. How-ever, if the unweighted capital ratio is used as abenchmark (in Switzerland, this ratio is not subject toformal regulation), the capital levels of the big banksare still low in an international comparison.

5 The previous editions of the Financial Stability Report are availableat www.snb.ch.


The indicators reflecting financial market valua-tions generally shore up the overall impression ofrobustness, which is conveyed by the banks’ highprofits and solid capital base. Looking at the bondyield spreads and the credit default swap prices, theSwiss banking sector appears robust both in a histori-cal and international comparison. However, accord-ing to share prices and the ratings of major agencies,it does not appear that the market considers thesoundness of the Swiss banks to be systematicallyabove the international average.

Finally, this overall impression of robustness isconfirmed by the SNB stress index.6 This indicatorcombines a number of variables representing poten-tial stress symptoms in the banking industry. Amongother variables, it takes into consideration the devel-opment of profits, share prices and risk premiums onbonds in the Swiss banking industry. The values takenby the index suggest that the stress experienced bythe Swiss banking sector in 2004 was very low com-pared with the historical average (cf. Graph below).Furthermore, there was no sign of any structuralimbalances whose unwinding would constitute amajor source of stress for the banking sector.

As far as the prospects for the future are con-cerned, we believe that the macroeconomic environ-ment and financial market conditions pose no majorthreat to the stability of the Swiss banking system.The outlook for 2005 indicates a slowdown in eco-nomic growth in most regions, albeit to a moderatedegree. In addition, the available indicators suggestthat there is currently little risk of contagion effectscausing a crisis in the Swiss banking sector. On the

one hand, the world’s main banking and insurancesectors also appear to be quite robust. On the otherhand, the exposure of the Swiss banking industry tohedge funds (cf. Box 1, p. 23) is modest. Finally, theanalysis of the development of the mortgage marketindicates – based on the decline in margins from theinterest-earning business – that the competition inthis area has intensified. However, none of our indi-cators suggest that the competitive pressures havehad a destabilising effect. The price rise in the Swissreal estate market, for instance, remained moderateboth by historical and international standards; anabrupt downward adjustment in this segment istherefore unlikely. Furthermore, in recent years, nei-ther mortgage claims nor the proportion of higher-risk second and third-rank mortgages in banks’ mort-gage portfolios grew strongly. It should beemphasised, however, that this does not exclude arelaxation of the lending criteria by banks in themortgage business – for instance, through anincrease in the loan-to-value ratios within the mort-gage categories. Such an increase cannot be observedby the SNB.

Nonetheless, two potential sources of riskshould be mentioned. Firstly, a sharp economicdownturn could negatively impact the quality of theloan portfolios and the stock markets. Secondly, inthe event of a sustained economic recovery, an unex-pectedly sharp hike in the still very low interest ratescould also affect the loan portfolio quality. Such ahike would lead to a sudden and significant increasein the debt burden of households and companies.Banks would also be directly affected by a steep

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

-3

-2

-1

0

1

2

3

Stress index*In standard deviations

Actual Forecast 90% confidence interval

Sources: Swiss Federal Banking Commission (SFBC), SNB, Thomson Datastream

*The higher the level of the index, the higher the level of stress in the Swiss banking sector. The index is expressed in terms of standarddeviations from its 1987–2004 average. A value above (below) zeroindicates that the stress is above (below) its historical average. The stressindex for the first quarter of 2005 is computed with provisional data.

*For a description of the underlying variables and the methodology, cf. Box 5.

6 For a detailed description of this indicator, cf. Chapter 6 (Stress indexfor the banking sector) and Box 5.


interest rate hike, as the interest rates on their assetstend to be fixed for a longer period than those ontheir liabilities. In view of the interest rate risk takenand the low level of provisions in the Swiss bankingsector, the negative consequences of an unexpected-ly sharp interest rate increase could be relatively far-reaching. However, our scenario analysis indicatesthat the Swiss banking sector’s capital base is suffi-cient to withstand a substantial deterioration both ineconomic and stock market conditions as well as asharp rise in interest rates (cf. Box 3, p. 29).

Financial market infrastructureWith regard to the clearing and settlement of

payments and transactions involving securities andother financial instruments, the Swiss financial sec-tor has a smoothly functioning financial market infra-structure that compares very favourably with those ofother countries in terms of safety and efficiency. Ofparticular significance for the stability of the finan-cial system are the payment and securities settlementsystems which are interlinked within the Swiss valuechain. These include the payment system Swiss Inter-bank Clearing (SIC), the securities settlement systemSECOM and the central counterparty x-clear. The firsttwo are long-established infrastructures. x-clear,which was launched in 2003, has hitherto also provedits worth. Given their architectures, all of these sys-tems contribute considerably to the minimisation ofsettlement risks. The link between SIC and SECOMpermits real-time settlement of securities transac-tions in accordance with the delivery-versus-paymentprinciple. The central counterparty x-clear furtherbenefits the participants of the electronic tradingplatform virt-x as it eliminates the individual coun-terparty risks.

Given their significance for the stability of theSwiss financial system, SIC, SECOM and x-clear areoverseen by the SNB. Since the revised National BankAct (NBA) entered into force on 1 May 2004, the SNBprepared for the oversight of these systems and isnow implementing the first practical steps necessaryfor an ongoing review of the compliance of the opera-tors of these systems with the minimum requirementsprescribed in the National Bank Ordinance (NBO). Themulti-currency payment system Continuous LinkedSettlement (CLS) is also of systemic importance to

the stability of the Swiss financial system – it enablesthe settlement of foreign exchange transactions in 15 currencies in accordance with the payment-versus-payment principle. The SNB exempted CLS from com-pliance with the minimum requirements, as the NewYork-based system operator, CLS Bank International,is already adequately overseen by the US FederalReserve. The Federal Reserve has regular contact withthe other central banks whose currencies are integ-rated into CLS, including the SNB.

The damage potential emanating from opera-tional risks in the financial sector is considerable.Strengthening the financial system’s resilience tocrises through improved contingency planning istherefore a vitally important objective – both for theindividual financial institutions and for the financialmarket infrastructures. In the event of major disrup-tions, the Swiss financial sector set itself the goal ofincreasing the resilience of business processes thatare considered to be critical for system stability, thusavoiding instabilities within the financial system as aresult of such disruptions. A special sector-wideanalysis on the existing measures for central finan-cial market infrastructures and critical system partici-pants gives an overall positive impression: the Swissfinancial sector is well-prepared, even for severe dis-ruptions. Nevertheless, there is room for improve-ment. To this end, appropriate recommendations arecurrently being drawn up and, once approved, willhave to be implemented by the individual financialinstitutions and infrastructure operators.

Data and data sourcesUnless otherwise stated, all data used in this

report come from internal statistics prepared by theSNB and the Swiss Federal Banking Commission(SFBC). The banking statistics are based on officialdata submitted by the individual banks. The data onthe big banks are analysed on a consolidated basis.For the other banks, an individual (non-consoli-dated) view is used.

This document is based on the data available asat 15 May 2005.

Part I: Banking sector



1 General conditionsThe analysis of the economic and financial envi-

ronment is based on economic activity, interest rates,the credit standing of borrowers in general and ofmajor insurance companies and foreign banks in par-ticular, the real estate market and the stock market.7

These risk factors have been singled out on the basisof an analysis of the sensitivity of the banking sectorto a range of economic variables and on the basis ofinformation on its exposure to specific sectors of the economy. Overall, it appears that during 2004 the Swiss banking sector operated in a favourableenvironment. The outlook for 2005 is relatively good,both with regard to the Swiss and the internationaleconomy.

Robust economic growth in 2004Economic growth picked up in Switzerland in

2004 and – with a 1.7% rise in GDP – exceeded thegrowth rates of the three previous years, thus bring-ing to an end the slowdown in growth that has beenongoing since 2001 (cf. Graph 1). GDP growth alsoclimbed in Europe8 – to 2.1% – after having slippedcontinuously for the past three years. In the US andJapan, meanwhile, the uptrend of recent years con-tinued unabated in 2004, with a GDP increase of 4.4%and 2.7% respectively.9

Interest rates still lowOverall, interest rates remained very low in

2004. At 2.7% and 0.5% respectively, the averageyield on ten-year Swiss Confederation bonds and theaverage three-month Libor were both well below theaverage of the last 15 years (cf. Graph 2). In fact, thereal three-month rate reached a negative value of–0.3%.10 While the three-month Libor was raised intwo steps in 2004 from 0.25% in January to 0.75% inDecember, the average yield on ten-year Swiss Con-federation bonds dropped from 2.85% to 2.38%. Inview of the current growth prospects (which have notdeteriorated in the medium term) and the inflationforecasts, the medium and long-term interest ratelevels recorded at the end of 2004 are low. This is thecase both in Switzerland and abroad. Long-terminterest rates also fell back in Europe and Japan,while the three-month Libor trended sideways. In theUS, both short and long-term interest ratesincreased. In a longer-term perspective, however,their levels are still low.11

Sources: OECD, SNB

7 Swiss banks are discussed in Chapters 2–68 European Monetary Union9 Source: OECD

10 Annual average of Libor minus annual inflation of 0.8%11 Source: Reuters

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

-2

-1

0

1

2

3

4

5

Growth in GDP Graph 1

%Real GDP Switzerland Real GDP EMU


Mixed signals regarding credit standing of borrowersIndicators present a mixed picture of borrowers’

creditworthiness. The low yield spreads between theindices for industrial and government bonds suggesta high credit standing. After having dropped steeplyin 2003, these yield spreads declined yet further inthe US and Europe in 2004 (cf. Graph 3), hitting ahistorically low level. At the end of 2004, the spreadsreached 0.51% in Europe and 0.92% in the US. Theinterest rate differential in Switzerland remained vir-tually unchanged at its currently low level (–2 bp to0.61%). The credit rating trend also indicates animprovement in credit standing. Of the companiesassessed in Moody’s global survey in 2004, 14% wereupgraded and 9% downgraded. The development ofthe bankruptcy figures in Switzerland contrasts withthis picture. In 2004, the bankruptcy rate rose slight-ly from 2.14% to 2.26% (cf. Graph 4), reaching a levelclose to the average of the last ten years. At the sametime, losses from concluded bankruptcy proceedingsincreased for the first time since 1998 (+32%). How-ever, it should be emphasised that half of this signifi-cant rise is due to one single bankruptcy case.12

Better assessment of credit standing of major foreign banksA number of indicators point to an improvement

in the credit standing of major foreign banks. Firstly,most of the biggest institutions generated higherprofits in 2004 than in the previous year. The sectorwas thus able to improve its results for the secondtime running. The increase in earnings was broadbased and the write-downs and provisions amongEuropean banks were lower overall. Secondly, thecredit ratings tended to increase in the bankingindustry. It should be noted, however, that the rat-ings of the larger Japanese and German banks, whichare below-average by international standards, didnot follow this positive trend. Thirdly, the develop-ment of credit default swaps (CDS) for bank debtsindicates an improvement in banks’ credit quality (cf.Graph 5). CDS prices exhibited a downward trend inEurope, the US and Japan. The decline was relativelyweak in the US, where the average prices for theseswaps were high in an international comparison inDecember 2004.13

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

0123456789

10

Swiss interest rates Graph 2

%

Yields on ten-year Swiss Confederation bondsYields on five-year Swiss Confederation bonds Three-month Libor

Sources: Reuters, SNB

12 Sources: Swiss Federal Statistical Office (SFSO) 13 Sources: Bloomberg, Moody’s Investors Service


2002 2003 2004 2005

0

20

40

60

80

100

120

Five-year senior credit default swap prices Graph 5

Basis pointsSwitzerland Germany US Japan UK

Premiums for credit protection on issuer bank (average of largest banks in the country)

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0

1

2

3

4

5

6

Bankruptcy rate in Switzerland Graph 4

%

Left-hand scale: bankruptcy rate (cases filed)Right-hand scale: losses (cases settled)

CHF billions

2001 2002 2003 2004 2005

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Yield spreads Graph 3Between corporate and government bonds

%EMU* US** CH***

Graph 3: Sources: SNB, Thomson DatastreamGraph 4: Sources: SFSOGraph 5: Source: Bloomberg

Graph 3:* Euro Aggregate Corporate and Euro Aggregate Government indices,

Lehman Brothers** US Aggregate Corporate Investment Grade and US Aggregate

Government indices, Lehman Brothers*** Yields (spot rates) for corporate bonds with a rating of at least

BBB- and on Swiss Confederation bonds, calculated by the SNB


Mixed signals for major insurance companiesIndicators paint a mixed picture with regard to

the creditworthiness of Swiss and foreign insurancecompanies. The major companies abroad reportedhigher profits for 2004. This positive result is onlypartially reflected in the credit-standing indicators,however. On the one hand, the number of credit rat-ing upgrades among life insurance companies signifi-cantly exceeded the number of downgrades. At thesame time, the development of CDS for insurancecompany debts suggests a slight improvement in thecompanies’ credit standing. The decline in CDS priceswas particularly pronounced in Europe, where theyare currently much lower than the international aver-age. On the other hand, the trend regarding creditratings points to a deterioration of the creditworthi-ness in the non-life sector, which has seen a sharpincrease in claims.

The Swiss insurance sector was also able toincrease its profits. Furthermore, market indicatorspoint to an overall improvement in the financialrobustness of Swiss insurance companies. In a long-term comparison, however, the credit ratings of thebig Swiss insurers are still low (cf. Box 2, p. 25).Looking at the CDS prices, it can be inferred that thecredit standing of Swiss insurers is similar to that oftheir international competitors, but lower comparedwith the Swiss big banks.14

Moderate increase in real estate prices in SwitzerlandReal estate prices rose 2.3% in Switzerland

in 2004 (cf. Graph 6). Although growth rates werehigher than those of consumer prices (0.8%), theycan nonetheless be considered modest and fell backslightly on the previous year. In 2004, real estateprices were still 14% below the peak level recorded in1989 and there was no sign of a speculative bubble inthe Swiss real estate market. However, substantialprice increases were observed in some regions. Thisapplies in particular to Geneva and the Lake Genevaarea, where property prices climbed in 2004 by11.5%, and also to Ticino (+5.8%).15

The situation in the Swiss real estate market dif-fers from that of the UK and US, where propertyprices increased in 2004 by 12% and 11% respective-ly. By contrast, prices for real estate dropped inJapan (–3%) and Germany (–2%).16

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

0

50

100

150

200

250

300

Real estate prices Graph 6

Switzerland Germany US Japan UK AustraliaAll prices indexed to year 1995 = 100

Sources: Bank for International Settlements (BIS), Wüest & Partner (Price index for single-family houses)

14 Sources: Bloomberg, Moody’s Investors Service, Standard & Poor’s

15 Source: Wüest & Partner (Price index for single-family houses) 16 Figures are year-on-year growth rates for Q3 2004. Source: BIS


Hedge funds are private pooled investment limited part-nerships which fall outside most of the rules and regulationsgoverning traditional investment funds. We have reached theconclusion that for the time being hedge funds do not pose a particular threat to system stability – despite their recentstrong growth and reduced profitability. Hence, the direct regu-lation of hedge funds does not appear necessary from a finan-cial stability perspective. Due to the rapid evolution of thehedge fund industry, however, its development and its linkageswith the banking sector need to be closely monitored.

Typically, hedge fund investors are wealthy, sophisticat-ed individuals or institutions. The market discipline provided bythis investor class constitutes a significant counterweight tothe hedge funds’ low degree of regulation. We have not identi-fied any serious shortcomings as regards transparency. It is im-portant that the clients and counterparties of the hedge fundshave access to the relevant information. This seems to be thecase in general. Therefore, the absence of any comprehensive,public data source (e.g. on the risk profiles and the positions of the individual hedge funds) is not problematic. Besides,there are a number of public databases which provide a goodoverview of the hedge fund industry.

Hedge funds are rather small – both on an individual andon an aggregated level – and they pursue relatively hetero-geneous strategies. According to various estimates, there arecurrently over 8,000 hedge funds managing assets of approx-imately USD 1,000 billion in total. At the end of 2003, onlyaround a dozen funds held more than USD 10 billion in assets,and none held more than USD 20 billion.17 By comparison, insti-tutional investors manage USD 25,000 billion in shares andbonds, the global market capitalisation of all shares and bondstotals around USD 20,000 billion each, and there are over 100banks worldwide whose assets exceed USD 100 billion. And atthe end of 2004, the trading books of individual large interna-tionally banks represented up to USD 500 billion. It is thus un-likely that the liquidation of individual hedge funds coulddestabilise large segments of the financial markets, in spite ofthe fact that they generate a large share of the turnover insome segments.

Moreover, as hedge funds do not play a role in traditionalfinancial intermediation (deposits and lending), the failure of in-dividual hedge funds has no direct bearing on those functions ofthe financial system that are crucial for the economy as a whole.

Box 1: Hedge funds – a threat to system stability? Based on the above factors, we consider it unlikely that

at this stage hedge funds are capable of causing a major finan-cial crisis, i.e., they do not pose a direct threat to financial sta-bility. The failure of a single or several hedge funds is more like-ly to be the symptom of a crisis triggered by other factors, suchas a deterioration in the macroeconomic environment. How-ever, there is a potential for hedge funds to indirectly affect financial stability, namely through exposures of large bankinginstitutions. For the time being, the linkages between hedgefunds and large banking institutions do not represent a particu-lar danger. Although the hedge fund business constitutes a significant source of income for Switzerland’s big banks, at pres-ent they play only a limited role compared with the total in-come, managed assets and risks of these banks. The banks keepa watchful eye on their exposure to hedge funds. Moreover,banks’ transactions with hedge funds are subject to the regula-tions and supervision by the Swiss Federal Banking Commission.In particular, the same capital adequacy requirements as forother market and credit risks apply for hedge fund exposure.

From a financial stability perspective, a direct regulationof hedge funds appears to be neither necessary nor useful forthe time being. The minimum solution – the compulsory regis-tration of hedge funds with an oversight authority – would pro-vide a false sense of security, as this would limit neither therisks associated with hedge funds nor their possible contribu-tion to systemic risk. Likewise, it is not necessary to requesthedge funds to report their positions, given their modest sizeand the sophistication of their main counterparties. Such re-porting would, moreover, give the misleading impression oftransparency, since the high volatility of trading portfoliosmakes it virtually impossible to provide a timely, representativepicture of these positions. And finally, direct restrictions re-garding leverage, investment possibilities or risk profile seemunreasonable in view of the limited systemic significance ofhedge funds and the sophistication of the counterparties.

Given the rapid evolution of the hedge fund industry, thedevelopments in this area must nevertheless be observed to ensure that hedge funds do not jeopardise system stability inthe future. Particular attention should be paid to the banks’ ex-posure to hedge funds and the way they handle this exposure.This is all the more important as the banks are signalling thatthey are under mounting pressure from hedge funds to relaxtheir lending standards.

17 The LTCM hedge fund, which almost became insolvent in 1998, was an exception: it had assets totalling USD 100 billion undermanagement.


1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

0

1000

2000

3000

4000

5000

6000

7000

0

5

10

15

20

25

30

35

Development of Swiss Performance Index (SPI) Graph 7Left-hand scale: SPI (monthly average)Right-hand scale: Volatility of SPI (standard deviation of daily returns, 250-day moving average, annualised)

Source: SNB

Comparatively favourable stock market climateThe stock markets performed relatively well in

2004. The Swiss Performance Index (SPI) rose overthe course of the year by 5.1% to 4,235 points (cf.Graph 7). The major foreign stock market indices,too, were higher at the end of 2004 than at thebeginning of the year. The European market (DJSTOXX 50) climbed 4.3%, the US market (S&P 500)gained 9.0% and the Japanese index (Nikkei 225)moved up 7.6%. In addition, the volume of Swissequities traded on the Swiss stock exchangeincreased by 17%, after having declined in the threeprevious years. At the same time, the uncertainty inthe markets abated. The volatility of the SPI – whichhad already decreased in 2003 – staged yet anothermarked decline in 2004 (–45%), reaching a levelwhich is close to the 15-year low.

Relatively good prospects for 2005The outlook for 2005 is quite positive for both

the Swiss and the international economy. Firstly, thesituation in the financial markets since the beginningof 2005 appears relatively stable. Stock market pricesrose in Switzerland and Europe, while they declinedsomewhat in the US and Japan. At the same time,both stock market volatility and the yield spreadsremained at their low levels; even though the latterexperienced a significant – but temporary – increase

in May. Secondly, GDP growth is set to continue,albeit at a somewhat slower pace than in 2004. For2005, the SNB puts GDP growth in Switzerland ataround 1.0% and growth rates in Europe, the US andJapan are expected to be within the region of 1.5%,3.5% and 0.9% respectively. Thirdly, there is no indi-cation of a price bubble having developed in eitherthe Swiss real estate market or the stock market. Asharp change in prices, which could potentiallydestabilise the banking sector, is therefore unlikelyat present. Finally, available indicators suggest thatthere is little risk of contagion effects causing a crisisin the Swiss banking sector. For one thing, theworld’s biggest banking and insurance sectors alsoappear to be quite robust. For another, the exposureof the Swiss banking sector to hedge funds (cf. Box 1,p. 23) is modest.

Notwithstanding these relatively positiveprospects, risks do exist. Firstly, the flattening of theyield curve implies that the market expects only amoderate interest rate hike in the medium term. Anunexpectedly sharp hike might lead to a significantdeterioration of the creditworthiness of borrowersoverall. Debtors in countries where there is a highrisk of a price adjustment in the real estate marketcould be particularly affected. Secondly, a slowdownin economic growth combined with a drop in shareprices cannot be ruled out.


18 Excluding the accounts held with independent pension funds 19 Insurance companies accounted for 1% of domestic equities, and banks for 2%20 Credit default swaps, insolvency indicators (SNB calculations)21 Moody’s and Standard & Poor’s

Problems experienced by insurance companies can havedestabilising effects on the financial system. Essentially, twotypes of effects can be distinguished:

Direct macroeconomic effects: The four major direct ef-fects of an insurance company’s default are (1) the discontinua-tion of insurance services, (2) a possible loss of savings, (3) the disruption of the insurance company’s function as a credi-tor and (4) price effects on asset markets. The first two of theseeffects are more significant, as they could weaken the economyin the event of an insurance crisis. The threat to, or loss of, sav-ings in particular is an important factor in this regard, since itwould adversely affect consumer confidence. According to ourestimate, savings18 placed with Swiss insurance companiesamounted to approximately CHF 230 billion at the end of 2003,compared with CHF 1,000 billion held with banks. However, sav-ings placed with insurance companies are used only to a verylimited extent to finance transactions of private householdsand companies. The role played by insurance companies withregard to supplying the national economy with liquidity istherefore very modest compared to that of banks, and themacroeconomic effects of an insurance crisis are consequentlyfar less severe than those of a bank crisis.

Effects on the banking system: Problems in the insurancesector can generate high macroeconomic costs if they con-tribute to, or even trigger, a destabilisation of the banking sys-tem. The extent to which an insurance crisis will spread to thebanking system depends on the significance of possible trans-mission channels. In this respect, three different cases can bedistinguished:

1 Banks as main owners of insurance companies within a fi-nancial conglomerate. In Switzerland, this applies toCredit Suisse Group (CSG). At the end of 2003, borrowedcapital at Winterthur Insurance equalled 446% of CSG’sequity capital. In extreme circumstances, problems ex-perienced by Winterthur – which belongs to this banc-assurance group – could impact on CSG to such an extent that the banking entity’s existence would be jeop-ardised.

2 Banks as creditors and shareholders of insurance com-panies. Swiss banks are important lenders to insurancecompanies. In individual cases, the amount of the credit

Box 2: Risks for the financial system emanating from insurance companies limit – expressed as a percentage of the correspondingbank’s equity capital – is not negligible. This does not,however, imply that real concentration risks exist. Banks’stakes in insurance companies are of minor importancecompared with their overall lending activities. The sameis true of credit default swaps.

3 Assets held both by banks and by insurance companies.Insurance companies held domestic bonds to the valueof CHF 63 billion, or 25% of the market capitalisation, atthe end of 2003. The banks accounted for a 10% share.The stakes in the other asset markets are considerablysmaller.19 Negative price effects might possibly be ex-pected from the bond markets, particularly from the mar-ket for government bonds, in which insurance companiesheld a 28% stake at the end of 2003. Government bondsare relatively easy to price, though, and it is improbablethat an insurance crisis would significantly weaken thecredit rating of the Swiss Confederation. It is thus un-likely that problems at insurance companies would betransmitted to banks through price falls of bonds andother assets.The extent to which the risks emanating from the insur-

ance sector can jeopardise the stability of the financial systemdepends, on the one hand, on the significance of the differentabove-mentioned effects in the event of an insurance crisisand, on the other hand, on the robustness of Swiss insurancecompanies, i.e. on the likelihood of a crisis. As regards the sig-nificance of the effects, our analysis shows that their transmis-sion within CSG has the potential to destabilise the financialsystem. The other effects could contribute to – but not actuallytrigger – a destabilisation.

Regarding the robustness of Swiss insurance companies,the available indicators show that it deteriorated drastically in2000 and 2001 before stabilising at a low level in the followingtwo years and moderately improving in 2004: the insurancesector was able to boost its profit in 2003 and 2004. In general,moreover, the market indicators20 signal an improvement in thefinancial robustness of insurance companies in 2003 and 2004.The insurance sector’s overall capital base and the ratings21 ofthe large Swiss insurance companies are, however, still lowcompared with 1999.


2 ProfitabilityThe Swiss banking sector published positive

results for 2004. While profits grew compared withthe previous year, profitability – measured by returnon assets – remained practically unchanged at a rela-tively high level. Much of the renewed rise in earningsis attributable to the growth in income from commis-sions and services. By contrast, the development ofnet interest income was well below average. Giventhe higher lending volume, this indicates a decline inthe interest margin. In addition, the cost-incomeratio remained virtually stable. While the outlook for2005 is relatively good, growth in profits is expectedto be rather modest.

Renewed increase in earningsNet profit in the Swiss banking sector rose in

2004 by 14.3% to CHF 19.8 billion. The cantonalbanks reported a particularly sharp increase of30.6%. Only among the commercial and stockexchange banks did earnings show a slight fall(–1.3%). Given that the earnings growth at the bigbanks was accompanied by a steep increase in thebalance sheet total, the return on assets (ROA) of theSwiss banking industry remained almost unchangedat 0.56% (2003: 0.57%). The ROA of the big banksdropped slightly from 0.53% to 0.51%, while that ofthe cantonal banks climbed from 0.39% to 0.50% (cf.Graph 8). Another reason for the fall in ROA at the bigbanks is the accounting changeover at CSG from Swiss

GAAP to US GAAP. Its ROA for 2003 would have beenfar lower if it had been stated according to US GAAPinstead of Swiss GAAP. The return on equity, by con-trast, increased both across the entire banking sectorand at the big banks. Finally, only 22 banks posted aloss for 2004 – the lowest number in the last nineyears. Of these banks, there was one regional bankand 21 foreign banks, accounting altogether for only0.24% of the Swiss banking sector in terms of assets.

Gross profits in the Swiss banking industry roseby 12.3% to CHF 33.8 billion.22 Unlike in 2003, how-ever, the rise in 2004 was not due to further cost-cut-ting measures, but to higher earnings. Overall, earn-ings grew by 11.3% to CHF 94.3 billion. Thecontributions of the various components varied, how-ever (cf. Graph 9). While income from commissionbusiness and services grew by CHF 3.4 billion (8.3%),net interest income fell slightly by CHF 0.14 billion(–0.5%). Along with other indicators, the stagnationof net interest income together with rising lendingvolumes points to a moderate decline in the interestmargin. This may be due to an increase in competitivepressure, which is exerted in part by the big banks. Inthis bank category, both net interest income (–1.7%)and the interest margin dropped considerably. At thecantonal banks, net interest income fell by 1.1%,while at the regional and Raiffeisen banks itincreased by 1.1% and 3.4% respectively against abackground of higher lending volumes (these rosemoderately at the regional banks, but recorded asharp rise at the Raiffeisen banks).

1999 2000 2001 2002 2003 2004

0.0

0.2

0.4

0.6

0.8

1.0

Return on assets Graph 8

%Big banks Cantonal banks Regional banks Raiffeisen banks Banking sector

Sources: SFBC, SNB 22 Gross profit corresponds to net profit before taxes, extraordinaryincome and expenditure, write-downs and provisions, as well as depre-ciation of fixed assets.


Finally, net trading income developed very differently in the various bank categories. The bigbanks recorded a rise of 23.6%, while at the Raiffeisen banks and commercial and stock exchangebanks these earnings contracted by 14.8% and 6.9%respectively. Overall, net trading income increased byCHF 1.5 billion, or 15.8%. Considering the low level ofvolatility and the rather moderate price rise in theasset markets, this increase is remarkable.

Costs in the Swiss banking sector climbed againfor the first time since 2001 and, at CHF 60.4 billion,were 10.8% higher than in 2003. This rise is primari-ly attributable to the growth of the big banks in par-ticular. The operating expenses of the big banks roseby 13.4%, while those of the regional banksdecreased by 0.2%. The cost-income ratio in theSwiss banking industry receded slightly from 64.5%to 64.4%. The ratio across all the bank categoriesremained relatively constant. Only the Raiffeisenbanks reported a small increase from 55.1% to 58.4%(cf. Graph 10).

A further reason for the rise in the net profit isthe decline in write-downs, provisions and losses.These narrowed from 0.32% of total lending (cus-tomer claims and mortgage loans) to 0.22%, or CHF 2.3 billion. Write-downs, provisions and lossesdecreased in all bank categories and are now very lowby historical standards. On the one hand, this trendcontrasts with the rise in the number of bankruptciesin Switzerland, while on the other hand, it is consist-ent with the relatively favourable economic environ-ment and the positive development of other creditstanding indicators.

1999 2000 2001 2002 2003 2004

0

20

40

60

80

100

Earning components development Graph 9Banking sector

CHF billionsInterest income Commission income Trading income Other income Total income

Sources: SFBC, SNB


OutlookThe prospects for the profitability of the Swiss

banking sector in 2005 are relatively good. However,a number of factors suggest that the growth in prof-its is likely to be rather modest. Firstly, profitabilityis already high. Secondly, in recent years the largerbanks in particular cut costs significantly. However,the cost-cutting appeared to bottom out in 2004 andthe potential for further reductions seems to havedwindled somewhat. Thirdly, write-downs, provisionsand losses are at a level that leaves very little roomfor any downward movement. Thus, any growth inprofits must be generated first and foremost by higher earnings. This in turn depends on the rate ofgrowth in the Swiss economy, on the timing and mag-nitude of change to the interest rate levels and onthe extent to which competition – especially in thelending market – intensifies.

1999 2000 2001 2002 2003 2004

0

20

40

60

80

Cost-income ratio Graph 10

%Big banks Cantonal banks Regional banks Raiffeisen banks Banking sector

Sources: SFBC, SNB


Box 3: Scenario analysisThe scenario analysis described in this box uses micro-

econometric methods and a set of macroeconomic and individ-ual bank variables to reveal systematic relations between themacroeconomic environment and profitability in the Swissbanking sector. The objective of the analysis is threefold: (i) toidentify macroeconomic factors which are systematically linkedto the profitability of the banking sector; (ii) to simulate thepresent and future profitability of the banking sector on the ba-sis of these variables; (iii) to provide an additional instrumentto assess the resilience of the Swiss banking sector. The sce-nario analysis involves two steps. In the first step, three essen-tial components of bank earnings – net interest income, provi-sions, and net earnings from trading and commissions – areregressed on a set of macroeconomic variables and individualbank characteristics in a panel regression. The set of macro vari-ables includes short and long-term interest rates, real GDPgrowth, real estate prices, the unemployment rate, and the re-turn on the Swiss stock market. The parameters obtained fromthe regression, which are estimated over the 1987–2004 period,reflect the sensitivity of the banking sector’s earning compo-nents to changes in the macroeconomic environment. Those pa-rameters which are statistically significant are then used, in thesecond step, to simulate the banking sector’s profitability un-der four macroeconomic scenarios: Firstly, an interest rate sce-nario characterised by an increase of the short and long-termnominal interest rate by 200 bp to a level of 300 bp and 450 bp,respectively. Secondly, a recession scenario characterised bynegative GDP growth (–1%), a relatively high unemploymentrate (5%) as well as a decline in real estate prices (–10%).Thirdly, a stock market scenario where the SPI is assumed todecrease by 30%. And, finally, a scenario which combines theshocks specified under scenarios 1 to 3. The size of the shocksused in each scenario is in line with the extreme variationsobserved during the period considered for the estimation.

Based on the first-step regression analysis it appears,firstly, that net interest income is rather insensitive to macro-economic developments. The expected negative relation be-tween interest rate changes and the interest margin is support-ed by the data; however, it is weak and appears to havereversed in recent years. These results are consistent with theevidence from banks’ reporting on their direct interest rate risk(cf. Chapter 3). Secondly, bank provisions are positively relatedwith interest rate and unemployment rate on the one hand andnegatively related with GDP growth and real estate prices onthe other hand. Finally, the results show that there is a positivecorrelation between stock market prices and trading and com-mission income and a negative correlation between interestrates and trading and commission income. The results from the

second-step simulation are reported in the table below. Thetable contains two elements: Firstly, the profit for the bankingsector measured as a percentage of the level of excess capitalfor each scenario. Secondly, the difference with respect to thebaseline scenario which roughly represents the status quo. Forexample, according to the second scenario – which implies a200 bp interest rate increase – the profit of the banking sectorwould amount to 17% of excess capital. Compared to the base-line scenario, this implies a decrease of profit representing 16%of excess capital.

As can be seen from the table, the resilience of the Swissbanking sector to potential macroeconomic shocks appears tobe relatively high. For instance, an interest rate shock wouldlead to a decrease in profitability, through its impact on net in-terest income, provisions and the value of the bond portfolio.Under such a scenario, however, the banking sector would stillbe profitable. Similarly, the recession scenario would entailsubstantially higher provisions and hence reduced profits with-out, however, affecting the banking sector’s capitalisation. Incomparison, the consequences of a stock market crash would berelatively severe. Under this scenario, the banking sector as awhole would experience losses. However, a stock market crashdoes not by itself constitute a serious threat to the bankingsector, as the losses entailed are small when compared to thesize of the sector’s excess capital. Finally, the scenario assum-ing a joint occurrence of a strong interest rate increase, a re-cession and a stock market crash, would lead to substantiallosses and a considerable decrease of excess capital in thebanking industry.

The scenario analysis is subject to a number of limita-tions. In particular, stress tests based on statistical inferenceassume that the observed structural relation is stable. For ex-ample, the speed of adjustment of the banking sector’s expo-sure to a particular risk factor during a period of stress is as-sumed constant. As a consequence, the figures reported in thetable would overestimate the impact of a macroeconomic sce-nario, should the speed of exposure adjustment be higher nowthan during the period considered for the estimation. Anotherlimitation lies in the fact that the analysis does not account forpossible non-linearities in the influence of macro variables onbanks’ profitability. In particular, synergy effects may amplifythe impact on profitability of each macroeconomic variablewhen various shocks occur simultaneously. Because the periodused for the estimation never saw the joint occurrence of astrong interest rate increase, a recession and a stock marketcrash, the model may underestimate the real effect of such acombination of shocks. Despite these reservations, the scenarioanalysis seems to corroborate the current assessment that theSwiss banking sector is quite robust with regard to reasonablemacroeconomic shocks.

Scenario Profits Difference with respect to(in percent of excess capital) the baseline scenario

(percentage points)

1. Baseline scenario* 33%2. Interest rates increase (+200 bp; parallel shift) 17% –16%3. Recession (GDP growth: –1%) 20% –13%4. Stock market crash (SPI: –30%) –4% –37%5. Combined scenario (2. to 4. combined) –33% –66%*The baseline scenario approximately corresponds to the status quo.

Estimated profits for 2005 (total banking sector)


3 RisksIn terms of the proportion of non-performing

loans and of write-downs and provisions, the creditrisk of Swiss banks declined again from the previousyear and is now at a low level. Market risk and inter-est rate risk have risen slightly from their modest2003 levels. In short, total risk in the banking sectorappears to be relatively low. There is a danger, how-ever, that an unexpectedly sharp hike in the interestrates could trigger a rise in credit risk as debtorsencounter payment difficulties.

Credit risk declining againCredit risk measures the risk of default by the

counterparty, in other words, the risk that the coun-terparty will fail to make the agreed interest andrepayment instalments in full. New write-downs andprovisions made by banks may be used as an indicatorfor this risk. Write-downs and provisions reflect theaverage quality, i.e. credit standing, of the loan port-folios.

Write-downs and provisions for default risk as apercentage of total lending (customer claims andmortgage loans) declined from 2.3% at year-end2003 to 1.7% at the end of 2004. Moreover, the shareof total lending accounted for by non-performingloans23 dropped from 2.1% to 1.5%. Overall, there-

1999 2000 2001 2002 2003 2004

0

1

2

3

4

5

6

Non-performing loans Graph 12As a percentage of total lending

%Big banks Cantonal banks Regional banks Raiffeisen banks* Banking sector

1999 2000 2001 2002 2003 2004

0

1

2

3

4

5

6

Write-downs and provisions for default risks Graph 11Total, as a percentage of all lending


Graphs 11 and 12: Sources: SFBC, SNB Graphs 11 and 12: *Statistics for the Raiffeisen banks only available from 2001

23 Non-performing loans are claims against customers and mortgageloans where interest payments are at risk or no longer expected to bemade. A loan is considered to be non-performing if payments (includ-ing interest payments) related to this loan are more than 90 days overdue.


fore, the average quality of loan portfolios improvedbetween year-end 2003 and year-end 2004. Write-downs and provisions for default risk and non-per-forming loans vary between approximately 0.5% and3% of total lending, depending on the category ofbanks (cf. Graphs 11 and 12).

In 2004, new write-downs and provisions – as aproportion of total lending volume – have also regis-tered a marked decline compared with the previousyear. Write-downs in the banking sector as a wholeamounted to 0.22% (2003: 0.32%). The new write-down and provisioning requirements were thus belowthe average for the previous nine years (0.72%). Bro-ken down by categories of banks, this year’s propor-tion of new write-downs and provisions was 0.13% oftotal lending at the cantonal banks, 0.17% at the bigbanks, 0.20% at regional banks and 0.05% at theRaiffeisen banks.

The decline in credit risk was due in large part tothe efforts undertaken in recent years to boost loanquality. Bad debt was reduced by streamlining lend-ing portfolios, improving the quality of loans (e.g. bydemanding higher collateral) and implementing amore cautious lending policy. For instance, the shareof collateralised loans rose again in the past year andnow accounts for 83% (end-2003: 82%, average ofthe past 10 years: 78%). Moreover, the share of low-risk first-rank mortgages24 in domestic mortgageclaims rose again: depending on the bank category,these mortgages now account for 90% to 94% of totalmortgage claims.

The credit risk indicators discussed above arelargely guided by the present and the past. In otherwords, they reflect future developments only to a lim-

ited extent. Even though some more future-orientedindicators also suggest that credit risk is now lower,25

a deterioration cannot be ruled out. Should interestrates go up significantly, some debtors are likely toencounter payment difficulties. Estimates based onour scenario analyses (cf. Box 3, p. 29) suggest thata 200 bp rise in interest rates will reduce the bankingsector’s profitability by one-sixth. While this overalleffect also comprises direct interest rate change andvaluation risks (cf. the next two sections), the mainreason for the decline in profits is higher provision-ing and write-down requirements. In interpreting thescenarios, one must bear in mind that the period overwhich the parameters of the models were estimateddoes not include any periods when interest rates wereas low as they are at present. Consequently, it is con-ceivable that the effects of an interest rate increaseare in fact stronger than the model – which was esti-mated on the basis of higher interest rates – wouldimply.

In a longer-term perspective, it appears that thevolume of lending has remained within reasonablelimits in recent years. Firstly, between 1994 and2004, domestic mortgage claims grew at an averagerate of 3.3% (2004: 4.3%), whereas real estate pricesfell by an average of 0.3% p.a. over the same period(cf. Graph 6). There is no sign of a speculative realestate bubble accompanied by high growth in mort-gage claims, as there was in the late 1980s. Secondly,there has been a general reduction in total customerclaims in recent years. At the end of 2004, domesticclaims were 11% lower than in 1998, while foreignclaims remained at their 1998 level (cf. Graph 13).This suggests that the banks have pursued a cautious

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

100120140160180200220240260280

Customer claims (secured and unsecured) Graph 13Banking sector

CHF billionsDomestic Foreign

Sources: SFBC, SNB

24 First-rank mortgages are claims that do not exceed 2/3 of themarket value of residential real estate or 1/2 of building land and com-mercial real estate. The limit is just 1/3 for large industrial commercialproperties or industrial real estate.

25 Among the big banks, for instance, the ratings distribution of UBS’sloan portfolio tended to improve year-on-year. Credit Suisse Group’seconomic risk capital (ERC) for credit risk is lower than in the two pre-ceding years. Source: Annual reports.


lending policy overall and have not expanded lendingaggressively by lowering lending standards.26 Accord-ingly, neither mortgages nor customer claims seem toshow any structural imbalances that would reflect adivergence between the development of the lendingactivity and the economic fundamentals. This assess-ment is confirmed by the analysis of the stress index(cf. Box 5, p. 43), which does not currently identifyany macroeconomic imbalances in lending or in realestate prices. It should be noted in this context that,firstly, these statements refer to the banking sectoras a whole and to the entire country. Problems atindividual banks or local imbalances cannot be dis-counted. For instance, growth in mortgages in theLake Geneva region has been well above the nationalaverage for some years now and could thus indicateregional overheating. And secondly, it is conceivablethat the banks might relax their lending criteria byallowing higher lending limits within individual mort-gage ranks or by issuing more generous real estateappraisals. Due to a lack of data, the development ofthese variables does not, however, enter into ouranalysis. Should these variables have in fact changed– e.g. as a result of stiffer competition in the mort-gage business – our analysis based on the other indic-ators would underestimate the credit risk in the Swissbanking sector.

Direct interest rate risk moderateA direct interest rate risk exists if there is a seri-

ous mismatching between the repricing maturities27

of a bank’s assets and liabilities. Banks typically useshort-term liabilities to refinance long-term loans.As a result of such maturity transformations, interestrates on assets may be fixed for a longer period thaninterest rates on liabilities. If a bank finds itself insuch a position, a rise in interest rates reduces thepresent value of assets more significantly than thepresent value of liabilities, thus reducing the net pre-sent value of the bank. The interest rate risk statisticscompiled by the SNB for the SFBC measure the expo-sure of individual banks to changes in interest rates.Essentially, the change in the present value of indi-vidual on-balance and off-balance-sheet items result-ing from a change in interest rates is calculated. Thesum of the changes in the present value of bothassets and liabilities shows the change in the net pre-sent value of the banks.

The measurement of interest rate risk is partial-ly based on hypotheses. For certain positions, such assight deposits, savings accounts and traditional vari-able-rate mortgages, the interest rate lock-in periodis not explicitly specified in the contract. The devel-opment of interest rates for these positions dependson both the bank’s policy and on market pressures.Such pressure is linked in particular to the clients’options with regard to early withdrawal or earlyrepayment.

Sources: SFBC, SNB 26 The sharp rise in loans to foreign customers in 1996–1998 is attrib-utable to the big banks, which were realigning their strategic focus andexpanding their international presence in this period.27 Time interval until interest rates can be readjusted.

< -20%-20--15%

-15--10%-10--8%

-8--6%-6--4%

-4--2%-2-0%

0-2%2-4%

4-6%6-8%

8-10%10-15%

15-20%>20%

0

30

60

90

120

150

0

20

40

60

80

100

Interest rate risk Graph 14Change in net present value (NPV) as a percentage of equity assuming a 200-bp interest rate rise (December 2004)Left-hand scale: number of banks Right-hand scale: cumulative proportion of aggregate total assets of all banks

%


28 With risk constant and depending on the composition of the port-folio, this adjustment caused the published VaR figures to fall by up toa quarter. Despite a rise in risk during the year under review, the re-quired capital resulting from this risk model is lower at year-end 2004than the figure produced by the former model at the end of 2003. Source: UBS, media release, 13 October 2004.29 These include the two big banks in particular. As an alternative tothis model approach, under the current capital adequacy rules a bankcan calculate the required capital by weighting and adding up the indi-vidual trading positions without taking into account the correlationbetween the positions.

Our primary analysis of the interest rate statis-tics rests on banks’ assumptions with regard to theinterest lock-in period of all their various positions.This basic analysis suggests that the Swiss bankingsystem as a whole is well hedged against the risk ofinterest rate changes. If the general level of interestrates were to rise by 200 bp, the aggregate result forall banks would be a reduction in the net presentvalue corresponding to 4.6% of available capital(end-2003: 4.1%). For those bank categories tradi-tionally exposed to interest rate risk (big banks, can-tonal and regional banks and Raiffeisen banks), thefigure is between 3% and 8%. At most banks, the riskis close to this mean, but at others it deviatessharply. In terms of their share of the aggregate balance sheet total, however, these outliers are oflimited systemic importance (cf. Graph 14).

Besides, the National Bank performs a comple-mentary interest rate risk analysis which is based onstandard hypotheses regarding the interest rate lock-in period of some positions, in particular savingsdeposits and traditional variable-rate mortgages.While our standard hypotheses are close to the lock-in period assumed by the average bank in the sample,they are significantly shorter than the average lock-in periods assumed by the larger banks. Based on thishypothesis, with a 200-bp rise in interest rates, theaverage sensitivity to interest rates is between 6%and 16% of available capital, depending on the bankcategory. This sensitivity is 1.5 to 2 times higher thanthat measured on the basis of the banks’ hypotheses,but it can still be qualified as moderate.

Finally, it should be noted that the interest raterisk only takes into account the valuation risk result-ing from fluctuations in interest rates. A significantrise in interest rates will additionally result in liquid-ity or solvency problems for debtors, in particularthose with variable-rate loan agreements, which inturn means a higher risk for the bank (cf. previoussection). This risk, however, is reflected in a highercredit risk and not in the interest rate risk statistics.

Slightly higher market riskMarket risk is the risk that changes in market

prices will generate profits or losses. This price riskmainly affects banks’ trading books, financial assetsand unconsolidated stakes in other companies. Wetake the minimum capital requirements derived fromthese items as a measure of market risk.

In absolute terms, market risk in the Swissbanking sector rose by 4% year-on-year. This risk alsoincreased slightly compared to total capital require-ments, edging up from 12.3% to 12.6%. A breakdownby bank category shows that the risk was 9.4% at thecantonal banks (2003: 9.4%), 12.7% at the big banks(2003: 12.3%), 4.8% at the regional banks (2003:4.9%) and 2.6% at the Raiffeisen banks (2003:2.3%). Market risk directly related to the banks’ trad-ing activities showed a similarly marginal change andremains in the range of previous years. UBS is anotable exception here, in that its market risk, meas-ured by value-at-risk (VaR), rose by just under 20%.Since UBS adjusted its risk model last year, this sig-nificant increase in risk is not, however, reflected inthe capital requirements.28

The direct valuation risk, i.e. the risk that thebanks’ own securities portfolio could depreciate,seems low relative to the credit risk. However, thefigures need to be put into context on two counts.Firstly, banks whose capital requirements are calcu-lated on the basis of a VaR model29 profit from thecurrently low market volatility (cf. Graph 7). With thecomposition of the portfolio unchanged, the lowermarket volatility, the lower the VaR. A generalincrease in volatility would therefore push up risk andrequired capital. Secondly, indirect risks whichemanate from the financial markets are not included.For example, the performance of asset managementand investment banking operations depends heavilyon the varying financial market conditions. Conse-quently, in addition to the valuation risk in the bal-ance sheets, financial market risks have an impact onbank revenues.

Our scenario analyses (cf. Box 3, p. 29) providea guide to the overall impact of market risks, i.e. thesum of the direct and indirect risks mentioned here.The estimates show that a 30% slump in the Swissequity market would result in the banking sectorgoing into the red.


4 Capital baseExcess capital increased considerably in 2004 as

a result of good annual results. However, theimprovement at the big banks was modest comparedwith the rest of the banking sector, despite high prof-its. From a historical perspective, all bank categorieshave a high level of excess capital. Taking an interna-tional view, the big Swiss banks are leading in termsof risk-weighted capital ratios, but bring up the rearin terms of unweighted capital ratios.

< 0% 0-20% 20-40% 40-60% 60-80% >80%

0

15

30

45

60

75

0

20

40

60

80

100

Excess of capital in percent of required capital (individual banks) Graph 16

%

1994 2003 20041994 2003 2004

Left-hand scale: percentage of all banksRight-hand scale: cumulative percentage of aggregate total assets of all banks

%

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

0

10

20

30

40

50

60

70

80

Excess capital in percent of required capital Graph 15


Graphs 15 and 16: Sources: SFBC, SNB Graph 15: *A significant proportion of capital at the Raiffeisen banks comprises

the members’ obligation to pay in additional capital. As of 1995, onlypart of this can be included in eligible capital, hence the sharp drop incapital at the Raiffeisen banks.

Rise in risk-weighted capital ratiosThe risk-weighted capital ratios rose overall in

2004. On an aggregate level, excess capital as a per-centage of required capital rose by 6% to 67%. Theincrease was particularly marked at the cantonalbanks: their excess capital increased by 26% to 73%.The rise was slightly lower at the regional banks(+11% to 61%) and the Raiffeisen banks (+6% to79%). At the big banks, however, excess capital as apercentage of required capital was basicallyunchanged (+1% to 46%) (cf. Graph 15). Similarly,the core capital (BIS Tier 1 capital) ratio of the bigbanks experienced only a small change from 11.8% to12.0% on average.


30 Cf. Chapter 2 (Profitability)31 Unlike Switzerland, the US has capital adequacy regulations that stipulate risk-weighted and unweighted capital ratios. Further information can be found atwww.fdic.gov/regulations/laws/rules/2000-4400.html.

Swiss banking legislation prescribes minimum capital ad-equacy ratios (cf. arts. 11–14 Banking Ordinance). Essentially,capital backing is required for all on-balance-sheet assets, off-balance-sheet operations and other open items in the tradingbook and elsewhere. These items are of a diverse nature, how-ever, and the underlying risks vary depending on the counter-party and collateral provided. To take account of this, the vari-ous items are risk-weighted. Of these risk-weighted items, 8%must be backed by capital at all times (required capital). How-ever, the SFBC can relax or tighten the regulations in specificcases (cf. art. 4 para. 3 Banking Act). Cantonal banks with astate guarantee are permitted to reduce required capital by upto 12.5% (cf. art. 13 (b) Banking Ordinance).

The eligible capital used to calculate capital adequacycomprises three components: core capital, supplementary capi-tal and additional capital. Core capital comprises paid-up equitycapital, reserves, retained earnings brought forward and, undercertain conditions, profit of the current year. Supplementarycapital comprises hidden reserves, subordinated debt papersand certain hybrid instruments (e.g. mandatory convertiblebonds). Additional capital comprises unsecured, subordinatedand fully paid-up liabilities that are subject to a lock-up clausewhich prevents the payment of interest and repayment of theprincipal if this violates the capital adequacy requirements.

If banks have more eligible capital than required, theyare said to have excess capital. The risk-weighted capital ratiocomprises eligible capital as a percentage of risk-weighted as-sets. The unweighted capital ratio comprises eligible capital asa percentage of total assets.

Box 4: Regulatory framework Rise in excess capital mainly due to increasein eligible capitalThe improvement in risk-weighted capital ratios

was due to a rise in eligible capital. Thanks to highprofits30, eligible capital was up 4% year-on-year inthe banking sector as a whole. In particular, the can-tonal banks (+9%) and Raiffeisen banks (+8%) great-ly strengthened their capital base. The only categorywhere the rise was modest – compared to the highprofits – was the big banks (+1%). This is attributableto share buy-back programmes and high dividendpayments. Required capital remained virtuallyunchanged at the big banks, cantonal banks andregional banks. This was particularly surprising at the big banks, as their total assets increased by 20% in 2004. However, the rise mainly related to receiv-ables with a risk weighting of 0%. As a result, totalrequired capital for balance-sheet assets actuallydeclined despite the sharp increase in total assets.The Raiffeisen banks were the only category in whichrequired capital increased (+5%), in line with the risein total assets (+4%).

By historical standards, the Swiss banks havestrong risk-weighted capitalisation. In the past tenyears, the capitalisation of most banks has improvedconsiderably (cf. Graph 16; bars). This applies to bothsmall and large banks: measured by assets, the mar-ket share of those banks with high excess capital hasrisen (cf. Graph 16; lines). At year-end 2004, all bankshad excess capital amounting to at least 20% ofrequired capital.

Looking at unweighted capital ratios, the pictureis less uniform. The increase in eligible capitalaccompanied by an only slight rise in total assets isclearly reflected in the unweighted capital ratios forthe cantonal, regional and Raiffeisen banks; theyrose from 7.0% in 2003 to 7.5% in 2004 at the can-tonal banks, from 7.4% to 7.6% at the regional banksand from 6.9% to 7.2% at the Raiffeisen banks. Bycontrast, the unweighted capital ratio at the bigbanks decreased from 2.9% to 2.5% in 2004(–15.9%) due to strong growth in total assets. Thisdivergent development is part of a longer-term trend:at year-end 2004, proportionately more banks hadhigh unweighted capital ratios than in 1994 (cf.Graph 17; bars). However, their importance – interms of their share of the banking sector’s totalassets – has declined sharply (cf. Graph 17; lines). InSwitzerland, there are no regulatory restrictions on theratio of capital and unweighted assets.31 The regula-tory capital adequacy requirements refer exclusivelyto risk-weighted assets (cf. Box 4). Nevertheless, the


unweighted capital ratios are important when assess-ing the solidity of a bank’s capital base. Since not allrisks can be covered fully by capital adequacy require-ments, risk-weighted capital ratios do not contain allinformation of relevance when judging solvency.32

Unweighted capital ratios can round out the pictureby indicating the bank’s level of indebtedness andthe buffer available to absorb shocks, regardless ofhow risky its business was considered to be ex ante.

Divergent trends in risk-weighted and unweighted capital ratios at big banksAt the big banks, the two capital ratios have

diverged in recent years. While the risk-weightedratios are essentially rising, the unweighted ratiosare declining. In other words, capital coverage oftotal assets is declining, yet these banks are able toreport a rise in excess capital compared to requiredcapital. This trend has led to a discrepancy betweenthe Swiss big banks’ risk-weighted and unweightedcapital ratios, which is particularly large in an inter-national comparison. On the one hand, when com-pared with 50 of the largest international banks inthe US, Japan and Europe, the Swiss big banks haveabove-average risk-weighted capital ratios. Measuredby the BIS total capital ratios, CSG ranks top(16.6%), while UBS ranks third (13.6%). On the otherhand, their unweighted captial ratios are low: lookingat BIS total capital as a percentage of total assets,they bring up the rear (CSG (3%) is 47th, while UBS(2.1%) ranks 50th out of 50). The discrepancy betweentheir risk-weighted and unweighted capital ratios istherefore exceptionally high (cf. Graph 18).33

33 Sources: Bloomberg, annual reportsSources: SFBC, SNB

32 The revision of the Basel Capital Accord, which forms the basis forstandards on risk-weighted capital, was triggered by a desire to reflectthe various banking risks more accurately.

< 4% 4-6% 6-8% 8-10% 10-12% >12%

0

15

30

45

60

75

0

20

40

60

80

100

Unweighted capital ratios (individual banks) Graph 17

%

1994 2003 20041994 2003 2004

Left-hand scale: percentage of all banksRight-hand scale: cumulative percentage of aggregate total assets of all banks

%


This discrepancy is attributable to the fact thatthe ratio of risk-weighted assets (including off-bal-ance-sheet items) to total assets is very low. Thecomparatively low unweighted capital ratios of theSwiss big banks cannot be explained by (i) differ-ences in international accounting standards, (ii) dif-ferences in the accounting treatment of the repur-chase value of derivatives, or (iii) different volumesof securitisation. Moreover, it is only due to a smallextent to reverse repurchase transactions.34 If reverserepurchase transactions and liquid assets werededucted from total assets, the unweighted capitalratio for the big banks would be 3.4% rather than2.6%. However, that is still low by international stand-ards (6.3% on average). The main reasons for thisdiscrepancy are the comparatively low proportion ofoff-balance-sheet business at the Swiss banks and itslow average risk weighting. It is also partly due to thelow average risk weighting of on-balance-sheet trad-ing inventories and loans at the Swiss big banks. Inother words, according to the Basel Capital Accordrisk-weighting scheme, off-balance-sheet items andtrading and loan books at Swiss banks are consideredto be less risky than those of their foreign competi-tors.

Scenario analysis and market indicators suggest capital base is adequate overallDetermining the adequacy of the capital base

would require a complete identification and assess-ment of all the risks to which banks are exposed.Since we do not have such an accurate yardstick, webase our analysis on a set of indicators. In particular, tocomplete the picture provided by the (risk-weightedand unweighted) capital ratios, we also conduct ascenario analysis (cf. Box 3, p. 29). According to thisanalysis, the Swiss banks seem to have a sound capi-tal base. This suggests that the banking sector shouldbe well able to withstand even larger shocks. All bankcategories should be able to continue to reportexcess capital, even in an unfavourable scenario fea-turing higher interest rates, a recession and a slumpin the financial markets. A deterioration in operatingconditions on this scale would nevertheless trimexcess capital in all categories and could seriouslyweaken some individual banks. Finally, we also use amarket assessment to evaluate the capital adequacyof the banking sector.35 The results obtained are some-what less favourable than suggested by the scenarioanalysis. According to the market assessment, thesoundness of the Swiss big banks is not as good asindicated by the risk-weighted capital ratios, but isbetter than the unweighted capital ratios suggest.

34 In a repurchase transaction, the recipient of funds sells securities to the provider of the funds. At the same time, an agreement is enteredinto to repurchase securities of the same type and volume from theprovider of the funds at a subsequent point in time. From the fund pro-vider’s viewpoint (in this case, big banks), this is a reverse repurchasetransaction. Economically, the transaction constitutes a loan securedby securities.35 Cf. Chapter 5 (Market assessment)

Sources: 2003 and 2004 annual reports

*Comprises the five largest banks in North America, Japan and all European countries, with total assets of over USD 100 billion (“The Banker”, July 2004).

CSG

UBS

02468

1012141618

02468

1012141618

Capital ratios of major international banks* Graph 18

%

Unweighted ratiosWeighted ratios (BIS total capital)Mean of unweighted ratiosMean of weighted ratios


5 Market assessment The market assessment of the soundness of a

bank is reflected in yield spreads, share prices andcredit ratings. These indicators suggest that the situa-tion in the Swiss banking sector is essentially good.At present, the market seems to assess the risk ofdefault by Swiss banks as low. In an internationalcomparison, the position of the big banks differs:while UBS ranks among the top banks, CSG is in mid-field.

Spreads on bank bonds and CDS prices remain low The yield spreads between bank bonds and sov-

ereign bonds or the prices of credit default swaps(CDS) are indicators of the market’s assessment ofthe soundness of banks. The higher the credit risk forthe holder, the greater the spread between the corre-sponding bond and a risk-free sovereign bond and thehigher the price of a CDS.

Having declined in 2003, the spread betweenthe bank bond and sovereign bond indices essential-ly moved sideways at a low level in 2004. The sametrend can be observed at most banks. By contrast, afurther significant decline was registered at BanqueCantonale Vaudoise and Banque Cantonale deGenève, where spreads had increased stronglybetween mid-2001 and end-2002 (cf. Financial Sta-bility Report 2003 and 2004). Spreads at all banks arelow – in some cases very low – compared with theaverage for the past seven years (cf. Graph 19).36

Prices of CDS suggest that the market currentlyrates the risk of default by the Swiss big banks as verylow (cf. Graph 20). The premiums for UBS are also lowrelative to other major international banks. CSGranks somewhere in mid-field following a sharp de-viation at year-end 2002. Prices are declining at virtu-ally all major Swiss and foreign banks.37

2001 2002 2003 2004 2005

0

50

100

150

200

250

Spreads Graph 19Between bank and Swiss Confederation bonds*

Basis pointsZürcher KB BC Vaudoise BC de Genève Raiffeisen banks Banking sector

*Average spread of all available securities satisfying the followingconditions: fixed coupons, no options, CHF denominated, residualterm of at least two years. End-month calculations.

36 Source: Thomson Datastream37 Source: Bloomberg

Sources: SNB, Thomson Datastream


Insolvency indicators derived from equity prices are lowShare prices provide an insight into the current

situation and future profit prospects of a bank. Theyreflect, in particular, the market valuation of thebank’s assets and of the risks embedded in thoseassets. These values, which can be derived from abank’s share prices using the option pricing theory,can then be used to assess the probability that thevalue of a bank’s assets falls below the value of its lia-bilities over a given time horizon. In other words, theoption pricing theory can be employed to derive theprobability of insolvency priced into shares and hence,allows the construction of an insolvency indicator.

Graph 21 shows the insolvency indicator for thetwo Swiss big banks and the average for a sample ofbig banks worldwide. The higher the indicator, thehigher the implied risk of insolvency as seen by themarket. Having peaked at year-end 2002, the risk ofinsolvency for both the Swiss and the foreign banksdeclined substantially by end-2003 and stabilised atthis level in 2004. Based on share prices, it appearsthat the market currently considers the soundness ofthe Swiss banks to be broadly in line with the inter-national average.38 However, caution is called forwhen interpreting the insolvency indicator, as its cal-culation is based on a number of simplifying assump-tions.39

1998 1999 2000 2001 2002 2003 2004 2005

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Insolvency indicator Graph 21

Standardised unitsUBS CSG Average of largest banks worldwide*

2002 2003 2004 2005

0

20

40

60

80

100

120

140

Five-year senior credit default swap prices for UBS and CSG Graph 20Premiums for credit protection on issuer bank

Basis pointsUBS CSG Average of largest banks worldwide*

Graphs 20 and 21:*Comprises a sample of the world’s largest banks from North America,

Japan and Europe according to “The Banker” (July 2004).

38 Sources: Bloomberg, Thomson Datastream, SNB39 In particular, normal distribution of bank revenues is assumed. If the actual distribution is different and bank specific, the insolvencyindicator may provide misleading figures. Moreover, in the short andmedium run, bank share prices and thus the insolvency indicator can beinfluenced by factors that do not necessarily reflect fundamentals.

Graph 20: Source: BloombergGraph 21: Sources: Bloomberg, SNB, Thomson Datastream


Ratings for big banks show a mixed pictureOf the 338 banks in Switzerland, 23 currently

have a rating from Moody’s, Standard & Poor’s and/orFitch. Their ratings remained unchanged at a relative-ly high level in 2004. The rating agencies also issuean outlook showing the anticipated medium-termtrend in their ratings. With the exception of anupgrade for two institutions, this outlook remainedunchanged at all banks in 2004. Overall, the outlooksreported by the rating agencies anticipate stable rat-ings in the medium term.40

On an international perspective, the picture forthe two Swiss big banks is mixed. According to the“Financial Strength Ratings” (Moody’s) and “Individ-ual bank ratings” (Fitch), UBS ranks in the upper andCSG in the lower mid-field (cf. Graphs 22 and 23).41

The ratings thus show a less positive view of thesoundness of the Swiss big banks than indicatorssuch as CDS prices and, above all, risk-weighted capi-tal ratios (cf. Chapter 4, Capital base).42


0

5

10

15

20

25CSG UBS

Individual bank ratings* (Fitch) Graph 23

Number of banksFitch

<C- C- C C+ B- B B+ A- A

0

5

10

15

20

25CS UBS

CSFB

Financial strength ratings* (Moody’s) Graph 22

Number of banksMoody’s

40 Sources: FitchRatings, Moody’s Investors Service, Standard & Poor’s41 The Financial Strength Ratings (Moody’s) and the Individual bankratings (Fitch) reflect an assessment of a bank’s strength on a stand-alone basis. They exclude considerations of external (government orparent company) support. 42 Sources: Bankscope, Moody’s Investors Service. If there is no ratingfor a group, the rating of the largest subsidiary is taken. CSG does nothave a Financial Strength Rating from Moody’s. However, CS and CSFBhave the same credit rating as CSG.

Graph 22: Source: Moody’s, May 2005Graph 23: Source: Bankscope, May 2005

Graphs 22 and 23:*Comprises a sample of the world’s largest banks from North America,

Japan and Europe according to “The Banker” (July 2004), providedthey are rated by Moody’s, Standard & Poor’s and Fitch. If a bankholding company is not assigned a financial strength rating, thecorresponding rating of its largest affiliate is taken instead.


*The higher the level of the index, the higher the level of stress in the Swiss banking sector. The index is expressed in terms of standarddeviations from its 1987–2004 average. A value above (below) zeroindicates that the stress is above (below) its historical average. Thestress index for the first quarter of 2005 is computed with provisionaldata.


Sources: SFBC, SNB, Thomson Datastream

6 Stress index for the banking sectorThe previous chapters of this report cover dif-

ferent aspects of the banking sector, all of which arepotentially relevant for its stability. In this chapter, wecombine these pieces of information within a “stressindex” measuring the current degree of instability inthe Swiss banking sector. We also develop a forecast-ing model for this index, using a set of macroeco-nomic and financial variables reflecting potentialeconomic imbalances. This allows the identificationof potential sources of future instability. Box 5(Stress index) on p. 43 outlines the methods used toproduce the stress index and its forecast.

According to this indicator, 2004 was a particu-larly calm period in the banking sector (cf. Graph 24).The level of stress remained relatively stable through-out the year and was low by historical standards. Theaverage level of stress recorded in 2004 was the low-est observed since the index was introduced (1987).Moreover, aside from capital variations (temporarydip in the first half of the year) and interbank loans(decline in the fourth quarter), all variables includedin the index pointed to low levels of stress for 2004(cf. Graph 25).

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

-3

-2

-1

0

1

2

3

Stress index* Graph 24In standard deviations

Actual Forecast 90% confidence interval


Our forecasting model suggests that the level ofstress should remain below average in 2005,although it would probably increase in the next 12months (cf. Graph 24). Such an increase would prin-cipally be a correction: since the index is currentlybelow the level forecasted by the model, it wouldbring it into line with the forecasts. The stress indexshould remain below average, as most of the financialand macroeconomic variables which influence it – thehousing price index, GDP, the investment ratio andthe credit ratio – are below or close to their long-term equilibrium level. A sudden downward correc-tion in these variables – and thus a rise in stress –therefore seems unlikely. The only real exception hereis the stock market index.

*The higher the intensity of an individual crisis symptom (e.g. thesharper the decrease of banks‘ share prices), the higher the level ofthe stress index. A value above (below) zero indicates that theintensity of an individual crisis symptom is above (below) its historicalaverage. The stress index for the first quarter of 2005 is computedwith provisional data.


Sources: SFBC, SNB, Thomson Datastream

2001 2002 2003 2004 2005

-4

-3

-2

-1

0

1

2

3

4

Stress index - contribution of individual stress symptoms to total stress* Graph 25In standard deviations

Share prices Bond spreads Interbank dep. Profitability Capital Provisions Watchlist Branches


Construction of the stress indexThe index is a continuous indicator of the level of stress

experienced by the Swiss banking sector at a given date. It isbased on a set of variables – including market data, balancesheet data, non-public data of the supervisory authorities andstructural data – which all represent possible symptoms ofstress in the banking sector.

These symptoms are:• a fall in the banks’ stock price index• an increase in the banks’ bond yield spreads• a fall in interbank deposits• a decrease in the banks’ profitability • a decrease in the banks’ capital• an increase in the banks’ provisioning rate• the share of total assets held by banks listed on the

regulator’s watchlist• a decrease in the number of banks’ branches

The higher the intensity of the individual stress symp-toms, the higher the level of the stress index. To build the in-dex, the eight variables described above are first normalisedand then aggregated with identical weights. The index is ex-pressed in terms of standard deviations from its historical aver-age. A positive (negative) value indicates that the stress isabove (below) its historical average.

Because the index is based on a large spectrum of poten-tial symptoms of instability, it should appropriately reflect thedifferent types of stress experienced by the banking sector. Thevalues taken by the index between 1987 – the starting date ofthe index – and 2004 are consistent with this conjecture. Theindex peaked three times and each peak corresponds to a peri-od of significant stress for the Swiss banking sector with differ-ent sources and symptoms: (i) the early 1990s, which werecharacterised by a real estate crisis in Switzerland, (ii) 1998,when the Russian and LTCM crisis occurred, and (iii) the2001/2002 period, which was characterised by a stock marketcrash and an economic slowdown. More generally, the indexshows that economic downturns or falls in stock or housingprices generate higher levels of stress in the banking sector.

Use of the stress index in forecastingPast experience suggests that banking crises tend to fol-

low the build-up of macroeconomic and financial imbalances.The (sudden) correction of these imbalances, such as reces-sions or stock market crashes, may eventually cause situationswhich generate stress in the banking system. The forecastingmodel presented here makes use of this fact and is based on theobservation of macroeconomic imbalances.

The forecasting model includes five macroeconomic andfinancial variables that are reliable predictors of banking crisesaccording to several studies: the share price index, the housingprice index, gross domestic product (GDP), the investment ratio(investment/GDP) and the credit ratio (private credits/GDP).The measure of imbalance is defined as the gap between thevariable and its trend, where the trend is computed using the

Box 5: Measuring and forecasting stress in the Swiss banking sector – methodology43

Hodrick-Prescott filter. For example, a positive credit ratio gapmeans that credits are growing faster than their sustainablerate. This could be interpreted as a sign of lower lending stand-ards of the banks. This imbalance will start to unwind whenborrowers find it more difficult to service their debt (e.g. be-cause of an interest rate rise, a recession or a drop in assetprices). For the banks that are engaged in lending business,both the share of non-performing loans and the provisioningincrease, which eventually leads to higher levels of stress. Forthe other four macroeconomic and financial variables, one canthink of similar mechanisms in which the correction of an exist-ing imbalance could lead to higher levels of stress.

The forecast is based on a linear regression of the stressindex on the past gaps of the macroeconomic and financialvariables:

where is the forecast of the stress index, h the fore-cast horizon the gaps of the k variables with a lag of zk. Apositive gap – i.e. the build-up of an imbalance – signals that afuture correction is likely to take place, which could lead to ahigher level of stress for the banks. In addition, different com-binations of the gaps have also been used as explanatory vari-ables. The regression is estimated with an autoregressive errorterm (AR(1)).

The results comprise forecasts for one to four quartersahead. The lag between the stress index and the gap is 25 quar-ters for the housing price index, 19 quarters for the share priceindex, 15 quarters for the investment ratio and 5 quarters forGDP. The model also uses combinations of the credit ratio with thestock price index, the housing price index and GDP. The model’sresults are relatively robust to changes in the lag structure.

With an average R2 of 56.4% over the four forecast hori-zons, the fit of the model can be considered as fairly good. Theout-of-sample error ratio (percentage of errors in the forecastof the direction of the stress index evolution) equals 20.9%,suggesting satisfactory predictive power.

Limitations of the methodAlthough the forecasting model for the stress index per-

forms relatively well, it is subject to several limitations. First,the macroeconomic and financial imbalances are computed us-ing a rather simple and mechanical Hodrick-Prescott filter. Themain advantage of this approach is that it does not imposemuch structure on the model. However, it may not guaranteethe most efficient use of the information available to assessthe magnitude of the imbalances. Second, different studieshave emphasised that banking crises are complex phenomena,which may involve non-linear interactions between the vari-ables. Even if the model takes into account some degree ofnon-linearity by combining some variables, it might miss somemore complex interactions between them. Finally, other non-macroeconomic/financial factors that are not included in themodel (e.g. deregulation), may also influence the level ofstress experienced by the banking sector.

43 This box is based on Monnin (2004), “Measuring, Explaining and Forecasting Stress in the Swiss Banking Sector”, Swiss National Bank,mimeo.

Part II: Financial market infrastructure



1 IntroductionSafe and efficient financial market infrastruc-

tures are a key prerequisite for a stable financial sys-tem. Alongside stock exchanges, the financial marketinfrastructures mainly comprise clearing and settle-ment systems for payments and for transactions insecurities and other financial instruments (subse-quently referred to as payment and securities settle-ment systems). Of particular interest are those pay-ment and securities settlement systems which areconsidered to be important to the stability of a coun-try’s financial system. The chief feature of these sys-tems is that they may trigger or channel the spread ofa systemic crisis and thus jeopardise the stability ofthe financial system.

At the centre of the analysis in this part of theFinancial Stability Report are thus those risks inher-ent in payment and securities settlement systemswhich might give rise to financial instability. Forinstance, severe operational failures or malfunctionsin systemically important infrastructures may causewidespread credit or liquidity problems for a largenumber of participants. Also, credit or liquidity prob-lems affecting one participant could spread to othersthrough the system. To reduce the likelihood andimpact of such systemic events, adequate organisa-tional and technical measures to mitigate operationalrisk as well as suitable rules and procedures thatallow confining the propagation of credit and liquid-ity risks through payment and securities settlementsystems are needed.

The risks inherent in payment and securitiessettlement systems do not change constantly as aresult of cyclical fluctuations or market trends.Rather, the safety of a financial market infrastructuredepends mostly on structural factors where change israre and gradual. The main events that directly affectthe infrastructures’ safety include the introduction ofnew payment and securities settlement systems andfundamental changes to established infrastructures.Major developments in the regulatory environmentcan also have an impact on financial market infra-structures. Such innovations are often a response tounderlying trends, such as a change in risk conditionsor risk awareness or a change in the needs of marketparticipants.

The statement made in last year’s Financial Sta-bility Report – namely, that Switzerland has verywell-functioning infrastructures and that safety andefficiency are high by international standards –therefore remains valid. The key elements of theSwiss post-trade value chain, i.e. the Swiss InterbankClearing (SIC) system for payments, the securitiessettlement system SECOM and the central counterpar-ty x-clear, have proven their functional efficiencyover the years, and their architectures and pro-cedures contribute to reducing settlement risks and,ultimately, systemic risk.

To promote the objective of safe and efficientfinancial market infrastructures, the Swiss NationalBank (SNB) oversees payment and securities settle-ment systems. Chapter 2 of this part deals with theSNB’s oversight activities. In particular, it explainswhy the National Bank Act (NBA) mandates the SNBto focus its oversight activities on systemically import-ant infrastructures and why SIC, SECOM, x-clear andthe multi-currency payment system ContinuousLinked Settlement (CLS) were designated to be of systemic importance. Moreover, it is described howthe SNB oversees these systems in practice, bothdomestically and in cooperation with foreign centralbanks or authorities in the case of cross-border sys-tems. Chapters 3 and 4 provide some backgroundinformation on two significant developments that havea direct bearing on the safety and efficiency of sys-temically important infrastructures. First, Chapter 3provides an overview of the Swiss financial sector’scurrent efforts to strengthen the resilience of thefinancial system by improving preparations for deal-ing with major disasters. Chapter 4 reviews majorinternational developments in the design of large-value payment systems. Innovations such as theintroduction of offsetting mechanisms or centralbanks’ policies of broadening the range of collateralthey accept in their liquidity provision operationsallow to achieve earlier finality of payments (andthus lower risks) or to economise on liquidity (andthus lower cost).


2 Oversight of systemically important infrastructuresPayment and securities settlement systems have

attracted greater attention over the last two decades.This reflects the significant growth in the values oftransactions cleared and settled, the growing techno-logical complexity of many systems and the conse-quent concern that systemic risk could increaseshould the design of key systems not adequatelyaddress various payment and settlement risks. Alongwith the awareness of mounting risks, central banks’endeavours to contain these risks – without compro-mising efficiency – have gained momentum. Accord-ingly, many central banks have strengthened theiroversight arrangements. The term “oversight” typi-cally refers to those central bank activities that pro-mote the objectives of safety and efficiency in a cur-rency’s payment and settlement infrastructure bymonitoring existing and planned arrangements,assessing them against the objectives and, in thelight of this, inducing any necessary change.

The new National Bank Act (arts. 19–21 NBA)contains the principles for the oversight of paymentand securities settlement systems by the SNB. Itempowers the SNB to impose minimum requirementson the operation of those systems from which risksfor the stability of the Swiss financial system mayemanate. The minimum requirements and otherimplementation provisions on oversight are set out inthe National Bank Ordinance (arts. 19–39 NBO). Thesmooth functioning of the systemically importantinfrastructures is a key prerequisite for the imple-mentation of monetary policy and for the successfuldevelopment of the national economy. In otherwords, systemically important payment and securitiessettlement systems are essential for the supply ofliquidity to the economy.

Since the revised NBA entered into force on 1 May 2004, the SNB has determined which systemsare important for the Swiss financial system and imple-mented the first practical steps for continuously veri-fying the corresponding system operators’ compli-ance with the minimum requirements. Which systemshave been classified as systemically important andhow they are overseen is described below.

Evaluation of systemic importanceThe systemic importance of a payment or secu-

rities settlement system is assessed primarily basedon the effects which the disruption of operations orpayment or delivery difficulties experienced by indi-vidual system participants can have on the stabilityof the Swiss financial system. The main focus is thuson the potential threat to the Swiss financial sys-tem’s stability.

The SNB evaluated the systemic importance ofall payment systems that settle at least CHF 25 billionper financial year and all securities settlement sys-tems. The analysis was based on the criteria set out inart. 20 NBO, which were broken down into moredetailed specifications and weighted according totheir importance. The criteria (in the order of theirweighting) are explained below. Table 1 lists theevaluated systems and some of the criteria taken intoconsideration.

– Type of transactions: Systems that clear or settlefinancial market transactions are extremelylikely to be of key importance to the Swissfinancial system. By contrast, systems that pri-marily settle retail payments cause inconven-ience to the user in the event of failure, butthey entail only minor risks for the stability ofthe financial system.

– Links with other systems: If a system is linkedwith other systems, there is a risk that shockscan be transmitted to other financial marketinfrastructures. Depending on the type of links,however, the consequences may vary. Forinstance, operational problems of a retail pay-ment system that settles its interbank liabilitiesin a large-value payment system do not signifi-cantly affect the functioning of the latter. Bycontrast, the failure of a payment system that islinked with a securities settlement system forthe delivery-versus-payment settlement ofsecurities transactions is likely to severelyimpair this securities settlement system.


– Alternative clearing and settlement arrange-ments: If a system has at least one alternativethat can be employed immediately and withoutincurring any additional settlement risks, thesystem poses hardly any risk to financial stabil-ity.

– Value and number of transactions: The grossvalue (on peak days in particular) is directlyrelated to the damage potential and is a signifi-cant indicator of a system’s importance. Thenumber of cleared or settled transactions mustalso be taken into consideration, but it is lessrelevant overall.

– Group of participants: In particular, systemswhose direct participants are financial inter-mediaries have the potential to destabilise thefinancial system.

– Currencies: From the viewpoint of the Swissfinancial system, the question of whether a sys-tem clears and settles in Swiss francs must betaken into account.

SIC, SECOM, x-clear and CLS designated as systemically importantThe SNB considers the following four systems to

be systemically important: the payment system SwissInterbank Clearing (SIC), the securities settlementsystem SECOM, the central counterparty x-clear andthe multi-currency payment system ContinuousLinked Settlement (CLS). The first three – SIC, SECOMand x-clear – are integrated in the Swiss Value Chainand, together with the electronic stock exchangeSWX, form the core of Switzerland’s financial marketinfrastructure. These systems are used almost exclu-sively for clearing and settling financial market trans-actions. Participants include various financial inter-mediaries with an important role to play in the Swissfinancial system. Operational problems occurring inone of these systems, and causing liquidity shortagesamong participants as a result, could have a negativeimpact on the functioning of the other linked sys-tems. With the exception of x-clear, the transactionvalues in these systems are so large that a systemfailure could jeopardise the economy’s supply ofliquidity. Moreover, if a system experiences opera-tional problems – particularly where SIC and SECOMare concerned – no alternative settlement arrange-ments exist for the settlement of payments or secu-rities transactions. Although there are alternatives toCLS and x-clear, it must be assumed that these couldbe employed only after a delay and with considerablyhigher risks. The systemic importance of x-clear is

primarily attributable to the concentration of variouscounterparty risks. The risk management of x-clearmust ensure in particular that no domino effectoccurs in the event of default by a participant.

Due to their systemic importance, the operatorsof SIC, SECOM and x-clear have to comply with theminimum requirements stipulated in arts. 22–34NBO. CLS was exempted from compliance with theminimum requirements as the New York-based opera-tor, CLS Bank International, is already adequatelyoverseen by the Federal Reserve (art. 21 NBO).

Retail systems not systemically importantThe other systems that were evaluated – DTA,

LSV, euroSIC and PostFinance – were classified by theSNB as not systemically important. While euroSICalso settles financial market transactions, the trans-action volumes are relatively small overall. The othersystems mainly settle retail payments. While thetransaction volumes are very high, the actualamounts processed are relatively low. Although thetransactions cleared via DTA and LSV are settled inSIC, these systems nonetheless have no critical in-fluence on the large-value payment system withregard to either operational problems or participants’liquidity bottlenecks. The same applies to Post-Finance, which is also connected directly to SIC forcertain payments. With these systems, moreover, it is possible to switch at short notice to other retailpayment systems. In consideration of the abovefacts, these systems do not pose any potential threatto the stability of the financial sector.

As a consequence, their operators do not haveto comply with the SNB’s minimum requirements.However, the SNB continuously monitors develop-ments in the area of retail payments and will periodi-cally review the importance of these systems.


System type Type of transaction Value of Number oftransactions in transactions1

CHF billions1

SIC Large-value payment system Financial market transactions 164 816,700and retail payments

SECOM Securities settlement system Financial market transactions 36.9 74,000CLS Multi-currency Financial market transactions 1,8002 133,000 2

payment system 823 7,300 3

x-clear Central counterparty Financial market transactions 1.4 19,000euroSIC Large-value payment system Financial market transactions 3.9 9,500

and retail paymentsLSV Retail payment system Retail payments 0.3 148,500DTA Retail payment system Retail payments 0.9 218,600PostFinance(EZAG + Postcard) Retail payment systems Retail payments 4.0 1,662,7001 Daily averages for 20042 Transactions of all currencies settled in CLS3 Only transactions in Swiss francs

Table 1: Evaluated payment and securities settlement systems


Three-step oversight methodologyThe SNB’s oversight focuses on ensuring compli-

ance of the systemically important infrastructureswith the minimum requirements specified in theNational Bank Ordinance (NBO). In order to guaran-tee effective and efficient oversight, a three-stepoversight methodology has been developed and iscurrently being implemented. The three steps can besummarised as (1) monitoring, (2) assessment and(3) inducing change. These three steps form a con-tinuous process. In formulating its oversight methodo-logy, the SNB greatly benefited from a constantexchange of views on oversight issues with other central banks within the Committee on Payment and Settlement Systems (CPSS, cf. Box 6) under theauspices of the Bank for International Settlements(BIS). The SNB also ensures that its oversight is inaccordance with internationally agreed principles ofcentral bank oversight (cf. Box 7, p. 50).

The oversight methodology is uniformly appliedacross the systemically important infrastructures SIC,SECOM and x-clear. This helps to ensure that over-sight standards are applied in a consistent and com-parable manner. As both SIS SegaInterSettle AG (theoperator of SECOM) and SIS x-clear AG are alsolicensed as Swiss banks and supervised by the SwissFederal Banking Commission (SFBC), the SNB co-operates closely with the SFBC to avoid gaps or du-plications in its oversight efforts. In the case of SIC,the SNB is responsible for performing certain func-tions such as determining participation requirements,managing settlement accounts and leading crisismanagement. As a consequence, the SNB oversees itsown organisational units responsible for providingthese operational tasks. To avoid potential conflictsof interest, the units responsible for oversight andoperations are organisationally separated and reportto different members of the Governing Board.

Under the auspices of the BIS, the Committee on Pay-ment and Settlement Systems (CPSS) serves as a forum for cen-tral banks to monitor and analyse developments in domesticpayment, settlement and clearing systems as well as in cross-border and multi-currency settlement schemes. Through theircollective efforts to analyse and promote awareness of risksand to develop minimum standards or best practices, the CPSSand its member central banks have played a leading role in pro-moting efficient and robust payment and settlement arrange-ments. Their continuing work in this area reflects the need tocontinuously monitor and improve risk management in thesesystems, which are essential to the stability of financial mar-kets and the global economy. CPSS members are the centralbanks from the Group of Ten countries (G-10) as well as theHong Kong Monetary Authority, the Monetary Authority ofSingapore and the European Central Bank. The CPSS reports tothe G-10 Governors.

The CPSS has published several important internationalstandards, among them the “Core Principles for SystemicallyImportant Payment Systems” (2001), the “Recommendationsfor Securities Settlement Systems” (2001) and the “Recommen-dations for Central Counterparties” (2004). The latter two werejointly prepared by the CPSS and the International Organisationof Securities Commissions (IOSCO). Beside these standards, theCPSS also regularly publishes reports on specific topics relatedto payment and settlement systems. The most recent reportsare “Central Bank Oversight of Payment and SettlementSystems” and “New Developments in Large-Value PaymentSystems” (cf. also Chapter 4).44

44 For these and other CPSS publications, cf. www.bis.org/cpss.

Box 6: Committee on Payment andSettlement Systems (CPSS)


Monitoring relies on four sources of information In order to carry out effective oversight and to

assess compliance with minimum requirements, theSNB must have a good understanding of the relevantpayment and securities settlement systems, includingtheir policies, operations and processes, potentialrisks as well as planned developments. Therefore,monitoring serves to gather information and toenhance the SNB’s understanding of the systems itoversees. Monitoring focuses on areas that potential-ly pose the greatest threat to financial stability.These areas typically include governance arrange-ments, IT systems and operations (especially infor-mation security), financial risk management and out-sourcing relationships.

Specifically, the SNB relies on four different sources of information:

– Reporting: The system operators provide rele-vant documentation and key indicators on aregular or ad-hoc basis. Relevant documenta-tion includes system rules, organisational rules

and regulations, business continuity plans andother static information. Key indicators typical-ly contain transaction figures, financial posi-tions, risk exposures and information systemperformance data.

– Self-assessment: A self-assessment, comple-mented by a questionnaire, is completed by thesystem operators on an annual basis. In theself-assessment, the system operators indicatewhether they consider themselves to be in com-pliance with the minimum requirements.

– Audits: Internal or external auditors are com-missioned by the system operators to conductaudits in areas of specific importance from theperspective of financial stability. The results ofthese audits are shared with the SNB. For exam-ple, such audits are conducted on a regularbasis in the areas of financial risk managementand information security.

– Meetings: Specific developments as well asresults from other monitoring activities are dis-cussed in meetings with the system operators.

In its recently published report “Central Bank Oversightof Payment and Settlement Systems”, the CPSS explains whyand how central banks oversee payment and securities settle-ment systems. Overall, there is substantial common ground inoversight practices among G-10 central banks. Based on thiscommon ground, the CPSS has developed two sets of oversightprinciples to help central banks organise and conduct effectiveoversight. The first set consists of five general oversight princi-ples applicable to all oversight arrangements. They read as fol-lows:

– Transparency: Central banks should set out publicly theiroversight policies, including the policy requirements orstandards for systems and the criteria for determiningwhich systems these apply to.

– International standards: Central banks should adopt,where relevant, internationally recognised standards forpayment and settlement systems.

– Effective powers and capacity: Central banks should havethe powers and capacity to carry out their oversight re-sponsibilities effectively.

– Consistency: Oversight standards should be applied con-sistently to comparable payment and settlement sys-tems, including systems operated by the central bank.

– Cooperation with other authorities: Central banks, in pro-moting the safety and efficiency of payment and settle-ment systems, should cooperate with other relevant cen-tral banks and authorities.The second set of principles is concerned with interna-

tional cooperative oversight between central banks and otherauthorities. These principles build on the well-established Lamfalussy principles for netting arrangements. The CPSS hascarefully reviewed the experiences gained since 1990 when the

Box 7: General principles of central bank oversightLamfalussy principles were adopted, and concluded that theyare still universally applicable. However, the applicability of thefive cooperative oversight principles was extended to all cross-border and multi-currency payment and settlement systems.

– Notification: Each central bank that has identified theactual or proposed operation of a cross-border or multi-currency payment or settlement system should informother central banks that may have an interest in the pru-dent design and management of the system.

– Primary responsibility: Cross-border and multi-currencypayment and settlement systems should be subject tooversight by a central bank which accepts primaryresponsibility for such oversight and there should be apresumption that the central bank where the system islocated will have this primary responsibility.

– Assessment of the system as a whole: In its oversight of asystem, the central bank or authority with primary re-sponsibility should periodically assess the design andoperation of the system as a whole in consultation withother relevant central banks and authorities.

– Settlement arrangements: The determination of the ad-equacy of a system’s settlement and failure-to-settleprocedures in a currency should be the joint responsibil-ity of the central bank of issue and the central bank orauthority with primary responsibility for oversight of thesystem.

– Unsound systems: In the absence of confidence in thesoundness of the design or management of any cross-border or multi-currency payment or settlement system,a central bank should, if necessary, discourage use of thesystem, for example by identifying its use as an unsafeand unsound practice.


Assessment of compliance with minimum re-quirementsThe insights gained from the monitoring activ-

ities form the basis for the assessment of the opera-tors’ compliance with the minimum requirements. Theminimum requirements are based on internationalstandards, particularly the “Core Principles for Sys-temically Important Payment Systems”, the “Recom-mendations for Securities Settlement Systems” andthe “Recommendations for Central Counterparties”(cf. Box 6, p. 49). In addition, the SNB relies on com-monly accepted industry standards, for example inthe area of information security.

To allow for a thorough assessment, the mini-mum requirements are put in concrete terms indetailed, system-specific control objectives. Compli-ance with these control objectives is assessed accord-ing to a scale with four degrees of compliance(observed, broadly observed, partly observed, non-observed). To account for developments of interna-tional standards and best practices in the area of pay-ment and securities settlement systems, the controlobjectives are reviewed and, if required, updated on aregular basis.

Different measures to induce changeIf the SNB concludes in the assessment that the

operator of a payment or securities settlement systemis not in compliance with the minimum requirements,it issues a recommendation to the operator. In therecommendation, a clear and convincing case forchange is provided. If the situation is not remedied,the SNB can issue a legally enforceable instruction.Before imposing these measures, the SNB gives boththe Swiss Federal Banking Commission and the opera-tor an opportunity to put forward their views regardingcompliance with the specific minimum requirement inquestion.

If an instruction is not followed, other correc-tive measures such as notification of other author-ities, issuing of a public statement on non-compliancewith the minimum requirements, or withdrawal ofaccess to the SNB’s sight deposit accounts can beemployed.

In addition, the SNB may induce a desiredchange within a system by commenting on specificissues or developments. This is typically done if a sys-tem is considered to be in compliance with the actualminimum requirements, but there is still room forimprovements to achieve best practice in a criticalarea.

Cooperative oversight of cross-border systems Operators of payment and securities settlement

systems which are systemically important from aSwiss perspective can be located outside Switzerland.Conversely, it is possible that systemically importantinfrastructures domiciled in Switzerland are a sourceof systemic risk for foreign financial markets. For thepurpose of overseeing systems from which risks forthe financial system emanate and which are inter-national in their scope and origin, the SNB is em-powered to cooperate with foreign supervisory andoversight authorities and to exchange informationwith them (art. 21 NBA). Furthermore, the operator ofa systemically important infrastructure which is domi-ciled abroad can be discharged in whole or in partfrom compliance with the minimum requirements,provided the system is subject to standards commen-surate with the minimum requirements establishedby the SNB and the relevant foreign oversightauthorities are willing to collaborate with the SNB(art. 21 NBO).


The SNB currently engages in cooperative over-sight for Continuous Linked Settlement (CLS) and forx-clear. CLS Bank International, the operator of CLS,is domiciled in the US and the Federal Reserve acts asauthority with primary responsibility for oversight.Oversight of CLS is carried out in accordance with theprinciples for international cooperative oversight (cf.Box 7, p. 50). In its role as overseer with primaryresponsibility, the Federal Reserve consults withother central banks whose currencies are integratedin CLS, including the SNB. Consequently, CLS was dis-charged in whole from compliance with the minimumrequirements laid down in the NBO.

SIS x-clear AG, the operator of x-clear, is domi-ciled in Switzerland. However, since x-clear is alsorecognised as an Overseas Clearing House by theFinancial Services Authority (FSA), the authorityresponsible for supervision of Overseas ClearingHouses in the United Kingdom, there is a need forcooperation between Swiss and UK authorities. TheSNB and the Swiss Federal Banking Commission(SFBC) act as authorities with primary responsibilityfor oversight and supervision, and cooperate with theFSA. The guidelines for this cooperation have beenestablished in a Memorandum of Understanding(MoU). In this document, the involved authoritiesagree on the principles and modalities of the co-operation in order to promote effective oversight andsupervision. Specifically, the MoU governs the com-munication between the involved authorities, theexchange of information and confidentiality issues aswell as practical matters, such as the frequency ofmeetings.


3 Business continuity planning in the Swiss financial sectorThe operational risks, i.e. vulnerabilities arising

from inadequate or failed internal processes, peopleand systems or from external events, have generallyincreased in the financial sector over the past fewyears. This is due in particular to the increasingdependence on rapidly changing technologies and,partly, also to the changed assessment of risks aris-ing from deliberate acts. At the same time, the dam-age potential of operational events has increased.The reasons for this development are the consolida-tion process within the financial sector on the onehand, and intensifying integration of the financialmarkets and interconnection of infrastructures on theother. Strengthening the resilience of the financialsystem by means of improved business continuityarrangements is thus an increasingly important issueboth for the financial sector itself and for financialauthorities. In this respect, particular emphasis islaid on disaster recovery and business continuityplans to ensure that financial institutions and criticalinfrastructures are able to operate on an ongoingbasis in the event of severe business disruptions.Business continuity planning (BCP) on the one handentails preventive measures aimed at preventing anydisruptions or minimising their likelihood and, on theother hand, contingency measures to overcome dis-ruptions as smoothly as possible.

Sector-wide task force to assess BCP arrangementsIn autumn 2003, a task force under the direc-

tion of the SNB started to review and assess the busi-ness continuity plans of the major financial partici-pants with a view to strengthening the BCParrangements in the Swiss financial sector. The taskforce consisted of representatives from the SNB aswell as from the central financial market infrastruc-tures (SIS, SWX and Telekurs), the major participantsin these infrastructures (Credit Suisse, UBS and Post-Finance) and the Swiss Federal Banking Commission.The sector-wide cooperation with regard to contin-gency planning reflects the awareness that whileeffective business continuity arrangements must beplanned and implemented by the individual compa-nies, they need to be coordinated to some degree.Sufficient coordination between all parties involvedis the only way to ensure that critical functions of thefinancial system can be maintained or quicklyresumed even after severe disruptions.

In January 2005, the SNB submitted a task forcereport for consultation. Besides listing the arrange-ments currently in place, it also contained a numberof specific recommendations. The task force will pub-lish its final report in the second half of 2005. Inaddition to recommendations, it will include state-ments on the form of regulation that should be cho-sen for the implementation of these recommenda-tions. The main elements of the consultation reportare described below.

Recovery and resumption requirements forcritical business processesThe task force focused its analysis on the two

business processes that were considered to be espe-cially critical for system stability and which must bemaintained without fail even in exceptional situa-tions: (1) liquidity supply of the financial system withcentral bank money by means of repos concluded withthe SNB and (2) large-value payment transactionsbetween the commercial banks in Swiss InterbankClearing (SIC). The corresponding measures taken bythe individual institutions were analysed with regardto physical and staff resources in the IT and businessdivisions.

The consultation report first presents a scenariowith a severe disruption causing the failure of animportant operational centre and incapacitating itsstaff. Some key requirements are then defined forthis scenario, notably with regard to the maximumtolerable downtime until the resumption of normaloperations. These requirements can be summarised asfollows:

– In the event of disruptions, the central infra-structures which are crucial for the critical busi-ness processes must be able to resume the criti-cal business activities within two hours andwithout losing any confirmed transactions.

– In the event of disruptions, critical system par-ticipants must be able to resume the definedcritical business activities within four hours.

– In the event of disruptions, the other systemparticipants must be able to resume operationswithin 24 hours.


A significant aspect is that the participants arefree to decide how to keep within the stipulated timelimits. They may, for example, employ alternativeprocessing mechanisms in the short term to processthe most important transactions. This would allowthem more time to restore normal operations.

The fact that the central infrastructures, thecritical and the other system participants have to ful-fil different requirements reflects their respectivesignificance for the maintenance of the critical busi-ness processes “liquidity supply” and “large-valuepayments”. The failure of a central infrastructurebrings the financial centre to a de facto standstilland can cause widespread credit and liquidity prob-lems among market participants. It is thus only lo-gical that these infrastructures must meet very highrequirements. However, the precautionary measurestaken by the central infrastructures to maintain thecritical business processes are only of limited benefitunless the (critical) system participants are function-ing as well. Therefore, they too must satisfy highrequirements.

Good preparations, but room for improvementOverall, the central infrastructures and the indi-

vidual system participants have made extensivepreparations to ensure that the critical businessprocesses are maintained even in exceptional circum-stances. There is still room for improvement in vari-ous respects, though, notably in the following areas:

– The precautionary measures are better suited to the failure of physical components (such asbuildings or hardware) than for events thataffect staff availability.

– Tests in the IT area are carried out mostly on aprepared basis and, in some cases, do not con-sider staff-specific aspects.

– The physical distance between the primary andsecondary operation centres is often belowaverage by international standards.

– The sector-wide alarm and crisis organisationshould be improved so that, in the event of dis-ruptions, problem and process-specific commu-nication can be coordinated among all partiesinvolved.


4 Developments in large-value payment systems The design of large-value payment systems

(LVPS) changes over time. These changes, which aregenerally due to technological progress, are driven byparticipants’ requirements and the regulators’endeavour to reduce risks and costs in LVPS. At thesame time, to the extent that there is a trade-offbetween risks and cost, innovations aim at improvingthe risk-cost equilibrium. The most important devel-opments that have taken place in LVPS over the lastcouple of years are analysed in a recent report by theCommittee on Payment and Settlement Systems(CPSS).45 Based on this analysis, the major trends andinnovations in LVPS of CPSS member countries aresummarised below.

Reduction of settlement risksOne of the distinct trends of the last decade is

that LVPS have increasingly moved towards the provi-sion of intraday finality. LVPS that are based ondeferred net settlement (DNS) still exist, but theirnumber has clearly decreased over the past few years.In a DNS system, payment orders are accumulatedthroughout the day and only the net amount is set-tled typically at the end of the day. While netting hasthe advantage that relatively little liquidity is need-ed, the deferred settlement has the drawback thatfinality of settlement is only achieved at the end ofthe day, thus creating settlement risk. There havebeen several approaches to reduce the settlementrisk associated with deferred settlement. The first isto protect systems with enough financial resources inorder to ensure final settlement even in the eventthat the participant with the largest net debit posi-tion fails to meet its payment obligation. The secondis to introduce real-time gross settlement (RTGS),which has been done in most of the CPSS membercountries. However, a third alternative has beendeveloped which achieves continuous intraday final-ity in LVPS without settling on a RTGS basis. The LargeValue Transfer System (LVTS) in Canada, for instance,still settles just once at the end of the day like a con-ventional DNS system. Yet it provides real-time intra-day finality like an RTGS system, because certainty ofsettlement is achieved by a combination of collateralposted by participants and a guarantee by the Bankof Canada for the residual exposures.

Reduction of liquidity needs and costs in RTGS systemsRTGS systems have become the most common

type of large-value payment systems in CPSS membercountries. In Switzerland, the Swiss Interbank Clear-ing (SIC) system was introduced as early as in 1987.RTGS systems continuously process and settle pay-ment orders in real time on a transaction-by-transac-tion basis. Since finality is achieved intraday, they donot bear the settlement risks of DNS systems. How-ever, real-time gross settlement has the disadvantagethat it is more costly in terms of liquidity. In order toaddress this issue, new features have been imple-mented to reduce the liquidity needs and to allowparticipants to better control and manage the pay-ment process. Moreover, central banks extend intra-day credit to participants of RTGS systems atfavourable conditions.

One major innovation has been the implementa-tion of LVPS that combine design features of DNS andRTGS systems. Whereas the classic hybrid systemsapply netting or offsetting routines at short, but dis-crete time intervals, a number of LVPS have recentlycombined this feature with real-time settlement func-tionality. Examples of such systems are CHIPS in theUnited States, PNS in France and RTGSplus in Germany.These systems first attempt to settle a payment orderon a gross basis. If immediate settlement is notpossible due to a lack of settlement balances, thepayments are stored in a queue, and then the systemchecks whether the queued payments can be offsetbilaterally or multilaterally. While in net settlementonly the net amounts are transferred, offsetting com-prises the gross execution of single payments simul-taneously within one legal and logical point in time.Different kinds of algorithms exist to execute the off-setting mechanism. Usually, relatively simple bilater-al algorithms tend to be applied in real-time, whilemore complex multilateral algorithms are employedintermittently at short intervals. The application ofcontinuous netting or offsetting has mostly beenmade feasible by significant technological progressin the area of computing power. Features that werepreviously too expensive to be implemented havebecome affordable over time. A fairly simple bilateraloffsetting mechanism has been implemented in SIC.It is automatically triggered should the system not beable to settle any payment within 20 seconds due toinsufficient or unevenly distributed funds on the par-ticipants’ settlement accounts.

45 Cf. CPSS (2005), “New developments in large-value payment sys-tems”, www.bis.org/cpss.


In response to demand among participants forbetter possibilities to manage their liquidity and pay-ment flows, most of the LVPS nowadays provide abroad range of real-time information on the settle-ment process, such as detailed information on queuesand account balances. This information is crucial tothe reasonable employment of the various interactivecontrol features that enable the participants to fine-tune the settlement process. Such controls include,for instance, the possibility to change the location ofa payment in the queue, the scheduling of the releaseof an instruction, the setting of bilateral and multi-lateral credit limits to control the outflow of funds aswell as the reservation of funds on the settlementaccount.

While the introduction of netting and offsettingfeatures tends to reduce the amount of liquidityneeded to settle payments, central banks’ collateralpolicy influences the costs of this liquidity. Sincemost central banks extend credit only against collat-eral, the type of collateral that LVPS participants canuse is an important factor in determining the oppor-tunity costs of holding collateral. In the last fewyears, several central banks have substantially broad-ened the range of collateral they accept in theirliquidity provision operations, thus lowering theopportunity costs of liquidity. Central banks from theEurosystem, for instance, accept euro-denominatedcollateral cross-border within the European MonetaryUnion. Furthermore, a few central banks, such as theFederal Reserve or the SNB now also accept collateraldenominated in foreign currency. The fact that55–70% of the SNB’s outstanding repo transactionsare typically secured by euro-denominated collateralreflects the participants’ strong demand for thesefacilities.

Increased cross-border activityAnother block of developments in LVPS can be

attributed to expanding cross-border activities of thefinancial sector. Examples are the creation of TARGETin the European Union in order to interlink thedomestic RTGS systems, and the emergence of newlarge-value payment infrastructures in countrieswhere a foreign currency plays an important role. InSwitzerland, euroSIC was created in order to processeuro payments. Similarly, a US Dollar and a EuroClearing House Automated Transfer System (USDCHATS and Euro CHATS) were introduced in HongKong for the processing of US dollars and eurosrespectively. These systems are basically replicas ofthe local LVPS, thus displaying more or less the samedesign, but settling on the accounts of a correspond-ent bank. Finally, Continuous Linked Settlement(CLS) was introduced in 2002. CLS is a private sectorsystem that specialises in the settlement of foreignexchange transactions on a payment-versus-paymentbasis. This settlement mechanism eliminates creditrisk (Herstatt risk) in the settlement of foreignexchange transactions of the 15 most actively tradedcurrencies. Although settlement members pay theirobligations via the domestic LVPS, settlement takesplace in commercial bank money on the books of CLSBank.

Financial Stability ReportBericht zur FinanzstabilitätRapport sur la stabilité financière

2005

financial stability report 2005 - swiss national bank...financial strength ratings* (moody’s)...

Documents