bank of canada financial...sisted, with a heightened level of tension in chart 1 total assets:...

B A N K O F C A N A D A

FinancialSystem Review

December 2002

Members of the Editorial Committee

Charles Freedman, Chair

Clyde GoodletDavid Longworth

Tiff MacklemDinah Maclean

John MurrayRon Parker

Graydon PaulinGeorge Pickering

James PowellDenis SchutheBonnie Schwab

Jack SelodyRobert Turnbull

The Bank of Canada’s Financial System Review is published semi-annually. Copies maybe obtained free of charge by contacting

Publications Distribution, Communications Department, Bank of Canada, Ottawa,Ontario, Canada K1A 0G9.Telephone: (613) 782-8248; e-mail: [email protected]

Please forward any comments on the Financial System Review to

Public Information, Communications Department, Bank of Canada, Ottawa,Ontario, Canada K1A 0G9Telephone: (613) 782-8111, 1-800-303-1282; e-mail: [email protected]

Web site: http://www.bankofcanada.ca

Bank of CanadaDecember 2002

Contents

Foreword

Developments and Trends ........................................................... 1

Introduction ............................................................................................. 3

Highlighted Issues ..................................................................................... 4

The Macrofinancial Environment .................................................................. 10

The Financial System ................................................................................. 17

Report ...................................................................................... 27

Systemic Risk, Designation, and the ACSS ...................................................... 29

Policy and Infrastructure Developments ....................................... 37

Introduction ............................................................................................. 39

The CLS Bank: Managing Risk in Foreign Exchange Settlements .......................... 41

The Impact of Participant Outages on Canada’s Large Value Transfer System ......... 45

Understanding Intraday Payment Flows in the Large Value Transfer System ........... 49

Research Summaries .................................................................. 53

Introduction ............................................................................................. 55

The Financial Services Sector: An Update on Recent Developments ....................... 57

Canadian Foreign Exchange Market Liquidity and Exchange Rate Dynamics .......... 59

Financial Structure and Economic Growth: A Non-Technical Survey ..................... 63

Banking Crises and Contagion: Empirical Evidence ........................................... 67

he financial system, which consists of financial institutions, financial markets, and clearingand settlement systems, plays an important role in a nation’s economy. Sound and efficientfinancial systems can make a significant contribution to economic growth. But as we haveseen in many countries in recent years, a financial system that is not soundly based can make

problems that originate elsewhere in the economy much worse and can even be the source ofproblems itself.

The Bank of Canada is one of several federal and provincial agencies and organizations in Canadathat promote the safety and efficiency of our financial system. Each one brings a specialized exper-tise to this activity. The Bank contributes a broad perspective that reflects the major activities inwhich it is engaged. As the monetary authority, the Bank brings a macroeconomic or systemwidepoint of view to issues concerning the financial sector, as well as extensive knowledge of financialmarkets. As the source of ultimate liquidity to the financial system (and, thus, the lender of last re-sort) the Bank is acutely aware of stresses that can develop in the system during times of financialturbulence. And, as the overseer of clearing and settlement systems that could pose significant riskto the financial system, the Bank has developed expertise in the design and operation of arrange-ments to control this risk. Finally, as fiscal agent for the Government of Canada, the Bank has a par-ticular interest in well-functioning government debt markets.

The Bank devotes considerable resources to assessing developments and trends in both domesticand international financial systems. As the oversight agency for key clearing and settlement systemsand as a participant in interagency committees, the Bank is also very involved in the policy and in-frastructure developments that affect the financial system. Bank staff conduct extensive research inthese areas, which enables us to better understand ongoing developments and to contribute to goodpolicy-making. The goal of the Financial System Review (FSR), which will be issued semi-annually bythe Bank, is to share with financial system participants and the Canadian public the Bank’s research,analyses, and judgments on various issues and developments concerning the financial system. Ourhope is that through this publication, the Bank will contribute to greater understanding of such is-sues and promote a more informed discussion of policy and developments in Canada and abroad.

With the publication of the FSR, the Bank of Canada is taking another step towards increased open-ness and transparency. Indeed, a number of central banks and the IMF already issue similar reports,which have proven useful to those with an interest in the financial sector in their countries andaround the world, and have improved awareness and discussion of developments in the financialsystem. We hope that the FSR will contribute to such a dialogue in Canada and that, working to-gether, the public and private sectors can enhance the efficiency and stability of the Canadian finan-cial system.

Charles FreedmanDeputy Governor and Chairof the Editorial Committee of theFinancial System Review

T

Foreword

Developments

and

Trends

Notes

The material in this document is based on information available to 4 December unlessotherwise indicated.

The phrase “major banks” in Canada refers to the six largest Canadian commercialbanks by asset size: the Bank of Montreal, CIBC, National Bank of Canada, RBC FinancialGroup, Scotiabank, and TD Bank Financial Group.

Financial System Review

Introduction

This section of the Financial System Review exam-ines the recent performance of the Canadian finan-cial system and the factors, both domestic and inter-national, that are influencing it. Related topics ofparticular interest are discussed under “HighlightedIssues.”

The Canadian financial system has coped wellduring the recent period of increased financialstress. Over the past two years, the global finan-cial system has been affected by a number ofshocks in conjunction with a weakening eco-nomic environment in the industrialized econ-omies. The Canadian financial system has notbeen unaffected. It has faced a period of slowerdomestic economic growth, a deterioration inthe quality of corporate credit, and substantiallosses by investors (including losses on invest-ments made by financial institutions). Revela-tions concerning corporate malfeasance(primarily in the United States) also shook

Key Points

• The Canadian financial system hasdemonstrated impressive resilience inthe face of a significant degree of globalfinancial stress.

• Revelations of corporate malfeasancehave prompted worldwide efforts toimprove transparency and buttress con-fidence in corporate accounting andgovernance.

• An important degree of uncertainty per-sists within the global economic andfinancial system, indicating a continuingneed to closely monitor and analyzefinancial developments.

market confidence. Together, these factors havecontributed to heightened uncertainty and, attimes, have been reflected in above-averagelevels of volatility in financial markets.

In the face of these global developments, theCanadian financial system has been resilient.Aided by strong financial foundations andimproved diversification relative to previousperiods of increased stress, Canadian financialinstitutions as a group remain in a sound posi-tion. Financial markets, despite periods of in-creased volatility relative to historical norms,functioned relatively well while maintainingreasonably good levels of liquidity. In addition,efforts are underway by both public and privatesector entities, in North America and elsewhere,to improve transparency and rebuild confi-dence in corporate accounting standards andgovernance practices.

A key element going forward is the increaseduncertainty that continues to persist within theglobal economic and financial system. Pastshocks will continue to affect the financialsystem, and global economic, financial, andgeopolitical uncertainties remain. In this envi-ronment, it will be important to closely moni-tor and analyze developments within thefinancial system as well as its overall perfor-mance. To improve understanding of how thefinancial system operates during periods ofabove-average stress, it is also important toencourage discussion and analysis of ongoingfinancial innovation (e.g., the steadily broad-ening array of financial instruments), both inCanada and globally.

3

Developments and Trends

Chart 1 Total Assets: Canadian FinancialInstitutions

Source: Bank of Canada

Can$ billions (log scale) Can$ billions (log scale)

600

800

1000

1200

1600

1800

1400

2000

150

200

250

300

350

400

450500

1992 1994 1996 1998 2000 2002

Banks(left scale)

Life insurers(right scale)

Chart 2 Real GDP Growth

Year-over-year percentage change

* Member economies of the OECDSource: Thompson Financial Datastream and Statistics

Canada

Canada

0

1

2

3

4

5

6

0

1

2

3

4

5

6

1998 1999 2000 2001 2002

Industrializedeconomies*

1

2

3

4

5

Chart 3 Tech Bubble

Indexed to 2 January 1998 = 100

Source: Thompson Financial Datastream

50

100

150

200

250

300

350

50

100

150

200

250

300

350

1998 1999 2000 2001 2002

NASDAQ

Dow JonesIndustrials

Highlighted Issues

Issues discussed in this section include the im-pressive resilience of the Canadian financialsystem in the face of increased global economicand financial stress, and the ongoing interna-tional and Canadian response to concernsover corporate accounting and governancestandards.

The Resilience of the CanadianFinancial System

The Canadian financial system, including finan-cial institutions, has expanded significantly overthe past decade (Chart 1). During the past twoyears, however, the global financial system hasexperienced an unusual series of adverse shocks,placing strains on national financial systemsaround the world. In the face of elevated levelsof financial stress, the Canadian financial sys-tem as a whole has shown itself capable of effec-tively meeting the current challenges. Canadianfinancial institutions entered the recent periodof economic slowing in a much-improved posi-tion compared with earlier stressful episodes,owing partly to strong profitability and im-proved capital positions within the banking sec-tor. Nevertheless, heightened levels of pressureon the financial system are likely to continueover the immediate future.

The recent adverse economic and financialshocks not only directly contributed to lossesfor investors and profit declines for financialinstitutions generally, but also contributed toa weaker global economic climate (Chart 2).Among these shocks was the bursting of theso-called “tech bubble,” which began in 2000.Prospects for telecommunications and Internet-related companies deteriorated dramaticallyand, as is evident in the NASDAQ index, werereflected in their falling equity valuations(Chart 3). This has led to significant sectorallosses for bank lenders and unusually highlosses for bond and equity holders.

In addition, the 11 September 2001 terroristattacks in the United States affected more thanshort-term prospects for economic growth.They also led to substantial global claims on in-surance companies and created difficulties anduncertainty for specific industries (e.g., airlinesand tourism). Geopolitical concerns have per-sisted, with a heightened level of tension in

4


Chart 4 Bond Spread (EMBI+)

Basis points (over U.S. Treasury bonds)

Source: J.P. Morgan

0

1000

2000

3000

4000

5000

6000

7000

8000

0

1000

2000

3000

4000

5000

6000

7000

8000

1998 1999 2000 2001 2002

Brazil

Argentina

Chart 5 Global Corporate Defaults

* Annualized figures based on the first three quartersSource: Standard & Poor’s

US$ billions

0

40

80

120

160

200

240

280

0

40

80

120

160

200

240

280

1982 1987 1992 1997 2002*

By number of issuers(left scale)

By debt amount(right scale)

Chart 6 Volatility of Equity Indexes

10-day annualized historical volatility

Source: Thompson Financial Datastream and Bank ofCanada calculations

%

0

10

20

30

40

50

60

0

10

20

30

40

50

60

2002

S&P/TSX

FTSE

S&P 500

areas such as the Middle East, the Persian Gulf,and Southeast Asia.

Markets have also been affected by mounting fi-nancial difficulties in several Latin Americancountries, and bond spreads have risen as a re-sult (Chart 4). In January 2002, Argentina effec-tively defaulted on its sovereign debt. Morerecently, markets have shown concern over thefinancial stability of Brazil.

In December 2001, the U.S. energy companyEnron declared bankruptcy, an event notablefor the surprising rapidity with which it occurred.Subsequent revelations about questionable cor-porate accounting and governance practices atEnron and at other companies damaged inves-tor confidence in financial statements and ledto a reassessment of the financial strength of anumber of corporations.

These events, along with the sluggish globaleconomy, contributed to a heightened level ofcorporate fragility. Defaults on corporate bondsglobally rose to unprecedented levels in 2002(Chart 5). Other sources of concern alsoemerged, such as the potential need for corpora-tions to make “top-up” contributions to definedbenefit pension plans because of diminished(or negative) returns on the plans’ equityportfolios.

During the summer of 2002, further dramaticrevelations concerning questionable corporateaccounting and governance practices againshook market confidence in the quality of fi-nancial statements and corporate management.The decline in global equity markets, com-pounded by a reallocation of investor portfoliostowards less-risky assets, contributed to dimin-ishing optimism regarding the economic out-look (particularly in the United States and,more recently, in Europe). Reflecting an elevat-ed level of uncertainty, financial markets, glo-bally and in Canada, were characterized byheightened volatility relative to typical levels(Charts 6 and 7).1

In this environment of increased uncertainty,changing perceptions of risk associated withfinancial assets and declining investor will-ingness to take on risk (i.e., rising risk aver-sion) contributed to the movements in equityand bond prices. Spreads on high-yield

1. While the recent observed levels of volatility areunusual, they are not unprecedented.

5


Chart 7 2-year Government of Canada Bonds:Yield and Volatility

* 10-day historical volatilitySource: Thompson Financial Datastream and

Bank of Canada calculations

%

2.5

3.0

3.5

4.0

4.5

0

1

2

3

4

2002

Yield(left scale)

Volatility*(right scale)

%

Chart 8 Spreads Between Canadian Corporateand Government Bonds

Basis points

* Large changes can occur owing tocompositional shifts

Source: Scotia Capital

Overall-termhigh yield*

0

200

400

600

800

1000

1200

1400

0

200

400

600

800

1000

1200

1400

2001 2002

Mid-term BBB

Long-term A

Chart 9 Provision for Credit Losses:Major Canadian Banks

* Assets on a taxable equivalent basisNote: Figures for 2002 are based on the first three

quarters.Source: Canadian Bankers Association and Bank of

Canada

Can$ millions

0

1000

2000

3000

4000

5000

6000

7000

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1991 1993 1995 1997 1999 2001

Level(left scale)

Per cent ofaverage total assets*

(right scale)

(i.e., relatively risky) bonds rose particularlysharply (Chart 8).

Inevitably, all of these developments have ad-versely affected the revenues and profitability ofCanadian financial institutions. More generally,losses have been widely distributed amongholders of financial assets. For example, thedownward trend in equity values has directly af-fected merchant banking portfolios, the invest-ment portfolios of insurance companies andpension funds, and the wealth of individual in-vestors. The deterioration in credit quality hasmost directly affected banks through increasedloan losses, but non-financial companies havealso experienced losses on vendor financing.Bondholders have also faced higher default lev-els (although declining medium- and long-term interest rates for higher-grade bonds haveprovided an offset in the form of higher pricesover the recent period).

Overall, the Canadian financial system hascoped well with the events and related pressuresof the past two years. Banks, which account forover one-half of the assets held in Canada bythe financial services sector, have responded tothe corporate sector difficulties arising fromthese shocks by adding significantly to theirloan-loss provisions (Chart 9). Insurance com-panies have also provisioned against losses ontheir portfolio investments. This has been ac-complished without significant impact on thecapital base of Canadian financial institutionsas a whole. Banks have also sought to managetheir exposure to credit risk in the current cli-mate, and some major banks have announcedlonger-term plans to reduce the proportion ofcapital allocated to corporate lending. Despiteheightened levels of volatility relative to histor-ical norms, financial markets continued to func-tion relatively well, setting prices of financialassets while maintaining reasonably liquidmarkets.

The resilience of the Canadian banking systemis especially welcome when compared with thedifficulties that the sector faced during earliercyclical downturns (e.g., in the early 1990s,which included the associated downturn in thecommercial real estate sector). Several factorshave contributed to this. Various measures haveenhanced the banks’ overall robustness (e.g.,loan-loss provisioning practices have been ad-justed to improve their timeliness and includethe use of general provisions), and banks now

6


Chart 10 Non-Performing Loans: Major Banks

As a percentage of total loans*

* Net of loan-loss allowances** Impaired loans less allowancesSource: Office of the Superintendent of Financial

Institutions

1990 1992 1994 1996 1998 2000 2002-1

0

1

2

3

4

5

6

-1

0

1

2

3

4

5

6

Net impaired loans**

Gross impaired loans

Chart 11 Tier 1 Capital Ratio: Major Banks

Net Tier 1 capital/Total risk-weighted assets

Source: Office of the Superintendent of FinancialInstitutions

%

6.5

7.0

7.5

8.0

8.5

9.0

9.5

6.5

7.0

7.5

8.0

8.5

9.0

9.5

1994 1996 1998 2000 2002

Chart 12 Size of Global Credit-Risk TransferMarkets

* Figure for 2002 is estimated© BBA Enterprises Ltd. 2002

All rights reserved.

0

500

1000

1500

2000

0

500

1000

1500

2000

1997 1998 1999 2000 2001 2002*

US$ billions

have more diversified revenue sources and port-folios. Benefiting from diversification, the levelof non-performing loans has risen but remainswell below earlier peaks (Chart 10). The averageTier 1 capital ratio of the major banks hasmoved upwards over the course of the 1990sand remains high despite current pressures(Chart 11).2

How financial systems in Canada and elsewhereperform during periods of increased financialstress may be affected by changes brought aboutby financial innovation.3 For example, innova-tive financial instruments involving the transferof credit risk (Chart 12) have expanded themeans available for managing risk exposure.These instruments, which allow credit risk (i.e.,the potential for losses when extending credit)to be treated essentially as a tradable commodi-ty, facilitate the adjustment of risk profiles in away that better reflects institutions’ ability to ab-sorb risk.4 At the same time, markets for someof these instruments are relatively new. Particu-larly during periods of above-average financialstress, it is important to encourage assessmentsof the effects of such financial innovations tohelp in understanding their implications.

Responding to Deficiencies inCorporate Accounting andGovernance

A series of revelations last summer increasedconcerns about the use of improper accountingmethods in corporate financial statements.These events also reinforced questions aboutwhether senior corporate executives and boardmembers had acted in the best interests ofshareholders. The perception that market confi-dence in corporate behaviour and financialstatements had been severely impaired added asense of urgency to the efforts within officialand private sector bodies to buttress confidencein the standards for corporate accounting andgovernance. This was particularly important inan environment where other factors were

2. Tier 1 capital represents the highest quality of capitalthat underpins the banks’ activities.

3. For a discussion of innovation among financial-service providers, see the article by Freedman andGoodlet in this Review (page 57).

4. See Box 5 on page 20 for a more specific discussion ofthese financial instruments.

7


Box 1

Selected Organizations

AASOC Auditing and Assurance StandardsOversight Council (Canada)

AcSB Accounting Standards Board(Canada)

AcSOC Accounting Standards OversightCouncil (Canada)

CICA Canadian Institute of CharteredAccountants

CSA Canadian SecuritiesAdministrators

CPAB Canadian Public AccountabilityBoard

FASB Financial Accounting StandardsBoard (United States)

FSF Financial Stability Forum

IAIS International Association ofInsurance Supervisors

IASB International AccountingStandards Board

IDA Investment Dealers Association(Canada)

IOSCO International Organization ofSecurities Commissions

OECD Organisation for Economic Co-operation and Development

OSFI Office of the Superintendent ofFinancial Institutions (Canada)

PCAOB Public Company AccountingOversight Board (United States)

SEC Securities and ExchangeCommission (United States)

TSX Toronto Stock Exchange

already contributing to increased risk aversionon the part of investors.

The range of international activity in this areahas been extensive. In the United States, wheremuch of the corporate malfeasance originallycame to light, government legislation in theform of the Sarbanes-Oxley Act (signed into lawon 30 July) introduced a broad array of changesdesigned to improve corporate governance andraise confidence in financial statements. The Actincluded a new oversight board (the PCAOB) tomonitor the U.S. accounting industry. See Box 1for a list of relevant organizations.

Securities regulators, stock exchanges, and gov-ernment authorities have also been examining awide range of business activities with respect totheir impact on investor confidence. For exam-ple, the combination of investment bankingactivities and stock research in the same firm,together with the distribution of shares underinitial public offerings (IPOs), has come underclose scrutiny because of perceived conflicts ofinterest.

There has also been an increased focus on inter-national co-operation. The IASB and the U.S.FASB recently issued a memorandum describ-ing their joint efforts to bring about greater con-vergence in global accounting standards. InOctober, the FASB released a proposal for a newprinciples-based approach to setting U.S.accounting standards. Also in October, IOSCOreleased a set of principles designed to guidenational securities regulators with respect tocompany disclosure and auditor independenceand oversight. This broad re-examination ofcorporate governance and accounting has beensupported by international bodies such as theFSF and the OECD.

These events have raised a range of issues forboth financial and non-financial Canadiancompanies, and for the relevant oversightbodies in Canada (see Box 2). While efforts toincrease financial transparency, allowing inves-tors to better assess risk and financial sound-ness, will be beneficial, careful evaluation overtime will be important to avoid unintendedeffects from regulatory and related changes.

8

9


International developments have raised a numberof issues for Canadian authorities, regulators, andaccounting bodies (among others). In some cases,Canadian firms are directly affected by foreign in-itiatives. For example, they may have their shareslisted on U.S. stock exchanges and are thus affect-ed by some of the measures contained in the Sar-banes-Oxley Act.1 More fundamentally, in thecurrent environment of diminished investor con-fidence, it appears unlikely that the status quo is aviable option. In assessing which changes toadopt, however, some provincial regulatory agen-cies and stock exchanges have argued that cautionis necessary, suggesting that not all of the newrules introduced in the United States, for example,may be appropriate in the Canadian context. Asample of recent initiatives in Canada follows.

Auditing Standards

Perhaps the key development in Canada to datehas been the announcement by the CSA, OSFI,and CICA of a new oversight board, the CPAB. TheBoard will oversee the introduction of new rulesinvolving more rigorous inspection of the audi-tors of public companies, auditor independence,and new quality-control requirements for audit-ing firms. It will have the ability to imposesanctions against auditors who do not meet therequired standards. In September, the CICA pro-posed new guidelines to assess and buttressauditor independence. These were adapted frominternational rules, as well as from the new SECrequirements in the United States. The CICA hasalso announced a new oversight body for the set-ting of auditing and assurance standards (theAASOC).

Accounting Standards

Specific accounting issues are also being ad-dressed. The CICA has proposed new guidelineson the treatment of financial “special-purpose ve-hicles” in financial statements. The treatment ofstock options has been a particularly contentiousissue, both internationally and domestically. TheCanadian AcSOC has publicly indicated that itsupports in principle companies expensing stockoptions. More recently, the AcSB has announcedthat it will draw up explicit proposals to that ef-fect. Indeed, some Canadian firms have alreadyindicated that they will begin to expense stockoptions on their financial statements.

Governance Issues

Partly at the behest of securities regulators, stockexchanges have been examining their corporate-governance requirements for listed companies.For example, the TSX has announced changes forlisted companies with regard to board composi-tion. It has also supported the proposal that CEOscertify the accuracy of their companies’ financialdisclosures.

A contentious issue has been that of potential con-flicts of interest in firms that offer an array of fi-nancial services. Some provincial regulators arecurrently pursuing this issue with the major Cana-dian securities firms. In April, the IDA proposed anew set of guidelines to address conflicts of inter-est, but a consensus has not yet developedaround these proposals. Some provincial regula-tors have recently proposed increased penaltiesfor firms and individuals who violate securitiesregulations.

Finally, in November, the CICA released a newframework for the management discussion andanalysis component of corporate reports.

1. While this involves a relatively small number ofCanadian publicly listed firms, they neverthelessrepresent approximately one-half of the marketcapitalization of Canadian firms. There is ongoingdiscussion as to whether Canadian companies andtheir auditors will receive an exemption from someelements of the new U.S. rules.

Box 2

The Regulatory and Accounting Response in Canada


10

The MacrofinancialEnvironment

Global Environment

Over the past six months, the global environ-ment has been marked by heightened uncer-tainty and downward revisions to estimates ofthe strength of the global recovery. Consensusprojections for growth this year in the industri-alized economies, after picking up sharply fol-lowing the surprisingly rapid rebound duringthe first part of the year, stabilized and thenreversed course in the summer (Chart 13).

A number of factors contributed to the pullbackin expectations for growth. These included agenerally unexpected weakening in U.S. growthin the second quarter, the uncertainty created byrevelations (primarily in the United States) con-cerning deficient corporate accounting and gov-ernance, and the downturn in global equitymarkets.

These developments occurred in an environ-ment where corporations were already underfinancial pressure. This was exemplified by theincrease in credit-rating downgrades relative toupgrades for U.S. and European firms (Chart14), with elevated levels of downgrades con-tinuing into the third quarter. The telecommu-nications sector has been responsible for arelatively large proportion of problem issuers.Reflecting the higher level of global financialstress, financial institutions have also accountedfor a relatively large proportion of credit down-grades. This is particularly the case for insurancecompanies which, worldwide, have come underpressure from higher claims and reduced invest-ment income.5

In some foreign economies, prices in the resi-dential real estate market have been rising rap-idly (e.g., Australia, the United Kingdom, and,to a lesser extent, the United States). Althoughthis has provided an offset to the decline inhousehold wealth arising from lower equityprices, such increases will eventually slow andmay, in some cases, reverse.

5. Standard & Poor’s estimates that the global insur-ance/reinsurance industry has experienced a netreduction in capital of approximately US$200 billionover the past several years.

Chart 13 Evolution of Consensus Estimates forAnnual Growth in 2002

Industrialized economies*

* North America, Western Europe, and JapanSource: Consensus Economics

%

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

2001 2002

Chart 14 Corporate Downgrade/Upgrade Ratio

Note: Figures for 2002 are based on the first threequarters.

Source: Moody’s

1995 1997 1999 20010

1

2

3

4

5

0

1

2

3

4

5

WesternEuropean

firms

U.S. firms


Table 1

Total Canadian Bank Claims on SelectedLatin American Countries ($ millions)*

* Includes local claims of subsidiaries

2002Q2 2001Q2

Mexico 21,465 22,241

Chile 4,947 4,536

Argentina 2,467 **

** Does not reflect the sale of Quilmes Bank byScotiabank (in the third quarter of 2002)


4,798

Brazil 2,039 2,267

Peru 558 493

Venezuela 537 508

Uruguay 96 110

Chart 16 Indicators for Brazil

* Increase indicates a depreciationSource: Thompson Financial Datastream and

J.P. Morgan EMBI+

Basis points Brazilian real/US$

500

1000

1500

2000

2500

3000

2.0

2.5

3.0

3.5

4.0

4.5

2002

Brazilian bond index(spread relative to U.S. Treasury bonds)

(left scale)

Brazil/U.S.exchange rate*

(right scale)

Chart 15 Crude Oil Prices


US$/barrel

16

18

20

22

24

26

28

30

32

34

16

18

20

22

24

26

28

30

32

34

2001 2002

Ongoing geopolitical concerns have been an ad-ditional source of uncertainty. These includetensions in the Middle East and the possibilityof military action in the Persian Gulf. Althoughthese concerns have induced some additionalvolatility in oil prices, price levels have movedup only moderately from the beginning of theyear (Chart 15).

Emerging marketsEconomic growth in emerging Asia has re-mained relatively robust, although showingsome recent signs of weakening. The region hasexperienced only modest increases in interestrate spreads. Although the picture is lesssanguine in Latin America (Box 3), emerging-market debt and equities generally have recent-ly rallied as risk aversion appears to havediminished somewhat.

Economic and financial problems in Argentina,which effectively defaulted on its sovereign debtin early January, have persisted. The Argentineauthorities and the International Monetary Fundhave so far been unable to reach a successfulconclusion to their discussions on a new finan-cial program. In November, Argentina fell intoarrears on its World Bank debt.

The financial situation is particularly volatile inBrazil, owing to uncertainty over the new gov-ernment’s economic program. This has beenreflected in a sharp deterioration in Brazilianfinancial indicators, although they have recent-ly improved (Chart 16).

Although some smaller Latin American coun-tries have also come under increased financialstress, Mexico, assisted by a strengthened policyframework, has been largely unaffected byregional developments. The exposure ofCanadian banks to Latin America is largelyconcentrated in Mexico (Table 1).

Japan and EuropeIn Japan, near-term prospects for growth in do-mestic demand remain relatively poor, owingpartly to low consumer confidence and persis-tent excess capacity. The weak economic perfor-mance of the economy since the early 1990s hasbeen associated with persistent problems in theJapanese financial sector. Japanese banks havedisposed of about ¥90 trillion (approximatelyCan$1.1 trillion) in non-performing loans overthe past decade. Nevertheless, non-performing

11


12

Many economies in Latin America have experi-enced deteriorating financial market conditionsthis year, as evidenced by rising bond spreads(Chart 1). The region has also experienced a sharpslowdown in economic activity, with some econ-omies sustaining significant contractions in realGDP. Argentina’s financial crisis is not yet re-solved and has led to a reduction in output of16 per cent in the first half of 2002 compared withthe first half of 2001. Brazil and Uruguay also re-main vulnerable, despite large financial packagesfrom the IMF. Uncertainty regarding the new Bra-zilian government’s economic program is con-tributing to financial market volatility in thatcountry, while concerns about the viability ofthe Uruguayan banking system are underminingfinancial conditions there. Brazil experienced nogrowth in the first half of 2002 compared with thefirst half of last year, while Uruguay’s GDP hascontracted nearly 8 per cent.

Domestic political uncertainties in Venezuela,Ecuador, and Colombia have contributed to finan-cial market volatility in these countries and toslowing growth as investment deteriorates.

While Mexico and Chile have experienced somefinancial sector turbulence, it has been on a muchsmaller scale than that seen in other Latin Ameri-can economies. An adverse terms-of-trade shock(low copper prices and high oil prices) for Chileand the hesitant U.S. recovery help explain the fi-nancial market volatility in these countries. Both

economies are expected to grow modestly thisyear.

Consensus forecasts indicate that growth in the re-gion will rebound to 2.3 per cent next year from-0.6 per cent this year (Chart 2). An important fac-tor driving expectations of a recovery is a bottom-ing out of the recession in Argentina and modestgrowth rebounds in Brazil and Mexico, as theirmajor export markets pick up.

There are two important risks to the outlook inLatin America. One is the strength of the recoveryin industrial countries, particularly in the UnitedStates. The second is whether Brazil’s new govern-ment will be able to restore lasting confidence andreduce vulnerabilities there. Continued difficul-ties in Brazil would negatively affect its closesttrading partners in the region (Argentina,Uruguay, and Paraguay).

The direct economic and financial links betweenCanada and Latin America are relatively small.Within Latin America, Canada is most closelylinked with Mexico and Chile, currently two of thehealthiest economies in the region. Also, given therelatively small size of the Latin American econo-my and the structure of its trade, even a further de-terioration of demand conditions would probablyhave a relatively small impact on global commod-ity prices. Nevertheless, the need to increase loan-loss provisions on their Latin American exposureis one factor behind increased provisioning levelsat Canadian banks this year.

Box 3

Economic and Financial Developments in Latin America

Chart 1 Spreads on Selected Emerging-MarketBonds (EMBI+)

Source: J.P. Morgan

Basis points

0

1000

2000

3000

4000

5000

6000

7000

8000

0

1000

2000

3000

4000

5000

6000

7000

8000

2001 2002

Brazil

Chile

Argentina

Mexico

Uruguay

Chart 2 Real GDP Growth: Latin America

Annual growth rate

Notes: 1. Latin America in this chart includesArgentina, Brazil, Chile, Colombia,Mexico, Peru, and Venezuela.

2. Forecasts for 2002 and 2003 are fromConsensus Forecasts

-2

0

2

4

6

-2

0

2

4

6

1990 1992 1994 1996 1998 2000 2002


Chart 17 Non-Performing Loans: Japan

* Change in definition with respect to non-performingloans affected the 2002 level

Source: Bank of Japan and Japanese Financial ServicesAgency

0

5

10

15

20

25

30

35

10

15

20

25

30

35

40

45

1992 1994 1996 1998 2000 2002

Bank disposal ofnon-performing loans

(left scale)

Non-performing loans*(right scale)

¥ trillion ¥ trillion

Chart 18 Japanese Stock Market (Nikkei)



60

80

100

120

140

160

180

200

60

80

100

120

140

160

180

200

1998 1999 2000 2001 2002


20

30

40

50

60

70

80

90

100

110

20

30

40

50

60

70

80

90

100

110

2001 2002

FTSE Europeaninsurance sector

FTSE 100

Chart 19 European Insurance Sector:Equity Prices


loans still on the banks’ books remain at highlevels (Chart 17).

In a report published by the Bank of Japan, thepersistent high level of bad loans was attributedto newly generated non-performing loans fromcorporate restructuring. In addition, the reportpointed to problems at banks created by the lowlending margins and the fact that unrealizedgains on bank shareholdings which, at one time,acted as a financial reserve for banks, have evap-orated as a result of the decline in equity prices(Chart 18). With the shift to mark-to-marketaccounting last year, market risk pertaining tobank shareholdings is now considered to be asignificant destabilizing factor. To help mitigatethis, the Bank of Japan announced that it wouldbuy up to ¥2 trillion of equities from domesticbanks, with purchases beginning in late No-vember. The impact of this measure is not yetclear.

Despite past efforts, the overall level and qualityof the capital of Japan’s major banks is consid-ered to compare poorly relative to their interna-tional peers. Action to support the financialsector has been hampered by a lack of domesticconsensus, although the government recentlyannounced a set of proposals aimed at reducingproblem loans in the banking system.

The European banking sector began the most re-cent credit cycle in a sounder position than atthe beginning of the previous decade. Neverthe-less, weaker economic growth and deterioratingcredit quality have placed banks under in-creased pressure. This has been particularly truein Germany, where bank sector profitability hadbeen comparatively weak. The credit ratings ofseveral major European banks were reduced inrecent months, and bank spreads have widened.

The European insurance sector has also comeunder increased pressure, partly as a result ofthe decline in equity prices. Despite significantstrengths as a group, European insurance com-panies tend to invest a relatively large propor-tion of their assets in equities and thus haveexperienced significant investment losses (add-ing to those on their bond portfolios). This,together with higher insurance payouts, hasadversely affected their financial strength and isreflected in the substantial decline in the equityvaluation of the sector (Chart 19). The capitallevels of large European reinsurance companieshave also been adversely affected by developments

13


Chart 20 Real GDP Growth: United States


-2

0

2

4

6

8

-2

0

2

4

6

8

1998 1999 2000 2001 2002

Year-over-yearpercentage change

Quarterly growth at annual rates

Chart 21 U.S. Corporate Debt and Net InterestPayments


%

600

700

800

900

1000

1100

1200

10

15

20

25

30

35

40

1980 1985 1990 1995 2000

Corporate debtto cash flow(left scale)

Net interest paymentto cash flow(right scale)

%

Chart 22 U.S. “Adversely Rated” SyndicatedLoans

As a per cent of commitments

Source: U.S. Federal Deposit InsuranceCorporation Shared National Credit Review

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

1991 1993 1995 1997 1999 2001

over this period, and some companies have hadtheir credit ratings lowered.

United StatesU.S. real GDP growth rebounded in the thirdquarter of this year (Chart 20). Despite this, therecent softening in business and consumer con-fidence and the slowing in manufacturing out-put have contributed to continued uncertaintyabout the persistence of the current recovery.

In this environment, business conditions re-main difficult. U.S. corporations continue to beaffected by a relatively high pace of credit down-grades (recently including Ford and GeneralMotors), and corporate profits remain underpressure.6 Nevertheless, aggregate corporatedebt and net interest payments relative to cashflow have moved down from their recent peaks(Chart 21).

In spite of the elevated level of corporate de-faults and a continued deterioration in syndi-cated loan quality (Chart 22), U.S. banks havebeen able to maintain a high level of profitabil-ity. Commerical banks earned US$68.6 billionin the first three quarters of the year, more than10 per cent above earnings for the same periodin 2001. Non-performing loans have continuedto rise, but the overall coverage ratio of reservesto non-performing loans has remained wellabove 100 per cent (Chart 23).

Some large U.S. banks that are heavily involvedin investment banking have fared relativelypoorly. Activity levels in the investment bank-ing area have diminished, reducing revenuesand, in some cases, leading to substantial reduc-tions in staffing. The credit ratings of severallarge institutions have been downgraded. Onerisk that is difficult to evaluate is the extent ofthe legal risk of litigation and investigation thatfinancial firms face in the United States in theaftermath of the recent financial and accountingimproprieties on the part of some corporations.

Reflecting developments in the global industry,the financial position of U.S. insurance compa-nies has deteriorated. Declining investment rev-enues and higher insurance payouts have hurtprofits. Property and casualty insurance firmshave been most heavily affected, and a portion

6. Standard & Poor’s reported that U.S. companiesincurred 157 credit-rating downgrades in the thirdquarter and only 40 upgrades.

14


Chart 23 U.S. Banks

Source: U.S. Federal Deposit InsuranceCorporation Quarterly Banking Profile

US$ billions %

20

30

40

50

60

70

80

90

100

40

60

80

100

120

140

160

180

200

1998 1999 2000 2001 2002

Loan-loss reserves(left scale)

Coverage ratio(right scale)

Non-current loans(left scale)

Chart 24 Real GDP Growth: Canada

Source: Statistics Canada

-2

0

2

4

6

8

1998 1999 2000 2001 2002

Quarterly growthat annual rates

Year-over-yearpercentage change

-2

0

2

4

6

8

Chart 25 Canadian Consumer Confidence

Indexed to 1991 = 100

Source: The Conference Board of Canada

100

105

110

115

120

125

130

100

105

110

115

120

125

130

1998 1999 2000 2001 2002

of the sector’s largest companies are believed tobe currently under-reserved against losses. Thissuggests that some insurers, in conjunction withadjusting premiums, will need to take measuresto boost reserves.

Canadian Developments

Domestic factors that have an important influ-ence on developments in the Canadian finan-cial system include the state of the Canadianeconomy, the financial position of the house-hold and corporate sectors, and developmentswithin specific industrial sectors (e.g., the tele-communications sector).

Canadian economyCanada experienced a robust economic expan-sion in the first three quarters of 2002 (Chart 24).7

The growth of household spending has beenquite strong. In addition, the resumption of in-ventory accumulation, marked gains in exports,and a recovery in business spending on machin-ery and equipment all helped to support eco-nomic expansion during this period. Theconsensus view calls for Canadian economicgrowth to exceed that of other G-7 economies inboth 2002 and 2003.

There are, nevertheless, a number of uncertain-ties in the current environment. Global eco-nomic, financial, and geopolitical develop-ments (discussed in previous sections) havecontributed to a weaker near-term outlook forthe expansion of the global economy (relativeto consensus expectations at the beginning ofthe summer) and therefore of Canada’s exports.Spending by corporations and householdscould also be adversely affected by the increaseduncertainty associated with concerns aboutcorporate governance and the unsettledgeopolitical climate.

Household and corporate sectorsThe financial condition of the household sectorcontinued to be relatively healthy in 2002. Withstrong employment gains, consumer confi-dence has remained at a fairly high level despitea slight easing since mid-year (Chart 25).

7. The Bank of Canada’s interpretation of global anddomestic economic trends is published every springand autumn in the Bank’s Monetary Policy Report.

15


Chart 26 Personal Sector Debt

* Does not include principal repaymentsSource: Statistics Canada

12

14

16

18

20

6

8

10

12

14

1980 1984 1988 1992 1996 2000

Householddebt to total assets

(left scale)

Householddebt-service ratio*

(right scale)

%

Chart 27 Growth in Real House Prices*

Year-over-year percentage change

* Deflated by total CPISource: Statistics Canada, Royal LePage

-25

-20

-15

-10

-5

0

5

10

15

20

25

-25

-20

-15

-10

-5

0

5

10

15

20

25

1980 1985 1990 1995 2000

New houses

Existing houses

Chart 28 Residential Mortgage Loans in Arrears3 Months or More

Source: Canadian Bankers Association and Bank ofCanada

Thousands of loans %

0

2

4

6

8

10

12

14

16

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

19861988 1990 1992 1994 1996 1998 2000 2002

As a percentageof total residential

housing loans(right scale)

Level(left scale)

Personal sector indebtedness has risen sharplyover the past decade, stabilizing in recent yearsat about 110 per cent of personal disposable in-come. Reflecting the higher level of householdassets, however, the household debt-to-assetratio has been more stable, exhibiting only amodest upward trend over the past several years(Chart 26). Household debt-service costs haveremained low throughout the current cycle, aid-ed by past declines in consumer and mortgageinterest rates.

With real estate assets accounting for about one-third of household wealth in Canada, the cur-rent strength in housing prices has providedsome offset to the loss in wealth arising fromweaker equity prices (Chart 27). Residentialmortgage arrears have remained relatively flat inlevel terms and have fallen as a percentage oftotal mortgages outstanding, despite the recentcyclical slowdown in economic growth(Chart 28).

The financial condition of the overall non-financial corporate sector also improved in thefirst half of 2002. However, with the volatility infinancial markets, business confidence, espe-cially that of large firms, fell appreciably in thethird quarter. Indeed, the proportion of largefirms that were unsure whether it was a good orbad time to undertake capital spending to ex-pand plants (one measure of uncertainty) hasrisen to a high level.8 As a result, firms have con-tinued to make more use of cash balances andless use of borrowings to help fund operationsand capital spending.

Corporate credit quality, after deteriorating fora number of quarters, has shown some signs ofstabilizing, supported by an improvement incorporate profits (Chart 29). The overall levelof credit-rating downgrades compared withupgrades has remained elevated throughoutthe year (Chart 30), and the number of firmsassigned a negative outlook by credit-ratingagencies remains high. Nevertheless, therehave been no additional reductions to below-investment-grade status among large firms inrecent months.

IndustryAlthough the financial position of the overallnon-financial corporate sector is relatively

8. The Conference Board of Canada, Index of BusinessConfidence, Autumn 2002.

16


Chart 29 Pre-Tax Corporate Profits

As a per cent of GDP

Source: Statistics Canada and Bank of Canada

4

5

6

7

8

9

10

11

12

13

4

5

6

7

8

9

10

11

12

13

1990 1992 1994 1996 1998 2000 2002

Chart 30 Corporate Issuers Downgrade/Upgrade Ratio

Excludes financial sector

* Based on the first three quartersSource: Moody’s

0

2

4

6

8

0

2

4

6

8

1995 1996 1997 1998 1999 2000 2001 2002

*

Chart 31 Return on Equity: AutomotiveManufacturing


%

0

5

10

15

20

25

30

0

5

10

15

20

25

30

1998 1999 2000 2001 2002

1988-2001 average

healthy, some industries (amounting to about15 per cent of total GDP) experienced signifi-cant financial stress, especially in the secondhalf of 2001. These include automotive manu-facturing and air and truck transport, as well asmost high-technology industries and producersof non-energy commodities.

However, with the overall economic recovery inCanada during 2002, activity increased substan-tially in most of the financially vulnerableindustries. As a result, there was some improve-ment in their financial situations. These gainsoccurred for automotive manufacturing(Chart 31), for air and truck transport, andproducers of metallic products.

Even so, the near-term financial outlook hasworsened for other industries experiencing fi-nancial stress. Levels of profitability are still verylow for many telecom equipment manufactur-ers (Chart 32) and telecom service firms,9 andnumerous companies are being forced to under-take further major restructuring of their opera-tions (Box 4). Activity in telecom equipmentmanufacturing decreased further in recentmonths, and most observers now expect that asignificant recovery in activity will not takeplace until 2004.

In addition, short-term financial prospects formany forest products companies remain weak.Prices are still very low, and lumber prices havefallen considerably since mid-year. The imple-mentation by the United States of high dutieson imports of lumber from Canada has addedto the downward pressure on profitability inthis industry.

The financial situation in Canada’s farm sectorhas also deteriorated. In particular, many grainand livestock producers are experiencing finan-cial stress as a result of the Prairie drought andvery low hog prices.

The Financial System

Financial Markets

Canadian markets have been influenced by glo-bal financial developments. Financial marketsworldwide experienced an above-average levelof volatility during the summer in response to a

9. The telecom services sector includes phone compa-nies, wireless carriers, and cable companies.

17


18

Box 4

Developments in Canada’s Telecom Sector

Canada’s telecom sector experienced very stronggrowth in the second half of the 1990s, reflectingthe impact on demand of marked reductions inrelative prices and booming export markets. Thiswas followed last year by a collapse in export de-mand and a marked weakening in profitability.

It is useful to put the recent developments in Can-ada’s telecom sector in an international context.The commercialization and mass-market adop-tion of the Internet during the second half of the1990s (and also the deregulation of the local U.S.telephone market) contributed to the belief thatthere were spectacular opportunities for growthworldwide. Firms in all parts of the global telecomsector substantially increased their capital outlaysduring this period, mainly with the intention ofbuilding additional national and global commu-nications networks based on new fibre-optic tech-nologies. A relatively high share of this rise inbusiness investment was financed through debtissuance.

While the growth of demand since the late 1990swas strong, it nonetheless fell well short of expec-tations and of the growth of capacity. With thebuildup of considerable excess capacity, prices fellmarkedly. Rates of profitability fell drastically,and some telecom companies failed. Other firmsresorted to questionable accounting procedures tohelp boost apparent revenue growth, leading tomany of the recently revealed accounting scan-

dals. Access to capital markets has subsequentlybeen curtailed, investment expenditures havebeen scaled back dramatically, and more telecomfirms have entered bankruptcy proceedings.

Canadian manufacturers of telecom equipmenthave been hard hit by these developments andhave been forced to drastically cut levels of activityand employment (Chart 1). Even with theseadjustments, most of these companies are stillexperiencing substantial financial losses. Mostindustry observers currently expect that a markedincrease in capital outlays by the global telecomsector will not take place until 2004, so that thereturn to profitability in Canada’s telecom manu-facturing industry may be delayed for some timeto come.

Canada’s telecom services industry has facedmany of the same problems as its foreign counter-parts. With substantial debt-financed increases incapital spending in 2000 and 2001, this industryexperienced a marked increase in its debt-equityratio. Revenue growth has fallen well short ofexpectations, and profitability has been very weaksince mid-2001 (Chart 2). Indeed, some wirelessservices providers are having particularly seriousfinancial difficulties. Moreover, with recent down-grades of much of the industry’s debt, most of thelargest companies have announced major restruc-turing of their operations and measures to reduceindebtedness.

Chart 1 New Orders: Communication-Equipment Manufacturing

Indexed to 2000Q3 = 100

20

40

60

80

100

120

20

40

60

80

100

120

1998 1999 2000 2001 2002Source: Statistics Canada

Chart 2 Operating Revenues and Net Profits:Canadian TelecommunicationsServices

* Annualized figure based on first two quartersSource: Statistics Canada

1988 1990 1992 1994 1996 1998 2000 2002*

Can$ billions Can$ billions

-10

0

10

20

30

40

50

-0.5

0

0.5

1.0

1.5

2.0

2.5Operating revenues

(left scale)Net profits(right scale)


Chart 32 Return on Equity: Electronic andComputer Manufacturing


%

-20

-15

-10

-5

0

5

10

15

-20

-15

-10

-5

0

5

10

15

1998 1999 2000 2001 2002

1988-2001 average

Chart 33 Yield for 10-Year U.S. BenchmarkBond

Month average


3

4

5

6

7

8

9

3

4

5

6

7

8

9

1990 1992 1994 1996 1998 2000 2002

%

Chart 34 Canadian Corporate Spreads byCredit Quality (5-10 years)

Basis points (over government securities)


BBB

0

50

100

150

200

250

300

0

50

100

150

200

250

300

1998 1999 2000 2001 2002

A

AA

range of factors that included revelations of de-ficient corporate accounting and governancepractices, geopolitical tensions, and changingperceptions about the economic outlook. Inthis uncertain environment, market participantsin Canada and elsewhere reacted with increasedsensitivity to new pieces of economic informa-tion. Risk aversion continues to be an importantfactor influencing markets, although recentfinancial indicators suggest that marketparticipants are becoming less risk averse.

Given the elevated level of loan losses, marketsfor credit-risk transfer, which continue to evolverapidly, have attracted considerable attention(Box 5). Despite the relative newness of some ofthese financial instruments, they appear to haveoperated largely as expected in the currentenvironment.

Fixed-income credit marketsIn fixed-income markets, international develop-ments have been significant. U.S. governmentbonds, which often benefit from “safe haven”flows during periods of increased turbulence,saw their yields approach historical lows in thethird and fourth quarters (Chart 33). Develop-ments in credit markets suggest that there hasbeen a shift to higher-quality assets, both corpo-rate and government, although recent evidenceindicates this effect is becoming less important.It is interesting to note that the risk premiumbetween bank and government credit in theUnited States (as represented by the spread be-tween the Treasury and fixed-rate portions of afixed-floating swap) has not increased, in con-trast to the experience during some previousturbulent periods (e.g., the autumn of 1998,during the problems associated with Long-TermCapital Management). Although other factorswere also present, this suggests that recent de-velopments have not markedly raised concernsover the longer-term health of the U.S. bankingsector.

In Canada, corporate credit spreads (over gov-ernment bonds) have risen from the lowsreached in the first half of the year, although de-clining somewhat in recent months (Chart 34).Although funding costs were near historicallows for many high-quality issuers, the uncer-tain economic environment and financial vola-tility led a number of firms to emphasize debtreduction. Canadian corporate debt issuancefell as financial market volatility peaked in July

19

20


Box 5

Evolution of Financial Markets for Transferring Credit Risk

Techniques for credit-risk transfer (CRT), al-though not new, have recently attracted signifi-cant attention with the development of marketsfor credit default swaps (CDSs) and collateralizeddebt obligations (CDOs). From the perspective ofa financial institution, company, or investor, CRTinstruments can be attractive because they allowthem to alter their financial-risk profile, either byincreasing their risk exposure (i.e., acquiring risk)or reducing it. Certain types of CRT, such as secu-ritization, can also represent an alternative sourceof funding for institutions that carry loans as as-sets on their balance sheets.

Significant developments in the CRT market in-clude the introduction of primary market syndica-tion of bank loans in the 1960s and the develop-ment of a secondary market in bank loans in theUnited States in the 1980s. Also in the UnitedStates, the 1970s saw the packaging and sale ofloan portfolios (securitization) to fund the financingof residential mortgages.

In recent years, the use of securitization and spe-cial-purpose vehicles (SPVs) to create asset-backedsecurities (ABSs) has spread to credit card receiv-ables, car loans, and commercial paper, amongother areas. Increasingly, originating institutionsdo not necessarily expect to hold on to assets, suchas mortgage loans or credit card receivables, tomaturity. Instead, these are seen as assets to bebundled and sold in order to finance otheractivities.

Collateralized debt obligations (CDOs) are similarto ABSs. The underlying assets for the CDO aretypically corporate bonds or loans, and the asso-ciated SPV uses two or more tranches of debt. The“junior” tranche absorbs any initial losses by theSPV, thereby insulating the more “senior”tranches. In a CDS, one counterparty agrees tocompensate the other if one or more specifiedcredit events (such as a debt default or a credit-rating downgrade) occur during the life of a con-tract, in exchange for a fee or premium. Themarket or standardized, “single-name” CDS iscurrently the most active and liquid area of thecredit derivatives markets. Finally, a “synthetic”CDO is created when an underlying portfolio of

high-quality debt securities is combined withCDSs to replicate the risks and returns of a port-folio of higher-yielding securities. CDSs andsynthetic CDOs can be used to transfer credit riskwithout transferring the underlying assets or, inthe case of loans, without even informing theborrower.

One implication of this activity is that risk expo-sure can be shifted more easily among financialsystem participants. But only limited informationis available on how this has influenced aggregaterisk exposures within the financial system. One re-cent survey suggests that banks, securities houses,and hedge funds tend to be the most active buyersof credit protection (which, alone, would reducetheir risk exposure), while these institutions, to-gether with insurance and re-insurance compa-nies, are the most active sellers (see chart).

Breakdown of Participants in Credit DerivativesMarkets

Showing institutions buying and selling credit protection

© BBA Enterprises Ltd. 2002All rights reserved.

0

10

20

30

40

50

60

0

10

20

30

40

50

60

BanksInsurance cos.

Re-insurers

Securities houses

Hedge funds

Mutual funds

Pension funds

Corporates

Government

Protectionpurchased

Protection sold

Per cent of total firms


Chart 35 Corporate New Issuance


Can$ billions

0

2

4

6

8

10

12

14

0

2

4

6

8

10

12

14

2001 2002

U.S-dollar debt issued byCanadian corporations

Canadian-dollar debt issuedby Canadian corporations

Chart 36 Equity Indexes



65

70

75

80

85

90

95

100

105

110

115

2002

S&P/TSX

DJIA

FTSE

Nikkei

65

70

75

80

85

90

95

100

105

110

115

Chart 37 S&P/TSX Index Level andVolatility

* 10-day annualized historical volatilitySource: Thompson Financial Datastream and

Bank of Canada calculations

Index level(left scale)


5500

6000

6500

7000

7500

8000

0

10

20

30

40

50

2002

%

and August, and many issuers chose to remainout of the market.10 Assisted by diminishedmarket volatility, bond issuance has recoveredsomewhat (Chart 35).

Equity marketsAlthough global equity prices had been fallingsince May, the pace of decline accelerated inmid-June, and international markets subse-quently posted broad-based losses (Chart 36).The market declines themselves added to fearsabout the resiliency of consumer confidenceand spending and, in turn, the outlook for con-tinued economic recovery. Most recently, equi-ty markets in Canada and elsewhere havemoved back from their lows, although in thecontext of above-average levels of volatility(Chart 37). The Canadian S&P/TSX index isnonetheless still about 15 per cent below its lev-el during the first part of the year.

The observed volatility in financial markets maybe linked to uncertainty among market partici-pants regarding appropriate equity valuationsin the current climate. The declines in equitymarkets have tended to move price-earnings ra-tios in Canada and elsewhere (based on trailing,i.e., actual, earnings) down towards their histor-ical trend values. However, uncertain prospectsover the future growth in company earnings,and in the economy more generally, suggestthat the volatility in equity market values maypersist.

Foreign exchange marketsThe Canadian-dollar exchange rate also reflect-ed the volatility that affected fixed-income andequity markets during the summer (Chart 38).Volatility in the Canadian/U.S. dollar exchangerate (which has typically been quite modest rel-ative to that of other major currencies againstthe U.S. dollar), has since returned to more nor-mal levels. The Canadian dollar has generallyremained in a range of 0.63 to 0.64 U.S. dollarsover the past several months. The U.S. dollar,which had moved down against the yen andeuro through the summer, has regained some ofits lost ground, particularly against the yen.

10. Similar to the United States, swap spreads on bankand government bonds in Canada have alsoremained relatively stable.

21


Chart 38 Exchange Rate and Volatility

* 10-day annualized historical volatilitySource: Thompson Financial Datastream and Bank

of Canada calculations

Can$/US$ %

1.45

1.50

1.55

1.60

1.65

1.70

1.75

1.80

0

2

4

6

8

10

12

14

2002

Exchange rate(left scale)


Chart 39 Provision for Credit Losses:Major Canadian Banks

Calculated from quarterly financial statementspublished by the major Canadian banksSource: Bank of Canada

Can$ millions

0

1000

2000

3000

4000

5000

0

1000

2000

3000

4000

5000

1998 1999 2000 2001 2002

Chart 40 Corporate Exposure of Banks

Per cent of total capital


0

100

200

300

400

500

600

0

100

200

300

400

500

600

1990 1992 1994 1996 1998 2000 2002

Equity

Credit

Total corporateexposure

Financial Institutions

The Canadian banking sector has responded tothe weakening in the overall quality of its assetbase. In the second half of fiscal 2002, the majorbanks again made large additions to loan-lossprovisions (Chart 39). Total loan-loss provi-sions in fiscal 2002 increased substantiallyrelative to a year earlier.

Specific problem areas for banks have beentheir loans to the telecommunications sector,involvement in the U.S. syndicated loan mar-ket, exposure to Latin America, and losses ontheir merchant banking portfolios. Banks havebeen working to reduce their exposure to thetroubled telecommunications sector. It never-theless remains at about $18 billion, althoughthe banks have provisioned against a substan-tial portion of this exposure. While the troubledU.S. merchant energy and power-generationsector will also contribute to losses, the impactis expected to be relatively limited. Banks havealso been affected by the decline in equityvaluations, which has become more importantto them in recent years in view of the trendincrease in bank holdings of equities (Chart 40).This increase reflects activities such as theirventure capital and merchant banking opera-tions, and investment dealer inventories ofequities.

The more difficult credit environment andincreased provisioning has placed downwardpressure on bank profits, and their return onequity has deteriorated, with some banks report-ing quarterly losses in the second half of fiscal2002 (Chart 41). The credit ratings of two majorCanadian banks were lowered by rating agen-cies in the autumn.

Canadian banks are nevertheless benefitingfrom the fact that their financial position as theyentered the current period of credit deteriora-tion was much improved relative to that of theearly 1990s. Their aggregate capital base hasremained strong in the face of recent develop-ments. Despite an increase in impaired loans,the extent to which they are covered by allow-ances (the “coverage ratio”) remains at muchhigher levels than was the case in previous cycli-cal slowdowns (Chart 42). Several banks haveindicated that, over the next several years, theyintend to place less emphasis on corporate lend-ing. They plan to reduce the capital devoted tosuch lending by approximately one-third, and

22


Chart 41 Return on Equity of Major CanadianBanks

Source: Canadian Bankers Association

%

0

5

10

15

20

25

0

5

10

15

20

25

1980 1985 1990 1995 2000

Chart 42 Major Canadian Banks:Asset Quality

* Relative to loans and acceptances less allowances** Total allowances/Gross impaired loansSource: Office of the Superintendent of Financial

Institutions

0

2

4

6

8

10

40

60

80

100

120

140

1990 1992 1994 1996 1998 2000 2002

Gross impaired loans ratio*(left scale)

Coverage ratio**(right scale)

%

Chart 43 Equity Price Indexes



60

70

80

90

100

110

120

130

60

70

80

90

100

110

120

130

2001 2002

S&P/TSX

Banks and trusts

reinforce areas such as retail banking andwealth-management operations. Bank sectorequity values have fared relatively well com-pared with broader equity indexes in Canada(Chart 43).

The Canadian insurance industry is well capital-ized and, as a whole, was not heavily exposed toclaims arising from the 11 September terroristattacks last year. Nevertheless, the sector faces anumber of challenges. Investment income, animportant source of revenue for the industry,has been adversely affected in the current envi-ronment. In addition, insurance companies facelosses on their extensive corporate bond port-folios (e.g., those involving firms in the NorthAmerican telecommunications, merchant ener-gy, and airline and aerospace sectors). Severalfirms have announced significant provisions forcredit losses.

The return on equity can vary substantiallyacross the industry. While remaining relativelyrobust among some of the major life insurers,the return on equity of the aggregate insurancesector has tended to decline, falling in the sec-ond quarter (Chart 44). The situation amongproperty and casualty insurers has been com-paratively difficult. Within this relatively heter-ogeneous group of companies, recent increasesin premiums and tighter underwriting criteriahave been at least partly offset by intense costpressure from claims, particularly from the autoinsurance segment of the industry.

The securities industry has experienced a reduc-tion in operating profits from the peak levelsachieved in 2000 (Chart 45). Revenues fromequity trading fell sharply in the second quarter,a development that likely continued into thethird quarter. Firms that primarily serviced re-tail clients have been particularly hard hit. Incontrast, firms serving institutional clients havefared better, as rising issuance of income trustunits relative to a year earlier (Chart 46) andhigher equity financings in the second quartercontributed to investment banking revenues.Commission revenues have been under down-ward pressure as a result of reduced investoractivity. Another indicator of reduced investoractivity has been the lower demand for margindebt by clients. Nevertheless, profits have beensupported by firms’ efforts to reduce costs,which have included significant staff reductionsin some areas.

23


Investor concern in the current environmenthas also been reflected in declining investmentin mutual funds. The industry has experiencednet redemptions in recent months (Chart 47),and some smaller funds are being shut down orconsolidated.

Payment, Clearing, and SettlementSystems

Systems designed to clear and settle paymentsand other financial obligations in Canada(Box 6) have undergone some key changes.

Recent developmentsIn September 2002, the Continuous Linked Set-tlement Bank (CLS Bank) began commercialoperations. The CLS Bank, based in New York,is an international system for the settlement offoreign exchange transactions. It settles trades inseven currencies, including the Canadian dol-lar. The CLS Bank has been designated underthe Payment Clearing and Settlement Act(PCSA). It is an important initiative that will re-duce risk in the settlement of foreign exchangetransactions.11

The Governor of the Bank of Canada has takenthe decision not to designate the AutomatedClearing Settlement System (ACSS) under thePCSA. The ACSS is primarily a system for small-er-value retail payments. In conjunction withthe decision that the ACSS does not currentlypose systemic risk, effective 2 December 2002,the Bank of Canada reduced the interest ratespread that is applied to the ACSS by 50 basispoints, to 300 basis points. Thus, when theACSS settles each day, the Bank will extend col-lateralized overdraft loans to settle positions inthe ACSS at the target rate plus 150 basis pointsand will pay the target rate of interest less 150basis points on deposits arising from ACSS set-tlement. The target rate is the midpoint of the50-basis-point-wide operating band for theovernight interest rate that is used to implementmonetary policy.

A $25 million cap on paper items eligible forclearing in the ACSS will be implemented by theCanadian Payments Association (CPA) in Feb-ruary 2003 (with a six-month grace period).Currently, it is estimated that about 35 per cent

11. For more information on the CLS Bank, see the arti-cle by Miller and Northcott in this Review (page 41).

24

Chart 44 Insurance Industry Profitability:Return on Equity


-2

0

2

4

6

8

10

12

-2

0

2

4

6

8

10

12

1987 1989 1991 1993 1995 1997 1999 2001

%

Chart 45 Operating Profits: Securities Industry

Source: Investment Dealers Association of Canada

Can$ millions

-200

0

200

400

600

800

1000

1200

-200

0

200

400

600

800

1000

1200

1990 1992 1994 1996 1998 2000 2002

Chart 46 Canadian Income Trust Units

New issuance

Source: Investment Dealers Association of Canada

500

1000

1500

2000

2500

3000

3500

500

1000

1500

2000

2500

3000

3500

2001 2002

Can$ millions


Chart 47 Net New Sales of Mutual Funds

Excluding reinvested distributions

Source: The Investment Funds Institute of Canada

-2

0

2

4

6

8

10

-2

0

2

4

6

8

10

All funds

1998 1999 2000 2001 2002

Total common shares

Can$ billions

Box 6

Payment, Clearing, andSettlement Systems inCanada

An essential component of the financial systemis a robust set of arrangements to clear and settlepayments and other financial obligations. Giventheir central role in settling financial transac-tions, clearing and settlement systems have thepotential to pose systemic risk should they fail tooperate as expected. Under the Payment Clearingand Settlement Act (PCSA), the Governor of theBank of Canada designates systems that have thepotential to pose systemic risk.1 The Bank ofCanada has responsibility for the oversight ofsuch domestic systems and shared oversight withother central banks for international systems thatinvolve the Canadian dollar.

The Bank currently oversees two domestic sys-tems: the Large Value Transfer System (LVTS) forthe exchange of payments, operated by the Cana-dian Payments Association (CPA), and the DebtClearing Service (DCS) for the clearing and set-tlement of securities transactions. The DCS is op-erated by the Canadian Depository for Securities(CDS). The Bank of Canada shares oversight ofthe Continuous Linked Settlement Bank (CLSBank) with other central banks, including theU.S. Federal Reserve, which is the lead overseer.Based in New York, the CLS Bank is an interna-tional system for the settlement of foreign ex-change transactions, involving seven currenciesincluding the Canadian dollar.

Other systems, while not judged to have the po-tential to pose systemic risk, are nevertheless im-portant to the financial system. These includesettlement systems such as the Automated Clear-ing Settlement System (ACSS), which settlesmainly smaller-value retail payments, and theCanadian Derivatives Clearing Corporation(CDCC), which clears and settles exchange-traded interest rate and equity derivative con-tracts in Canada.

The Bank of Canada supplies services to theLVTS, the DCS, the CLS Bank, and the ACSS byproviding settlement assets and liquidity, as wellas collateral and settlement-agent services. TheBank of Canada also provides contingency ar-rangements for these settlement systems.

1. The Minister of Finance must also be of the opinionthat such designation is in the public interest.

of the ACSS’s daily payment flows of about$21 billion are accounted for by payments over$25 million. The CPA decision will encouragethe migration of large-value items to the LVTS,which has payment flows of about $114 billionper day. The LVTS has stronger risk controls andprovides users with intraday finality and cer-tainty of settlement.12

The Canadian Depository for Securities (CDS)is working to implement a new system for thesettlement of debt, equities, and some U.S.-dollar-denominated securities. The new system,called CDSX, is expected to be available in early2003, although equities are not expected to bebrought into the system until some time later.Once the new system is fully operational, theDebt Clearing Service (DCS) and the book-based Securities Settlement Service (SSS), whichcurrently settles equities, will both cease to exist.While the risk controls for the CDSX are basedon those of the DCS, a thorough review andredrafting of system rules is underway.

Operational events in payment, clearing,and settlement systemsPayment, clearing, and settlement (PCS) sys-tems in Canada generally operate withoutmajor disruption or lengthy systems or tele-communications outages. The LVTS has a one-hour Disaster Recovery Plan (DRP) that allows

12. For more information on the ACSS, see the article byNorthcott in this Review (page 29).

25


operations to resume if primary systems fail.The DCS has a two-hour DRP and LVTS parti-cipants are also expected to be able to resumeoperations within two hours of a system outage.Longer outages would be of greater concernbecause they would be more likely to lead tonegative impacts on system participants or tosignificant settlement delays.13

The Bank of Canada monitors problems that oc-cur as indicators of the reliability of these sys-tems. Serious events in PCS systems, if they wereto occur, could affect financial stability becausethese systems underpin much of the financialactivity in the economy.

13. For further discussion of participant outages, see thearticle by McPhail and Senger in this Review (page 45).

26

Report


Reports address specific issues in some depth.The report in this issue of the FinancialSystem Review discusses the decision not to designate the Automated Clearing Settle-ment System as a systemically important system, as well as some of the research contrib-uting to that decision.

Systemic Risk, Designation, and the ACSSCarol Ann Northcott

learing and settlement systems operatevirtually unnoticed in our daily livesand yet are crucial to a well-functioningfinancial system and economy. By cre-

ating tight linkages between financial institu-tions they also provide a way to transmit risk.Therefore, if not well designed, some clearingand settlement systems have the potential topose a serious risk to the financial system.

Because of their importance, the Bank of Cana-da is responsible, under the Payment Clearingand Settlement Act (PCSA), for identifying anddesignating for its oversight those systems thatare seen to have the potential to pose a systemicrisk (Goodlet 1997). Three systems are currentlydesignated under the PCSA, and the Bank ofCanada has recently examined a fourth, the Au-tomated Clearing Settlement System (ACSS), toassess whether it too should be designated.

The first part of this note sets out the frameworkwithin which an assessment of the risks posedby the ACSS took place. This is the basis onwhich the Governor formed the opinion thatthe ACSS does not currently pose systemic riskto the financial system and, hence, that it neednot be designated at this time. The second partdescribes the model used for the analysis of riskexposures in the ACSS. The results of this workwere an important contributing factor to thedecision not to designate the system.

Designation and the PCSA

The Governor of the Bank of Canada may des-ignate under the PCSA those clearing and set-tlement systems that could be operated in sucha way as to pose systemic risk.1 Once a systemhas been designated under the Act, the Bank ofCanada is responsible for its regulatory over-

1. The Minister of Finance must also be of the opinionthat such designation is in the public interest.

C
sight, and the system is expected to meet min-imum standards set out by the Bank to controlsystemic risk. These standards currently incor-porate those advocated by the Bank for Inter-national Settlements (BIS) in its Core Principlesfor Systemically Important Payment Systems (BIS2001).
The PCSA defines systemic risk in the context ofclearing and settlement systems as the risk thatthe inability of one participant to meet its obli-gations to the system could cause

• other participants in the system to be unableto meet their obligations when due;

• financial institutions in other parts of theCanadian financial system to be unable tomeet their obligations when due; or

• the clearing and settlement system’s clearinghouse or the clearing house of another clear-ing and settlement system to be unable tomeet its obligations when due.

Determining whether a clearing and settlementarrangement could have such serious implica-tions for the financial system can be quite diffi-cult. There are, however, certain characteristicsthat would make a system more likely thanothers to pose such a risk.

The larger the payments processed by a sys-tem, the larger the potential exposures forparticipants. Therefore, systems that processmainly large-value (often called wholesale)payments come under particular scrutiny. Thesize of the potential exposures relative to par-ticipants’ capital is also important, since thiswill largely determine whether participantscan absorb the potential exposures. As well,those systems that play a central role in sup-porting transactions in either the financialmarkets or in the economy more broadlycome under particular scrutiny, since a failure

29

Report

within such a system could have broader im-plications for the financial system.2

Three systems are currently designated underthe PCSA. The Large Value Transfer System(LVTS) is an electronic funds-transfer systemused especially for large-value and time-criticalpayments. The average daily value of all paymentsin 2002 (to November) was $114 billion. TheDebt Clearing Service (DCS) clears and settlesdebt securities, and the CLS Bank settles foreignexchange transactions. Both the DCS and theCLS Bank are not only integral to financial mar-kets, but also process very large transactions.The DCS processes approximately $100 billionto $150 billion per day, and the CLS Bank pro-cesses approximately US$400 billion per day.All three designated systems are explicitlylinked such that the smooth functioning of eachis essential to the smooth functioning of theoverall financial system. For example, paymentobligations arising from the settlement of debtsecurities and foreign exchange transactions arecleared and settled through the LVTS. Clearly,all three systems are systemically important inthe Canadian financial system.

Over the past year, the Bank of Canada hasexamined a fourth payments system, theAutomated Clearing Settlement System, todetermine whether or not it too should bedesignated under the PCSA.

The Automated ClearingSettlement System

The ACSS is a multilateral net settlement systemowned and operated by the Canadian PaymentsAssociation (CPA). It is an uncollateralized sys-tem that does not provide its participants withreal-time information or real-time, risk-controltools to manage their exposures each day. Par-ticipants in the ACSS are therefore exposed toboth liquidity and credit risk on a daily basis.

Credit risk is assumed throughout the day be-cause while value is credited to client accounts,final settlement does not occur until the nextday. But this value may not be received, eitherbecause of insufficient funds in the account onwhich the payment item was drawn or, moreimportantly, because a participant in the ACSSdefaults on its obligation to the system.

2. For more details on these considerations, see Bank ofCanada (1997).

30

In the event that a participant does default on itsobligation, other participants (the “survivors”)face a credit exposure on two fronts. First, theyare exposed for the value already deposited inclient accounts, which will now not be receivedfrom the defaulter. The larger the value theyexpected to receive, the larger this exposure.Second, survivors may be required to pay addi-tional value to ensure that the system completessettlement.3

Participants face liquidity risk on a daily basisbecause of uncertainty in determining theirmultilateral net position in the ACSS. This un-certainty is exacerbated in the event of a partici-pant default.

In the ACSS, therefore, it is the survivors that areexposed if a participant defaults, which createsthe potential for systemic risk.

The Designation Decision

In conducting its review of the ACSS, the Bankof Canada examined the ACSS and the broadenvironment in which it operates to determinewhether the ACSS has the potential to pose asystemic risk. The main considerations were therole of the ACSS in financial markets, the size ofexposures taken on by participants in the system,and whether such exposures were manageable.

Prior to the introduction of the LVTS in Febru-ary 1999, the ACSS was the only system to clearand settle interbank payments. At that time, itwould have been considered to have the poten-tial to pose a systemic risk. It processed a veryhigh volume of large-value payments, leadingto substantial exposures for participants. It wasalso crucial to financial markets, since it pro-cessed payment obligations arising from foreignexchange transactions and from the settlementof debt and equity securities. Once the LVTSbegan operations, payment obligations fromforeign exchange transactions and from thesettlement of debt securities in the Debt Clear-ing Service immediately moved to this well-risk-

3. These exposures result from ACSS rules, whichrequire the defaulter to return (unwind) certain pay-ments to other ACSS participants and which stipulatehow any shortfall is allocated among survivors. Therecovery of value from client accounts and from theestate of the defaulting participant will reduce lossesfrom these exposures. These issues are discussed fur-ther in the second part of this article.


Chart 1 Average Daily Value: ACSS

Can$ billions

10

20

30

40

50

60

70

80

10

20

30

40

50

60

70

80

1997 1998 1999 2000 2001 2002

proofed system. This greatly reduced the ACSS’simportance to financial markets.

The migration of payment obligations arisingfrom the settlement of foreign exchange anddebt securities transactions to the LVTS greatlycontributed to the substantial decrease in thevalue of payments processed through the ACSS(see Chart 1). Other wholesale payments havealso migrated, with the result that the ACSS isbecoming increasingly characterized by thehigh volume of small payments that it settles.This has led to a great reduction in the expo-sures taken on by participants. This trend is fur-ther underpinned by the CPA’s active supportfor the continued migration of wholesale pay-ments. Indeed, the CPA recently announced acap on the value of payments that can be pro-cessed through the ACSS. This cap is set at$25 million. It is expected that once this capis put in place in February 2003, the value ofpayments in the ACSS, and hence potentialexposures, will decrease further.

With the decrease in exposures brought on bymigration to the LVTS, it is believed that partic-ipants’ exposures in the ACSS are now small rel-ative to their capital. This judgment was reachedlargely on the basis of research done within theBank, which used a model of the ACSS to esti-mate potential exposures using actual data fromthe system. This work is discussed in the secondpart of this report.

The Governor of the Bank of Canada is of theopinion that the ACSS does not pose a systemicrisk to the Canadian financial system at thistime. This is based on a broad range of consid-erations, including a downward trend in thevalue of payments in the ACSS and thecommensurate decrease in the potential expo-sures taken on by participants in the system.Hence, the ACSS will not be designated at thistime.

Nevertheless, in keeping with its responsibilitiesunder the Payment Clearing and Settlement Act,the Bank of Canada will continue to monitordevelopments that may necessitate a re-evalua-tion of the risk potential of the system. For ex-ample:

• An important risk factor is the value of itemssent through the system. The migration ofvalue to the LVTS is expected to continuewith the CPA initiative to impose a cap onindividual payments processed through the

31

Report

system. Were this initiative not to succeed,however, or if value were to rise in the ACSS,the system might need to be re-assessed.

• New legislation has recently opened accessto the payments system to three new classesof institutions: life insurance companies,money market mutual funds, and securitiesbrokers. The first two are constrained to beindirect clearers. The exposures in the ACSSwill be re-assessed if these new entrantsbecome an important part of the system.

Understanding Risk inthe Automated ClearingSettlement System

One factor considered by the Bank in its desig-nation discussions was the potential size ofexposures that could arise in the ACSS andwhether such exposures could be successfullymanaged by participants. To assist the discus-sion, a model of the ACSS, incorporating itsunique design and risk characteristics, was de-veloped to estimate the size of exposures thatcould occur under various conditions. Theresults contributed to the Governor’s assess-ment that the ACSS does not currently pose asystemic risk.4

The Nature of Risks in the ACSS

The ACSS is a net settlement system that clearsand settles paper-based payments (such ascheques) and certain electronic payments (suchas debit card payments) in Canada. Financialinstitutions that are members of the CPA canparticipate in the ACSS as either direct or indi-rect clearers. Direct clearers have a settlementaccount at the Bank of Canada across which allobligations in the ACSS are settled. Indirectclearers access the system through a direct clearer.

A variety of payments are made every day. Forexample, a tenant paying rent may write acheque drawn on an account at Bank A, whichthe landlord then deposits in a chequing ac-count at Bank B. Bank B will enter the value ofthe cheque drawn on Bank A into an ACSS ter-minal.5 Once all the payment items are entered

4. For more information on this research and furtherresults, see Northcott (2002).

5. Bank B pulls the value from Bank A—the ACSS is a“debit-pull” system.

32

into the system, the ACSS nets the payment obli-gations.6 Each direct clearer in the ACSS willeither owe money to other participants in thesystem or be owed money—its multilateral netposition. Obligations for payments enteredthroughout the day are settled by each directclearer making or receiving a single paymentequal to its multilateral net position. This isaccomplished through a debit or credit to itssettlement account at the Bank of Canada.

Participants in the ACSS are exposed to risk dai-ly. Direct clearers may owe or be owed moneyon any given day and cannot precisely forecastwhich will occur. This uncertainty carries a de-gree of liquidity risk. Through the netting pro-cess, direct clearers extend credit to one anotherthroughout the day. Since the ACSS does nothave real-time information technology, directclearers cannot control to whom they are ex-tending credit during the day, nor how much,and thus are exposed to credit risk.

A default occurs in the ACSS when a participantthat owes money at the end of the day cannotmeet its obligation. In this event, the defaulterreturns (i.e., unwinds) certain payment items tothe other participants (the survivors). As a gen-eral rule, the defaulter returns items that requireit to pay the survivors, while keeping items thatrequire the survivors to pay money to it, the de-faulter. Once this is done, the net positions ofall direct clearers are recalculated. If the default-er continues to owe money after the unwindingprocess, the amount needed to bring its posi-tion to zero (the “shortfall”) is divided on a pro-portional basis among the survivors, based ontheir original dealings that day with the default-er.7 Therefore, a survivor’s final position on theday of default comprises both its revised net po-sition and its share of any shortfall.

6. This amount will include the payments entered intothe ACSS by all the direct clearers for their indirectclearers as well.

7. If all items are returned by the defaulter, it would nec-essarily be owed funds after the unwinding process.However, certain items cannot be returned, eitherbecause they are no longer in the defaulter’s posses-sion, or because it is prohibited in the ACSS rules.Therefore, it is possible that a defaulter could con-tinue to owe funds after the unwinding process.


Measuring Risk Exposures in theACSS

The unwinding of payments and the allocationof a shortfall enable the defaulter, and thus thesystem, to complete settlement. These actionsalso expose survivors to liquidity pressures andcredit losses. One approach to measuring theseexposures is discussed below.

The immediate settlement concern is one of li-quidity—whether survivors can meet their finalpositions on the day of default, thereby allow-ing the ACSS to complete settlement. Moreover,after the unwinding process, participants aresubject to a liquidity “surprise,” since they mayowe more funds or be due less funds than theyhad anticipated. If a survivor owes funds, itmust find the liquidity to cover the (larger-than-expected) obligation. Therefore, its settlement li-quidity exposure is equal to its obligation to thesystem.8 If a survivor is owed funds, and so doesnot pay into the system, its settlement liquidityexposure is zero.

Survivors’ credit exposure appears on two fronts:the value of payment items sent to the defaulterthat are unwound plus their share of any short-fall. Potentially, the survivor has depositedfunds into clients’ accounts for which it will notreceive value from the defaulter because of theunwinding process. From our previous exam-ple, Bank B may have deposited the funds intothe landlord’s account expecting to receive thefunds from Bank A. But, if Bank A defaults andthe cheque is returned to Bank B, Bank B mayrealize a loss if it cannot recover the value fromthe landlord’s account. Credit exposure is re-duced if the survivor can recover value fromclients’ accounts and if it can recover some ofits credit loss from the estate of the defaultinginstitution.

These two exposures, liquidity and credit, con-tribute to the possibility of systemic risk—therisk that the inability of one participant in thepayments system to meet its payment obliga-tion will cause other participants to be unableto meet their obligations. In the context of theACSS, an important factor in assessing systemic

8. Settlement liquidity exposure is defined here as theamount the survivor must cover in order to allow theACSS to settle. It does not consider liquidity pressuresthat the participant may have outside of the ACSSbecause of the default.

risk rests on the size of survivors’ exposuresfollowing an initial default and whether suchexposures can be managed.

Estimating the Potential forSystemic Risk in the ACSS

A model of the ACSS that incorporates its designand risk characteristics was developed to esti-mate the potential for systemic risk (contagion).An initial participant default is simulated in themodel using bilateral payments data from theACSS. Liquidity and credit exposures for the sur-vivors are calculated based on assumptions re-garding the following three factors, which areset at the beginning of the simulation and re-flect a particular state of the world:

• the proportion of payment items received bythe defaulter that are returned to the survi-vors in the unwinding process

• the proportion of value that a survivor canrecover from client accounts, following theunwinding process

• the proportion of a credit loss that survivorscan eventually recover from the estate of thedefaulting institution

To determine whether there is contagion withinthe context of the model, a rule is used to definewhen a survivor “fails” because of the initialdefault. The rule used is that a survivor subse-quently defaults (a “knock-on” default) if itscredit and liquidity exposures resulting from theunwinding process and allocation of a shortfallare both larger than its ability to cover them (seeBox 1). Assumptions regarding two factors per-taining to survivors’ ability to cover exposuresare made at the beginning of the simulation:

• the proportion of a survivor’s Tier 1 capitalthat can be used to cover the credit exposure

• the proportion of a survivor’s liquid assetsthat is available to cover a liquidity exposure

If there is a knock-on default, a further unwind-ing takes place, and the process continues untilall the remaining survivors can cover theirexposures. Systemic risk is measured as thenumber of knock-on defaults experienced for agiven initial default.

This process is carried out using each directclearer that is initially in a net debit position asthe initial defaulter. The simulation can then be

33

Report

Box 1

Defining a ContagionThreshold

To determine whether there is contagionfrom the initial default (i.e., systemic risk),a rule must be set that defines when a sur-vivor “fails” because of the initial partici-pant default.

To define such a threshold, we first defineilliquidity and insolvency in the context ofthe model. A survivor with a liquidity expo-sure must have enough liquid assets avail-able to cover the exposure in order tocomplete settlement in the ACSS. If the sur-vivor’s exposure is greater than its availableliquid assets, it is “illiquid.” A survivor witha credit exposure must be financially soundenough to be able to withstand the creditloss while continuing to function; other-wise it is “insolvent.”

To define contagion, consider four cases. Ifa survivor is liquid and solvent followingthe unwinding process and allocation ofany shortfall, it will fulfill its settlement ob-ligation. If it is illiquid but solvent, it is as-sumed that the central bank will advancethe funds necessary to complete settlement.If an institution is liquid but insolvent, itwill complete settlement, since it will beable to meet its obligation to the system. Ineach of these cases, despite the default ofthe initial participant, the survivor is ableto meet its obligation to the system. If,however, the survivor is illiquid and insol-vent, it will not be able to meet its obliga-tion to the system. Therefore, a knock-ondefault (contagion) occurs if, following adefault and unwinding of payments, a sur-vivor is illiquid and insolvent.

performed changing the five assumptions toreflect different states of the world (see Box 2).

Data and Results

For each of the 231 days in the data set (August2000 to June 2001 inclusive), the bilateral valueof items sent between each of the 12 directclearers in the ACSS is used to determine creditand liquidity exposures. Over this period, theaverage daily value of items sent through theACSS was $20.6 billion.

To determine a participant’s ability to cover aliquidity exposure, a portfolio of liquid assets isconstructed from monthly or quarterly balancesheet data. To determine its ability to cover acredit exposure, Tier 1 capital (as reported to theOffice of the Superintendant of Financial Insti-tutions) is used.

Result 1: There is very limitedpotential for systemic risk in theACSS.

When the model is run under assumptions thatrepresent a “normal” state of the world, we findno evidence of contagion over the period stud-ied. That is, for each participant default simulat-ed, all other participants are able to handle theirexposures without subsequently defaulting (asdefined in the model). More importantly, this isthe case even when individual assumptions areadjusted to reflect a much more risky environ-ment; for example, assuming there is no recov-ery from the estate of the defaulting institution.

Assumptions are then made in the model thatconsider an extraordinarily unlikely environ-ment—an “extreme” state. Even here, on aver-age, not even one knock-on default occurs.There is, however, the potential for contagion ifcertain direct clearers default on certain days,depending on the configuration of exposures.In some highly unusual cases, the initial defaultcauses all direct clearers (excluding the centralbank) to subsequently default.

Result 2: A uniform decrease inthe value of payments sent by allparticipants leads to a decreasedrisk of contagion.

As discussed earlier, the less payment value pro-cessed through a system, the lower the potentialexposures and the lower the risk of contagion in

34


Box 2

State-of-the-WorldAssumptions

Under the “normal state,” it is assumedthat survivors use 100 per cent of theirTier 1 capital to cover a credit exposure andthat 50 per cent of their liquid assets areavailable to cover a liquidity exposure. Thedefaulter returns payment items represent-ing 50 per cent of the value of the itemsit received, and survivors can recover 50 percent of the returned items from clientaccounts. In addition, survivors recover anet present value of 75 per cent of theircredit loss from the estate of the defaultinginstitution. From this normal state, oneassumption at a time is altered to considercontagion under a variety of conditions,from benign to risky.

The “extreme state” is a worst-case world. Itis assumed that survivors can use only 10 percent of their current Tier 1 capital levels tocover a credit exposure and only 10 percent of their liquid assets are available tocover a liquidity exposure. It is assumedthat the defaulter returns all the paymentitems it originally received, but survivorscannot recover any of this value from clientaccounts. Finally, survivors cannot recoverany loss from the estate of the defaultinginstitution. Each of these assumptions isconsidered extremely unlikely, let alonetheir occurring simultaneously.

See Northcott (2002) for more details onthese assumptions.

the event of a failure. A trivial case, of course, iswhere the value goes to zero: a system cannotpose risk if it is not used. What is interesting,however, is how quickly risk is reduced as thevalue sent through the system falls. Once again,using parameters reflecting an “extreme” state, a25 per cent decrease in the value of items sentthrough the system leads to a dramatic decreasein the maximum number of knock-ons that oc-curs from 10, the largest number possible, to 2.9

The overall results are highly encouraging, sinceit appears to require a confluence of extraordi-nary conditions for the ACSS to give rise to con-tagion. Given this, and given that no contagionis observed under a range of normal conditionsfrom benign to very risky, the research supportsthe view that the exposures in the ACSS at thistime are manageable by participants. This posi-tion is strengthened by initiatives to furthermigrate value to the LVTS.

References

Bank of Canada. 1997. Guideline Related to Bankof Canada Oversight Activities under the Pay-ment Clearing and Settlement Act. Availableat www.bankofcanada.ca/en/guide97.htm.

Bank for International Settlements (BIS). 2001.Core Principles for Systemically ImportantPayment Systems. Committee on Paymentand Settlement Systems Report No. 43.Basel: BIS, January.

Goodlet, C. 1997. “Clearing and SettlementSystems and the Bank of Canada.” Bank ofCanada Review (Autumn): 49–64.

Northcott, C.A. 2002. “Estimating SettlementRisk and the Potential for Contagion inCanada’s Automated Clearing SettlementSystem.” Bank of Canada Working PaperNo. 2002-41.

9. Since no knock-on defaults are produced under arange of normal conditions, an extreme scenario isused to demonstrate the effect of decreasing value inthe system.

35

Policy and

Infrastructure

Developments


Introduction

he financial system and all of its variouscomponents (institutions, markets, andclearing and settlement systems) are sup-ported by a set of arrangements, including

government policies, that influence its structure andfacilitate its operation. Taken together, thesearrangements form the financial system’s infra-structure. Experience has demonstrated that a keydeterminant of a robust financial system is theextent to which it is underpinned by a solid, well-developed infrastructure. This section of the Reviewhighlights work in this area, including that relatedto relevant policy developments.

For this issue, the articles in this section focuson the payment, clearing, and settlement sys-tems used by Canadian financial institutions.These systems are at the core of the financial sys-tem, providing the links through which majorparticipants can transfer financial instrumentsand make payments between themselves safelyand reliably.1

The CLS Bank: Managing Risk in Foreign ExchangeSettlements describes a major international in-itiative to create a private sector institution thatoffers simultaneous settlement for both curren-cy legs of foreign exchange transactions, therebyreducing the risk of non-payment. The CLSBank, which became active in September, is in-itially operating in seven currencies, includingthe Canadian dollar. The Bank of Canada sup-ports the CLS Bank by providing it with bankingservices that facilitate its interaction with Cana-dian financial institutions. The Bank of Canadaalso contributes to the oversight of the CLSBank, with particular responsibility for itsCanadian-dollar operations.

1. Information on clearing and settlement systems inCanada can be found on the Bank of Canada’s Website at http://www.bankofcanada.ca/en/payments/mainpage.

T
The Large Value Transfer System (LVTS), operat-ed by the Canadian Payments Association, isone of Canada’s most important paymentssystems, processing an average total value of$114 billion each business day. Although theLVTS is an extremely reliable system, on rareoccasions a significant outage in the computeror telecommunications systems of a participantcould disrupt the flow of payments in the LVTSand adversely affect liquidity in the paymentssystem. The Impact of Participant Outages onCanada’s Large Value Transfer System examineshow participant outages can affect paymentflows and offers suggestions to help minimizetheir impact.
One factor hampering the study of the opera-tion of payments systems has been a lack ofrobust “benchmarks” against which to assesspayment flows through them. In UnderstandingIntraday Payment Flows in the Large Value TransferSystem, the early stages of building a set ofbenchmarks for payment flows within the LVTSare discussed. These preliminary benchmarksare compared with actual payment flows on11 September 2001, providing some insightinto the ability of the LVTS to operate underdifficult circumstances.

39


The CLS Bank: Managing Risk in ForeignExchange SettlementsPaul Miller and Carol Ann Northcott

he foreign exchange market is the largestfinancial market in the world, with anaverage daily turnover of approximatelyUS$1.2 trillion (BIS 2002). Partici-

pants in this market take on significant risks insettling their transactions. Indeed, these risksare so significant that exposures created by dis-ruptions in settling these transactions have thepotential to pose systemic risk.1

The CLS Bank International was created to ad-dress foreign exchange settlement risk, particu-larly its most significant component, credit risk.It does this by providing a form of payment-versus-payment settlement for foreign exchangetransactions, virtually eliminating credit risk forcounterparties settling through the system.

The CLS Bank began operations on 9 September2002. It is a significant contribution to the globalfinancial system generally and to the Canadianfinancial environment specifically, since theCanadian dollar is one of seven currencies thatcan be settled through the system.2

Foreign Exchange SettlementRisk

Foreign exchange traders engage in variouskinds of transactions that involve exchangingone currency for another. But once a deal hasbeen struck, how does the actual exchange takeplace? To understand how a typical foreign ex-change transaction is settled (without the CLSBank), consider the following example involv-ing two banks. Bank A is based in Japan and is a

1. Systemic risk in this context is often defined as therisk that the failure of one participant in a financialsystem to meet its required obligations will causeother financial institutions to be unable to meet theirobligations when due.

2. For more information on the topics discussed here,see Miller and Northcott (2002).

T
participant in the Japanese large-value pay-ments system, BOJ-NET. Bank B is based inCanada and is a participant in the Canadianlarge-value payments system, the LVTS. Bank Aand Bank B enter into a foreign exchange trans-action when Bank A sells yen to Bank B forCanadian dollars. How is this transactionsettled?
Bank A will pay Bank B the yen through theBOJ-NET. Since Bank B is not a participant inthe BOJ-NET, it must engage a bank that is aparticipant to receive the payment on its behalf.This is Bank B’s correspondent, or “nostro,”bank. Likewise, Bank B will pay Canadian dol-lars to Bank A through the LVTS via Bank A’snostro bank.

Foreign exchange trades are two-way transac-tions: each counterparty pays one currency andreceives another in return. One source of risk forcounterparties arises when payments systemsare in different time zones. In the above exam-ple, Bank A pays out the yen through the BOJ-NET before the Canadian payments system isopen. If Bank B defaults in the interim, Bank Awill have paid out the yen but will not receivethe Canadian dollars. This is often termed“principal risk,” a type of credit risk. As well, be-cause of limitations on current information-management practices, it could be several daysfrom the time a counterparty initiates the pro-cess to pay the “sold” currency until it knowswith certainty whether it has received the“bought” currency, subjecting it to liquidity riskand replacement risk if the bought currency ar-rives later than expected. Finally, given thatcountries have different legal and regulatory re-gimes, legal risk may also be a factor in the eventa counterparty fails to deliver a currency. Allrisks that arise in the settlement of foreign ex-change transactions comprise foreign exchangesettlement risk, with credit risk being the mostsignificant component.

41

Policy and Infrastructure Developments

Types of Risk

Banker risk The risk that the bank where asettlement account is held couldbecome insolvent.

Credit risk The risk that a counterparty willnot settle an obligation for fullvalue, either when due or at anytime thereafter. This includes prin-cipal risk, the risk that a counter-party could pay the currency soldwithout receiving the currencybought (BIS 2001).

Legal risk The risk of loss because of theunexpected application of a law orregulation, or because a contractcannot be enforced (BIS 2001).

Liquidity risk The risk that a counterparty willnot settle an obligation for fullvalue when due but will settle atsome unspecified time thereafter(BIS 2001).

Operationalrisk

The risk that deficiencies in infor-mation systems or in internalcontrols, human errors, or man-agement failures will cause orexacerbate credit or liquidity risks(BIS 2001).

Replacementrisk

The risk that a counterparty to anoutstanding transaction will fail toperform on the settlement date.The resulting exposure is the costof replacing, at current marketprices, the original transaction(BIS 1996).

Systemic risk The risk that the failure of one par-ticipant in a financial system tomeet its required obligations willcause other financial institutionsto be unable to meet their obliga-tions when due (BIS 2001).

The CLS Bank

Based in New York City, the CLS Bank is de-signed specifically for the settlement of foreignexchange transactions. Seven currencies can cur-rently be settled through the system: the Austra-lian, Canadian, and U.S. dollars, the euro, theyen, the Swiss franc, and the pound sterling.3

The CLS Bank virtually eliminates the credit riskassociated with settling foreign exchange trans-actions. It does this by providing a payment-versus-payment arrangement, settling bothsides of a transaction simultaneously acrossaccounts that financial institutions (settlementmembers) hold at the CLS Bank.4 So, if thetransaction from our previous example is settledin the CLS Bank, Bank A and Bank B receivetheir expected currencies simultaneously intheir respective settlement accounts at the CLSBank. Counterparties do not give up the soldcurrency without receiving something in return.

Settlement members pay currencies that areowed to the CLS Bank’s accounts, which areheld at central banks, through domestic pay-ments systems. Currencies that are due to settle-ment members are paid out by the CLS Bank inthe same way.

Risk Management in theCLS Bank

The simultaneous settlement of foreign ex-change transactions across the books of the CLSBank means that the settlement asset for foreignexchange transactions is an intraday claim onthe CLS Bank. For this to be acceptable to partic-ipants and to the central bank community, theCLS Bank must be virtually risk-free. Therefore,risk-management controls are applied to eachtrade before it is settled to protect the CLS Bankfrom credit and liquidity risk. First and fore-most, although each settlement member willowe some currencies and be owed other curren-cies over the course of settlement, the balancein each member’s settlement account at theCLS Bank over all currencies must always bepositive. There are also limits on how much a

3. More currencies are expected to be added in thefuture.

4. Financial institutions can participate in the CLS Bankin various ways, but only settlement members holdsettlement accounts at the CLS Bank.

42


settlement member can owe in aggregate acrossall currencies, and how much it can owe in aparticular currency.

To protect itself from legal risk, the CLS Bankhas obtained legal opinions that the finality oftransactions settling across its books can be sup-ported in the legal systems of all jurisdictionswith currencies settling through the system. Aswell, all payments to the CLS Bank from settle-ment members are made through payments sys-tems that provide intraday finality.5 The CLSBank holds these payments in accounts at cen-tral banks, ensuring that the CLS Bank is pro-tected from banker risk. Finally, the CLS Bankhas an explicit plan to address operational risk.

For participants in the CLS Bank, the risk-management controls and other arrangementsensure that, in virtually all circumstances,participants will receive either the currencytransacted for or a refund of the amount theycontributed, even if another participant defaultson its payment obligations. That is, participantsare protected from credit risk arising from thefailure of another participant.6 Nevertheless, inthe event of a failure, participants do continueto be potentially exposed to liquidity risk andreplacement risk, although it is expected thatthese risks are manageable.

The CLS Bank and theCanadian Financial System

The CLS settlement cycle takes place during theNorth American overnight period, normallyfrom 1 a.m. until 6 a.m. ET. The approved pay-ments system for the Canadian dollar is theLarge Value Transfer System (LVTS), and theDebt Clearing Service (DCS) will continue to beused to support LVTS collateral operations. Cur-rently, only one Canadian bank is a settlementmember, the Royal Bank of Canada, with someothers intending to enter the system as settle-ment members in the future.

The Bank of Canada plays three key roles withrespect to the CLS Bank in the Canadian finan-cial system.

5. Intraday finality indicates that once a payment hasbeen accepted within a payments system, the receiverhas irrevocable access to the funds that same day.

6. Only under the most extreme conditions does someelement of credit risk remain. See Miller and North-cott (2002).

• To mitigate major disruptions caused by theoperational failure of a Canadian settlementmember, a nostro agent, or the LVTS, theBank of Canada is prepared to assist, if nec-essary, by entering payments directly acrossthe CLS Bank’s and participants’ settlementaccounts with the Bank of Canada.

• The Bank of Canada acts as banker for theCLS Bank, providing it with two main ser-vices. First, the Bank of Canada provides asettlement account to the CLS Bank. Second,the Bank of Canada makes and receives pay-ments through the LVTS on behalf of theCLS Bank.

• The CLS Bank is subject to regulation by theBoard of Governors of the Federal ReserveSystem in the United States. Supported bythe Federal Reserve Bank of New York, theBoard is therefore the lead overseer of thesystem and consults with the central banksof those countries whose currencies will set-tle in the CLS Bank, including the Bank ofCanada. In addition, the Governor of theBank of Canada has designated the Cana-dian-dollar operations of the CLS Bank forBank of Canada oversight under the Pay-ment Clearing and Settlement Act. The Bankof Canada is satisfied that the system meetsthe standards that the Bank has set for desig-nated systems.

Conclusion

Through the co-operative efforts of private sec-tor financial institutions, central banks, and theoperators of domestic payments systems, theCLS Bank has been created to address foreignexchange settlement risk—particularly creditrisk, which use of the CLS virtually eliminates.The world’s largest foreign-exchange-dealing in-stitutions are shareholders of the CLS Bank, andit is expected that most will interact directly orindirectly with it. Growing participation has thepotential to position the CLS Bank as the dom-inant global mechanism for settling foreignexchange transactions.

43


References

Bank for International Settlements (BIS). 1996.Settlement Risk in Foreign Exchange Transac-tions. Report prepared by the Committeeon Payment and Settlement Systems ofthe Central Banks of the Group of TenCountries. Basel: BIS, March.

———. 2001. Core Principles for SystemicallyImportant Payment Systems. Committee onPayment and Settlement Systems ReportNo. 43. Basel: BIS, January.

———. 2002. Central Bank Survey of ForeignExchange and Derivatives Market Activity in2001. Basel: BIS, March.

Miller, P. and C. A. Northcott. 2002. “CLS Bank:Managing Foreign Exchange SettlementRisk.” Bank of Canada Review (Autumn):13–25.

44


The Impact of Participant Outages onCanada’s Large Value Transfer SystemKim McPhail and David Senger

ach business day, about 15,000 paymentmessages, with a total value averaging$114 billion, flow through Canada’sLarge Value Transfer System (LVTS). The

Bank of Canada and 13 deposit-taking financialinstitutions participate directly in the LVTS.1 Itis owned and operated by the Canadian Pay-ments Association (CPA).

The LVTS functions smoothly because, on mostdays and for most participants, inflows and out-flows tend to be roughly offsetting. This, togeth-er with the legally enforceable netting ofpayments, as well as intraday borrowing backedby collateral held at the Bank of Canada, reduc-es participants’ intraday liquidity requirements.If an LVTS participant was unable to send pay-ments to other participants because of an out-age of its own internal systems, the paymentflows of other participants and of the LVTS as awhole might be disrupted and could be madeonly at greater expense (because of increasedcollateral requirements).

Lengthy outages in the computer systems ortelecommunications systems of LVTS partici-pants are infrequent. Between June and August2002, seven outages occurred. Four were re-solved fairly quickly. However, one lasted oneand a half hours, and two lasted for just overtwo hours. The potential effects of disruptions tothe flow of payments in the LVTS increase withthe length of an outage. It is therefore importantthat participants have reliable backup systemsthat they can switch to quickly if primary

1. For more information on Canadian payments sys-tems and the structure of the LVTS, see the Bank ofCanada’s Web site at http://www.bankofcanada.ca/en/payments/mainpage. The LVTS is a multilateralnetting system. Payments made during the daythrough the LVTS are final and irrevocable. The risk-control mechanisms in the LVTS ensure that it will beable to complete settlement in all circumstances atthe end of each day.

E
systems fail. It is also important that proceduresare in place to deal with participant outages inorder to limit their impact on the paymentssystem as a whole.
In this article, a simple illustrative model is usedto describe how a participant outage affects thepayment flows of other participants and of theLVTS as a whole. The model is then used to pro-vide an indication of the effect of an actualparticipant outage on the LVTS. The proceduresthat are currently in place to deal with the po-tential problems raised by participant outagesare also described. When an outage occurs, it isimportant that these procedures be implementedquickly.

How Does a Liquidity DrainOccur?

Consider an outage that prevents a participantfrom sending payment instructions to the LVTS.Payments sent to that participant by other par-ticipants will continue to pass through the LVTSuntil those participants take specific action todelay payments or until sending additional pay-ments would violate the LVTS’s risk controls.These payments will be recorded as a “credit” tothe position of the problem participant. If thisposition becomes sufficiently large, substantialliquidity could be drained from the system.

The LVTS has two separate payments streams. Inthe first stream (called Tranche 1 or T1), thesender, in effect, fully collateralizes each pay-ment sent through the system. In this article, wefocus on the second stream (Tranche 2 or T2),because it accounts for about 90 per cent of thevalue of payments sent through the LVTS andbecause it is the stream for which the issue ofliquidity drains is most relevant. To supportpayment flows, T2 relies on intraday creditextended between participants, a collateralpool, and robust risk controls rather than on

45


Table 1

Risk Controls and the MultilateralNetting Mechanism in T2 of the LVTS:An Example

BCL granted to:Sum

A B C D E

BCLgrantedby:

A x 30 50 60 70 210

B 25 x 60 50 70 205

C 45 60 x 300 300 705

D 60 75 250 x 500 885

E 65 60 250 500 x 875

Sum ofBCLs 195 225 610 910 940

X

Systempara-meter

0.24 0.24 0.24 0.24 0.24

=

T2NDC 47 54 146 218 226

full collateralization by the sender of a pay-ment. Payments sent via T2 are as protectedfrom risk as those sent by T1. When sufficientcredit is available to them, LVTS participantsgenerally choose to send payments via T2because the collateral requirements are lower.

To contain the risk in the T2 stream, each pay-ment sent via T2 during the day must pass cer-tain risk controls. A hypothetical example,outlined in Table 1, uses five financial institu-tions to demonstrate how the risk controls andmultilateral netting mechanism in T2 function.Two types of risk controls (explained below) areapplied to each payment sent through T2: bilat-eral credit limits (BCL) and T2 multilateral netdebit caps (T2NDC).

Each participant can grant each other partici-pant a BCL. The BCL granted by one participantto a second participant represents the maxi-mum net debit (or negative) position that thesecond participant is allowed to incur with re-spect to the first. This BCL can also be viewed asthe maximum positive balance that the first par-ticipant will allow with respect to the second.For example, in Table 1, A has granted a BCL of30 to B and a BCL of 50 to C. Thus, B’s bilateralnet debit position with respect to A cannotexceed 30 and C’s negative balance with respectto A cannot exceed 50.

The first step in calculating a participant’sT2NDC is to add the BCLs granted to that par-ticipant by all other participants (e.g., for A, thisis equal to 25 + 45 + 60 + 65 = 195). This sumis then multiplied by a “system parameter” tocalculate each participant’s T2NDC. (In Table 1,this is 0.24, the system parameter currently usedin the LVTS.) The T2NDC represents the maxi-mum allowable T2 negative position that re-sults from one participant’s flow of payments toand from all other participants. In the case of A,for example, the T2NDC is 47.

Because of the offsetting nature of payments ina multilateral netting system, a relatively smallT2NDC (i.e., much smaller than the sum of theBCLs) can support a large number of payments.The greater the power of multilateral netting is,the more the sum of the BCLs can be scaleddown by the system parameter without impair-ing the smooth flow of payments through theLVTS. The CPA has chosen a small system pa-rameter (which results in smaller T2NDCs) thatstill allows payments to flow smoothly, because

46


this reduces the collateral requirements of LVTSparticipants.

Suppose that, at the beginning of the day, par-ticipant A (a small financial institution thatgrants and receives relatively small BCLs) is un-able to send payment messages because of atechnical outage, but continues to receive pay-ments from other participants. In this example,B can send a maximum of 30 (the BCL) to A,C can send a maximum of 50, and so on. Thus,participant A can drain 210 in liquidity fromother participants—i.e., the sum of BCLs grant-ed by A. Participants B, C, D, and E, however,each retain the ability to send payments to eachother (e.g., given that B has sent 30 to A andsince B’s T2NDC is 54, B can still send up to24 to C, D, and E). The outage at participant Adrains liquidity from other participants, butthey retain the ability to send and receive T2funds.

Now, suppose participant E (a large financial in-stitution that grants and receives relatively largeBCLs) has an outage. The BCL that E has grantedto A is 65; however, A’s ability to send 65 to E isconstrained because its T2NDC is smaller thanthe BCL. Participant A can send a maximum of47, its T2NDC, to E. The same situation appliesfor B, C, and D. In this worst-case scenario, Ehas drained all T2 liquidity from other partici-pants because their T2NDC prevents them frommaking any payment to any other participant.

The Potential Impact on theLVTS of Participant Outages

Both large and small financial institutions par-ticipate in the LVTS. If a small LVTS participantexperiences an outage, and other participantscontinue to send payments to the problem par-ticipant until their BCL or T2NDC is reached(i.e., a worst-case scenario), that participantcould drain about 15 per cent of the T2 liquidityof other participants. An outage at one of thelarge participants in the LVTS, however, couldtheoretically drain about 85 per cent of T2liquidity from the system.

In practice, this worst-case scenario is unlikelyto ever occur because other participants wouldeventually stop sending payments to the prob-lem participant. Nevertheless, if there was anoutage when a large participant had alreadybuilt up a large positive balance in the LVTS, asubstantial liquidity problem would result,

because that participant would be unable to re-cycle liquidity back to other participants. If thatparticipant continued to receive LVTS fundswithout being able to send LVTS payments for aconsiderable length of time, it would continueto drain liquidity. In actual practice, an outagelasting several hours at a large LVTS participantmight quickly drain on the order of 30 to40 per cent of the total T2 liquidity that existsin the LVTS. Other participants would still beable to divert payments from the T2 stream toT1, but this is much more expensive because itrequires more collateral.

How Does the CPA Limit theConsequences of ParticipantOutages?

The LVTS has several mechanisms in place toaddress this issue. They are designed to makethe consequences of a participant outagemuch less severe than the worst-case scenariosdescribed above.

First of all, there is an expectation among LVTSparticipants that participants should be able toresume payment operations within two hoursof a technical failure, although this is not cur-rently incorporated into the LVTS rules. Thisshould limit the length of time during which aparticipant with a problem could drain fundsfrom other participants. The Bank of Canadahas noticed a tendency among LVTS partici-pants with outages to prefer to try to restoretheir primary systems, rather than switching tobackup systems, since they hope that the pri-mary-system outage can be resolved within twohours. However, if primary systems cannot berestored fairly quickly, an outage could persistfor several hours before a decision is taken totransfer operations to backup systems. Addi-tionally, once this decision is made, it can takeup to two hours to begin operations at backupfacilities. Thus, a stronger incentive for partici-pants to resume processing within two hours,perhaps by incorporating this requirementwithin LVTS rules, might be beneficial.

Equally important, under the CPA rules, anLVTS participant with a technical outage isrequired to notify the system operator immedi-ately. The system operator then notifies otherparticipants, so that they can choose to tempo-rarily stop sending payments to the affected

47


participant until the problem is resolved. By do-ing this, other participants can monitor and pre-serve their liquidity.

As noted above, lengthy participant outages areinfrequent, but they do sometimes occur and,on rare occasions, it may be difficult to resolvethe problem in a reasonable length of time. Useof reliable backup processing capabilities thatcan restore payments processing within a maxi-mum of two hours is important. Moreover,tighter domestic and international require-ments regarding time-sensitive payments areshortening the acceptable duration of partici-pant outages.2 When a participant outage doesoccur, it is important that the participant followthe CPA rules and notify the CPA promptly inorder to prevent the buildup of liquidity at thefailed participant and a corresponding drain ofliquidity from other participants. This willminimize the impact of such outages on thepayments system as a whole.

2. See “The CLS Bank: Managing Risk in ForeignExchange Settlements,” on page 41.

48


Understanding Intraday Payment Flows inthe Large Value Transfer SystemLindsay Cheung

he Large Value Transfer System (LVTS) isthe key mechanism in Canada for settlinglarge-value and time-sensitive payments,such as those involved in settling foreign

exchange transactions, since it is the only elec-tronic transfer system in Canada that processespayments in real time and with intraday finalityand irrevocability. Major disruptions affectingthis system could therefore have potentially se-vere ramifications for the financial system. Anunderstanding of the normal patterns of intra-day payment flows in the LVTS will enable us toquickly assess and monitor the impact of anintraday disruption to the system. This articlepresents a preliminary benchmark for theseintraday flows using data provided by theCanadian Payments Association (CPA).1

Although the benchmark is still preliminary,since it is derived from a very limited amountof data, it is used to assess the impact of theevents of 11 September 2001 on the Canadianpayments system.

Data

Two weeks’ worth of hourly aggregated paymentvolumes and values sent between 28 Januaryand 1 February and between 11 and 15 February2002 were used to derive the benchmark. Statis-tical analysis of total data flows shows that vol-ume increases on the first two and the last fivebusiness days of each month, as well as at mid-month and on Fridays. Value increases on thefirst two and the last three business days ofeach month and at mid-month, but falls on

1. We would like to thank the Canadian Payments Asso-ciation for providing the data and for agreeing to itsuse in this article, as well as for their comments.

T
Tuesdays.2 The intraday payment flows were ad-justed by scaling them to remove these system-atic effects on aggregate daily volume and value.This method assumes that the intraday patternis not altered by either the business day or theday of the week.
Intraday Pattern

The LVTS allows participating financial institu-tions to exchange payments, either for them-selves or for their clients, between 8 a.m. and6 p.m. every business day, and it begins settle-ment at 6:30 p.m.3 Some LVTS payments aretime sensitive because of the time-critical natureof client payments, deadlines associated withthe settlement of other systems, or because ofpayment flows involving the Government ofCanada. More important to the overall intradayLVTS flows, the CPA’s guideline on the Timing ofPayment Messages states that each participant,excluding the Bank of Canada, should completea certain percentage of its daily payment flowsaccording to the following schedule:

2. Both daily volume and value also drop on all U.S.national holidays but rise on the business day imme-diately thereafter. In addition, the levels of volume andvalue fall on every first Monday in August, since it is aholiday in all provinces except Quebec. Thus, whilethe LVTS is open on that day, there are significantlyfewer payments. Volume and value increase on dayswhen the Government of Canada pays interest onmany of its bonds. These fall on the first business dayin June, September, and December. These paymentsgenerate increased activity in the LVTS.

3. To support the overnight operation of the Continu-ous Linked Settlement (CLS) Bank, the LVTS is nowopen at 1 a.m. every business day for payment pro-cessing. In particular, the period between 1 a.m. and8 a.m. is reserved for payments related to the CLSBank. The impact of transactions involving the CLSBank on the intraday payment flows of the LVTS stillneeds to be assessed.

49


To reduce the need for borrowing from, or toavoid holding deposits at, the Bank of Canadaovernight, participants may exchange paymentswith each other between 6 p.m. and 6:30 p.m.in order to to even out or “flatten” their surplusor deficit positions.4 This is called the presettle-ment period.

Volume

Data on hourly payments volumes show a sta-ble intraday pattern from one day to another(Chart 1a), with standard deviations that varybetween 10 to 20 per cent during various hoursof the day. The highest volume occurs duringthe first hour of operation and averages about30 per cent of total daily volume. This is becauseparticipants enter many previously “known”payments in their internal systems overnight,which are then automatically transmitted to theLVTS for processing when it opens at 8 a.m.

The volume falls sharply between 9 a.m. and10 a.m., remains flat between 10 a.m. and3 p.m, and then increases slightly between3 p.m. and 4 p.m. This rise in volume is associ-ated with the completion of most clientpayments before 4 p.m. As suggested by theguideline, about 60 per cent of total dailypayment volume is typically completed before1 p.m. Hourly volume declines slightly between4 p.m. and 5 p.m. to about 1,000 payments andto about 300 payments for the following hour(as participants complete any remainingtransactions, such as Settlement Exchange

Hours of operation Volume(per cent)

Value(per cent)

Before 10 a.m. local time 40 25

Before 1 p.m. local time 60 60

Before 4:30 p.m. EasternTime 80 80

4. The spread between the rates that the Bank of Canadacharges for lending and pays for overnight depositsforms the “operating band” for the overnight rate ofinterest. The Bank conducts its monetary policy bysetting a target for the overnight rate that is at the cen-tre of the band.

50

Transactions).5 The day typically ends with5 to 7 payments during the presettlement period,when participants exchange a small number ofpayments to flatten their positions. Overall, thehourly payment volume follows the CPAguideline (indicated by the horizontal lines inChart 1b).

Value

Intraday payment values exhibit a more volatilepattern (Chart 2a), with standard deviationsvarying between 20 to 30 per cent. Although thehighest hourly volume occurs during the firsthour of operation, value does not peak at thistime. On average, 20 per cent of the total dailypayment value is completed before 10 a.m.,slightly less than the 25 per cent contained inthe CPA guideline (indicated by the horizontalline in Chart 2b).

Hourly payment value tends to increase slightlybetween noon and 1 p.m. This is partially due tothe settlement of the federal government Re-ceiver General (RG) morning auction and therelease of overnight deposits in the AutomatedClearing Settlement System (ACSS). By 1 p.m.,about 50 per cent of the daily payment valuehas been processed; the CPA guideline is 60 percent.6 The largest spike in hourly value emergesbetween 4 p.m. and 5 p.m., when participantssettle the Debt Clearing System (DCS)7 and the

5. Settlement Exchange Transactions are transactionsbetween direct clearers in the Automated ClearingSettlement System (ACSS) and direct participants inthe LVTS. They are used to correct the dislocation ofpayment flows between the two systems. In short, aparticipant who is long in the LVTS and short in theACSS would swap with another participant who isshort in the LVTS and long in the ACSS.

6. One possible explanation for why participants arenot meeting the guideline could be the greater con-centration of larger-value payments towards the endof the day compared with the level at the time theguidelines were originally established. This concen-tration includes payments for the settlement of DCSand government-related items. The CPA plans torevisit the guidelines.

7. The DCS is a real-time trading system for Govern-ment of Canada and most provincial governmentbonds and bills, as well as for money market instru-ments and corporate bonds. This system is owned bythe Canadian Depository for Securities, which usesthe Bank of Canada as its settlement agent. The DCSsettles via the LVTS between 4 p.m. and 5 p.m. everybusiness day.


Chart 1a Hourly Payments Volume Chart 1b Cumulative Payments Volume

%

0

1000

2000

3000

4000

5000

0

1000

2000

3000

4000

5000

9:0 13:0 17:0

0

20

40

60

80

100

0

20

40

60

80

100

Average percentage oftotal daily volumeactually completed

CPAguideline

9:00 11:00 13:00 15:00 17:00 19:00

Chart 2a

Can$ bill

0

2

4

6

8

10

12

14

16

18

9:0

Chart 3a

0

1000

2000

3000

4000

5000

9:0

Note: Ea

Numberof payments

Numberof payme

0 11:00

28 Jan29 Jan30 Jan31 Jan

Hourly Pay

ions

Impact of 1Hourly Pay

11 SepUpperLower

0 11:00 13:0

ch data point

nts

0 15:00

1 Feb11 Feb12 Feb

ments Value

:

1 September oments Volum

t 01 levellevel

0 15:00 17:0

(in Charts 1

0 19:00

13 Feb14 Feb15 Feb

Chart 2b Cumulative Payments Value

%

0

20

40

60

80

100

0

20

40

60

80

100

Average percentageof total daily valueactually completed

CPAguideline

9:00 11:00 13:00 15:00 17:00 19:000

2

4

6

8

10

12

14

16

18

28 Jan29 Jan30 Jan

0 11:00
31 Jan1 Feb11 Feb
13:00 15
12 Feb13 Feb14 Feb
00 17:00

ne

0

a, 2

15 Feb

19:00

51

Chart 3b Impact of 11 September onHourly Payments Value

Can$ billions

0

1000

2000

3000

4000

5000

19:00

11-Sept 01Upper boundBenchmark meanLower bound

9:00 11:00 13:00 15:00 17:00 19:000

2

4

6

8

10

12

14

16

18

20

22

0

2

4

6

8

10

12

14

16

18

20

22

11 Sept 01Upper levelLower level

a, 3a, and 3b) represents total activity over the past hour.


RG afternoon auction. By this time of the day,the LVTS has already processed about 95 percent of total daily value, exceeding the CPA tar-get of 80 per cent. Hourly payment value de-clines sharply between 5 p.m. and 6 p.m. andcontinues to decline during the presettlementperiod, as participants exchange only a fewpayments.

Average Value per Payment

Average value per payment is lowest between8 a.m. and 9 a.m. It increases for the next threehours, peaks at noon at $10 million, andreturns to the $7 million to $8 million levelbetween 1 p.m. and 4 p.m. Average value perpayment rises substantially between 4 p.m. and6 p.m. and spikes significantly during the pre-settlement period. This spike occurs becauseparticipants are evening out positions bymaking only a few, but possibly very large,payments.

Assessing the Impact of11 September 2001

In this section, the intraday benchmark is usedto assess how the Canadian payments systemwas affected by the terrorist attacks in the Unit-ed States on 11 September 2001. To do so, thehourly intraday payments data for 11 September2001 are plotted against the benchmark(Charts 3a and 3b).

On that day, both volume and value were oper-ating normally before 10 a.m. Between 10 a.m.and noon, they fell below their lower level(minus one standard deviation). In response tothe slowdown in payment flows, the Bank ofCanada announced at 1:30 p.m. that therewould be a liquidity injection of $1 billion onthat day (raising excess settlement balances inthe LVTS from the typical $50 million).8 As aresult, the volume and value of payments recov-ered and rose above the upper level (plus onestandard deviation) between 1 p.m. and 2 p.m.

Volume started to decline between 2 p.m. and3 p.m. and remained below the lower level

8. For additional details on the Bank of Canada’sactions with respect to financial markets at that time,see “Actions Taken in Canada to Deal with PossibleDisruptions to the Financial System,” TechnicalBox 2, Bank of Canada Monetary Policy Report,November 2001, p. 17.

52

prior to the presettlement period. In contrast,value rose above the upper level again between4 p.m. and 5 p.m. This increase in value mighthave been triggered by the release of extra li-quidity committed by the Bank through the RGafternoon auction. During the presettlementperiod, volume and value were also higher thannormal.

For the day as a whole, volume and value wereoperating at about 90 and 100 per cent of thebenchmark, respectively.

Summary and Future Research

This is only a preliminary analysis of normalLVTS intraday payment flows. More work is un-doubtedly necessary because the benchmark isderived from a limited amount of intraday data.Accordingly, we plan to collect additional intra-day data to more fully explore the underlyingfactors in order to understand how they influ-ence intraday payment patterns. In addition,future consideration should be given to devel-oping real-time access to intraday payment data,which would allow ongoing monitoring, as wellas an immediate assessment when major dis-ruptions in the payments system occur. Regulardata on intraday flows could also be used to as-sess the impact of structural changes to paymentflows, such as those caused by the introductionof the CLS Bank in September 2002 and themigration of payments exceeding $25 millionfrom the ACSS to the LVTS as a result of the capto be introduced starting in February 2003.

Research

Summaries


Introduction

ank of Canada staff undertake research de-signed to improve the overall knowledge andunderstanding of the Canadian and interna-tional financial systems. This work is often

pursued from a broad, system-wide perspective thatemphasizes linkages across the different parts of thefinancial system (institutions, markets, and clearingand settlement systems). Other linkages of impor-tance may include those between the Canadian fi-nancial system and the rest of the economy, as wellas those with the international environment, includ-ing the international financial system. This sectionsummarizes some of the Bank’s recent work.

This issue of the Financial System Review high-lights research that addresses the changingstructure of the Canadian financial services in-dustry, as well as aspects of the structure andfunctioning of the Canadian foreign exchangemarket. Linkages are also emphasized. These in-clude the relationship between the structure ofthe financial system and economic growth andthe links between the banking systems of differ-ent countries when a serious disturbance arisesin one of them.

Innovation and change in the provision offinancial services has been a hallmark of theindustry for many years. The Financial ServicesSector: An Update on Recent Developments reviewsdevelopments in the Canadian financial servicessector. With a number of forces continuing tospur rapid innovation, the pace of change has, ifanything, accelerated. Within this context, Ca-nadian financial service providers are examin-ing the profitability of individual business linesmore closely and developing new ways todeliver financial products.

The Canadian foreign exchange market involvesa number of financial institutions and partici-pants, together with large trading flows. InCanadian Foreign Exchange Market Liquidity andExchange Rate Dynamics, a market-microstruc-ture approach is used to examine how market

B
participants use certain types of informationand, in turn, how this affects the exchange rateand the underlying liquidity in the foreignexchange market. This approach can also beused to study the effect of the risk-managementpractices of banks on market liquidity.
It is readily accepted that a well-functioning fi-nancial system is an important contributor toan environment of sustained economic growth,but there is less agreement on which specificstructure of the financial system contributes togrowth most strongly. In Financial Structure andEconomic Growth: A Non-Technical Survey, the au-thors identify the two main types of financialstructure and discuss whether one of them ismore beneficial to long-run growth than theother. They conclude that different structurescan, in fact, be complementary, and that it is theoverall level and quality of financial servicesthat is most important. To enhance the latter,policy-makers should focus on pursuingsupportive legal, regulatory, and other policyreforms.

Fortunately, serious disturbances in the Canadi-an banking system are rare. Nevertheless, froma global perspective, post-war history has pro-duced a number of banking crises in both theindustrialized and emerging economies. Whenproblems arise in the banking sector of onecountry, how are the banking systems of othercountries affected? Banking Crises and Contagion:Empirical Evidence examines this issue based onan empirical model of contagion and finds thatcontagion is more likely to occur between twocountries if they share similar macroeconomiccharacteristics. It is important that policy-makers understand how information from theoccurrence of a banking crisis affects thebehaviour of market participants.

55


The Financial Services Sector:An Update on Recent DevelopmentsCharles Freedman and Clyde Goodlet

n updated review of the Canadian fi-nancial industry shows that it contin-ues to experience significant changes.1

In an earlier Bank of Canada technicalreport, the driving forces behind the develop-ments that had been taking place over the previ-ous decade or so were examined, and some ofthe challenges that these forces would pose forfinancial service providers (FSPs) were indicat-ed. The key factors identified were technologicalchange, the changing nature of competition inthe financial services sector, and changes inhousehold demographics. The challenges facingthe financial services industry were discussedunder two main headings—the importance ofsize and the choice of the range of services andproducts that an FSP would provide.

This update builds on the previous work and, inparticular, highlights the role of economies ofscale and scope,2 mergers and concentration,the strategies being followed by FSPs, and therole of changes in information technology onservice delivery. Developments in these areascontinue to pose significant challenges for FSPsas they attempt to develop strategies to main-tain their profitability and long-run viability.While change in the financial sector is not new,the current period is noteworthy because of thepace and the scope of change, which appear tobe greater than ever.

Canadian financial service providers continueto search for ways to operate at an efficient scalein their back-office activities. They are followingthree different strategies to achieve this scale:

1. This note summarizes the recently published Bank ofCanada Technical Report No. 91, Freedman andGoodlet (2002), which updates Technical ReportNo. 82, published in 1998.

2. Economies of scale and scope refer to the possibilitythat a firm will realize a reduction in the cost of pro-ducing goods and services as a result of an increase inthe size or breadth of its activities.

A
(i) creating or building it; (ii) buying it; or(iii) borrowing it. Because of technologicalchange, the optimum scale of activities in manyback-office operations has increased. As a result,some FSPs are trying to gain the largest marketshare in Canada in particular activities (for ex-ample, transactions processing). Other FSPs areexiting these same areas, having decided thatthey will not be able to achieve a sufficient sizeof operations to be efficient. They are then pur-chasing these services from low-cost providers.Specific examples of back-office activities wheretechnological change has significantly increasedthe scale at which FSPs must operate to be effi-cient include credit card processing and pay-ment processing activities, such as debit cardacceptance services.
With regard to involvement in existing and newfinancial instruments, FSPs continue to empha-size the need for each product or service to beprofitable. Some FSPs have rigorously assessedthe profitability of each business activity in anattempt to allocate balance sheet resourcestowards activities of high strategic value andsustainable profitability. A consequence of suchassessments is that the FSP will exit areas that donot meet the test; for example, selling non-coresubsidiaries or getting out of certain lending ac-tivities. This development has been facilitatedand accompanied by an unbundling of activi-ties. One example of unbundling is the furtherseparation of loan-origination activities fromthe ongoing credit-risk exposure to the borrow-er, resulting from the development and spreadof various credit-risk-transfer arrangements.

At the same time, some FSPs have announcedstrategies that involve the rebundling of prod-ucts and services, particularly where economiesof scope are significant. This can be seen in theareas of consumer lending and corporate lend-ing. For example, some banks are linking theirwillingness to extend corporate loans to

57

Research Summaries

customers to the readiness of those customersto undertake their capital market business (suchas underwriting) with the bank.

Electronic money, which was introduced anumber of years ago with great fanfare, has beenshown to be technically feasible but not eco-nomically viable at this time. The potential rev-enues from a fully functioning arrangementappear to be insufficient to offset the high costsof establishing a national infrastructure capableof supporting such a scheme. Expectations of arapid deployment of electronic money schemes,either using stored-value cards or network mon-ey, have all but disappeared.

Mechanisms used to deliver financial servicesand products continue to evolve. A broaderrange of delivery channels has been developed,including expanded use of Automated BankingMachines, computer banking, and the use of theInternet, to handle routine, low-margin finan-cial transactions. Nevertheless, branches contin-ue to play a very central role in the plans ofFSPs, but their nature is changing (a strategycharacterized as “bricks and clicks”). Branchstaff must now have different qualifications, bebetter trained, and have access to much betterinformation technology. Branches are also be-ing opened on the premises of non-financialcompanies. Some FSPs are placing increasingemphasis on the revenues to be earned from thedistribution of financial services or products(their own and others) and from the develop-ment and operation of Web-based auction sites.But there continue to be significant barriers tothe use of information technology by FSPs inthe innovation of products and services andtheir delivery channels.

With regard to the size of institutions, it is im-portant to distinguish between the businesslines of FSPs and the size of a financial institu-tion as a whole. The recent literature seems tosuggest that economies of scale in a number ofbusiness lines extend further than previous em-pirical work had indicated. Evidence of this isseen in the growth of financial firms that spe-cialize in a small number of product areas (theso-called “monolines”). These firms exploitscale economies in process-intensive or infor-mation-intensive areas such as credit card pro-cessing. The growing importance of outsourcingin certain areas is also in part a recognition thatsignificant scale economies exist. The benefitsfrom the overall size of a financial institution

58

come from somewhat different sources, such asan increased possibility of economies of scopein institutions with multiple business lines andthe ability to engage in activities that requiremore capital. In addition, diversification acrossbusiness lines can lead to smoother revenueflows.

The prevalent view is that Canadian markets forfinancial services are too small for even the larg-est FSPs to operate in at an efficient scale in cer-tain lines of business. Large Canadian FSPsbelieve that they must operate as North Ameri-can entities. Indeed, there are a number of re-cent examples of Canadian FSPs implementingsuch a strategy. The key questions for these FSPsare the extent of the economies of scale in theirvarious areas of specialization and, where theeconomies of scale are important, whether theFSPs can achieve the size necessary to realizethem and to be competitive with the very largeFSPs in the United States. Regulatory restric-tions may limit the ability of FSPs to realizethese economies. Finally, there continue to bequestions regarding the importance of econo-mies of scope or synergies. In the non-financialsector, there have been waves of conglomera-tion and divestiture as views about the benefitsand costs of size change. It will be interesting tosee whether the financial sector experiences asimilar pattern.

References

Freedman, C. and C. Goodlet. 2002. The Finan-cial Services Sector: An Update on RecentDevelopments. Bank of Canada TechnicalReport No. 91.


Canadian Foreign Exchange MarketLiquidity and Exchange Rate DynamicsChris D’Souza

liquid financial market is a market inwhich supply and demand can bematched at a low cost. It is importantthat there be ample liquidity in the Ca-

nadian foreign exchange market, since a poorlyfunctioning foreign exchange market will createadditional costs for companies engaged in inter-national trade or investments, thereby adverselyaffecting the economy. One way to analyzeliquidity in the foreign exchange market isthrough market-microstructure research, whichfocuses on market arrangements and practices.

The foreign exchange (FX) market in Canadaconsists of a network of financial institutionslinked together by a high-speed communicationssystem. The participants in the FX market includedealers, customers, and brokers. Dealers contin-uously supply bid and ask quotes to both cus-tomers and other dealers. Through the course ofthe day, they stand ready to buy and sell foreignexchange, thus providing liquidity to the market.Brokers in the FX market are those intermediarieswho match the best buy and sell orders of deal-ers. Unlike dealers, who sometimes take specula-tive positions, brokers act as pure matchmakers.In Canada, most of the actual trading in the for-eign exchange spot market is handled by the topCanadian banks through their foreign exchangeoperations. Customers in these markets are thosefinancial and non-financial corporations thatneed foreign currencies for financing interna-tional trade, investing overseas, hedging cross-currency transactions, or pursuing short-term in-vestment opportunities, and those corporationsthat have supplies of foreign exchange.

Dealers in the foreign exchange market dis-tinguish between two types of trades: customertrades1 and proprietary trades. For customers,

1. Customers include the Bank of Canada, commercialclient businesses, and non-dealer financial institu-tions. Canadian chartered banks also trade with eachother as part of the interdealer market.

A
the ability to complete trades quickly is impor-tant when adjusting their FX position. Dealersprovide this fundamental element of liquidityto the market by trading with the customer. Inproviding these liquidity services, however,dealers may take on an undesired amount ofexposure to foreign exchange risk. Proprietarytrades are trades on a dealer’s own account toadjust its own FX portfolio position. Thesetrades are undertaken to help dealers managetheir exposure to foreign exchange risk in a prof-itable fashion. Proprietary trades are typicallyundertaken on the basis of available informa-tion about likely changes in foreign exchangerates. To achieve their foreign exchange invest-ment objectives, dealers must be assured thattheir information is at least as good as that oftheir trading counterparty, since trading againstbetter-informed traders is a losing proposition.
One way that dealers gauge market information isto observe order flow. One measure of order flowis the aggregate value of buy orders relative to sellorders that have been completed or that are“queued up” for future trades. An excess quantityof net buy (sell) orders for the Canadian dollarsuggests that other market participants have a pos-itive (negative) impression about the future pros-pects of the Canadian dollar based on availableinformation. Dealers acquire order-flow informa-tion from customer trades and through their com-munications with brokers. In this last case, theymay have electronic access to broker “screens,”which contain a part of the order flow.

Order flow is one key component of the market-microstructure approach,2 and is found to ex-plain a large proportion of the short-term (daily,weekly, or even quarterly) variation in nominalexchange rates. In contrast to traditional modelsof the exchange rate, which rely on factors such

2. See O’Hara (1995) for a review of market-microstruc-ture models.

59

Research Summaries

as interest rates, money supply, rates of infla-tion, gross domestic product, the trade accountbalance, and commodity prices to explain ex-change rate movements, order flow focuses onchanges in the market expectations of changesin these factors. Therefore, it often performsmore favourably than the individual factorsthemselves in empirical studies of short-term FXrate movements.

While recent market-microstructure studies ofthe FX market have had some preliminary suc-cess empirically in explaining exchange ratemovements using order-flow information, theunderlying determinants of order flow and thebehaviour of the dealers who provide liquidityto this market have not been tested explicitly.According to the microstructure view, liquiditywill be affected by the institutional features andinformation flows of the foreign exchange mar-ket. It is therefore of interest to examine wheth-er access to private information via customertrades, and the management of their own FXpositions affect a dealing bank’s willingness tosupply liquidity to the Canadian FX market.

Some researchers (e.g., Lyons 1997) have ar-gued that customer trades are the catalyst forprofitable dealer strategies. Our work suggeststhat dealers behave similarly in response to alltypes of trades, independent of where tradesoriginate. This indicates that valuable privateinformation about the fundamentals that mayaffect the value of the exchange rate is not ob-tained only from individual customer trades.Instead, it appears that dealers use private in-formation about their inventories, which areaffected partly by their own customer orders, asa profitable avenue for speculation in the inter-dealer market.

This has direct implications for liquidity in theFX market. Providing liquidity to customers al-lows dealers the opportunity to speculate in theinterdealer market. The more profitable suchspeculative opportunities are, the more compet-itive dealers will become in attracting customerorders. Consequently, the spreads between theirbid and ask quotes for customer trades will besmaller. Furthermore, our work suggests that li-quidity in the Canadian FX market is not affect-ed by the type of trading with the dealer, i.e., itoccurs from both customer trades and trades inthe interdealer market.

Daily hedging and risk-management practicesof banks with dealing operations can also be

60

examined. Information about each dealer’s nettrading position over the course of a day, in bothspot and forward contract FX markets, suggeststhat financial institutions operating in the FX mar-ket behave in a similar way when managing theirexposure to market risk. In particular, dealingbanks do not fully hedge their spot market risk.The amount of hedging depends on market vola-tility, the magnitude of banks’ risk exposure,and their comparative advantage in bearingrisk, especially compared with their customers.

There are various sources of comparative advan-tage for dealing banks in bearing risk: First,reciprocal agreements between dealing banksguarantee that these market-makers have access toliquidity. Customers, however, do not have thissame access. Second, banks allocate capital acrossbusiness lines in order to diversify risk and return.This allows intermediaries to bear risk with a high-er tolerance than the customers at non-financialinstitutions that may be specialized in relativelyfew business lines.3 Hedging by dealers is foundto depend on the overall risk-bearing capacity ofdealers in the market and on each dealer’s individ-ual access to order flow. Analysis suggests that li-quidity provision in spot and derivatives marketsis determined interdependently, since prices inthese markets are correlated, and dealers are ableto hedge risk across markets.

Interpreting FX trading data through market-mi-crostructure models helps characterize some ofthe factors that determine liquidity in the FX mar-ket. Results obtained to date suggest that dealerswith greater opportunities for profitable specula-tion and a larger appetite for risk will providegreater liquidity to the market. More generally, afocus on the institutional features of the marketthat determine its dynamics is critical to under-standing market liquidity. Policy-makers and re-searchers can use the tools developed in the fieldof market-microstructure finance to determine theeffects of various factors on liquidity. These factorsinclude the increased utilization of electronic bro-kering systems, a declining number of reciprocalagreements among dealing banks to provide li-quidity, possible consolidation of dealing banks,and the greater participation of foreign dealers inthe Canadian FX market.

3. D’Souza and Lai (2002) show that a decentralizedcapital-allocation function can reduce the overall riskof a financial institution with business lines that havecorrelated cash flows.


References

D’Souza, C. 2002a. “A Market MicrostructureAnalysis of Foreign Exchange Interventionin Canada.” Bank of Canada WorkingPaper No. 2002-16.

———. 2002b. “How Do Canadian Banks ThatDeal in Foreign Exchange Hedge TheirExposure to Risk?” Bank of Canada Work-ing Paper No. 2002-34.

D’Souza, C. and A. Lai. 2002. “The Effects ofBank Consolidation on Risk Capital Allo-cation and Market Liquidity.” Bank ofCanada Working Paper No. 2002-5.

Lyons, R. 1997. “A Simultaneous Trade Modelof the Foreign Exchange Hot Potato.” Jour-nal of International Economics 42: 275–98.

O’Hara, M. 1995. Market Microstructure Theory.Cambridge, Mass.: Blackwell Business.

61


Financial Structure and Economic Growth:A Non-Technical SurveyVeronika Dolar and Césaire Meh

n any modern economy, a primary eco-nomic function of the financial system (fi-nancial markets, intermediaries such ascommercial banks, and payments sys-

tems) is to transform household savings intoproductive investments. This function can beseparated into three basic subfunctions: the mo-bilization of savings, the acquisition of infor-mation, and the management of risk. Financialmarkets (stock and bond markets) and interme-diaries (which include banks, insurance compa-nies, and mutual funds) are two alternativetypes of agents that perform more or less thesame functions but in different ways and withdifferent degrees of success. Financial systemsthat rely mainly on the former are deemed mar-ket-based, while those that rely mainly on thelatter are called intermediary-based.

Many researchers have presented evidence fromcross-country, industry-level, firm-level, andtime-series studies to show that financial devel-opment exerts a positive impact on long-runeconomic growth. This raises an importantquestion: Which specific types of financial sys-tems are more growth-enhancing? There arefour competing views of financial structure andits relationship to long-run economic growth:

• The intermediary-based view asserts that inter-mediary-based systems are more growth-promoting than market-based systems. Thisis mainly explained by the fact that closerelationships between intermediaries andfirms reduce information costs and easefinancing constraints on firms, with positiveramifications for investment spending andeconomic growth.

• The market-based view argues that market-based systems encourage long-run economicgrowth better than intermediary-based sys-tems. This view stems primarily from thefact that markets, by allowing people withsimilar views to join together to finance

I
projects, are effective at financing new tech-nologies which, in turn, boost economicgrowth.
• According to the financial services view, theissue is not intermediaries versus markets,but rather the creation of an environmentfor the optimal functioning of intermediar-ies and markets, or for the efficient provisionof financial services generally, regardless ofthe mixture of intermediaries and markets.What matters for growth is the overall leveland quality of financial services and not thedistinction between markets and intermedi-aries.

• The law and finance view, a subset of thefinancial services view, also rejects the inter-mediary-versus-market-based distinction,but instead emphasizes that legal and regu-latory systems play the key role in determin-ing growth-fostering financial services. Forexample, a well-developed legal system thatenforces property rights and contractsreduces the cost of external financing bylowering the costs of acquiring informationabout firms. This increases external financ-ing and enhances economic growth.

According to the first two views, markets and in-termediaries are substitutes, whereas the last twoviews stress the complementarity of markets andintermediaries in providing growth-promotingfinancial services.

Which of these competing views of the linkbetween financial structure and growth are con-sistent with the data? Investigating the linkbetween financial structure and long-run growthinvolves complex relationships, and it is there-fore not surprising that there are no straight-forward conclusions. A survey of the literature,however, suggests that there is more empiricalsupport for the financial services and the lawand finance views than for either the intermediary-

63

Research Summaries

based view or the market-based view. Themajority of empirical researchers on this topicargue that financial structure (the degree to whichthe financial system of a country is intermediary-based or market-based) is not important forexplaining differential growth rates across econ-omies. For example, countries do not grow faster,and firms’ access to external financing is not sys-tematically easier, in either system. This conclu-sion is in line with the broad empirical analysisof financial structure and economic growth byDemirgüç-Kunt and Levine (2001), who use themost complete existing data set and a variety ofeconometric methods and yet consistently findthat financial structure is not important for eco-nomic development. They argue that “through adiverse set of analyses, the answers are surpris-ingly clear…. Overall financial development[efficient financial services, well-developedintermediaries and well-functioning markets]matters for economic success, but financialstructure per se does not seem to mattermuch” (p. 12).

Another reason to view markets and intermediar-ies as complements is that intermediaries are keyparticipants in markets, and they tend to play asupporting role in ensuring that financial mar-kets function properly. Investors need consider-able expertise to participate in financial markets,which makes their participation costly in termsof time and money. Financial intermediarieshelp to reduce these costs. More precisely, bybundling investors’ funds together, the costs ofparticipation in markets for each investor aresmaller (economies of scale). That is, because of theeconomies of scale, there is a reduction in thecost per dollar of investment as the size of trans-actions increases. In facilitating participation infinancial markets, financial intermediaries con-tribute substantially to the effective functioningof markets. The most obvious example of a fi-nancial intermediary that emerged because ofeconomies of scale and that supports markets isthe mutual fund. Because the mutual fund buyslarge blocks of stocks or bonds, it can take advan-tage of lower transactions costs. This argument issupported by Allen and Gale’s (2001) survey offinancial systems. They present evidence on theownership of corporate equities in the U.S. econ-omy. They find that in the year 2000, householdsheld less than 40 per cent of corporate equities,while intermediaries, particularly pension fundsand mutual funds, held over 40 per cent of totalcorporate equities. They conclude that “it is no

64

longer possible to consider the role of financialmarkets and financial institutions (intermediar-ies) separately. Rather than intermediating direct-ly between households and firms, financialinstitutions have increasingly come to intermedi-ate between households and markets, on the onehand, and between firms and markets, on theother” (p. 1).

We argue that the relationship between finan-cial structure and financial stability provides an-other reason for focusing on the need for bothwell-developed intermediaries and markets. Inthe event of a crisis in one system, the other sys-tem can perform the function of the “sparewheel.” Greenspan (1999) advocates this viewand argues persuasively that

What we perceived in the UnitedStates in 1998 may reflect an impor-tant general principle: multiple alter-natives to transform an economy’ssavings into capital investment act asbackup facilities should the primaryform of intermediation fail. In 1998in the United States, banking replacedthe capital markets. Far more often ithas been the other way around, as itwas most recently in the United Statesa decade ago. When American banksstopped lending in 1990, as a conse-quence of a collapse in the value ofreal estate collateral, the capital mar-kets were able to substitute for the lossof bank financial intermediation.Interestingly, the then recently devel-oped mortgage-backed securities mar-ket kept residential mortgage creditflowing, which in prior years wouldhave contracted sharply. Arguably,without the capital market backing,the mild recession of 1991 could havebeen far more severe (p.1).

These arguments suggest that it is not a questionof markets versus intermediaries but of marketsand intermediaries.

This implies an important policy message. Poli-cy-makers should focus their attention on legal,regulatory, and other policy reforms that en-courage the effective functioning of both mar-kets and intermediaries, rather than concerningthemselves with the degree to which theirnational financial system is market-based orintermediary-based.


References

Allen, F. and D. Gale. 2001. “ComparativeFinancial Systems: A Survey.” Manuscript,New York University.

Demirgüç-Kunt, A. and R. Levine. 2001. Finan-cial Structure and Economic Growth: ACross-Country Comparison of Banks, Mar-kets, and Development. Cambridge, Mass.:MIT Press.

Dolar, V. and C. Meh. 2002. “Financial Struc-ture and Economic Growth: A Non-Tech-nical Survey.” Bank of Canada WorkingPaper No. 2002-24.

Greenspan, A. 1999. “Do Efficient FinancialMarkets Mitigate Financial Crises?”Remarks before the 1999 Financial MarketConference of the Federal Reserve Bank ofAtlanta. http://www.federalreserve.gov/boarddocs/speeches/1999/19991019.htm

65


Banking Crises and Contagion: EmpiricalEvidenceEric Santor

inancial deregulation and the global inte-gration of markets have heightened theawareness of the potential fragility ofbanking systems in the face of external

crises. From a global perspective, banking crisesare numerous: Glick and Hutchison (1999)document 90 crises since 1975 across a sampleof 90 developing and developed countries.1

High-profile events such as the U.S. Savings andLoan, Mexican, Scandinavian, East Asian, andArgentinian crises reinforce this perception.Unfortunately, despite considerable efforts toempirically model the nature of banking crises,current analyses do not provide uniform con-clusions regarding the determinants of crises.2

Likewise, little is known with respect to thepresence and effect of contagion across bankingsystems.

This leaves several questions unanswered in theempirical literature:

• First, does the theoretical literature on bank-ing crises and contagion provide suitabletestable hypotheses with respect to the like-lihood of a banking crisis when interbankmarkets exist, and can these hypotheses beempirically assessed?

• Second, given the limitations of the data,can the onset of a banking crisis be accu-rately predicted?

1. Banking crises are defined when one or more of thefollowing events occur: the ratio of non-performingloans to total assets is greater than 10 per cent, thecost of rescue operations is more than 2 per cent ofGDP, and/or banks are nationalized, a bank holiday,or a guarantee of deposits, or loan losses and theerosion of bank capital exceed defined thresholds.

2. Given the current emphasis of the InternationalMonetary Fund and central banks on constructing“stress indicators” and “early-warning systems” toquantify the potential risks in the financial system,it is important to be confident of the methods ofempirical assessment used in these processes.

F
• And third, conditional on the ability torobustly predict banking crises, can the exist-ence of contagion be assessed? Moreover,does the occurrence of a crisis in one marketallow the prediction of crises in other mar-kets, over and above the effects of macroeco-nomic interconnections?
With regard to the first of these questions,contagion can be defined in terms of “funda-mental” and “informational” channels.Fundamentals-based contagion is used to de-scribe shocks that affect markets because ofcommon components, such as changes in U.S.interest rates, the price of oil, or the growth rateof the OECD countries (Dornbusch, Park, andClaessens 2000). These shocks lead to conta-gion because of the normal interdependence ofbanks and real-side markets. Information-basedcontagion occurs when the onset of a crisis inone market leads investors to re-assess the risksassociated with investments in other markets,regardless of whether or not there are any real-side linkages between the respective markets.The subsequent impact on asset prices fromchanges in investor behaviour can negatively af-fect the balance sheets of banks and, ultimately,the stability of the banking system.

In both cases, there are several pathways bywhich these shocks can lead to banking criseswithin and across banking systems. It is there-fore interesting to consider how contagion ismodelled in the theoretical literature andwhether the predictions can be empirically test-ed. For instance, Allen and Gale (2000) showthat the likelihood and effect of contagion de-pends on the degree of interbank market com-pleteness; i.e., the extent to which banks areinterconnected with other banks. But the datarequired to assess their model simply do not ex-ist. Alternatively, Chen (1999) shows that thefailure of one bank can lead to the failure ofother banks simply because of informational

67

Research Summaries

contagion. This suggests that crises can be prop-agated without any real-side links amongbanks. This notion of informational contagioncan be empirically assessed.

With regard to the second question, the empiri-cal literature on banking crises does not ade-quately address the issue of how to choose anappropriate sample of countries to test hypoth-eses. Most studies arbitrarily choose the sampleof crisis and non-crisis countries, neglecting thepotential impact of sample selection on the ap-propriate estimation procedure. It would bepreferable to pay particular attention to the con-struction of the cross-country sample: match-ing-method techniques should be used toconstruct a suitable control group analogue tothe set of crisis countries. This would allow theprobability of the occurrence of a banking crisisand of banking-system contagion to be quanti-fied more accurately. Sample selections in previ-ous studies introduced bias into the estimates ofthe probability of the occurrence of a bankingcrisis, because of differences between the char-acteristics of the crisis and non-crisis countrygroups.

Finally, in terms of the third and final question,given a clearly defined empirical benchmark, anempirical model of contagion can be estimated.Following Ahluwalia (2000), it is possible toconstruct contagion indexes to capture the no-tion of “informational contagion,” reflectingthe extent to which a country shares macroeco-nomic characteristics with a country that previ-ously experienced a banking crisis: the indextakes positive values proportional to the degreeof similarity. The contagion index does not re-quire the respective countries to share any real-side links; rather, the empirical specificationsuggests that the information associated withthe crisis leads to changes in investor behaviourthat may affect banks’ balance sheets. This al-lows a simple empirical test to be conducted:Do lagged values of the contagion index accu-rately predict the occurrence of a banking crisisin the current period, conditional on macroeco-nomic fundamentals? The analysis indicatesthat the probability of a banking crisis increaseswhen countries have characteristics similar tothose that have experienced a crisis, regardlessof the degree of actual economic linkagesbetween the respective countries.

In conclusion, the implications of these resultsare intriguing. If the fundamentals are

68

controlled, then the occurrence of a bankingcrisis in the previous period in one country pre-dicts the onset of a banking crisis in anothercountry, if the countries have similar macroeco-nomic characteristics. This suggests that infor-mational contagion plays a larger role thanpreviously suspected, since the onset of a crisisis related to the information provided by theinitial crisis event, over and above macroeco-nomic effects. The existence of informationalcontagion raises many issues for policy-makers.In particular, institutions that oversee, super-vise, or regulate financial institutions need toaccount for the process by which informationfrom the occurrence of one banking crisis affectsthe behaviour of market participants, and howmarket completeness propagates or mitigatesthe transmission of macroeconomic effects.

References

Ahluwalia, P. 2000. “Discriminating Conta-gion: An Alternative Explanation of Con-tagious Currency Crises in EmergingMarkets.” International Monetary FundWorking Paper No. WP0014.

Allen, F. and D. Gale. 2000. “Financial Conta-gion.” The Journal of Political Economy 108:1–33.

Chen, Y. 1999. “Banking Panics: The Role ofthe First-Come, First-Served Rule andInformation Externalities.” The Journal ofPolitical Economy 107: 946–68.

Dornbusch, R., Y. Park, and S. Claessens. 2000.“Contagion: How It Spreads and How ItCan Be Stopped?” World Bank. Photo-copy.

Glick, R. and M. Hutchison. 1999. “Bankingand Currency Crises: How Common Arethe Twins?” Center for Pacific Basin Mon-etary and Economic Studies. WorkingPaper No. 99-07. Federal Reserve Bank ofSan Francisco.

Santor, E. “Banking Crisis and Contagion:Empirical Evidence.” Bank of CanadaWorking Paper (forthcoming).

bank of canada financial...sisted, with a heightened level of tension in chart 1 total assets:...

Documents