April 2006 Global Financial Stability Report: Chapter I. Global

Previous issues of the Global FinancialStability Report (GFSR) have analyzedand assessed how the global financialsystem recovered from various shocks,

including the bursting of the equity bubble in2000–01 and the debt crises in a few emergingmarket (EM) countries. They spelled out indetail how cyclically favorable conditions andstructural changes have made financial inter-mediaries much stronger. The positive assess-ment contained in the September 2005 GFSRthat “the global financial system has yet againgathered strength and resilience” has beenvalidated by recent developments. However, anumber of cyclical challenges appear to begathering on the horizon, which necessitate amore nuanced view of the financial outlookfor the remainder of 2006 and beyond.

As this report has argued earlier, globaliza-tion and financial innovations have advancedthe scope for capital markets to channel creditto various users in the economy. In particular,the emergence of numerous, and often verylarge, institutional investors and the rapidgrowth of credit risk transfer instruments haveenabled banks to manage their credit riskmore actively and to outsource the warehous-ing of credit risk to a diverse range of investors.A wider dispersion of credit risk has “derisked”the banking sector, which still occupies a strate-gically important role in the economy, in partbecause of its role in the payments system. It iswidely acknowledged, meanwhile, that holdingof credit risk by a diverse multitude of investorsincreases the ability of the financial system as awhole to absorb potential shocks. This con-trasts with a situation where a small number of

systemically important banks bear the brunt ofmaking provisions for nonperforming loans.1 Itis true, as mentioned below, that the details ofwho holds which risk and in what amount areless transparent outside the banking systembecause of less stringent reporting require-ments. On balance, however, it is the wider dis-persion of risks, as such, that increases theshock-absorbing capacity of the financial sys-tem. As with a reinsurance system, the riskdiversification and dispersion aspects mattermore than the precise details of who is theultimate risk bearer.

Beyond risk diversification, the unbundlingand active trading of risk, including throughcredit derivative markets, seem to have cre-ated an efficient, timely, and transparent pricediscovery process for credit risk. Instead ofwaiting to learn about the possible deteriora-tion of credit quality through infrequentreports of nonperforming loans on banks’balance sheets, counterparties in the markets,as well as supervisors, can now monitor indica-tors of credit quality in real time (see ChapterII). All these structural changes, takentogether, have made financial markets moreflexible and resilient. As former U.S. FederalReserve Chairman Alan Greenspan said:“These increasingly complex financial instru-ments have contributed to the development ofa far more flexible, efficient, and henceresilient financial system than the one thatexisted just a quarter-century ago.”2

At the same time, this “brave new world” ofmodern capital markets creates its own set ofrisks and challenges. As mentioned above,these include a lower level of information

1

CHAPTER IGLOBAL FINANCIAL SYSTEM RESILIENCE IN THE FACEOF CYCLICAL CHALLENGES

1By way of example, the annual number of U.S. bank failures has fallen to a very low level in the past 10 years—aperiod that witnessed a number of major shocks that in the past could have caused great anxiety among banks andtheir supervisors.

2See Greenspan (2005).

about the distribution of risk to and amongthe nonbank financial institutions, whichincreases the potential for unpleasant surprisesfrom the less regulated market segments. Moreimportant, this new market-based environmentfor credit risk is predicated much more on theavailability of liquidity in all crucial areas ofthis market. Any potential liquidity disturbancecould amplify market corrections (see ChapterII). In addition, operational risks, such asdelays in credit derivative trade confirmations,assignments of contracts to third parties, andcontract settlements, have been identified asweaknesses, and remedial actions are beingundertaken by market participants to addressthem.

In the area of emerging market debt, manycountries—especially the large and systemi-cally important ones—have substantiallyimproved the structure of their sovereign debtand their domestic capital markets. At thesame time, their investor base, both interna-tional and domestic, has expanded andbecome more diverse. All together, these

changes have made EM countries moreresilient to external shocks (see Chapter III).

In parallel, cyclical factors have also beenvery favorable over the past few years. Lowinterest rates and an abundant supply ofliquidity have supported a solid global eco-nomic recovery and set in motion a search foryield. Banks and corporations implementedcost-cutting restructurings in response to pre-vious over-leveraging and to competitive pres-sure more generally. As a result, corporateearnings have recovered strongly in the pastthree years, and corporate balance sheets havestrengthened beyond expectations. Balancesheets of the household sector in major coun-tries have also improved since 2001, becauseof the rise in house prices and the recovery ofinternational equity markets. For example, networth of the U.S. household sector has recov-ered to close to the all-time high reached in1999 (Table 1.1).3 Benefiting from thesedevelopments, banks in many countries—especially the large internationally active insti-tutions—currently enjoy very strong financial

CHAPTER I GLOBAL FINANCIAL SYSTEM RESILIENCE IN THE FACE OF CYCLICAL CHALLENGES

2

Table 1.1. Sectoral Balance Sheets(In percent)

2000 2001 2002 2003 2004 2005

United StatesBanking: NPL/total loans 1.1 1.4 1.5 1.2 0.9 0.8Banking: Return on equity 14.8 14.2 14.9 15.2 14.6 13.7Corporate: Debt/net worth 48.6 51.6 50.8 49.3 47.1 45.5Household: Net worth/disposable personal income 583.3 543.4 497.6 537.9 555.9 564.9

EuropeBanking: NPL/total loans 3.0 2.9 3.0 3.0 2.3 . . .Banking: Return on equity (after tax) 18.3 11.2 9.0 11.3 14.2 . . .Corporate: Debt/equity 67.3 71.3 74.2 72.1 70.3 . . .Household: Net worth/assets 85.3 84.6 84.3 84.4 84.3 . . .

Japan1

Banking: NPL/total loans 6.3 8.4 7.4 5.8 4.0 3.5Banking: Return on equity –0.5 –14.3 –19.5 –2.7 4.1 6.3Corporate: Debt/equity (book value) 156.8 156.0 146.1 121.3 121.5 108.2Household: Net worth/net disposable income 767.5 763.9 753.0 749.0 . . . . . .

Source: National authorities.Note: NPL = nonperforming loans. Expanded balance sheets and detailed notes may be found in Tables 7–9 of the Statistical Appendix.1Data are for fiscal years beginning April 1. Data on household nonfinancial assets and disposable income are only available through FY2003.

Data in FY2005 are for the first half of 2005.

3It is also important to note that households face many risks, such as abrupt movements in asset prices, includinghouse prices, that in turn might substantially affect net worth. Their debt service burden would rise as interest ratesincrease. Households have also taken more responsibility for their future financial needs, including retirement needs.

health: strong capital bases, good profitability,and good asset quality as reflected in their lownonperforming loan ratios (see Figure 1.1and Box 1.1). All in all, strong balance sheetsin the financial, corporate, and householdsectors have created substantial financial cush-ions in practically all major financial systems.

However, these favorable cyclical conditionswill not be permanent. At a time when policyinterest rates have been raised and credit qual-ity is expected to deteriorate somewhat, anumber of questions arise: To what extent, andhow fast, will cyclical conditions change? Howwill that affect asset reallocations and pricecorrections? How much cushion and supportwould the aforementioned structural changesin financial systems provide? Given the paucityof data in many areas and the quite recentnature of the underlying structural changes,the answers to these questions will have to betentative and qualitative in nature. While itwould be desirable to apply a “bottom-upapproach” by using extensive financial stabilityindicators, these are either not available oravailable only with a considerable time lag. Asis often the case with developments in finan-cial markets, waiting for conclusive empiricalevidence would take very long and woulddeprive policymakers of the chance to reactwithin a reasonable time span.

Chapter I analyzes the main cyclical risks inthe financial markets going forward, especiallythose stemming from higher interest ratesand/or higher inflation, a deterioration in thecredit quality of various debtors, and a suddenunwinding of global imbalances. In addition, ithighlights a number of policy conclusions.Chapter II discusses developments in thecredit derivative and structured credit markets,focusing on the implications for financial sta-bility and on potential influences on creditcycle dynamics. It argues that while thesedevelopments have helped to make the bank-ing and overall financial system more resilient,they present new challenges and vulnerabilitiesthat need to be better understood. Chapter IIIdescribes changes that have taken place in the

GLOBAL FINANCIAL SYSTEM RESILIENCE IN THE FACE OF CYCLICAL CHALLENGES

3

0

2000

4000

6000

8000

10000

U.S. banksEuropean banks

200520042003

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

200520042003

0

2

4

6

8

10

12

200520042003

Figure 1.1. Financial Indicators for U.S. and European Banks

Source: IMF staff estimates.

Net Income(In millions of euros or U.S. dollars)

Nonperforming Loans(In percent of total loans)

Tier 1 Capital Ratio(In percent)


4

Market and Credit Risk Indicators for the MatureMarket Financial System

This issue of the GFSR continues the use ofmarket risk indicators (MRIs) and credit riskindicators (CRIs) to review the evolution of risksin mature financial systems, and expands thescope of the CRI to include insurance compa-nies.1 During the past year, banking system risk,

as measured by the MRIs and CRIs, hasremained relatively low, and does not indicateany particular financial stability concerns, as sig-naled in the financial markets. Similarly, theinsurance company MRIs and CRIs seem toindicate that the property and casualty insurers,and the reinsurers, are sufficiently capitalizedand diversified to absorb the catastrophic 2005hurricane-related losses.

Mature Market MRIs

Throughout 2005, the VaR-beta for theportfolio of financial institutions fluctuatedin fairly narrow bands, suggesting that therehave not been any significant changes to theaggregate risk profile of these financial insti-tutions (see first figure). The VaR-beta doesrise rather sharply at the end of 2005, butthis relates to a sharp rise in equity prices,and not to any apparent financial stabilityconcerns.2

During 2005, the VaR-betas for the insurancecompanies continued to fluctuate in a fairlytight range (see second figure).3 However, inSeptember and October, the VaR-betas for thereinsurance and the property and casualtycompanies surged higher in the aftermath ofHurricanes Katrina and Rita. Though the prop-erty and casualty insurers’ VaR-betas settleddown by October, the reinsurers’ VaR-betas

Box 1.1. Financial Systems in Mature and Emerging Markets

Note: The main authors of this box are John Kiffand Yoon Sook Kim.

1The MRI index captures institution-specific risksbased on the Value-at-Risk (VaR) of portfolios com-prised of equities from three groups of institutions:large complex financial institutions (LCFIs), com-mercial banks, and insurance companies. VaR meas-ures the market capitalization–weighted potentialloss over a 10-day period at the 95 percent confi-dence level of a portfolio of equity securities. Thevariances and correlations used in the computationsare, at each point in time, daily estimates over a75-day rolling period, and they are obtained usingan exponential smoothing technique that givesmore weight to the most recent observations. Toisolate the risks to the specific institutions in ques-tion, we continue to use a methodology suggestedby Hawkesby, Marsh, and Stevens (2005) to removethe effects of global and domestic equity marketvolatility (VaR-beta). The CRI index measures theprobability of multiple defaults within the threeabove-mentioned groups of institutions, impliedfrom the market prices of credit default swaps. Thedefinition of LCFIs is the same as that suggested byHawkesby, Marsh, and Stevens (2005), and ourportfolio comprises ABN Amro, Bank of America,Barclays, BNP Paribas, Citigroup, Credit Suisse,Deutsche Bank, Goldman Sachs, HSBC Holdings,JPMorgan Chase & Co., Lehman Brothers, MerrillLynch, Morgan Stanley, Société Générale, and UBS.The commercial banks captured in the MRI areAustralia and New Zealand Banking Group, BancaIntesa, Banco Bilbao Vizcaya Argentaria, Bank ofEast Asia, Bank of Nova Scotia, CIBC,Commerzbank, Fortis Bank, HVB Group, INGBank, KBC Bank, Mitsubishi Tokyo Financial,Mizuho Financial, National Australia Bank, Nordea,Royal Bank of Canada, Royal Bank of Scotland,SanPaolo IMI, Santander Hispano Group,Skandinaviska Enskilda Banken, Sumitomo MitsuiFinancial, Svenska Handelsbanken, TorontoDominion, UFJ Holdings, UniCredito, Wachovia,and Westpac Banking Corp. The CRI focuses on asmaller group of such banks for which CDS quota-tions are available.

2As noted in the September 2005 issue of theGFSR, one of the potential flaws in the MRI is thatthe risk metrics of parametric VaR measures tendto increase with the volatility of the underlyingassets, regardless of whether the volatility is associ-ated with price increases or decreases. This andother MRI shortcomings will be addressed as wecontinue to develop our indicators in future issuesof the GFSR.

3Insurance companies captured in the MRIare Aegon, AIG, Allianz Group, Allstate, Aviva,AXA, Chubb, Friends Provident, Gruppo Generali,Hartford Financial Services Group, MetLife,Millea, Mitsui Sumitomo, Munich Re, PrudentialFinancial, Prudential PLC, Sampo, Skandia,St Paul, Swiss Re, and Zurich Re. The CRI focuseson 15 insurers for which CDS quotations areavailable.

GLOBAL FINANCIAL SYSTEM RESILIENCE IN THE FACE OF CYCLICAL CHALLENGES

5

continued to rise. The global property andcasualty insurance and reinsurance industryfaced large losses from Katrina.4 However, thelarge reinsurers are well capitalized and diversi-fied, and were perceived as able to absorbKatrina-related losses. In fact, some analystsnoted that the reinsurers may benefit from acontinuation of the strong or rising pricingenvironment, as a result of the significant hurri-cane activity during the second half of 2005,and their equity prices generally strengthenedthrough the end of 2005.

Mature Market CRIs

The large complex financial institution(LCFI) and commercial bank CRIs indicate thatthe probability of multiple defaults spikedsharply in May, as the market digested the Fordand GM downgrades, and the related volatilityin the structured credit markets (see third

figure).5 However, since then, default probabili-ties have declined steadily, showing no discern-able reaction to a number of significant defaults(Refco, Delphi, and Calpine) and the continu-ing deterioration of the health of the U.S. autosector. In the past, such events may have beenexpected to impact materially on the financialinstitutions represented in the MRIs and CRIs.However, it seems that the market perceives thatthese institutions are less exposed to such eventrisks, possibly based on better risk managementtechniques and tools available (see Chapter II).

The new insurance company CRIs also fol-lowed the same track (downward) since August2005 (see fourth figure), although there was a

1

2

3

4

5

6

7

8

LCFIs

Commercial banks

All financial institutions

VaR-Betas for Full Portfolio of Financial Institutions(In percent)

2002 03 04 05 06

Sources: Bloomberg L.P.; and IMF staff estimates.Note: LCFIs are large complex financial institutions.

0

2

4

6

8

10

12

14

Reinsurance

Property and casualty

Life insurance

VaR-Betas for the Insurance Industry(In percent)

2002 03 04 05 06

Sources: Bloomberg L.P.; and IMF staff estimates.

4Mitigating some of the losses related to Katrinawas the fact that much of the damage was floodrelated, which is covered by federal flood insuranceprograms and excluded from most homeownerinsurance policies.

5See Chapter II of the September 2005 GFSR formore detail on the CRI methodology. Basically, itreflects the probability of multiple defaults over atwo-year horizon, as imputed from a portfolio offive-year CDSs referenced to the 15 institutions inthe three baskets. The methodology is based on a“structural” model that requires two key inputs,aside from the individual institution risk-neutraldefault probabilities implied by their CDS pricelevels: the loss-given-default (45 percent) andinterinstitution equity correlation levels (50 per-cent for LCFIs and 30 percent for the commercialbanks and insurance companies).


6

September surge in the nonlife-specific (reinsur-ance, and property and casualty) CRI in thewake of Hurricane Katrina.

Developments in Emerging Market Banking Systems

Banking systems in emerging markets havegenerally strengthened overall as a result of theeconomic recovery and reforms. However, risksto financial stability may be increasing in somecountries because of rapid credit growth and ris-ing real estate prices. Large-scale intermediationof foreign inflows by major banks is contributingto credit expansion and is another source of sys-temic risk. The situation, however, varies consid-erably across regions and groups of countries.

In Asia, financial systems seem to havestrengthened following banking sector reformsand improved supervision, although problemspersist in a few countries where the banking sys-tems still suffer from structural weaknesses. Thecenter of gravity of growth in financial servicescontinues to shift toward the large and rapidlygrowing economies of India and China. Themain risks going forward are the following:• rapid growth of credit to households in a

number of Southeast Asian countries, espe-

cially for mortgages, and of sale to retail cus-tomers of complex structured products withlimited hedging possibilities;

• the dominance of state-owned banks in Indiaand China, with expanding bank balancesheets in the latter in the context of highnonperforming loans ratios; and

• continued corporate sector lending againstthe backdrop of weak governance andtransparency.In emerging Europe, rapid credit growth in

many countries, especially in Eastern Europe,driven by the expansion of large foreign bankscompeting for market share, poses the mainrisk. In addition, intraregional contagion riskhas also increased as these banks pursue com-mon credit expansion strategies and areexposed to the same risk factors. The authoritieshave implemented measures, such as higherreserve requirements and tighter prudential lim-its, to slow credit growth, but with mixed effectsthus far.

In Latin America, performance indicators forthe financial systems have improved, including

Box 1.1 (concluded)

0

1

2

3

4

5

6

LCFIs

Commercial banks

Probability of More Than One Default Among Portfolio of Financial Institutions(In percent)

2003 04 05 06

Sources: Bloomberg L.P.; and IMF staff estimates.Note: LCFIs are large complex financial institutions.

0

1

2

3

4

5

All insurance companies

Nonlife

Life

Probability of More Than One DefaultAmong Portfolio of Insurance Companies1

(In percent)

2003 04 05 06

Sources: Bloomberg L.P.; and IMF staff estimates.1The life- and nonlife-specific CRIs are each based on

five insurers, whereas the all-insurers CRI is based on 15 insurers, which is why the all-insurers CRI is higher.

composition of emerging market countries’sovereign debt and investor base, and gaugeshow these changes affect resilience to adverseshocks. It shows that the EM investor base isbecoming more diversified, with more long-term-oriented investors in both domestic andexternal debt markets and more foreigninvestors willing to invest in local currency EMdebt. Active debt management and develop-ment of local debt markets in several large EMcountries have contributed significantly tothese positive results.

Possible Cyclical Challenges FacingFinancial Markets

Cyclical developments, such as higher inter-est rates and the turning of the credit cycle,are likely to present a number of challenges tofinancial markets and institutions. The benignfinancial environment to date described abovehas reduced credit risk premiums and finan-cial volatility (Figures 1.2 and 1.3). Term riskpremiums (for long-term government bonds)

have also been low in major countries. The lowlevel of risk premiums is open to differentinterpretations—either the actual risks embed-ded in financial instruments have declined orinvestors’ risk appetite has increased, leadingthem to bid risk premiums down. In the for-mer case, the more stable macroeconomic cli-mate in the United States and the globaleconomy since the mid-1980s could explainsome of the decline in risk premiums.4 In thelatter case, such investor behavior could leadto a mispricing or underpricing of risk, whichthen might lead to abrupt corrections. Theanalysis in Box 1.2 (p. 10) suggests that thereis no solid evidence of a systemic underpricingof risk because of a change in investors’ riskpreferences. However, as cyclical conditionsbecome less favorable, volatility and ultimatelyrisk premiums could increase.

Interest Rate and Inflation Risks

The recent rise in short-term interest rateshas created a flat and, at times, mildly

POSSIBLE CYCLICAL CHALLENGES FACING FINANCIAL MARKETS

7

in countries emerging from financial crises oraffected by political turbulence. The favorabletrends include sound capital adequacy, sus-tained expansion of lending activity, rising prof-itability, and better asset quality. Financialmarkets performed well, as illustrated by thestrong performance of stock prices, narrowingof sovereign debt spreads, and upgrades of sov-ereign credit ratings.

In the Middle East, Central Asia, and Africa,high commodity prices are the main factors driv-ing developments. In oil-producing countries,high oil prices are supporting strong economicactivity and inflation of asset prices. The mainrisks stem from rapid credit growth in combina-tion with rising asset prices, low transparency,and political uncertainty in some countries. A

sharp reversal in oil prices could have adverseeffects on the financial systems in some of thesecountries. The financial systems of sub-SaharanAfrican countries have mostly continued tostrengthen, supported by strong economicgrowth and enhanced regulatory frameworks,but fragilities persist. Sociopolitical instabilitycoupled with weaknesses in the enforcementof prudential frameworks accounted for muchof the deterioration in the banking systems inparts of the West African Economic and Mone-tary Union and Communité Economique etMonétaire de l’Afrique Centrale regions. Moregenerally, the current high levels of excessliquidity coupled with the high nonperformingassets in most sub-Saharan countries are sourcesof vulnerability.

4See Ferguson (2005) and Bernanke (2006).

inverted yield curve in the United States and,to a lesser degree, in other major currencyareas. There are some concerns that such aflat yield curve environment could be a har-binger of slowing economic growth in the yearahead. In the past, an inverted yield curve inthe United States has been a reasonably good,but not always accurate, forward indicator ofrecessions to come. This time, a number offactors suggest that a flat yield curve does notnecessarily herald recession—in particular, thestill-low levels of real interest rates and well-anchored inflationary expectations that lessenthe need for aggressive monetary tightening(see Box 1.3, p. 13, for a detailed discussion ofthe implications of the flattening and possibleinversion of the yield curve).

As reflected by the moderate differentialsbetween nominal and inflation-linked govern-ment bond yields, inflationary expectations infinancial markets are currently still wellanchored (Figure 1.4). By the same token,market participants currently expect only mildand mixed movements in short-term rates inthe year ahead. Interest rate futures marketscurrently show that U.K. short-term rates areexpected to continue falling gently, U.S. ratesare expected to rise modestly before declininglater this year, while euro and Japanese ratesare expected to rise modestly (Figure 1.5).5

Consequently, most yield curves are expectedto remain essentially flat, and not becomeinverted to any significant degree and for anysustained period. So far, a flat yield curve hasnot caused difficulties for U.S.-based financialintermediaries. Reported financial results for2005 indicate continued strong profitability bybanks. Owing to a more diversified businessmix, they have been able to remain very prof-itable. By contrast, in the past, flat yield curvesreduced the earnings power and threatenedthe health of many of those banks whose busi-


8

0

100

200

300

400

500

0

200

400

600

800

1000

1200

1400

1600

2001

European high yield (right scale)European high grade (left scale)U.S. high yield (right scale)U.S. high grade (left scale)Japanese high grade (left scale)

02 03 04 05 06

Figure 1.2. Corporate Spreads(In basis points)

Source: Merrill Lynch.

0

5

10

15

20

25

30

2003

Equities

Interest rates

CurrenciesAverage of all threeasset classes

04 05 06

Figure 1.3. Implied Volatilities(In percent)


5Japanese policy rates are expected to stay at currentlevels and to rise modestly much later this year, follow-ing the Bank of Japan’s decision to exit its quantitativeeasing policy.

ness models were based much more on matu-rity mismatches.

However, if inflation expectations, for what-ever reasons, including further rises in oilprices, were to increase significantly for a sus-tained period of time, this would create head-winds in financial markets through severalchannels.• A general rise in short- and long-term

interest rates would most likely lead to aneconomic slowdown, with negative conse-quences for corporate earnings and creditquality, and credit spreads would widensubstantially.

• Bond portfolios would incur substantial val-uation losses for both domestic and interna-tional investors. Institutional investors, suchas pension funds and life insurance compa-nies, might also experience mark-to-marketlosses in the near term. However, their bal-ance sheets could improve in the mediumterm as the present value of their liabilitiesfalls and they invest new pension plan con-tributions or insurance premium income inhigher-yielding fixed-income instruments.

• Equity markets would come under pressure,especially since earnings growth in manymarkets has already been expected todecline, albeit from strong and double-digitrates in the past few years. However, anymarket correction is unlikely to be very sig-nificant given that market valuations, mea-sured by price-earnings ratios, are currentlyat around their long-term averages in mostcountries (Table 1.2)—meaning, they are


9

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

2001 02 03

JapanUnited States

Europe

United Kingdom

04 05 06

Figure 1.4. Ten-Year Inflation Expectations(In percent; nominal yields less inflation-indexed yields)

Source: Bloomberg L.P.

0

1

2

3

4

5

6

Eurodollar

Eurosterling

Euribor

Euroyen

2006 07 08 09 10

Figure 1.5. Short-Term Interest Rate Expectations(In percent; term structure of three-month LIBOR futures rates, as of February 10, 2006)


Table 1.2. Equity Valuations

Price-Earnings Ratios__________________________________January 1996–2005 1970–2005

2006 average average

Germany 17.5 25.6 18.2Japan 23.6 27.2 31.1United Kingdom 13.9 17.8 13.4United States 18.6 24.1 16.6

Developed Europe 15.3 20.3 . . .Emerging markets 15.0 16.6 . . .

Source: Morgan Stanley Capital International.


10

There is a widely held view that investors’appetite for risk has increased over the past fewyears, leading to higher prices for risky assetsand narrower spreads on credit and other riskyproducts. Investors appear to have shifted tomore traditionally risky assets, including emerg-ing market equities, while leverage has risen instructured loan and LBO markets, and flows intohedge funds have accelerated. Low interest ratesand abundant liquidity have been attributed tothis behavior, as market participants confrontedwith low rates on relatively safe assets havemoved toward riskier assets in a search for yield.

Based on this assessment, some market practi-tioners and public policymakers have cautionedthat such behavior could have resulted in a mis-pricing or underpricing of risk. And, in fact,there is good evidence that risk premiums forcredit products are quite low on a historicalbasis, making such products potentially vulnera-ble to a cyclical shift in volatility. However, theanalysis here suggests that investors’ overall atti-tude toward risk appears not to have changedappreciably, although relative price movementsmay indicate shifting perceptions of the relativeriskiness of specific asset classes.

Analysis of Market Premiums and PortfolioDevelopments

To evaluate the overall market risk premium, asimplified version of the capital asset pricingmodel is employed, with risk-return trade-offscomputed for the basket of risky assets alone, andthen for the risky basket plus the safe asset.1

Three points of comparison are taken—2000when the stock market was near its highs andthe Fed funds rate was near current levels,2003 when the Fed funds rate was near itslowest level of 1 percent, and January 2006with the Fed funds rate having risen to 4.5 per-cent. The change in the market risk pre-mium over the three periods is moderate—a small reduction in the expected return-riskratio from .15 to .14, followed by a decline

Box 1.2. Is the Market Underpricing Risk?

Note: The main author of this box is Chris Walker.1For each asset, the expected return is the contem-

poraneous market yield on the asset. This is the earn-ings yield for equities, the yield adjusted forexpected default for bonds, and long-run historicalreturns for commodities and real estate. Varianceand covariances with other asset classes are based onperformance over the most recent three-year period.The safe rate is the U.S. Fed funds target rate at thetime. Risky assets are U.S., European, Japanese, andemerging market equities; emerging market sover-eign bonds; a commodities index; an index for U.S.real estate; and a high-yield bond index.

14

12

10

8

6

4

2

0

Expected return

Investment Possibilities, January 2003(In percent)

0 5 10 15 20Portfolio standard deviation

Expected return of portfolio

25 30 35 40


Expected return from risky assets

Implied risk premium = 0.14}

14

12

10

8

6

4

2

0

Expected return

Investment Possibilities, January 2006(In percent)

0 5 10 15 20 25 30 35 40


Expected return from risky assets

Implied risk premium = 0.12

Portfolio standard deviation

Expected return of portfolio


11

to .12 (see first and second figures).2 However,the overall stability of the risk premium masked anumber of changes that may have occurred.

For most fixed-income assets, spreads havefallen to the low side of historical ranges and, insome cases, are close to historical lows. In 2003,most of the high returns to risky assets wereattributable to the high yields then available onsovereign emerging debt and corporate bonds.3

Since then, the EMBIG spread index for emerg-ing market bonds has fallen from 650 basispoints to about 200 basis points. The spreads forhigh-yield corporate bonds have declined by asimilar margin, from an average of about 500basis points to 150 basis points. Furthermore,spreads on such assets, which typically have dura-tions of several years, have fallen even furtherrelative to short-term rates, since term premiums(the difference between long-term and short-term yields) have dropped sharply.

However, expected returns have not declinedfor all major asset classes. Equity earnings yields—the ratio of earnings to share price—are withintheir historical range, both in absolute terms andas a spread to the risk-free interest rate. For exam-ple, the average earnings yield for the S&P 500 hasincreased from about 4 percent in January 2003 toabout 7 percent at present. Some analysts attributethe apparent unpopularity of equities to the lin-gering effects of the tech share–led crash of2000–01. Others argue that recent structuralchanges, such as new accounting standards forinsurance companies and pension funds, may haveinduced a relative shift from equities to bonds.

Volatilities—a measure of the riskiness ofassets—have dropped across a wide range ofassets, in some cases to historically low levels. Thisis true both for realized volatilities and for theexpected volatilities implied by options prices.The standard deviations of returns for corporatebonds, equities, commodities, and foreignexchange have all dropped, as indicated in themiddle section of the table above. For manyassets, recent observed volatilities (measured asthe standard deviation of the daily change inyield) are in the lowest one-quarter of the histori-cal distribution. In the equity market, the widelyused VIX index of implied volatility derived fromthe pricing of options on S&P 500 stocks is at a10-year low, and is at less than half of its 10-yearaverage. When measured in the aggregate, thetrend toward lower volatility is even more pro-nounced. While correlations among returns fromdifferent asset classes have not changed substan-tially, volatility cycles appear to have become syn-chronized across asset markets.4

2The best risk premium available at each time toan investor able to choose among different assets canbe determined by the angle of the straight line (the“capital market line”), which expresses the ratio ofexpected return to one standard deviation in thereturn. This is also known as the Sharpe ratio.

3These yields do not constitute pure expectedexcess returns, insofar as some share of the yieldmatches the risk-neutral expected loss. That is,investors effectively use some of the excess yield toprovision against expected default. The analysisattempts to compensate for this bias by adjustingfor the realized default over the period in the calcu-lation of risk premiums.

Excess Returns, Volatility, and Risk Premiums(In percent)

Emerging Emerging S&P Market Market Com-500 DAX Equities Bonds modities

Excess returnsJanuary 2000 n.a. n.a. 4.00 3.71 n.a.January 2003 2.75 2.75 3.00 4.72 3.02January 2006 2.70 3.00 3.83 2.10 n.a.

Price volatilityJanuary 2000 58.7 65.0 115.1 80.9 71.8January 2003 65.6 67.5 81.9 38.5 72.5January 2006 31.1 38.9 57.2 26.0 82.5

Risk premiumsJanuary 2000 0.00 0.00 0.04 0.05 0.00January 2003 0.04 0.04 0.04 0.12 0.02January 2006 0.09 0.08 0.07 0.08 0.00

Sources: Bloomberg L.P.; JPMorgan Chase & Co; MorganStanley Capital International; and IMF staff estimates.

Note: Expected excess returns are computed as spreads in per-cent to the risk-free rate for fixed-income instruments (adjustedfor default) and earnings yield minus risk-free rate for equities.Volatilities are expressed as standard deviation in annualized one-month returns. Individual risk premiums are calculated as the ratioof excess returns to one standard deviation in returns.

4This has clearly not always been the case. Forexample, when bond market volatility peaked in1994, equity volatility was quite low.


12

Reflecting the changing pattern of expectedexcess returns, the components of the “optimalrisky portfolio”5 have shifted from a basket con-sisting only of emerging market and corporatedebt to one now containing a mix of emergingmarket debt, corporate debt, emerging marketequities, and U.S. equities.

Looking only at the change in spread, it maybe tempting to conclude that risk aversion hasfallen, or that the appetite for risk has grown.But, to the extent that volatilities have declined,and are expected to remain low, risk premiums—computed as the ratio of expected returns torealized volatility—suggest a more nuanced sce-nario. On an individual asset basis, risk premi-ums have fallen for higher-yielding fixed-incomeinstruments, as illustrated in the bottom blockof the table. At the same time, however, risk pre-miums have risen quite sharply for equities. Thishas left the overall premium for market risk lit-tle changed.

What Are the Risks of a Market Correction in thePrice of Risky Assets?

The analysis has questioned the idea that therehas been a systemic mispricing of risk due to achange in investors’ risk preferences. Neverthe-less, there is some evidence that volatility mayhave a large cyclical component,6 suggesting thatdeclines in asset price volatility may prove lesspermanent than markets appear to expect. Thecyclical view holds that price volatility is low whenthe economy is running below capacity but picksup as aggregate supply becomes more inelasticand the range of possible outcomes for inflationand asset prices widens. This is typically reflectedin a rise in asset price volatility, particularly forequity prices, as illustrated in the third figure(above left). Accordingly, as the economic cyclecontinues to mature, volatility may rise, prompt-ing investors to shift out of risky assets and caus-ing the prices of those assets to adjust downwardsomewhat to compensate. The credit risk pre-mium may be the most susceptible to adjustment,particularly in view of the high correlationbetween equity volatility and credit spreads, asrepresented in the fourth figure (above right).

Box 1.2 (concluded)

200

400

600

800

1000

1200

10

20

30

40

50High-yield spreads

(basis points;right scale)

VIX index(percent;left scale)

VIX and High-Yield Spreads

1997 99 2001 03 05

Sources: Bloomberg L.P.; and Merrill Lynch.

0

50

100

150

200

250

300

Starting November 1982Starting May 2001Starting November 1990

Volatility and the Business Cycle

0 10 20 30 40 50 60

Number of months since peak recessionary volatility

Vola

tility

Inde

x

70 80 90

Source: Goldman Sachs.

October 2005

5In the standard capital asset pricing model, thisis the portfolio corresponding to the point of tan-gency between the two frontiers. Whatever theirrisk preference, investors optimizing the risk-returntrade-off hold some weighted combination of thisoptimal risky portfolio and the safe asset. 6See Goldman Sachs (2005).

13


Market attention has focused on the U.S.yield curve, which first inverted briefly (whenmeasured at 10-year less 2-year maturities) for afew basis points around year-end 2005 and hassince then remained essentially flat or mildlyinverted across much of the term structure.This flattening phenomenon has not been con-fined to the United States, as some othermature markets have experienced similar devel-opments to varying degrees. In the euro area,spreads between short- and long-dated maturi-ties have tightened in recent months, but notnearly to the extent seen in the United States.In the United Kingdom, the yield curve hasremained both flat and mildly inverted forsome time, while Japan’s term structure reflectsan accommodative monetary policy (see firstfigure above).

Because an inversion of a yield curve hasoften been a good forward indicator of reces-sions in the past, some market participants haveexpressed a concern that markets are signalinga significant slowdown in the United States.Indeed, the U.S. curve inverted before all of itssix recessions since 1960, most recently in August2000, just ahead of the March 2001 downturn; adisappointing reading on fourth quarter GDPhas only added to these concerns (see secondfigure below). In Germany, an inverted curve hasalso historically been a forward indicator ofrecession, but there the record is not as clear.While spreads between short- and long-termrates have narrowed considerably since peakingin early 2004, the German yield curve still retainsa positive slope (see third figure).

There are several reasons to suggest why con-cerns of an impending recession in the UnitedStates may be overstated, and why the recentenvironment is the result of other causes. First,historically, recessions generally have been pre-ceded by a steep and prolonged inversion.During past periods, the yield curve inverted toan average peak of more than 150 basis pointsand the average length of the inversion was over

one year (see first table). Such inversions werean indicator of a recession on average about11 months in advance. In contrast, the recentinversion was minimal and short lived, and theyield curve is expected to remain essentiallyflat in the period ahead (10-year less 2-year).Furthermore, despite fears about inversion, mosteconomic activity indicators and consensus fore-casts are pointing to sustained economic expan-

Box 1.3. Flattening and Inversion of the Yield Curve: Implications and Outlook

Note: The main authors of this box are PeterDattels and Ned Rumpeltin.

0

1

2

3

4

5United States

United Kingdom

Germany

Japan

Government Benchmark Yield Curves(In percent; as of February 10, 2006)

1month

3months

6months

2years

5years

10years

30years


–6

–4

–2

0

2

4

6

8

10

Real Federal funds rate

10-year minus3-month yield

United States: Slope of the Yield Curve(In percent)

1970 75 80 85 90 95 2000 05

Sources: Bloomberg L.P.; National Bureau of Economic Research; and IMF staff estimates.

Recessions


14

sion. In particular, the OECD’s index of leadingindicators for the United States signaled thateconomic activity accelerated throughout the sec-ond half of 2005. In the inversion episodes thatled to a recession, this indicator deterioratedmarkedly in the months prior to a downturn.

Second, past inversions—and subsequentrecessions—occurred mainly because of thedegree of monetary tightening needed to bringinflation under control. In the United States,this has typically taken a real Fed funds rate ofmore than 4 percent to push the economy into

recession. The present situation is different. Byraising the Fed funds target rate from 1 percentto 4.5 percent, the policy rate is within a rangeconsidered neutral, neither stimulating norinhibiting growth, with real Fed funds ratesmuch lower than the previous flattenings atabout 2 percent.

Third, the factors behind the flattening of theyield curve have changed, implying that aslightly inverted yield curve may be a less nega-tive signal than in the past:• Yield curve term premiums have diminished

as investors no longer demand as much com-pensation for risks of volatile or unexpectedinflation. Changes in realized inflation volatil-ity can often have a profound impact on termpremiums, given the particular sensitivity offixed-income investors to changes in theprice level (see fourth figure). Improved pol-icy transparency and central bank credibility,particularly at the U.S. Federal Reserve, havecontributed to lower and more stable infla-tion and better anchored inflationaryexpectations.

• Yields at the longer end of the curve have alsobeen influenced by rising demand for longer-term securities from several distinct investorclasses (see September 2005 GFSR). Notably,pension funds and insurance companies havebeen active purchasers of longer-durationassets, following changes in accounting stan-

Box 1.3 (continued)

–3

–2

–1

0

1

2

3

4

10-year minus1-year yield

Recessions

Germany: Slope of the Yield Curve(In percent)

1970 75 80 85 90 95 2000 05


History of Yield Curve Inversions and Recessions in the United States

Length of Lead Time Start of Yield Inversion Period Inversion Trough Was the Fed Did Recession to RecessionCurve Inversion1 (In months) (In basis points) Tightening? Follow? (In months)

January 1966 1 –3 Yes No . . .September 1966 5 –39 Yes No . . .December 1968 14 –42 Yes Yes 12June 1973 15 –179 Yes Yes 5November 1978 17 –279 Yes Yes 25October 1980 11 –357 Yes Yes 9July 1989 14 –3 Yes Yes 11August 2000 5 –63 Yes Yes 7

Average 10 –121 11.5Average (pre-recession) 13 –154 11.4Median 13 –53 10

Sources: Board of Governors of the Federal Reserve System; National Bureau of Economic Research; and IMF staff estimates.1Inversion computed using a 5-day moving average of the spread between 3-month bills and 10-year U.S. treasury notes.

15


dards and government regulations, that haveinduced these institutions to minimize the“mismatches” between the duration of theirassets and liabilities. Thus, they have started toshift portfolio allocations from equities tofixed-income assets, particularly those withmaturities that more closely match theincreasingly long maturities of their liabilities.

This is particularly the case in Europe, wherepension reform is more advanced than in theUnited States, and increasing demand forbonds with ultralong maturities has driven

0

0.5

1.0

1.5

2.0

2.5

3.0

–100

0

100

200

300

400

500Inflation volatility(percent; left scale) Slope of the yield curve

(basis points; right scale)

Inflation Volatility of the United States andSlope of the Yield Curve(3-month moving averages)

1985 89 93 97 2001 05

Sources: Bloomberg L.P.; and IMF staff estimates.Note: Inflation volatility is defined as the annualized 24-month

rolling standard deviation of yoy core CPI growth, while the slope of the yield curve is the yield spread between 10-year treasury notes and 3-month treasury bills.

0.25

0.50

0.75

1.00

1.25

1.50

3.25

3.50

3.75

4.00

4.25

4.50

50-year nominal yield (right scale)

50-year inflation-linked yield (left scale)

United Kingdom: Ultra-Long Gilt Yields: Nominal and Inflation Linked(In percent)

Sep. Oct. Nov. Dec.2005 2006

Jan. Feb.


0

200

400

600

800

1000

1200

0

50

100

150

200

250

300

High yield (right scale)

High grade (left scale)

Periods of flat or inverted yield curve1

Mortgage backed (left scale)

U.S. Credit Spreads and Slope of the Yield Curve(Option-adjusted spreads; in basis points)

1997 99 2001 03 05

Sources: Bloomberg L.P.; and Merrill Lynch.1Defined as a period when the spread between 10-year

and 3-month U.S. treasury instruments is less than 10 basis points.

Slope of the Yield Curve as a RecessionForward Indicator

1954–87 1988–2005

Constant 0.394799 0.248361[1.782] [0.775]

Yield curve slope (T – 4)1 0.50016 0.866319[2.940] [0.881]

Real GDP growth (T – 1) 0.7333 0.080028

R2 0.728073 0.786503Standard error of regression 1.495832 0.670346Durbin-Watson 1.163819 1.404759

Source: IMF staff estimates.Note: Percentage statistics are in brackets and are calculated

using Newey-West standard errors; the dependent variable isreal GDP growth (year-on-year, in percent).

1Spread between 10-year and 3-month U.S. treasuryinstruments.

not as stretched as they were in 2000 andare therefore less vulnerable to a “burstingof the bubble.”6

Under these circumstances, financial inter-mediaries would be stressed by a combinationof losses, and their currently strong balancesheets would be tested. At present, marketparticipants expect this to be a rather remoterisk, but it bears watching as the consequencesfor financial markets can be serious.

Turning of the Credit Cycle: Impact on CorporateCredit Markets

The credit cycle refers to fluctuations in thefinancial health or the balance sheet quality ofthe corporate sector that affect firms’ access

to, and cost of, credit. Variations in averagecorporate credit quality give rise to the needto write down credit spread products andadjust credit provisions by banks and otherholders of credit risk. These statements alsoapply to the household sector, or to any bor-rowers on capital markets. Historically, a turn-ing of the credit cycle against the backdrop oflow risk premiums is a normal cyclical devel-opment. While the credit cycle cannot beobserved directly, there are different metricsto gauge it—such as (1) changes in creditspreads in the corporate bond and creditderivative markets, (2) changes in the differ-ence between the number of credit upgradesand downgrades and the default rates,7

(3) changes in credit standards used by com-


16

nominal and real yields to very low levels (seefifth figure).Fourth, recycling of emerging market balance

of payment surpluses, including, particularly, inthe last year, petrodollars into U.S. financialassets, has led to sustained demand for dollar-denominated assets. At the same time, supply ofcorporate debt securities has been low, reflect-ing low world investment levels. These factorshave given rise to what appears to be almost a“scarcity premium” for longer-duration assets(see Box 1.6).

The structural changes and differences incyclical developments may have reduced thepredictive powers of the yield curve as a predic-tor of economic conditions, as suggested byempirical evidence. Prior to 1988, the slope ofthe yield curve was a statistically significant

indicator of future economic performance.Since that time, however, the same can nolonger be said, as the significance of a yieldcurve flattening has deteriorated significantly(see second table).

The flattening of the yield curve is often con-sistent with the prospect of a turn in the creditcycle. Spreads of investment-grade, high-yieldcredits and mortgage-backed securities tend towiden during inversions that come ahead ofcyclical turning points. At this turn, spreads onasset-backed securities appear to have widenedsomewhat earlier, possibly reflecting someuncertainty ahead regarding marginal borrowersin the mortgage markets, while corporatespreads saw some widening in 2005 associatedwith developments in the U.S. auto sector (seesixth figure).

Box 1.3 (concluded)

6The exceptions to the above assessment are the Middle East equity markets. These markets have rallied sharplyin the past two years. The overall market capitalization of the six stock exchanges in the Gulf region has more thandoubled in the past year to slightly more than $1 trillion. Average price-earnings ratios for these exchanges havereached 30–40. The rapid rise of these markets has been driven by the substantial oil price windfall, a portion ofwhich has been invested within the region. Local banks have been actively involved in providing brokerage servicesto individual investors in these equity markets. The swift rise of these markets, based on a relatively small number oflisted companies, harbors risk of a substantial correction—with potentially detrimental effects felt throughout thebanking system in the region.

7Ratio of the value of defaulted bonds to the value of all outstanding bonds.

17


mercial loan officers, (4) changes in the vol-ume of credit flowing to the corporate sector,and (5) changes in the quality of corporatebalance sheets.

Only credit spreads are readily available inreal time and cover an ever-widening array ofindividual names. While it is true that changesin credit spreads reflect only a collective mar-ket assessment of credit quality, there are nobetter or more timely indicators that arewidely available to market participants.Chapter II explains why and how movementsin credit spreads can serve as an early indica-tor of changes in credit quality and thus creditcycles.8 Indeed, Figure 1.2 shows that corpo-rate bond spreads have begun to widen sincethe second quarter of 2005, providing an earlysign of a turning of the credit cycle.

The differences between the number ofcredit upgrades and downgrades and defaultrates of various segments of the U.S., Euro-pean, and Japanese corporate bond marketsare available, but they are not as timely ascredit spreads. The major rating agencies havereported that the number of upgrades minusdowngrades is peaking in U.S. and Europeanbond markets but continues to improve inJapan (Figure 1.6).9 They expect the differ-ences to decline in the future. They alsoexpect default rates to rise moderately fromhistorically low levels. In particular, the rate ofdefault in the U.S. high-yield bond market isexpected to increase to 2.0–3.5 percent thisyear, and to almost double in 2007 to 4.5–5.0percent.

The tightening or loosening of credit stan-dards as constructed from surveys of seniorloan officers has usually preceded changes incommercial loan growth.10 Figure 1.7 showsthat bank loan officers have begun to tenta-

–175

–150

–125

–100

–75

–50

–25

0

25

50

75

–700

–600

–500

–400

–300

–200

–100

0

100

200

300

1987 89 91 93 95 97 99 2001 03 05

Europe(left scale)

Japan(left scale)

United States (right scale)

Figure 1.6. Corporate Upgrade Minus Corporate Downgrade Actions

Source: Standard & Poor’s.

–40–30

–20

–10

0

10

20

30

40

50

60

70

80

1990 92

Net tightening

Net easing

United States

Euro area

Japan

94 96 98 2000 02 04 06

Figure 1.7. Bank Lending Standards from Surveys of Loan Officers(In percent)

Sources: Bank of Japan; Board of Governors of the U.S. Federal Reserve System; European Central Bank; and IMF staff estimates.8See Zhu (2004) for a comparison of credit spreads

between the bond market and the credit default swap(CDS) market.

9See Hull, Predescu, and White (2004) for a discus-sion of the relationship between CDS spreads, bondyields, and credit rating announcements.

10See Lown and Morgan (2004).

tively tighten lending standards in the UnitedStates and, more recently, in the euro area.

If available in a timely manner, credit flowsshould clearly be an important indicator.Unfortunately, data on such flows are releasedonly after a long lag—for example, as of mid-February 2006, the U.S. Flow of Funds datawere available only for the third quarter of2005 (although some component data areavailable with shorter time lags). As such, flowdata can explain a past credit crunch but theydo not lend themselves to forecasting achange in credit quality.

Last but not least, changes in the quality ofcorporate balance sheets can also signal aturn in the credit cycle. Though, yet again,balance sheets can be analyzed in detail onlywith a considerable time lag. The level of pro-visioning in the balance sheets of commercialbanks is a good indicator, but it is also avail-able with a time lag. As a result, to arrive at aplausible assessment at an earlier stage, it isindispensable to look at some qualitative indi-cators on the health of corporate balancesheets. Listed below are several developmentsthat typically indicate a deterioration of cor-porate credit quality, that is, that the creditcycle is turning—they corroborate the evi-dence from credit spread widening, earlysigns of falling differences between upgradesand downgrades and rising default rates, anda possible tightening of lending standardsmentioned earlier.

In the past year or two, a number of corpo-rations have begun to reverse course instrengthening their balance sheets. As men-tioned in previous issues of the GFSR, thecorporate sector in many countries, mostnotably the United States, European coun-tries, and Japan, have significantly strength-ened their balance sheets since 2001,reflected in their historically strong financialpositions (Figure 1.8 ).11 Their ability to serv-


18

–6

–4

–2

0

2

4

6

8

Japan1

United States

EMU3

Aggregate

1980 84 88 92 96 2000 04

Figure 1.8. Nonfinancial Corporate Financing Gap(In percent of GDP)

Sources: Goldman Sachs; and IMF staff estimates.1Fiscal year.

1978 82 86 90 94 98 2002

Figure 1.9. U.S. Corporate Sector Earnings Interest Coverage Ratio

Source: Credit Suisse First Boston.

7

6

5

4

3

2

1

11For a detailed discussion of the net lending posi-tions of corporations in major countries, see IMF(2006, Chapter IV).

ice debt has been greatly strengthened. Forexample, the ratio of the earnings to interestcoverage for the United States is at a 28-yearhigh of 5.8 times compared with a long-termaverage of 4.1 times (Figure 1.9). However,more recently there has been a growing ten-dency to releverage balance sheets and takeactions that benefit shareholders at theexpense of creditors—such as higher divi-dend payouts,12 large share buybacks, andmerger and acquisition (M&A) activities(Figures 1.10, 1.11, 1.12, and 1.13 ).

In particular, leveraged buyouts (LBOs),facilitated by a significant increase in the vol-ume of LBO loans in the United States andEurope (Figure 1.14), could significantlyweaken the credit quality of the acquired com-pany. Recently, acquirers (usually privateequity funds and, increasingly, hedge funds)have adopted the practice of significantlyleveraging the balance sheets of the compa-nies they have just acquired so as to pay highdividends to themselves right away. This is incontrast with past practices according towhich acquirers spent about five years (if notmore) improving the profitability of theacquired company before doing an initialpublic offering (IPO) or a trade sale to realizetheir investments. As a result of a moreaggressive LBO style, in terms of both leverag-ing and a much shorter time frame, the creditquality of the acquired companies may deteri-orate sharply, typically leading to a multiple-notch downgrading by the rating agencies.For corporate bondholders, this type of idio-syncratic risk is more difficult to anticipatebecause it can materialize abruptly and tosome extent arbitrarily, compared with adeterioration of the company’s business overtime.

As the private equity funds have been ableto attract large amounts of funds in the past

19


1.50

1.75

2.00

2.25

2.50

2.75

3.00

3.25

1990 92 94 96 98 2000 02 04

Figure 1.10. Net U.S. Corporate Dividend Payments(In percent of GDP; 4-quarter moving average)

Sources: Board of Governors of the U.S. Federal Reserve System, Flow of Funds Accounts of the United States; and IMF staff estimates.

0

10

20

30

40

FY20061FY2005FY2004FY2003

Payout ratio

Dividend growth

Figure 1.11. Japanese Dividend Growth and Payout Ratio(In percent)

Sources: UBS; and IMF staff estimates.1Estimated.

12In the United States, higher dividends may alsoreflect changes in U.S. tax policy, which reduced theincome tax rate on dividend income.

year or two, and as they have reportedlyincreased the degree of leverage (from 3–4times earnings before interest, taxes, deprecia-tion, and amortization (EBITDA) to 6–8times), they have the financial resources totarget a much wider range of companies. Asan illustration, LBO debt raised in 2004–05 inthe U.S. market amounted to $200 billion, or5 percent of total corporate debt outstanding.Thus, despite strong balance sheets for thecorporate sector as a whole, this is one reasonfor the rise in idiosyncratic risk.

In more traditional areas, specific industriesthat are burdened by overcapacity—such asthe automobile and airline industries—areexpected to continue to shrink and cause spe-cific companies to suffer deterioration alongthe way. More generally, companies that—forother reasons—neglected to improve theirfinances could be exposed to pressure as cycli-cal conditions become less favorable.

In short, the turning of the credit cyclebrings about an increase in idiosyncratic risk,though against the backdrop of still-healthybalance sheets of the corporate sector as awhole. The most likely repercussion of thisdevelopment is a widening of credit spreadsfor specific firms (a stronger differentiationbetween credit quality on the part of investorsshould be healthy and not lead to a general“blowout” of credit spreads).

If such a company-specific deterioration incredit quality were to affect a very large“name” in the fast-growing market for creditderivatives and structured credit, and if it wereto materialize in an abrupt and unexpectedfashion, there is a possibility that this couldupset some of the complex correlations onwhich investment positions in these marketsare often predicated (see Chapter II). A sud-den drying up of liquidity as a result of unex-pected price dislocations could, in turn, leadto large-scale redemptions for hedge fundsactive in this area. As a potential next step insuch a chain reaction, hedge funds could feelunder pressure to liquidate other assets—inother words, start a contagion process.


20

0

100

200

300

400

500

600

Jan. Feb. Mar. Apr. May Jun. Jul.2005

United KingdomUnited States

Japan

Aug. Sep. Oct. Nov. Dec.

Figure 1.12. Announced Share Buybacks


0

500

1000

1500

2000

2500

3000

3500Other

Japan

Asia, excluding Japan

United States

Western Europe

1999 2000 01 02 03 04 05

Figure 1.13. Global Merger Activity(In billions of U.S. dollars)


Overall, however, the healthy corporate sec-tor balance sheet and the still-low default ratesshould provide firm anchors to the creditmarkets—the broad corporate spread shouldchange more moderately. This would enablethe self-correcting forces to operate: as somecredit spreads move beyond perceived funda-mental values, other investors would step inand, in the process, stabilize the markets.

Turning of the Credit Cycle: Impact on Housingand Mortgage Markets

The housing and mortgage markets also sig-nal a turning of the credit cycle. In the UnitedStates, indicators of housing activity havefallen in recent months in response to higherinterest rates. In particular, the MortgageBankers Association’s purchase index, a lead-ing indicator, has fallen by about 15 percentsince July 2005. Average house price increasesappear to have decelerated from an annualrate of 12.95 percent in the fourth quarter of2005.13

There are two related concerns by marketparticipants. First, cooling house prices couldreduce the volume of home equity withdrawaland therefore help to weaken personal con-sumption and economic growth. Accordingto IMF staff estimates, a slowdown of U.S.real house price appreciation from 10 per-cent last year to zero would reduce personalconsumption by around 0.5–1 percentagepoint. At the same time, this could allow thepersonal savings rate to rise somewhat, thuscontributing to a moderation of the U.S.current account deficit. For the U.S. house-hold sector, as an illustration, a 10 percentoutright decline in average house priceswould reduce household net worth by$1.9 trillion to $49.1 trillion—or from 564.9percent of disposable personal income to545.5 percent, still well above the low level

21


0

50

100

150

200

250

United StatesEurope

1999 01 022000 03 04 05

Figure 1.14. Issuance of Leveraged Loans(In billions of U.S. dollars)

Source: Bloomberg L.P.Note: Leveraged loans are floating-rate obligations issued to corporations

that rating agencies have typically assigned noninvestment-grade ratings, and that bear a yield in excess of 125 basis points over major benchmark rates such as LIBOR and EURIBOR.

13According to the latest release by the Office ofFederal Housing Enterprise Oversight (March 1, 2006).

reached in 2002.14 Even though much ofhousehold wealth is in relatively illiquidforms, such as housing and pension claims,this level of net worth is still important as acushion to help households absorb potentiallosses.

The second concern is that higher interestrates could raise the debt-servicing burden ofhomeowners—at present, the debt-service-to-income ratio for the household sector isalready at a high level of 13.5 percent. Thiscould worsen the credit quality of the mort-gage markets and result in losses for the lend-ing institutions. However, because 60 percentof U.S. mortgages are at long-term fixed rates,the increase in the debt-service burden wouldbe limited in the near term. The credit qualityof the traditional prime quality mortgage mar-ket is quite high—the delinquency rate is nearrecord lows and should not worsen too muchif the U.S. unemployment rate remains lowand personal income continues to grow.Moreover, U.S. mortgage lenders have distrib-uted large portions of their mortgages to awide range of investors in the mortgage-backed securities market, and many of thoseinvestors, such as life insurance companies,may be in a better position to hold mortgagerisks because of their longer-term liabilitiesand lower liquidity pressures than banks.Consequently, the impact on financial inter-mediaries active in the large prime qualitymortgage market is likely to be rather limited.

The main vulnerability in the U.S. mortgagemarkets at present lies more in the sub-primesegment of the market.15 Many sub-primeborrowers in these new instruments intend to

refinance fairly quickly to take advantage ofany house price appreciation and to avoidrefixing at higher rates. However, rising inter-est rates, a cooling-off in the U.S. housingmarket, and regulatory tightening, all ofwhich would make it more difficult for theseborrowers to refinance if they cannot other-wise qualify for a traditional mortgage, meanthat many will be trapped in the original resetterms of these mortgages. An estimated $140billion of such mortgages are due to berefixed in 2006, and $350 billion in 2007.16

The result is that the delinquency rate in thismarket segment will probably rise.Anticipating this, spreads on asset-backedsecurities (ABS) using sub-prime mortgagesand on collateralized debt obligations (CDOs)including or referencing such ABS havewidened substantially (Figure 1.15). Sincethese markets are new, there are concernsabout whether investors, especially newentrants, fully understand the risk and havethe capacity to adequately manage it.

While these developments in the U.S. hous-ing market bring risks that bear watching,especially the risk of a larger-than-expectedfall in personal consumption, there areprospects of a “soft landing,” judging from theexperience in the United Kingdom andAustralia. House prices in these countries—which increased by much more than in theUnited States—decelerated sharply fromaround 20 percent annual growth in 2003–04to practically zero for most of 2005. Sincethen, U.K. house prices have recovered, risingby 1.4 percent in January 2006.17 The mainimpact of the price adjustments seems to have


22

14Calculations based on Federal Reserve Board’s U.S. Flow of Funds, December 2005.15Sub-prime mortgages refer to mortgages granted to borrowers who otherwise could not qualify for a traditional

mortgage. This has been accomplished by lenders relaxing credit standards and/or documentation requirementsfor such borrowers and via new mortgage products designed to minimize the debt-servicing burden during an ini-tial period. These products offer features such as no down payments, variable rates to take advantage of lower short-term rates, and interest only and/or no amortization products. This market segment has grown rapidly in the past18 months, accounting for about 30 percent of new mortgage origination in 2005, even though it still represents asmall portion of the outstanding mortgage market in the United States.

16Lehman Brothers MBS and ABS Research (2006).17According to a 2006 press release of the Nationwide Building Society.

been a weakening of personal consumption,which contributed to the slowdown of theseeconomies. The impact on financial institu-tions and markets has been mild—a smalluptick in delinquency rates and increased badloan provisions, which were absorbed by lend-ing institutions without signs of stress.

Turning of the Credit Cycle: Impact on EmergingBond Markets

In contrast to the spread widening in cor-porate and mortgage markets in matureeconomies, spreads on EM internationalbonds have continued to trend to record lowlevels (Figure 1.16). As examined in previousissues of the GFSR, improvements in the fun-damentals of the EM countries and favorableexternal financial conditions have attractedgrowing capital inflows, contributing to theEM spread compression (see Chapter III).Improvement in the EM countries as a whole,of course, masks some individual countries’weaknesses—but by and large EM countriesare stronger. Of particular importance is thefact that EM countries have run large currentaccount surpluses for six consecutive years.Growing private sector capital inflows—particularly foreign direct investment (FDI)flows (see Box 1.4, p. 26)—combined with thecurrent account surplus, meant that the EMcountries had to accumulate a huge volume ofinternational assets, through both the publicand the private sectors (Table 1.3).

Focusing on determinants of the decliningEM spreads, an econometric analysis of EMbond spreads (Box 1.5, p. 28) shows that fun-damental factors (including economic, politi-cal, and financial developments) and externalfinancial variables (mainly volatility measuresand the Fed funds futures rate) have con-tributed about equally to explaining EMspread movements. The fundamental factorshave explained rather well the downwardtrend in EM spreads since 2001. The externalfinancial variables have explained the volatilityin the EM spreads. Consequently, if the funda-

23


0

100

200

300

400

500

Asset-backed security spread

Credit default swap spreadon asset-backed security

1997 2000 02 03 0501 04 06

Figure 1.15. U.S. Asset-Backed Spreads1

(In basis points)

Source: Bear Stearns.1For the Baa tranche of asset-backed securities composed of floating-rate home

equity lines of credit.

100

200

300

400

500

600

2005

EMBIG Latin America

EMBIG Europe

EMBIG Asia

EMBI Global

2006

Figure 1.16. Emerging Market Spreads(In basis points)

Sources: JPMorgan Chase & Co.; and Merrill Lynch.

mental improvements in EM countries remainintact, a worsening of the external financialvariables—for example, a rise in volatilitiesand/or in the Fed funds futures rate—wouldbe consistent with a widening of the EMspreads, but probably around the trend lineinstead of a reversal to the long-term mean. Inother words, improved fundamentals in EMcountries will most likely act to offset EMbond spread movements because of a worsen-ing of cyclical conditions.

This assessment goes hand in hand with thematuring of EM bonds to eventually becomea mainstream asset class for internationalinvestors. Specifically, the yield premiumsthat EM bonds have to offer over comparablyrated mature market (MM) corporate bondshave declined steadily since 2001 (Figure1.17). In addition, EM bond spreads havecontinued to narrow even after the FederalReserve began to tighten monetary policy inmid-2004, instead of widening as in the past.Thus, it appears that EM bonds are increas-ingly being evaluated using normal capitalmarket yardsticks, with concerns decreasingabout their special emerging market charac-teristics. As such, EM bond markets as a wholewill behave like any other securities markets inMM countries, that is, bonds will be pricedaccording to their creditworthiness. Theywould be subject to market correctionsbecause of changes in cyclical conditions, butprobably become less prone to the kind ofstress and crisis experienced in the past (or

“feast and famine,” as we called it in previousissues of the GFSR).

In particular, EM bond markets are likely tobe less vulnerable to near-term fluctuations inexternal financial market conditions, sinceEM countries have prefinanced more thanhalf of their 2006 external issuing plansbecause of their active debt managementoperations. Latin American countries havepractically completed their 2006 financingplans, and a few countries have begun to pre-finance their 2007 requirements (see theappendix on Emerging Market FinancialFlows). Some have begun to buy back theirexternal debt, causing EM spreads to tightenfurther as international investors begin to dealwith the prospect of a declining supply ofexternal sovereign debt in the face of theirstrong demand.

While the EM countries as a group havebecome net exporters of capital and hencetended to reduce their external financial vul-nerability, a number of countries, includingthose in Central and Eastern Europe, areexperiencing developments traditionally asso-ciated with increased external vulnerability—large current account deficits, rapid creditgrowth, and rising external debt. In inter-preting trends in the Central and EasternEuropean region, the offsetting and mitigat-ing factors (including membership in the EUand, in time, in the euro area), as well as theconsiderable differentiation across countries,must be borne in mind. In some countries, fis-


24

Table 1.3. Emerging Markets and Developing Countries: Current Account Balance and External Financing(In billions of U.S. dollars)

1996 1997 1998 1999 2000 2001 2002 2003 2004 20051 20061

Current account balance –85 –85 –115 –18 88 43 86 114 228 410 494External financing 339 424 296 250 243 178 203 269 454 395 393

Of which:Foreign direct investment 129 167 173 176 168 181 167 158 210 221 223Private debt 123 143 44 14 42 –45 –6 68 178 114 93

Asset accumulation 254 339 182 233 331 220 288 414 681 805 887By official sector (reserves) 81 54 –54 4 111 84 124 367 514 600 612By private sector 173 285 236 229 220 136 164 47 167 205 275

Source: International Monetary Fund, World Economic Outlook.1Data for 2005 are estimated and data for 2006 are forecast.

cal positions are broadly balanced and publicindebtedness relatively low. In these countries,as well as others, sizable current accountdeficits—and, hence, reliance on foreign sav-ings—are associated with rapid growth, consis-tent with their catch-up potential as newmembers of the European Union. Moreover,in some countries, falling reserve coverage ofshort-term debt obligations reflects, in part,borrowings by domestic subsidiaries from rep-utable and well-supervised parent banks.Nevertheless, risks are inherent in such situa-tions. In particular, where current accountdeficits reflect the inability to rein in fiscaldeficits, and public debt ratios are on the rise,strong policy efforts to contain these trendsare needed lest these countries face increasedmarket scrutiny when the environment turnsless friendly.

Global Imbalances and Capital Flows

There is broad agreement among policy-makers and market participants that a “disor-derly” unwinding of global imbalances could,and probably would, have very negative conse-quences for financial stability. If the necessaryfinancing of the U.S. current account deficitwere to require significant risk premiums tobe factored into prices for U.S. assets, theglobal economy and world financial marketswould be seriously affected. Given the increas-ingly global asset allocation process, spillovereffects in virtually all other asset marketsshould be expected. Also, once such a devel-opment were to gain momentum, it is notclear how soon countervailing forces wouldkick in. Given the dominance of U.S. financialmarkets in the global economy, it is not auto-matic that other non–U.S. dollar based assetclasses would automatically benefit. It is quitepossible in such a scenario that a “decou-pling” of non–U.S. dollar assets would be con-fined to government bond markets at best. Acombination of a weaker dollar, higher mar-ket interest rates, depressed equity markets,and widened credit spreads could create a

25


–200

0

200

400

600

800

1000

1200

1999 2001 03 05

B rated

BB ratedInvestment grade

Figure 1.17. Spread Differences Between Emerging Market and U.S. Corporates(In basis points)

Sources: JPMorgan Chase & Co.; and Merrill Lynch.


26

This box discusses recent trends in foreigndirect investment in emerging market countriesand reports on discussions with a private sectorcontact network.

Total FDI flows to EMs increased substantiallyin 2004 and 2005 (see the figure). The inflowsin 2004 amounted to $180 billion, 41 percenthigher than in 2003 and above the 2001 peak,although they were smaller as a share of GDP(2.7 percent rather than 3.1 percent).1 Foreigndirect investment in EMs is estimated to haveincreased by more than 10 percent in 2005.The increases in FDI have been fueled byimproved growth, higher commodity prices,improved business and investment climate, per-ceptions of reduced risks in EMs, and moreM&A activity in EMs. Equity financing repre-sented the bulk of FDI flows to EMs: 87 percentin 2002–04.2

Almost all regions experienced expansionsin FDI in 2004. The largest increases were inemerging Europe, Central Asia, and LatinAmerica. Flows to Asia also increased. In 2005,flows to emerging Europe, Central Asia, andAsia continued to increase strongly, but declinedsomewhat to Latin America. Flows to SouthAfrica increased dramatically because of a largebank acquisition.

Increasingly, EM firms are globalizing: out-ward FDI from EMs has increased rapidly overthe past several years. Some EMs that havebeen large recipients of FDI have become

sources of outward FDI. The World Bank’sforthcoming Global Development Finance reportwill include a fuller discussion of FDI flowsbetween EMs.

The IMF and the World Bank Group staffhave established an informal private sectorcontact network to discuss FDI in EMs to(1) improve surveillance and forward-lookingassessments of FDI flows to EMs and (2) seekinput for IMF–World Bank policy advice on FDIand the business climate.3 Discussions with anetwork of participants from EM and MM coun-tries focused on the following:

Box 1.4. Foreign Direct Investment in Emerging Market Countries

OtherEmerging EuropeEmerging AsiaEmerging Latin America

0

10

20

30

40

50

60

70

Foreign Direct Investment in Emerging Markets(In billions of U.S. dollars)

2001 02 03 04 05

Sources: National authorities; and IMF, InternationalFinancial Statistics.

Note: The main author of this box is Paul Ross.The box draws heavily from the work of the IMF–World Bank FDI group. The members of the groupare Charles Blitzer, Ceyla Pazarbasioglu, and PaulRoss (all IMF), and Thomas Davenport, JosephBattat, Dilek Aykut, and Zenaida Hernandez (allWorld Bank Group).

1As in the previous issues of the GFSR, the dataon and discussion of FDI inflows to EMs focus onthe 22 largest EM recipients, which account forabout 85 percent of FDI inflows to all developingcountries. The data for 2004 have been revisedupward substantially, because of revisions of rein-vested earnings in several countries.

2Based on the IMF’s World Economic Outlook data.

3See September 2005 GFSR, Box 2.5: “ForeignDirect Investment to Emerging Market Countries:An Asian Perspective.” The group comprises sen-ior executives from 40 private sector companiesand financial institutions. The companies operatein the following sectors: natural resources andmining, food and beverage, metals, machinery andmotor vehicles, financial services, electronicgoods, consumer goods, telecommunications,information technology, utilities, and retail.About half of the contact network participants areEM firms.

27


• The increase in FDI to EMs in 2004–05appears more to be a part of a secular trendrather than being driven by cyclical factors.With relatively high growth prospects, EMcountries as a group are seen as justifyingan increased share of the investment pie.Furthermore, perceptions of the riskiness ofEM investments are declining because ofimproved macroeconomic and structuralpolicies, and growing transparency andother steps taken to improving the businessclimate in many countries. In addition toincreases in FDI in EM countries, FDI flowsfrom firms located in these countries—directed to new businesses in both other EMsand MMs—have been increasing rapidly, andthat trend is widely expected to continue.

• The dominant strategic reason for most FDIinvestments is diversification into new mar-kets. Diversification strategies are being fol-lowed similarly by firms headquartered inemerging markets as well as by traditional FDIinvestors from mature market countries. Withthe focus being on potential market size andgrowth, interest in China remains extremelyhigh. However, investors also anticipateincreasing their businesses (or establishingnew ones) in other large EM countries.Efficiency-seeking FDI appears most concen-trated in emerging Europe, some SoutheastAsian countries, India (services more thanmanufacturing), and China. EM investors dif-fer in some ways from MM investors: theyhave a more regional outlook and some areconcentrating in niche markets for reasons ofscale familiarity.

• Controlling and managing FDI-related risksare an integral part of companies’ strategies.They actively manage their foreign currencyexposure arising from balance sheets andrevenues of overseas businesses. Managingthese risks incurs management and financialcosts; the cost of managing these risks cansometimes be a significant factor in locatingFDI projects. Investors continue to generallyprefer wholly owned subsidiaries but oftenteam up, at least temporarily, with local part-

ners, when they lack local knowledge, net-works, or contacts; where there are limitationson foreign ownership; or to espouse a more“local” image. Most investors seek manage-ment control.

• Financing of FDI investments is shifting tolocal sources. Equity investments most typi-cally continue to be financed by the parentcompany. However, debt financing increas-ingly is in local currency and sourced locally,and reliance on parent company guaranteesis diminishing. Many investors anticipate thatover time they will rely less on bank loans andmore on bonds as local capital marketsdevelop further. This development is impor-tant in that FDI and local capital marketdevelopment are supportive of each other.Foreign direct investors’ desire to increasetheir use of local capital markets is a positivesupply-side factor for development of thesemarkets that allows local institutionalinvestors to diversify their portfolios. Thus, itis likely that the development of local capitalmarkets will increase interest from foreigndirect investors.

• Investors look to the IMF and the WorldBank Group to continue to promote policyimprovements that would further facilitateFDI in EMs. In addition to the activitiesthat both institutions now undertake, inves-tors (those that undertake foreign directinvestment) asked that the IMF and theWorld Bank Group focus more on strength-ening the legal and regulatory framework forthe private sector, developing local capitalmarkets and private equity initiatives, dis-seminating more information on less wellknown (smaller) member countries, encour-aging public-private sector dialogue, and play-ing a greater role in investment disputeresolution.4

4Some participants wanted involvement from theIMF and the World Bank Group that goes beyondthe scope of the work of the International Centerfor Settlement of Investment Disputes, which ispart of the World Bank Group.


28

Over the past several months, spreads onemerging market debt have reached recordlow levels. This box sheds light on a number ofquestions concerning developments in emerg-ing markets. Are emerging market spreadsappropriately priced in relation to own-countryfundamentals and external conditions? Whathave been the improvements in fundamentals?How have external factors influenced spreads?Why have spreads narrowed even though theFed has tightened rates? How resilient wouldemerging market spreads be in the face of adeterioration in the external environment?

The Model

We explore the determinants of emergingmarket sovereign debt spreads using a simplifiedversion of the panel data model presented in theSeptember 2004 GFSR.1 The model sheds lighton spreads for individual countries and can beaggregated to predict the composite JPMorganEMBIG index. The data consist of monthlyobservations between January 1998 and Decem-ber 2005, in a panel for the countries in theEMBIG index.2 Spread determinants includecountry factors as well as external factors.

The dependent variable is the log of EMBIGspreads at time t. Lagged variables are notincluded in the model, which is estimated usingfixed effects. As proxies for country fundamen-tals, we use a set of economic, financial, andpolitical risk ratings, whereby higher ratings indi-cate better fundamentals and lower risk.3 The

ratings are updated on a monthly basis and areconstructed in a transparent manner with lessjudgment compared with credit ratings, whichare commonly used to proxy fundamentals. Tocapture the effects of external factors, we includean index of the implied volatility of the U.S. stockmarket, VIX, which we take as a proxy for inves-tors’ perception of global financial risk. We fur-ther include the yield on three-month Fed fundsfutures to gauge the effects of U.S. monetary pol-icy and international liquidity on emerging mar-ket spreads. The volatility of the Fed fundsfutures market is included to capture the extentto which uncertainty about future U.S. monetarypolicy affects emerging market spreads.4

In recent years, the economic and financialrisk ratings of the EMBIG countries have seenstrong improvements, reflecting the substantialaccumulation of international reserves in majoremerging markets, persistent current accountsurpluses, improved fiscal balances, containedinflation, and lower foreign debt to GDP (seefirst figure). The improvement in the politicalrisk rating for the EMBIG countries has beenless dramatic in recent years. In 2005, closelycontested elections in several countries inemerging Europe and the upcoming electioncycle in Latin America weighed on governmentstability and the political risk rating for theEMBIG countries as a whole.

Box 1.5. Main Drivers of Emerging Market Bond Spreads: Fundamentals or External Factors?

Note: The main author of this box is KristianHartelius.

1For a lengthier and more recent exposition ofthe research underlying this box, see Kashiwaseand Kodres (forthcoming).

2We exclude Argentina because of breaks in theseries related to debt restructuring. Due to shortdata series, we also exclude Serbia and Montenegro,Indonesia, Vietnam, Greece, and Qatar from themodel. The analysis thus includes 32 countries.

3The International Country Risk Guide, published bythe PRS Group, releases monthly ratings coveringthree types of risks: economic, financial, and politi-cal. The economic risk rating includes variablessuch as annual inflation, budget balance to GDP,

and the current account to GDP. The financial rat-ing includes variables such as foreign debt to GDP,net international liquidity as months of importcover, and a measure of exchange rate stability. Thepolitical variable includes various more subjectivemeasures of political risk, such as government stabil-ity, conflict, and bureaucracy quality. Each country isassessed on the same basis to allow for comparabil-ity. A higher rating indicates better fundamentalsand lower risk. The political variable takes a valuebetween 0 and 100, whereas the economic andfinancial variables range between 0 and 50.

4The volatility of the Fed funds futures market iscalculated as the standard deviation of the differ-ence between the yield on three-month-ahead Fedfunds futures and the Fed funds target rate. For adetailed explanation of the independent variables,see Kashiwase and Kodres (forthcoming).

29


Our proxy for investors’ perception of globalfinancial risk has seen a dramatic decline overthe past few years. From the beginning of 2003to the end of 2005, the VIX declined by morethan 50 percent to its lowest level since 1995(see second figure). During the same period, incontrast, the Fed funds futures rate increased by333 basis points as a result of the monetary pol-icy tightening by the Federal Reserve. The cross-correlations between the explanatory variablesare all low, except for the economic and finan-cial risk ratings, which tend to move together(their correlation is 0.68).

Results

The signs of the estimated coefficients con-firm the findings in the September 2004 issue ofthe GFSR. Improved fundamentals are associ-ated with lower spreads, whereas higherexpected volatility of U.S. equities, highervolatility on the Fed funds futures market, andtighter U.S. monetary policy are all associatedwith higher spreads (see the table below).5

In the aggregate, the model appears to followthe general movement of actual spreads. In thethird figure, the actual and the EMBIG spreadspredicted by the model for each country havebeen aggregated using the weights in theEMBIG index. The close fit of the model, anoverall R-squared of 0.57, implies that the recentlevels of EMBIG spreads are in line with what wewould expect, given historical relations between

30

32

34

36

38

40

58

60

62

64

66

68

70

72

Fundamental Factors, Aggregated with EMBIG Weights

1998 2000 02 04

Sources: JPMorgan Chase & Co.; The PRS Group; and IMF staff estimates.

Economic risk rating(left scale)

Political risk rating(right scale)

Financial risk rating(left scale)

Emerging Market Bond Spreads: Fixed-Effect Panel Regression Model

Percent Impact on the Spread by One Standard

Coefficient Standard Error t-statistic Deviation Increase

Dependent variable: Log of EMBIG spreads

Explanatory variablesEconomic risk rating –0.035328 0.003162 –11.17 –9.39Political risk rating –0.027701 0.001879 –14.74 –8.55Financial risk rating –0.038093 0.003209 –11.87 –10.65VIX 0.038216 0.001171 32.63 29.17Three-month-ahead Fed funds futures rate 0.044670 0.004065 10.99 9.87Volatility of three-month-ahead Fed funds

futures rate minus target rate 1.476950 0.138052 10.7 8.56

R2 within 0.57R2 between 0.58R2 overall 0.565

Number of observations 2,634

Sources: Bloomberg L.P.; The PRS Group; JPMorgan Chase & Co.; and IMF staff estimates.

5The estimates are robust to changing the modelspecification, the estimation techniques, as well asthe sample. The coefficients remain significant,and keep their signs and magnitudes, over a longersample starting in 1993, during which the variablesshow much less pronounced trends.


30

the variables. However, by the end of 2005,spreads were about 40 basis points lower thanpredicted by the model, and spreads have sincenarrowed further into 2006.

To assess the sensitivity of spreads to changesin fundamentals, we calculate the percentchange in the model spread that would resultfrom a one standard deviation increase in eachvariable, other things being equal (see thetable).6 A one standard deviation improvementin any of the three risk ratings leads to a drop inthe country spread by about 10 percent of theprevailing level. However, since the economicand financial risk ratings tend to move together,this measure is likely to understate the sensitivityof the model spread to changes in fundamentals.

By decomposing the model dynamics, we canget an insight into how important the recentimprovements in fundamentals have been forthe tightening of EMBIG spreads.7 Since thebeginning of 2003, the aggregated model spreadhas fallen by 422 basis points. Improved finan-

cial and economic risk ratings contributed 105basis points and 53 basis points of this droprespectively, while improvements in the politicalrisk rating contributed to a decline in the aggre-gate spread by 19 basis points. The impact ofimproved fundamentals has thus been less than50 percent of the total decline in spreads sinceJanuary 2003. However, a separate regressionusing only the three risk ratings as independentvariables reveals that fundamentals can explainmuch of the trend in EMBIG spreads over thefull sample (see fourth figure). The overallR-squared for the regression using only funda-mental factors is 0.5. The estimated coefficientsare of the same sign and order of magnitude asin the full model.

Turning to the external factors, the model sin-gles out investors’ perception of global financialrisk as the variable that has the largest impacton spreads. An increase in VIX by one standarddeviation causes spreads to increase by 29 per-cent of the prevailing level, other things beingequal. The corresponding effects of changes inthe Fed funds futures rate and the volatility inthe Fed funds futures market are both close to10 percent of the prevailing level of spreads.

Box 1.5 (concluded)

0

5

10

15

20

25

30

35

40

45

0

1

2

3

4

5

6

7

8

External Factors

1998 2000 02 04


Fed funds futures rate; 3-months(left scale)

VIX(right scale)

Volatility of Fed funds*100(right scale)

0

200

400

600

800

1000

1200

1400Actual

Model

Aggregated EMBIG Spreads, Actual and Model(In basis points)

1998 2000 02 04

Sources: Bloomberg L.P.; JPMorgan Chase & Co.; The PRS Group; and IMF staff estimates.

6For reasons of comparability, we use the stan-dard deviations of the risk ratings for each country.

7We calculate the effects of the different variableson each predicted country spread and aggregatethe effects using EMBIG weights.

vicious circle resulting in a sharp drop in riskappetite. It would go without saying that all“warehouses” of financial risk—banks andnonbanks alike—would see negative repercus-sions as to their earnings capacity and their

capital base. Hence, the question is, how likelyis it that such a scenario will unfold anytimesoon?

Most recently, global imbalances have con-tinued to widen, with the U.S. current account

31


When decomposing the dynamics of themodel, we also find that lower global risk hasbeen a major reason behind the recent tighten-ing of EMBIG spreads, with the VIX contribut-ing as much as 280 basis points to the declinesince the beginning of January 2003. HigherFed funds futures rates have had an upwardeffect on spreads in the model (57 basis pointssince January 2003), but the downward effects ofstronger fundamentals and lower perceivedglobal risk have dominated the effects of tightermonetary policy. Our measure of uncertaintyabout monetary policy has declined somewhatsince January 2003, which in the model con-tributes to the drop in the aggregate EMBIGspread by 22 basis points. A separate regressionusing only the three external factors as inde-pendent variables suggests that much of thefluctuation in EMBIG spreads can be explainedby the external variables (see fourth figure).The “R-squared within” for the regression usingonly external factors is 0.41, and the estimatedcoefficients are of the same sign and order ofmagnitude as in the full model.

Conclusions

EMBIG spreads may currently be too finelypriced, as suggested by the fact that spreadstoward the end of our sample were lower thanwhat the model predicted. However, the differ-ence between predicted and actual spreads wasonly 40 basis points by December 2005.

The analysis suggests that improved own-country fundamentals can explain much of thetrend in EMBIG spreads, whereas external fac-tors capture a lot of the volatility in emergingmarket bond spreads. The two sets of explana-tory variables complement each other and seemto be of about equal importance for explainingEMBIG spreads since 1998. The model shows

that improved fundamentals cannot fully explainthe record tight EMBIG spreads. The low volatil-ity environment of global financial markets is acrucial element in explaining the dramatic tight-ening of EMBIG spreads since January 2003.

How resilient would emerging markets be incase of a deterioration in the external environ-ment? With the end of the Fed’s tighteningcycle drawing near, the main threat in the exter-nal environment is an increase in global finan-cial market volatility. Assuming the linear modelcontinues to be valid, an increase in VIX by twostandard deviations (13.2 points in the sample)would result in an increase in the aggregatedspreads by 140 basis points, other things beingequal. Given that such an external shock wouldbring spreads merely to their level of summer2004, it appears that EMBIG spreads are wellanchored in fundamentals.

0

200

400

600

800

1000

1200

1400

1600

Aggregated EMBIG Spreads, Actual and Predicted by Separate Regressions(In basis points)

Actual

External factors only

Fundamentalfactors only

1998 2000 02 04

Sources: Bloomberg L.P.; JPMorgan Chase & Co.; The PRS Group; and IMF staff estimates.

deficit reaching about 6.5 percent of GDP.Conversely, current account surpluses haveincreased, in particular in EM countries—inAsia and even more so among oil-exportingcountries. In fact, the current account surplusof the EM countries is estimated to be around$500 billion in 2005 and 2006. So far, increas-ingly globalized and flexible financial marketshave smoothly intermediated capital flowsfrom surplus to deficit countries, mainly theUnited States. At present, there seems to be awillingness in the rest of the world to accumu-late U.S. assets—without any visible risk pre-mium attached (see Box 1.6). Motivationsrange from central banks acquiring foreignexchange reserves, in part to avoid the appre-ciation of their local currencies, to private sec-tor investors buying U.S. assets for a variety ofreasons (including retirement needs). In fact,the share of U.S. assets in total wealth of theworld has increased from about 5 percent20 years ago to around 17 percent in recentyears—accounting for 40 percent of globalannual income.18

There are several reasons why the UnitedStates is so attractive to international capitalflows. Some reasons are more structural, oth-ers might well change with cyclical develop-ments.• In an unprecedented way and not

matched—at least not in full—by any otherregion, the United States has created large,deep, flexible, sophisticated, and by andlarge well-regulated financial markets thatoffer a wide range of assets to meet differ-ent needs. In addition, the U.S. dollar is theworld’s major reserve currency. As such,U.S. financial markets have become the des-tination of choice for excess savings. Suchsavings increasingly originate from EMcountries, whose financial markets have notdeveloped enough to supply the volumeand range of assets required to meet the


32

0

100

200

300

400

500U.S. governmentsU.S. equitiesEuropean governmentsEuropean equitiesJapanese governmentsJapanese equities

Figure 1.18. Asset Class Performance, 1990–2006(In local currency)


1990 92 94 96 98 2000 02 04 06

18Caballero, Farhi, and Gourinchas (2006). Thisdevelopment also reflects the overall decline in homebias observed in many countries.

needs of the investors.19 While financialmarkets in EM countries are being devel-oped, it will take a long time for them tocatch up as the U.S. markets are also evolv-ing. Thus, the advantage enjoyed by theU.S. financial markets is an important struc-tural feature in the allocation of global sav-ings, and it explains why other high-growthcountries in the OECD area, such asCanada and Finland, have not attractedmore capital flows.

• In parallel, and for a number of years now,the United States has enjoyed a highertrend growth rate than the euro area andJapan, which also have relatively well-developed financial markets. This growthdifferential helps to explain why the euroarea and Japan do not attract similar vol-umes of capital inflows, especially on a netbasis. Over time, a higher trend growth rateshould produce better returns on financialassets—thus rewarding international invest-ments in U.S. assets. Indeed, since 1990,U.S. equity and fixed-income markets haveperformed well compared with similar mar-kets in Europe and Japan (Figure 1.18).However, since 2003, the U.S. assets havenoticeably underperformed their Europeanand Japanese counterparts in both equitiesand bonds (Figure 1.19). Furthermore, thesubstantial growth and interest rate differ-entials in favor of the United States com-pared with Europe and Japan are expectedto narrow somewhat over the year ahead. Itis not clear whether international investors’expectations of future returns on U.S. assets

33


80

100

120

140

160

180

200

220U.S. equitiesU.S. governmentsEuropean equitiesEuropean governmentsJapanese equitiesJapanese governments

2003 04 05 06

Figure 1.19. Asset Class Performance, 2003–06(In local currency)


19Caballero, Farhi, and Gourinchas (2006) constructa model according to which the United States hasgood quality financial assets and strong growth poten-tial, Europe and Japan have good assets but slowgrowth, while the emerging market countries havestrong growth and demand financial assets but areunable to produce them domestically. In such amodel, an equilibrium could be established where theUnited States has persistent current account deficitsand capital inflows (largely from the EM countries),and the dollar tends to be stronger and interest rateslower than implied by traditional economic models.


34

Oil exporter earnings this year have addedsubstantially to other official accumulation ofmature market assets. Part of the additional oilexport proceeds appears to have gone, at leastinitially, into offshore bank deposits, predomi-nantly in U.S. dollars. Another part of the pro-ceeds may have gone into U.S. treasury andagency securities bought from U.K. dealers. Onbalance, some global financial prices may havebeen influenced by the flows, but the diversityof investments apparently undertaken arguesfor little risk to systemic stability.

Oil exporter earnings have added substan-tially to the pool of global savings managed byofficial agencies. Depending on the methodof estimation, officially managed assets of thelarge oil-exporting nations may have risen by$300–$450 billion last year, at a rate comparableto the annual accumulation by Japanese auth-orities through early 2004, and to the accu-mulation by the Chinese authorities throughmid-2005. Official asset buying, which hasstopped for Japan but is ongoing for China,may have led to an aggregate accumulation of$600 billion in officially managed assets in theglobal financial system during 2005.

Oil exporter assets in mature markets arenot fully reported, creating an understate-ment of official transactions. Chinese officialasset buying is more fully reported than that ofthe oil exporters, but together these officialflows may be significantly understated in theU.S. balance of payments, which includes a$195 billion estimate for official reserves accu-mulated in the United States during 2005,down from $395 billion in 2004 (see the table).Some of the error may be included in thereported sharp rise in private sector investmentin U.S. bonds.1

Reported U.S. securities flows have beencapturing a diminishing share of official invest-ments. Total Japanese investment, includingprivate and official flows, overstated officialinvestments in U.S. markets, but not by muchif an adjustment is made for normal privateflows (see first figure). Chinese purchases ofU.S. securities, in contrast, have been signifi-cantly less than reported reserve accumulation(see second figure). For the oil exporters, esti-mates of additional earnings, due to higher oilprices, are barely reflected in reported pur-chases in U.S. securities markets (see thirdfigure).2

Where have the oil earnings gone? Using allavailable information, the best answer may bethat the additional oil earnings may have goneinto offshore bank deposits, U.S. treasury andagency securities bought directly and through

Box 1.6. Petrodollar Recycling and Capital Flows into the United States

Note: The main author of this box is LarsPedersen.

1The balance of payments presentation includesadjustment to monthly securities reports for flowsthrough offshore financial centers and acceleratedprepayment of asset-backed principal, among otheradjustments.

2Oil producer asset accumulation is approxi-mated as the difference between spot oil prices anda two-year average that may be used for budgetingpurposes, yielding an implied official accumulationrate of $200–$300 billion during 2005. Indepen-dently, staff estimates, based on World EconomicOutlook estimates (built-up from expert desk analy-sis and available official information), put thechange in oil exporter asset accumulation at$300 billion between 2003 and 2005.

U.S. Balance of Payments (In billions of U.S. dollars)

2003 2004 2005

Current account balance –520 –668 –790

Private financing 241 273 595Net direct investment –73 –145 89Net corporate equity –79 –23 –49Net bonds 329 350 486

U.S. corporate 224 243 317U.S. treasury and agency 104 107 169Foreign bonds –42 –19 –22

Net bank flows and residual 65 91 69Official financing 278 395 195

U.S. treasury and agency 225 311 151Other 53 84 44

Sources: U.S. Bureau of Economic Analysis; and IMF staffestimates.

Note: Data for 2005 are through Q3, annualized.

35


U.K. dealers, and a variety of other investmentsand debt repayments.

Short-term flows. Oil producers have depositedsubstantial sums at Bank for InternationalSettlements (BIS) reporting banks, mostly indollars. Deposits of residents in oil-producingnations at BIS banks rose by $140 billion inthe year to September 2005. Over the sameperiod, deposits of all monetary authoritiesin BIS reporting banks (including those ofthe oil countries) denominated in dollarsrose by $110 billion. The largest offshorecomponent of these dollar flows ($80 billion)is not part of the U.S. balance of payments,although near-perfect arbitrage between off-shore and onshore funding markets meansthese deposits effectively support the value ofthe dollar exactly as would an onshore deposit.3

Long-term flows into U.S. securities through theUnited Kingdom. Reported oil exporter buyingof U.S. securities goes disproportionately intoU.S. treasury and agency bonds, with compara-

tively little going to corporate bonds and equi-ties. Recorded flows diminished, however, to amere $40 billion pace in the year through

EquitiesCorporate bondsAgency securitiesTreasury securities

Purchases of U.S. Securities by Japan(12-month cumulative purchases; in billions of U.S. dollars)

2000 01 02 03 04 05 06

Japan official reserves accumulation (yoy)

Sources: U.S. Treasury Department, Treasury International Capital System; Japan Ministry of Finance; and IMF staff estimates.

350

300

250

200

150

100

50

0

–50

Purchases of U.S. Securities by China(12-month cumulative purchases; in billions of U.S. dollars)

2000 01 02 03 04 05 06

350

300

250

200

150

100

50

0

–50

EquitiesCorporate bondsAgency securitiesTreasury securitiesChina official reserves accumulation (yoy)

Sources: U.S. Treasury Department, Treasury International Capital System; Bloomberg L.P.; and IMF staff estimates.

Note: Purchases of U.S. equities by China over the period were negligible.

EquitiesCorporate bondsAgency securitiesTreasury securities

Purchases of U.S. Securities by Major Oil Exporters(12-month cumulative purchases; in billions of U.S. dollars)

Oil producers’ official reserves accumulation (yoy)1

2000 01 02 03 04 05 06

350

300

250

200

150

100

50

0

–50

–100

–150

Sources: U.S. Treasury Department, Treasury International Capital System; Bloomberg L.P.; and IMF staff estimates.

1Implied from changes in oil prices versus U.S. reported total net foreign security purchases.

3See, for example, McCauley (2005).

relative to Europe and Japan are shapedmore by long-term or by shorter-term per-formances. However, it is likely that the rate-of-return argument in favor of capital flowsto the United States will weaken over time ifthe recent underperformance of U.S. equi-ties and bonds persists.20 Such underper-formance could result from a turning of

the credit cycle in the U.S. corporate andmortgage markets, and could start to feedon itself, leading to overshooting, if so-called “stop-loss strategies” were to kick in.Naturally, there is a risk that if this over-shooting occurs in an abrupt way, leadingto widespread losses and an increase inmarket volatility, it could trigger a disor-


36

November 2005. However, rising U.S. securitiesflows booked through the United Kingdominclude a suggestive jump in treasury andagency bond sales to $180 billion in the yearthrough November, from $120 billion a yearearlier. The surge in U.S. transactions with U.K.residents in those securities favored by oilexporters suggests a significant portion of oilexporter portfolio flows has gone into U.S.markets through U.K. dealers.

Direct investment flows. Based on staff estimates(using World Economic Outlook data), a slimincrease of oil exporter direct investment abroadof $20 billion in 2005 suggests limited increasesin direct investment, although several countrieshave reportedly used part of their export earn-ings to take private equity stakes in foreign com-panies, both in the United States and elsewhere.Widespread anecdotal evidence suggests a partof the recent growth in the venture capital andprivate equity funds is supported by inflows fromthe Middle East.

Debt reduction. Some oil exporters have takenthe opportunity to pay off debts, followingRussia’s repayment of Paris Club debt totaling$15 billion.

Local market investments. Finally, some of the oilsurpluses appear to have also gone into realestate and equity markets in the Middle East,and into the world gold market. Many stockmarkets in the Middle East have seen heavyinflows this year, which have contributed topushing the key indices to lofty levels. Local real

estate markets are also showing spectacular lev-els of price appreciation.4

Impact of Oil Recycling

Investments of oil funds, where identified,have been concentrated in dollar assets. In par-ticular, offshore dollar bank deposits haveincreased substantially, providing support forthe dollar. Risk-free interest rates may also havebeen kept lower than otherwise because of therelatively conservative investment profile ofsome of the oil funds. (See World EconomicOutlook, April 2006, Box 2.3, for an estimate ofthe impact of oil earnings on U.S. governmentbond yields.) Identified investment flows do not,however, account for all of the oil earnings.

Looking ahead, the wide range of investmentsopen to oil exporters may be a source of systemicstability. The diffuse and varied paths by whichthe oil funds find their way into mature marketsargues that no single investor is in a position totake sudden disruptive action, although a slowdiversification from dollar bank deposits intoother asset classes, including those denominatedin other currencies, may be expected.

Box 1.6 (concluded)

4These investment options, of course, do notdirectly recycle windfall oil earnings back intomature financial markets. That can only happenafter an official investor buys regional assets, and thesellers then deposit the payment into mature finan-cial markets or use the funds to pay for increasedimports from the mature market countries.

20In fact, some prominent U.S. mutual funds have recently announced that they will reduce exposures to U.S.equities in favor of international equities.

derly change in the flow of capital to theUnited States.While the motivations of international

investors are not fully understood, it is reason-able to say the following:• Accumulation of U.S. dollar assets by the

rest of the world at the present scale cannotgo on forever—at some juncture, a satura-tion point would be reached after whichinternational investors would want to diver-sify their portfolios. From the U.S. perspec-tive, its external debt dynamic wouldbecome unsustainable.

• However, the international accumulation ofU.S. assets could go on for some time, if theunderlying regional pattern of saving andinvestment and the differences in the levelof development of international financialmarkets persist.

• Since such characteristics of market struc-tures change only slowly—absent a majorglobal shock or politically motivated assetreallocation—it would be highly speculativeto expect the current pattern of asset pref-erences to change abruptly and sharply inthe near future, even though such a possi-bility exists.21 Given the strong ties of anumber of Middle Eastern surplus countriesto the U.S. dollar, a shift of global currentaccount surpluses to this region would notsuggest any abrupt changes (see Box 1.6 inthis issue of the GFSR and Chapter II of theApril 2006 World Economic Outlook).Nevertheless, the expected narrowing of

favorable growth and interest rate differentials(for the United States) is likely to lead tosome moderation in the pace of foreign accu-mulation of U.S. assets. This would weakenthe dollar and push U.S. bond yields up some-what. Of course, there is a risk of market over-shooting during the transition period. If theprospect of more balanced growth around theworld becomes firmer, helping to reduce

global imbalances, financial markets could“front run” such developments by reallocatingassets away from the United States to areassuch as Europe and Japan. This could sharplyweaken the dollar and push up interest rates.

More important, sudden and negativedevelopments—such as military confronta-tion, major terrorist attacks, a sharp fall inthe supply of crude oil or other vital sourcesof energy, and maybe, more realistically, a sig-nificant rise in protectionism—could changethe rational framework for global asset alloca-tion and trigger a disorderly unwinding ofglobal imbalances. These uncertainties, how-ever, are difficult, if not impossible, to quan-tify. In a similar vein, and equally impossibleto quantify, would be an outbreak of the avianflu pandemic. Such an event could have a seri-ous disruptive effect on international financialsystems—especially the payment clearing andsettlement system—and the global economy(see Box 1.7).

Assessment of Financial Stability andPolicy Implications

To sum up, in the base case scenario of con-tinued growth and contained inflation, finan-cial systems—from a position of financialstrength—should be able to cope with theenvisaged cyclical risks rather well. For exam-ple, credit spreads would widen and volatilitywould increase only moderately, and marketadjustments would generally remain orderly.Obviously, if the unwinding of the variousimbalances were to signal lower-than-expectedgrowth, markets would react more sharply;but there is little evidence from the aboveanalysis to suggest that the expected or likelymarket corrections in the period ahead wouldlead to crises of systemic proportions.

The most likely cyclical setting for financialmarkets in 2006 could be defined as “not bad,

3737

ASSESSMENT OF FINANCIAL STABILITY AND POLICY IMPLICATIONS

21Central Banking Publications (2006) recently reported the results of a survey of 56 central banks holding $1.9trillion of reserves, concluding that “reserves managers did not indicate a strong desire to switch out of U.S. dollarholdings.”


3838

An avian flu pandemic could affect the globalfinancial system through (1) operational disrup-tions caused by a sharp increase in workerabsenteeism in the financial industry and (2)market disruptions and changes in capital flowsresulting from an increase in risk aversion. Themagnitude and duration of these disruptionswould depend on the severity of the pandemicand on the degree of preparedness.1

Operational Risks

Operational risks constitute the single largestset of risks in the event of a pandemic. Highabsenteeism could result in disruptions of thecritical functions and services of the financial sys-tem, including payments, clearing and settle-ments, trading, and IT and communicationinfrastructure. Such operational disruptionscould prevent transactions from being completedand obligations from being met. Moreover, dis-ruptions in one jurisdiction could spread to otherjurisdictions, leading to disorderly changes inasset prices and capital flows in countries not yetdirectly affected by the avian flu pandemic.

In recent years, financial institutions, centralbanks, and regulators have adopted businesscontinuity plans (BCPs) to deal with terrorismand natural disasters. However, planning for apandemic, with widespread absenteeism andhealth concerns, has been limited. A number oflarge financial institutions have now extendedtheir plans to deal with an avian flu pandemicby identifying noncore activities, and planningfor work from home, heavy demand for cash bythe public, and transport of key personnelwhose functions cannot be done from home.However, the level of preparedness variesgreatly across financial institutions and amongnational authorities.

Market Disruption Risks

The outbreak of avian flu could threatenglobal financial markets. A sharp increase in riskaversion could occur, resulting in increases indemand for liquidity, specifically for cash and forlow-risk assets. This “flight to quality” would leadto declines in equity values and a widening ofcredit spreads for both corporations and emerg-ing markets. Given that an avian flu pandemicwould be expected to spread rapidly around theworld, similar asset price adjustments are likelyto occur across regions. Although these effectsare likely to be temporary, and their magnitudewill depend on the severity of the pandemic,such asset price declines could put the balancesheets of some financial institutions under stressand challenge their ability to comply with regula-tory norms. Market disruptions could becomemore disorderly in the case of a breakdown inthe market infrastructure leading to limitedand/or intermittent trading.

A pandemic may also lead to a significant buttemporary reduction in net capital flows toemerging markets. Some capital flight from res-idents seeking safe havens could be expected.Based on the SARS experience, FDI plans maychange little, although the timing of majorinvestments may be postponed. A shift in riskpreferences could lead to modest portfoliooutflows, particularly from members with rela-tively high-priced equities, or with weaker pub-lic finances, and/or with current accountshighly dependent on commodity prices andexport of services. However, to varying degrees,members would be able to address temporarybalance of payments pressures by drawing onreserves, which for many countries are at ahistorical high.

Policy Responses

To minimize financial sector disruptions,national authorities, including regulators,should take preemptive measures and provideguidance on the contents of BCPs, outliningbest industry practice. They should reviewBCPs for adequacy and consistency. In addition,all banks and national authorities should test

Box 1.7. Financial Implications of an Avian Flu Pandemic

Note: The main authors of this box are CharlesBlitzer and David Hoelscher. The box was jointlyprepared by the International Capital MarketsDepartment and the Monetary and FinancialSystems Department of the IMF.

1See World Economic Outlook (April 2006) for adiscussion of the potential impact of a pandemicon the real economy and on global trade.

but not as good as the stellar year 2005.” Theearnings capacity for systemically importantfinancial intermediaries might be inferior tolast year’s, especially if default rates and creditspreads were to rise somewhat, equity marketswere to become softer, and flat yield curveswere to make carry trades of all kinds moredifficult. But it is difficult to make a case thatrealistic economic developments all by them-selves could—at least over a 6–12 month timehorizon—seriously affect the global financialsystem in a systemic way.

If cyclical changes were indeed to exposesome weaker, but nonsystemic, spots, suchas idiosyncratic credit risks or poor risk man-agement in a number of individual cases,from a macroprudential point of view thepresent time is an ideal occasion to let theself-correcting forces of the market work outprice dislocations by adhering to a strict no-bailout policy. Regulators and supervisorswould have an almost unique opportunity to

contain complacency and ultimately moralhazard. If certain investors, such as hedgefunds, were to experience difficulties, it wouldbe important to “let nature take its course”and let individual investors suffer losses. Thetransfer of risk, at least in part, to the house-hold sector has somewhat changed the natureof moral hazard from “too big to fail” forsome key financial institutions to whole mar-ket segments being “too important to fall.”22

Hence, the combination of solid resilience ofthe financial system in general and somepotential weak spots here and there exposedby cyclical deteriorations present authoritieswith an occasion to reinforce market disci-pline as a key parameter for the pricing of var-ious risks. At a time when markets maybecome weaker, especially in pockets of highvaluations and low risk premiums, it is impor-tant to let investors experience a two-way mar-ket. During a long period of stability,especially if accompanied by a generational

3939

ASSESSMENT OF FINANCIAL STABILITY AND POLICY IMPLICATIONS

their BCPs, assessing whether essential func-tions can continue over a sustained period inthe event of potentially very high absenteerates, ensuring that back-up equipment, datacenters, and telecommunication networks areadequate to deal with a surge in remote accessactivities (including online banking and workfrom home) and that institutions can meetsharp increases in liquidity preference. Somecountries have established emergency commit-tees composed of government, public health,central bank, and regulatory officials for deci-sion making in the event of an avian flu pan-demic. Finally, a communications strategy, bothwithin the country and with international con-tacts, should be developed. A good communica-tions strategy will help minimize marketoverreaction.

Authorities should be prepared to accommo-date a surge in liquidity demand and shock-related price increases. In particular, centralbanks will need to ensure adequate supplies ofcash notes and the capacity to deliver them. Tocalm markets and avoid forced selling intofalling markets, financial regulators may have toconsider a degree of prudential forbearance.For instance, liquidity requirements, solvencyregimes, audit regulations, and provisioningrequirements could be temporarily eased, andaudit requirements could be adjusted for awork-from-home environment. To contain assetprice deflation, regulators, including those regu-lating the insurance and pension industries, mayhave to consider temporary forbearance whereprudential limits are breached by an initialdecline in asset prices.

22See IMF (2005).

change among traders, asset managers, andrisk officers, the appreciation for the publicsector’s “tough love” approach—no bailout inany shape or form in case of financial risksmaterializing—might get lost and merits areminder.

The structural changes in the financial sectorlaid out above permit authorities to behavemore robustly in the case of asset repricings.With the exception of the mentioned “tailrisks,” one can expect the self-correctingforces of financial markets, especially theincreasing diversity of investors and theirinvestment horizons, to take care of mostrisks. Apart from the regular microeconomic“supervisory vigilance,” complemented by anumber of more specific and detailed recom-mendations mentioned below, and apart alsofrom the macroeconomic policy advice con-tained in the World Economic Outlook,23 it ishard to see what other macroprudential poli-cies could be recommended to further bolsterthe resilience of global financial stability. Asusual, of course, macroeconomic policies thataim for solid and well-balanced global growthalso underpin the strength and resilience ofthe financial system.

Capital markets—especially for new prod-ucts such as credit derivatives—have becomeever more important in financial intermedia-tion. As a consequence, and rightly so, inter-national financial regulatory efforts havefocused on improving standards for disclo-sure, transparency, corporate governance, andrisk management so as to help markets func-tion smoothly. Against that background, theproper policy focus at present should not beon developing new regulations that could sti-fle innovations, but rather on effective surveil-lance and supervision—in particular theimplementation of already-announced meas-ures, such as Basel II and the new interna-

tional financial accounting and reportingstandards. Given the complexity of financialmarkets and products, supervisory authoritiesneed to continually narrow the gap betweenthemselves and the ever-advancing financialmarkets; this is the most crucial prerequisitefor making informed and timely assessmentsof risks and vulnerabilities in financialsystems.

Supervisors should specifically encouragefurther improvements in the robustness ofmarket infrastructures. The recent supervisoryinitiative (see Chapter II) to prod market par-ticipants to take steps to reduce the confirma-tion backlog in credit derivative trades is awelcome example. Also to be welcomed arevarious private sector market initiatives,including work to develop a cash settlementprotocol for credit derivative contracts, recom-mendations by the Counterparty RiskManagement Policy Group II, the Group ofThirty (G-30) Plan of Action for GlobalClearing and Settlement dating back to 2003,and the G-30 study on “Reinsurance andInternational Financial Markets.” All recom-mendations should be implemented by mar-ket participants without delay. Marketparticipants and supervisors alike should con-tinue to improve corporate governance, dis-closure, transparency, and the integrity offinancial statements to reduce the scope for“surprises” about the true health of corporatebalance sheets—it was largely surprises, orunanticipated shocks, that caused many crisesin the past.

As the GFSR has emphasized in the past,sound macroeconomic policy, especially pru-dent fiscal policy and flexible exchange rates,is essential to reduce vulnerabilities. At thesame time, the recent evidence from EMcountries clearly points to active debt manage-ment programs as playing a significant role in


4040

23According to the April 2006 World Economic Outlook, “a coordinated package of policies across major regions—including measures to reduce the budget deficit and spur private savings in the United States; structural and otherreforms to boost domestic demand in surplus countries; and greater exchange rate flexibility in China and someother countries to allow necessary appreciations to take place—could significantly reduce risks.”

reducing vulnerabilities. The types of opera-tions that a wider group of countries couldconsider include using reserves above pruden-tial requirements to buy back external debt,market-based exchanges of foreign currencydebt for local currency debt, and graduallengthening of the yield curves even whenthey are modestly upward sloping. A numberof countries could benefit from improvedinvestor relations programs, includingthrough enhanced data transparency. And,finally, low-income countries, which are justbeginning to attract significant interest fromforeign investors, will need to consider howbest—including through avoidance of over-borrowing and development of debt manage-ment programs—to protect their balance ofpayments and financial sectors from perhapsswift changes in investor sentiment.

Emerging market countries, as a group andcertainly the large, systemically importantones, have better access to internationalinvestors than ever before, be it on interna-tional markets or increasingly in domesticmarkets. Again, emerging market countries,as a group and mostly large ones, have beennet exporters of capital to the rest of theworld for six consecutive years and manycountries have reduced the supply of theirexternal sovereign debt, and therefore canfacilitate access to international capital mar-kets by EMs with current account deficits.Idiosyncratic risk events could still materialize(i.e., small countries with weak fundamentalscould experience funding difficulties) but willnot easily trigger financial contagion thatwould threaten to curtail emerging marketcountries’ access to international capital mar-kets in general.

In addition to more traditional cyclical andstructural risks, financial systems at the globaland national levels are also vulnerable to eventrisks such as terrorism or an avian flu pan-demic. Such events have the potential to dis-rupt the normal operations of financialsystems and to undermine market confidence,with negative consequences for financial stabil-

ity. Largely because the probability and theseverity of these risks are highly uncertain, riskmanagers generally have paid less attention inassessing their balance sheet risks. On theother hand, event risk is an important compo-nent of business continuity planning and,since 2001, significant progress has been madeby financial institutions and national authori-ties to plan for terrorist events. However, con-cerns about an avian flu pandemic, whichwould create somewhat different challenges,are just coming to the fore. Such a pandemiccould affect the global financial system throughoperational disruptions caused by a sharpincrease in worker absenteeism, a surge indemand for liquidity and cash, and a declinein asset prices (see Box 1.7). Policymakerswill need to ensure that their financial sys-tems are adequately prepared for such dis-ruptions so that core financial services remainoperational.

As an urgent matter, regulators shouldprovide guidance on the content of businesscontinuity plans. They should ensure throughtesting that essential functions can continuein the event of significant absentee rates, thatinstitutions can meet a sharp increase inliquidity preference, and that there is ade-quate international coordination. Finally, acommunications strategy should be developedin advance of an avian flu pandemic to helpminimize market overreaction. In the eventof a pandemic, central banks will have toconsider the appropriate extent of theirliquidity operations, and regulators will haveto determine an appropriate degree of pru-dential forbearance.

Appendix: Emerging MarketFinancing Flows

Private capital flows to emerging marketsare estimated to have hit record highs in2005, surpassing the previous record of flowsin 1996 before the Asian crisis. The flowshave been supported by strong foreign directinvestment, record issuance in global primary

4141

APPENDIX: EMERGING MARKET FINANCING FLOWS

markets (in both external and local curren-cies), record issuance of syndicated loans,and strong flows into local currency second-ary markets (for both bonds and equities).This appendix reviews these developments.

Primary Issuance in External Markets

New issuance in emerging markets hit his-toric highs both in gross and in net terms in2005 (Figure 1.20 and Table 1.4). Grossannual issuance of bonds, loans, and equitieswas $406.4 billion in 2005, far surpassing thelevel of 2004 ($286.9 billion), which was itselfa record. The majority of new issuance was inbonds, as emerging market sovereigns andcorporates took advantage of unusually favor-able external financing conditions. Syndicatedloans also increased substantially, reflectingincreased activity by commercial banks inemerging markets. However, new issuance ofequities, starting from a lower base, grew themost in relative terms.

Gross bond issuance was at a record high in2005, reaching $182.2 billion and exceedingthe previous record for 2004 by 35 percent.Issuance was relatively evenly split betweensovereigns and corporates ($83.6 billion ver-sus $77.1 billion), and both grew substantiallyfrom their level in 2004. On a regional basis,Asia, Europe, and Latin America all sawgrowth, but Latin America saw the biggestincrease in issuance (78 percent), related tothe return of Argentina to capital marketsafter its debt restructuring early in the year(Figure 1.21). On a net basis, bond issuancewas also at a record high, reaching $116.8 bil-lion, surpassing the previous high of $96 bil-lion in 1997 (Figure 1.22). New net issuancewas evenly split between sovereigns and cor-porates. The most significant regional devel-opment was the very large increase in netissuance in Latin America in 2005 comparedwith previous years, again related to Argen-tina’s return to the bond markets. Marketpractice for sovereign issuance of bonds hasconverged to a broad acceptance of collective


4242

0

100

200

300

400

500

DecNovOctSepAugJulJunMayAprMarFebJan

2005

2004

2000

2002

2003 2001

Figure 1.20. Cumulative Gross Annual Issuance of Bonds, Loans, and Equity (In billions of U.S. dollars)

Source: Dealogic.

action clauses (CACs) in international issuesunder New York law (Box 1.8, p. 46).

Net issuance by corporates in internationalmarkets doubled in 2005, reflecting increasedrisk appetite by investors and a continuedreleveraging of balance sheets by emergingmarket corporates after the financial crises of

the late 1990s and early 2000s. The increasein net issuance by corporates started in 2003and has accelerated significantly since then.By region, Asia and Europe, Middle East, andAfrica (EMEA) dominate, with Korean andChinese firms (largely trading companies andbanks) dominating in the former, and

4343


Table 1.4. Emerging Market External Financing(In billions of U.S. dollars)

2004 2005 20061_____________________ ______________________ ______2000 2001 2002 2003 2004 2005 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Jan.

Gross issuance by asset 216.4 162.1 135.6 199.7 286.9 406.4 71.6 63.4 70.2 81.8 94.1 87.9 106.6 117.8 18.4Bonds 80.5 89.0 61.6 99.8 135.5 182.2 40.9 29.1 34.5 30.9 61.3 37.2 40.4 43.4 15.4Equities 41.8 11.2 16.4 27.7 45.2 78.2 13.9 10.2 5.6 15.5 10.5 17.4 23.0 27.3 1.7Loans 94.2 61.9 57.6 72.2 106.2 145.9 16.8 24.0 30.1 35.3 22.3 33.3 43.1 47.2 1.4

Gross issuance by region 216.4 162.1 135.6 199.7 286.9 406.4 71.6 63.4 70.2 81.8 94.1 87.9 106.6 117.8 18.4Asia 85.9 67.5 53.9 88.8 123.7 150.4 34.1 28.5 25.8 35.4 26.7 33.5 40.7 49.5 5.9Latin America 69.1 53.9 33.4 43.3 54.3 86.2 14.4 9.7 16.2 13.9 34.1 13.8 22.6 15.7 3.5Europe, Middle East, Africa 61.4 40.8 48.3 67.7 109.0 169.8 23.1 25.2 28.3 32.5 33.3 40.7 43.2 52.6 9.1

Amortization by asset 113.9 147.1 128.5 119.5 128.1 108.0 35.0 32.8 30.6 29.7 21.7 26.2 32.6 27.5 9.4Bonds 51.8 59.2 59.0 57.1 69.6 65.4 21.5 17.5 15.9 14.7 13.4 14.6 21.6 15.8 5.0Loans 62.1 88.0 69.5 62.4 58.5 42.6 13.5 15.3 14.7 15.0 8.3 11.6 11.0 11.7 4.3

Amortization by region 113.9 147.1 128.5 119.5 128.1 108.0 35.0 32.8 30.6 29.7 21.7 26.2 32.6 27.5 9.4Asia 56.6 66.2 55.7 45.5 49.8 38.7 13.2 12.9 11.8 11.8 8.1 5.9 11.4 13.4 3.8Latin America 32.3 45.6 40.8 40.4 46.7 37.1 12.3 13.4 10.2 10.9 7.7 10.4 11.1 7.9 3.9Europe, Middle East, Africa 24.9 35.3 32.0 33.6 31.6 32.2 9.5 6.6 8.6 7.0 5.9 9.9 10.1 6.3 1.6

Net issuance by asset 102.5 15.0 7.2 80.2 158.8 298.4 36.6 30.6 39.6 52.1 72.4 61.8 74.0 90.3 9.1Bonds 28.7 29.9 2.6 42.7 65.9 116.8 19.4 11.6 18.6 16.2 47.9 22.6 18.8 27.6 10.3Equities 41.8 11.2 16.4 27.7 45.2 78.2 13.9 10.2 5.6 15.5 10.5 17.4 23.0 27.3 1.7Loans 32.1 –26.1 –11.8 9.8 47.7 103.3 3.3 8.7 15.4 20.3 14.0 21.7 32.2 35.5 –2.9

Net issuance by region 102.5 15.0 7.2 80.2 158.8 298.4 36.6 30.6 39.6 52.1 72.4 61.8 74.0 90.3 9.1Asia 29.2 1.3 –1.8 43.3 73.9 111.7 20.8 15.6 13.9 23.6 18.6 27.6 29.4 36.1 2.0Latin America 36.8 8.3 –7.4 2.9 7.6 49.1 2.1 –3.6 6.0 3.0 26.4 3.4 11.5 7.8 –0.4Europe, Middle East, Africa 36.5 5.5 16.3 34.0 77.3 137.6 13.6 18.6 19.6 25.5 27.4 30.8 33.1 46.3 7.5

Secondary markets

BondsEMBI global

(spread in basis points) 735 728 725 403 347 237 414 482 409 347 373 297 235 237 210Merrill Lynch high-yield

(spread in basis points) 890 795 871 418 310 371 438 404 384 310 352 385 354 371 342Merrill Lynch high-grade

(spread in basis points) 200 162 184 93 83 92 94 97 91 83 93 95 89 92 90U.S. 10-year treasury

yield (percent) 5.12 5.05 3.82 4.25 4.22 4.39 3.84 4.58 4.12 4.22 4.48 3.92 4.33 4.39 4.52

(In percent)

EquityDOW –6.2 –7.1 –16.8 25.0 3.1 –0.6 –0.9 0.8 –3.4 –1.9 –2.7 –2.2 2.9 1.4 1.4NASDAQ –39.3 –21.1 –31.5 50.5 8.6 1.4 –0.5 2.7 –7.4 1.9 –8.2 2.9 4.6 2.5 4.6MSCI Emerging Market –31.8 –4.9 –8.0 51.2 22.4 30.3 8.9 –10.3 7.4 –0.2 1.8 3.0 17.0 6.8 10.9

Asia –42.5 4.2 –6.2 46.1 12.2 23.5 7.6 –12.2 4.2 –0.5 2.9 2.8 8.5 8.6 7.5Latin America –18.4 –4.3 –24.8 66.7 34.8 44.9 6.2 –9.2 16.6 –1.1 1.9 7.1 29.5 2.7 17.0EMEA –22.3 –20.9 4.7 51.9 35.8 34.9 13.2 –7.4 7.8 1.0 –0.4 0.5 27.1 6.6 13.5

Sources: Bloomberg L.P.; Capital Data; JPMorgan Chase & Co.; Merrill Lynch; Morgan Stanley Capital International; and IMF staff estimates.1Issuance data are as of January 31, 2006, close-of-business, London. Secondary market data are as of January 31, 2006, close-of-business,

New York.

Russian banks and oil companies in thelatter.

A large part of the issuance of bonds by sov-ereigns in 2005 represented prefinancing of2006 sovereign external issuance needs.Including issuance during the first month of2006 (about $8 billion), total financing alreadyundertaken to meet 2006 needs is estimated at$24.6 billion, against a total sovereign plannedissuance of $54 billion. Emerging market sov-ereigns had thus met almost half of theirissuance needs for the year by the end ofJanuary 2006. On a regional basis, most LatinAmerican countries had completed 100 per-cent of their issuance needs, and a number ofsovereigns were beginning to prefinance 2007,while emerging Asia had met 24 percent andemerging Europe about 37 percent.

Syndicated loan commitments also hit arecord high in 2005 (Figure 1.23). In grossterms new syndicated loans reached $145.9billion, substantially above the previous peakof $122 billion in 1997. Regionally, the largestincrease in syndicated lending has been chan-neled to the Middle East, which more thandoubled from $11.5 billion in 2004 to $36.3billion in 2005. European emerging marketsalso had a large increase, related to financingof Russian corporates. In general, cross-borderlending to emerging market corporates hasincreased substantially as commercial bankshave sought to build relationships in thesecountries to take advantage of expandingbusiness opportunities in other fee-generatingareas. In addition, some of the activity isrelated to the refinancing of previous projects,under which banks recycle the same moneywith the same clients at lower margins and alonger tenor.24 However, net issuance figuresshow that even taking this effect into account,flows were at historical highs, with net inflowsof syndicated loans in 2005 at $103 billion wellexceeding the previous high of $82 billion in1997.


4444

EMEALatin AmericaAsia

0

10

20

30

40

50

60

70

1994 96 98 2000 02 04

Figure 1.21. Gross Bond Issuance (In billions of U.S. dollars)

Source: Capital Data.

LoansEquitiesBonds

–20

0

20

40

60

80

100

Total

1997 99 2001 03 05

Figure 1.22. Quarterly Net Issuance (In billions of U.S. dollars)

Sources: Dealogic; and IMF staff estimates.

24Institute of International Finance (2006).

Equity issuance grew the most out of all pri-mary capital flows to emerging markets, rising73 percent to $78.2 billion in 2005 over thatof 2004 (Figure 1.24). As in past years, equityissuance was dominated by Asian countries,and in particular China, where initial publicofferings raised over $21 billion. ChinaConstruction Bank’s IPO, which took place inOctober, was the largest ever initial publicoffering by a bank and the largest IPO since2001. European equity issuance followed a dis-tant second, dominated by Russian IPOs. InLatin America, IPOs have been relatively moreactive in Brazil and Mexico.

Foreign Investor Flows into Local Markets

International investor interest in local mar-kets, in both equities and bonds, continues toaccelerate, pushed by the search for yield anddiversification. Investment in local marketinstruments has been facilitated by improvingfundamentals in many emerging market coun-tries, as well as the inclusion of local currencygovernment bonds from emerging markets instandard benchmark indices (the LehmanAggregate, for example) and the developmentof new index products (such as the JPMorganGBI-EM index) designed specifically to bench-mark funds investing in local market govern-ment bonds.25

Initially the move into higher-yielding localmarkets was spearheaded by speculativeinvestors. However, dedicated emerging mar-ket investors are increasingly setting up localmarket funds to invest in local cash bond mar-kets. Fixed-income investor surveys suggestthat investors are raising their benchmarkexposure to emerging market local debt to 30percent of assets. At the margin, this hasrequired that as much as 80 cents per dollarof new money be dedicated to buying local

4545


EMEA

Latin America

Asia

0

10

20

30

40

50

1994 96 98 2000 02 04

Figure 1.23. Syndicated Loan Commitments (In billions of U.S. dollars)

Source: Dealogic.

EMEA

Latin America

Asia

0

5

10

15

20

25

30

1994 96 98 2000 02 04

Figure 1.24. Equity Placements (In billions of U.S. dollars)

Source: Dealogic.

25For more on the drivers for local market invest-ments, see Chapter III. In addition, see IMF (2005,Box 2.3): “Foreign Investment in Local CurrencyInstruments: A Cyclical or Fundamental Phenomenon?”

currency investments. These purchases arelargely limited to the cash market because ofinvestor mandates that restrict the use ofderivative products.

The issuance of local currency externalbonds continues to be an important develop-ment. Many local currency bond markets aredifficult to access because of high transaction


4646

Market practice has converged toward broadacceptance of the use of collective actionclauses (CACs) in international sovereignbonds issued under New York law. With onlytwo exceptions, all sovereigns that have issuedunder New York law since May 2003—26emerging market countries and one maturemarket country—have included CACs in theirbonds (see the table below). In 2005, morethan 95 percent of new issues, in value, includedCACs, while the share of bonds with CACs ofthe outstanding stock of sovereign bonds ofemerging market countries climbed to around57 percent as of January 23, 2006.

Since September 2005, emerging marketcountries have continued with their established

practice of including CACs in their interna-tional sovereign bonds issued under NewYork law. Three emerging market issuers—Ecuador, Iraq, and Vietnam1—included theseclauses in their bonds for the first time, while12 other emerging market issuers—Brazil,Colombia, Indonesia, Korea, Lebanon,2

Panama, Peru, the Philippines, Poland, Turkey,Uruguay, and Venezuela—continued theirpractice of including these clauses in theirbonds governed by New York law. Italy was theonly mature market country to issue underNew York law. Jamaica was the only country

Box 1.8. Collective Action Clauses

Emerging Market Sovereign Bond Issuance by Jurisdiction1

2003 2004 2005 20062______________________ ________________________ ________________________Q1 Q23 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q24 Q3 Q4 Q1

With CACs1

Number of issues 9 31 10 5 25 19 19 15 18 38 12 24 6Of which:

New York law 1 22 5 4 14 12 12 13 11 22 10 19 5

(In billions of U.S. dollars)

Value of issues 5.6 18.0 6.4 4.3 18.5 15.9 10.7 9.1 22.3 34.1 11.9 15.3 8.5Of which:

New York law 1.0 12.8 3.6 4.0 10.6 9.5 6.5 7.7 11.1 19.2 10.4 12.8 7.3

Without CACs5

Number of issues 14 4 7 7 2 1 1 4 0 1 1 1 0

(In billions of U.S. dollars)

Value of issues 8.1 2.5 3.5 4.2 1.5 0.1 0.2 2.7 — 0.3 1.3 0.3 0.0

Source: Dealogic.1English and Japanese laws, and New York law where relevant.2Data as of January 23, 2006.3Includes issues of restructured bonds by Uruguay.4Includes issues of restructured bonds by Argentina and the Dominican Republic in their respective debt exchanges.5German and New York laws (includes reopenings of previously issued bonds without CACs).

Note: The author of this box is Luisa Zanforlin.

1The Vietnam bond includes only majorityrestructuring provisions.

2The Lebanon bond includes only majorityrestructuring provisions.

costs for both entry and exit. For example, anumber of Latin American countries (includ-ing Argentina and Colombia) require a mini-mum holding period in the country before

4747


that did not include CACs in its New Yorklaw–issued bonds.3

Following standard market practice, allissues under English and Japanese lawsincluded CACs. Emerging market issuersunder English law were Barbados, Hungary,Israel, Macedonia, Poland, Turkey, andUkraine, while the mature market issuerswere Austria and Sweden. Poland was the onlyissuer under Japanese law.

As before, the inclusion of CACs did nothave any observable impact on the price ofemerging market sovereign bonds governedby New York law. In recent discussions withthe IMF staff, market participants suggestedthat the favorable conditions in global finan-cial markets and generally sound economicpolicy stance of emerging market countrieshave made the inclusion of CACs in interna-tional bond contracts a neutral factor in thepricing of these bonds. They also noted that,since CACs have not played a major role inany restructuring of sovereign bonds gov-erned by New York law, the market has notbeen able to determine the value of the inclu-sion of CACs on bonds’ recovery prices.However, many market participants stressedthat CACs could still affect the pricing ofbonds if a sovereign debtor were to comeclose to a default or decide to use CACs in arestructuring of its bonds, particularly if themajority of this sovereign’s outstanding bondscontain CACs. In this vein, CACs could beviewed as an option in a bond that would bein-the-money only when a sovereign issuer fac-ing debt-servicing difficulties is near default.

0

100

200

300

400

500

0

300

600

900

1200

1500

2003 2004 2005

Brazil

Poland

TurkeyMexico

Colombia(right scale)

Figure 1.25. Nonresident Holdings of LocalGovernment Bonds(Rebased, January 2003 = 100)

Sources: National authorities; and IMF staff estimates.

–2

0

2

4

6

8

10

12

14

16

2004

Asia, excluding Japan

Latin America

EMEA

2005

Figure 1.26. Cumulative Net Inflows into Emerging Market Equity Funds (In billions of U.S. dollars)

Sources: Emerging Portfolio Fund Research, Inc.; and IMF staff estimates.

3Egypt did not include CACs in its September2005 bond issued under New York law, which isfully guaranteed by the United States withrespect to principal and interest.

local market investments can be liquidated,and others, such as Brazil, include foreignexchange transaction taxes for the liquidationof investments held within the country for lessthan 90 days. In addition, these countries andothers have capital gains and financial transac-tion taxes that make active trading in localbond markets expensive for foreigners. Oneway in which sovereign and corporate issuershave overcome this problem for their externalinvestor base is by issuing local currencybonds in global markets.

As more traditional investors move intoemerging market cash bond markets, therehas been a further search for yield in moreexotic local markets. African governmentbonds have become more sought after, forexample, with investor interest not just byspeculative money, but also from foreign deal-ers, in countries such as Kenya, Nigeria,Tanzania, Botswana, Zaire, and Malawi. Inaddition, interest in local corporate bondmarkets has increased, though these marketsare generally not very liquid.

Only incomplete data exist on foreigninvestor flows into local bond markets, butthey confirm evidence from investor surveysthat there is a significant new allocation intolocal markets. Data on holdings of local gov-ernment bonds by nonresidents show signifi-cant increases in Mexico, Brazil, Poland,Turkey, and Colombia, particularly starting in2004 and continuing through the end of 2005(Figure 1.25).

Foreign investment flows into emergingmarket equities have also increased signifi-cantly, pushing up the value of local stockindexes. With a paucity of marketable debt,Asia has traditionally dominated as a destina-tion for emerging market equity investors. Butflows into Latin America and EMEA have alsoaccelerated rapidly in 2005 (Figure 1.26). Atthe end of 2005, there was an acceleration ofequity investment flows into emerging Asiastock markets driven by global growth plays,the turn in the tech cycle, and expectations ofsome recovery of local demand in the region


4848

200

250

300

350

400

450

500

–10

0

10

20

30

40

50Weekly flows (left scale)1

Cumulative flows (from January 2004; left scale)MSCI Emerging Asia index (right scale)

2004 05 06

Figure 1.27. Flows into Asian Emerging Market Equities (In billions of U.S. dollars)

Sources: Bloomberg L.P.; and IMF staff estimates.1Includes flows into Indonesia, Korea, Philippines, Taiwan Province of China, and

Thailand.

–500

500

1500

2500

3500

4500

5500

60

80

100

120

140

160

180

200

220

2004 2005

Figure 1.28. Turkey: Foreign Investment in the Stock Market


Net monthly purchases (left scale; in millions of U.S. dollars)Cumulative net purchases (left scale; in millions of U.S. dollars)Equity index (December 2003 = 100; right scale)

4949

(Figure 1.27). Since the middle of 2004,equity valuations have moved in tandem withforeign investor inflows suggesting that for-eign investors are an increasingly importantsegment of the market.

Major equity markets in other regions havealso started to move in tandem with foreigninvestor flows, suggesting this may be a moregeneralized phenomenon for emerging mar-kets. Equity markets in Turkey, Brazil, andMexico all hit historical highs at the end of2005, with the data indicating that theseincreases in valuation were increasingly drivenby foreign investor inflows (Figures 1.28, 1.29,and 1.30).

The sharp rise in oil prices has also fueleda surge in equity markets in the Gulf Coop-eration Council countries. Among the moreimportant bourses in the region, the SaudiArabian equity index, which is the largestand most liquid, rose 105 percent in 2005and was up 15 percent by the end of January2006.

ReferencesBernanke, B., 2006, “Monetary Policy Report to the

Congress,” Testimony of U.S. Federal ReserveBoard Chairman before the Committee onFinancial Services, U.S. House ofRepresentatives, Washington, February 15.

Caballero, R.J., Emmanuel Farhi, and Pierre-OlivierGourinchas, 2006, “An Equilibrium Model of‘Global Imbalances’ and Low Interest Rates,”Department of Economics Working PaperNo. 06-02 (Cambridge, Massachusetts:Massachusetts Institute of Technology).

Central Banking Publications, 2006, RBS ReserveManagement Trends 2006 (London, February 13).

Ferguson, Roger W., 2005, “Asset Price Levels andVolatility: Causes and Implications,” remarksbefore the Banco de Mexico InternationalConference, Mexico City, November 15.

Goldman Sachs, 2005, “Is the Low Volatility RegimeComing to an End?” Global Economic Weekly,No. 05/40 (London, November).

Greenspan, A., 2005, “Economic Flexibility,”remarks before the National Italian AmericanFoundation, Washington, October 12.

REFERENCES

4000

6000

8000

10000

12000

14000

16000

18000

40

50

60

70

80

90

100

Bolsa stock market index (right scale)

Nonresident holdings(left scale; in billions of U.S. dollars)

2002 03 04 05

Figure 1.29. Mexico: Nonresident Holdings of Local Equities


0

5000

10000

15000

20000

25000

30000

35000

40000

–2000

–1000

0

1000

2000

3000

4000

5000

6000 Net monthly inflow (left scale; in millions of U.S. dollars)Cumulative inflow (left scale; in millions of U.S. dollars)Bovespa equity index (right scale)

2001 02 03 04 05

Figure 1.30. Brazil: Foreign Investment in the Stock Market



5050

Hawkesby, Christian, Ian Marsh, and IbrahimStevens, 2005, “Comovements in the Prices ofSecurities Issued by Large Complex FinancialInstitutions,” Working Paper No. 256 (London:Bank of England).

Hull, John, Mirela Predescu, and Alan White, 2004,“The Relationship Between Credit Default SwapSpreads, Bond Yields, and Credit RatingAnnouncements,” Journal of Banking and Finance,Vol. 28, pp. 2789–811.

Institute of International Finance, 2006, “CapitalFlows to Emerging Market Economies”(Washington), pp. 1–24.

International Monetary Fund, 2004, GlobalFinancial Stability Report, World Economic andFinancial Surveys (Washington, September).

———, 2005, Global Financial Stability Report, WorldEconomic and Financial Surveys (Washington,September).

———, 2006, World Economic Outlook, World Econo-mic and Financial Surveys (Washington, April).

Kashiwase, Kenichiro, and Laura Kodres,“Emerging Market Spread Compression: Is ItReal or Is It Liquidity,” IMF Working Paper(Washington: International Monetary Fund,forthcoming).

Lehman Brothers MBS, and ABS Research, 2006,“Interagency Guidance on Non-TraditionalMortgage Products” (January).

Lown, Cara, and Donald P. Morgan, 2004, “TheCredit Cycle and the Business Cycle: New Find-ings Using the Loan Officer Opinion Survey,”SIFR Research Report Series (Stockholm: SwedishInstitute for Financial Research, June 8).

McCauley, Robert, 2005, “Distinguishing GlobalReserves from Official Holdings in the UnitedStates,” Bank for International Settlements QuarterlyReview (September), pp. 57–72.

Zhu, Haibin, 2004, “An Empirical Comparison ofCredit Spreads Between the Bond Market andthe Credit Default Swap Markets,” BIS WorkingPaper No. 160 (Basel, August).

Documents

April 2006 Global Financial Stability Report: Chapter I. Global