Chapter

CHAPTER 1. ASSESSING RISKS TO GLOBAL FINANCIAL STABILITY

Author(s):
International Monetary Fund. Monetary and Capital Markets Department
Published Date:
April 2008
Share
  • ShareShare
Show Summary Details

Overall risks to financial stability have increased sharply since the October 2007 Global Financial Stability Report (GFSR). The crisis that originated in a small segment of the U.S. mortgage market has spread to broader cross-border credit and funding markets through both direct (via exposure to subprime mortgage markets) and indirect (via perturbations in banking and funding markets) channels. A broadening deterioration of credit is likely to put added pressure on systemically important financial institutions. The risks of a credit crunch have increased, threatening economic growth. In turn, the potential for spillovers to emerging markets has increased through funding channels and trade linkages.

Global Financial Stability Map

The global financial stability map (Figure 1.1) presents an overall assessment of how changes in underlying conditions and risk factors bear on global financial stability in the period ahead.1 Nearly all the elements of the map point to a degradation of financial stability, with credit and macroeconomic risks having deteriorated the most.

Figure 1.1.Global Financial Stability Map

Source: IMF staff estimates.

Note: Closer to center signifies less risk or tighter conditions.

Downside risks to the macroeconomy…

A significant increase in risks to financial stability stems from an increase in our assessment of macroeconomic risks. Since the October 2007 GFSR, concerns about the potential for a significant economic slowdown have been reinforced by a string of weaker-than-expected economic data and weaker confidence in the United States and other mature markets, underscored by a sharp dip in leading global growth indicators. The World Economic Outlook (WEO) baseline projection is for global growth to moderate to 3.7 percent in 2008. However at this juncture, the macroeconomic outlook is clouded by a great deal of uncertainty, and risks to the baseline case are skewed to the downside. The key risk to the economic outlook appears to be unfolding. In particular, the dislocations in credit and funding markets are beginning to restrict the overall provision and channeling of credit.

Downside macroeconomic risks that are concentrated in the U.S. economy have a significant impact on systemically important financial institutions that may spill over to global markets. Of particular importance for financial stability are the linkages between the real and financial sector, including the effects of credit or financial decelerators on the real economy, the extent of balance sheet adjustments, and the absorptive capacity of financial markets. Our analysis indicates that a contraction in the supply of private sector credit and market borrowings could bring a significant slowdown in U.S. output growth in the following several quarters, as some securitization markets are functioning poorly in the wake of the crisis and banks are seeking to repair their balance sheets (see the section entitled “Credit Squeeze or Credit Crunch?”). Europe is also at risk, given the size of bank losses and disruptions in bank funding and securitization markets.

…threaten a deeper and wider deterioration in credit beyond subprime mortgages, weakening the capital and funding positions of systemically important financial institutions.

The increase of macroeconomic risks contributes to raising our assessment of credit risks.2 This assessment reflects the potential for a sharper slowdown in U.S. and global growth, which, coupled with past credit indiscipline, has heightened strains on the capital of systemically important financial institutions.

Credit deterioration has widened beyond subprime mortgages, and mark-to-market losses have mounted as markets anticipate a more difficult economic and financial environment. Nonprime mortgage losses have continued to rise, while the credit performance of higher-quality residential mortgages, commercial mortgages, and consumer credit products has also begun to weaken (see the section entitled “Systemic Risks Have Risen Sharply”).

An area of specific concern is the leveraged segment of the corporate debt market. As flagged in prior GFSRs, weak credit discipline in the mortgage market had also figured in leveraged corporate financing in recent years, as reflected by elevated low-tier corporate debt issuance and the marked rise in covenant-lite loans, fewer creditworthy deals, and high leverage and price multiples in the leveraged buyout sector. Defaults have already begun to rise on U.S. and European high-yield corporate debt, albeit from historically low levels, as higher spreads and diminished liquidity have put pressure on stressed companies.

Difficulties faced by institutions that underwrite credit risk have exacerbated systemic concerns. Financial guarantors that sold credit enhancements on mortgage-related products containing subprime assets have come under pressure as losses on structured securities have mounted. This poses risks for the municipal bond market, where half of the market is insured by financial guarantors, and for banks and other markets that rely on insurance provided by financial guarantors.

Higher market and liquidity risks underscore the uncertainty surrounding economic and systemic spillovers…

Reflecting the exposure of systemically important financial institutions to credit markets and the potential rise in market losses, we have raised our assessment of market and liquidity risks (signifying higher risks to financial stability).3 Strains in interbank money markets have intensified since the October 2007 GFSR, and the composite indicator of funding and market liquidity risks indicates that pressures exceeded levels observed during the market turbulence in 1998. Coordinated central bank actions have eased some of the liquidity strains, but pressures in term money markets have recently intensified, reflecting growing concerns about counterparty credit risk. Meanwhile, volatility has continued to rise across major asset classes to a level comparable to earlier in this decade, reflecting uncertainty associated with the size and location of credit losses as well as valuations of structured products. This leaves financial institutions—most recently hedge funds—vulnerable to mutually enforcing funding and market liquidity spirals, in which investors sell assets to meet funding requirements, creating price declines, a loss of confidence, and further funding pressures (see Chapter 3).

…and risk appetite has continued to retrench, restricting flows of global capital and forcing a further deleveraging in the financial system.

Investor risk appetite has diminished partly owing to greater uncertainty over the economic outlook, but also in reaction to a loss of confidence in structured finance and a collapse in some funding markets, which has forced a broad deleveraging in the financial system and threatens a disorderly adjustment of markets and further strains on bank balance sheets.

Monetary policy easing has been offset by a tightening of financial conditions.

Since the October 2007 GFSR, real short-term interest rates have declined across a range of economies, owing to a combination of the easing in monetary policy and actions by global central banks. As a result of the weaker economic outlook, markets are pricing in even more monetary policy easing across a range of economies. However, the easing in monetary policy to date has been offset by the sharp repricing in credit and funding markets, resulting in slightly tighter monetary and financial conditions overall.4 The repricing has been triggered by tighter lending conditions across the major economies, making credit more difficult to access for corporates and households. Faced with the increasing probability of unintended balance sheet expansion and losses, banks have become increasingly reluctant to extend credit while securitization markets may remain impaired. Combined with widening spreads, this increases the risks to the economy of a credit crunch.

Emerging markets have so far been resilient, but strains are already evident in those economies most vulnerable to a repricing of credit risks and restricting of external funding.

Unlike past financial crises, emerging markets have remained relatively resilient, supported by solid fundamentals, prudent macroeconomic policies, and financial cushions built up over recent years. However, we have raised our assessment of emerging market risks, as the market turmoil has exacerbated vulnerabilities in a number of emerging markets—notably in some countries in emerging Europe that had relied excessively on foreign bank credit or wholesale funding to finance rapid domestic credit expansion (see the section entitled “Will Emerging Markets Remain Resilient?”).

The risk of potential funding pressures stemming from over-reliance on external portfolio inflows and bank loans was a key theme in the October 2007 GFSR (IMF, 2007a), and these risks have since become more pronounced. Broader emerging sovereign risks have also risen, albeit from historic lows, primarily due to deterioration in financial fundamentals. Markets are concerned that emerging economies will become increasingly linked to mature economies if the latter’s growth continues to slow.

Credit Deterioration—How Deep and Widespread?

The U.S. nonprime mortgage sector continues to deteriorate.5

As detailed in the April 2007 GFSR, the deterioration in the U.S. nonprime mortgage market initially reflected a combination of lax underwriting standards, “risk layering,” and adverse trends in employment and income in certain U.S. regions (IMF, 2007b).6 Since then, delinquency rates on subprime mortgage loans originated in 2005–06 have continued to rise, exceeding the highest rates recorded on any prior vintage (at comparable seasoning). Mortgages originated in 2007 are on track to perform even worse, based on their current trajectory. With declines in U.S. home prices, recent vintages will have lower (and possibly negative) equity cushions, a greater probability of becoming delinquent, and lower recovery rates on foreclosure. Within recent cohorts, the deterioration has been primarily associated with the least creditworthy borrowers defaulting on adjustable-rate mortgages (ARMs).7 Going forward, as initial “teaser” rates on ARMs expire, the rise in interest payments is likely to cause a further rise in delinquencies.8

Lax underwriting standards also played a role in higher-quality segments of the U.S. mortgage universe, but downward real estate prices and the employment rate are now the key drivers.

The same pattern of weakly performing recent vintages has emerged in higher-quality alt-A and nonagency prime (“jumbo”) sectors, although the degree of underperformance is much lower (Figure 1.2).9 Delinquencies on prime mortgages are more significantly driven by weakness in underlying economic fundamentals. 10 However, most prime borrowers have more equity cushion to withstand possible future headwinds, including interest rate resets. Even with the declines in nationwide home prices, on average, outstanding mortgage equity stands at 40 to 50 percent of home value on ARMs extended to prime borrowers, compared with less than 5 percent for subprime borrowers. Going forward, however, if home prices continue to fall and other macroeconomic fundamentals weaken, there is a risk of higher defaults on prime mortgages, especially on recent vintages. Reflecting the deterioration in the underlying collateral, prices have continued to slide on nonagency securitized mortgages (Figure 1.3).

Figure 1.2.Mortgage Delinquencies by Vintage Year

(60+ day delinquencies, in percent of balance)

Sources: Merrill Lynch; and LoanPerformance.

Figure 1.3.U.S. Mortgage-Related Securities Prices

Sources: JPMorgan Chase & Co.; and Lehman Brothers.

Note: ABX = an index of credit default swaps on mortgage-related asset-backed securities.

Some similar features are beginning to emerge in Europe, as housing cycles start to turn.

European housing and mortgage markets have unique characteristics that vary considerably from country to country. Signs of a downturn are becoming evident in certain European housing markets. Market pricing of property derivatives points to outright home price declines in the United Kingdom, following the U.S. trajectory with a one- to two-year lag (though with a shortage of participants seeking to take long positions, property derivative markets can be fairly illiquid, failing fully to reflect market views). In other over-extended markets (see Box 3.1 in the April 2008 WEO), industry analysts are also forecasting declines in home prices (Figure 1.4). In addition, in the United Kingdom a sizable share of mortgage loans face interest rates that will reset to higher levels this year, just at a time when lenders are tightening standards, adding another source of stress.11 Nevertheless, underlying collateral performance remains strong in Europe. As a result, recent prime delinquencies are trending in line or lower relative to prior vintages, and loss rates remain low. More conservative mortgage financing arrangements in European countries suggest effects of house price declines will likely be more muted than those in the United States.

Figure 1.4.U.S. and European House Price Changes

(Percent year-on-year)

Sources: Standard & Poor’s/Case Shiller; national authorities; and IMF staff estimates.

Note: Europe excluding the United Kingdom: unweighted average of Spain, Germany, Italy, Netherlands, Greece (from 1995), and Ireland (from 1997). Estimates are based on futures prices. Dashed lines are futures implied.

If growth slows in Europe, as predicted in the latest WEO, repossessions and write-offs will rise. Some analysts foresee a near doubling of repossessions in the United Kingdom, for example, pushing writedowns to 1.4 percent of total mortgages outstanding or around $32 billion, driven mainly by nonprime and high loan-to-value loans.12 Delinquency rates on UK nonconforming loans would therefore rise (Figure 1.5).

Figure 1.5.U.S. and UK Nonconforming Delinquencies by Mortgage Vintage Year

(In percent of balance)

Sources: Fitch Ratings; LoanPerformance; Merrill Lynch; and IMF estimates.

Note: UK delinquencies for 90+ days; U.S. delinquencies for 60+ days.

Spillovers have emerged in the U.S. commercial real estate sector, which is unlikely to remain insulated from a cyclical deterioration and tightening in financing conditions.

The $3.3 trillion commercial real estate market, like the residential market, has experienced rising property prices, rapid origination growth, and increasing securitization, and has also begun to show signs of strain (Figure 1.6). Property price appreciation has already slowed and securitization has stalled so far this year. Although product innovation and risk layering techniques have been less widespread, loan-to-value ratios have risen, debt service coverage ratios have dropped, and an increasing share of loans have been originated under looser standards.13 So far, delinquency and loss rates have remained low as rents have stayed high and vacancy rates low. However, the weaker U.S. economic outlook, combined with tighter lending standards, is likely to lead to increasing losses, particularly on recently originated loans. Commercial mortgage-backed security (CMBS) spreads have widened to near-record levels, even on the highest-rated tranches, implying market expectations for default and loss rates worse than any yet experienced in the U.S. commercial property market (Figure 1.7).14

Figure 1.6.Commercial Mortgage Borrowing and Real Estate Prices

Sources: Federal Reserve; and Standard and Poor’s.

Note: saar = seasonally-adjusted annual rate.

Figure 1.7.CMBX Spreads

(In basis points)

Source: JPMorgan Chase & Co.

Note: CMBX = an index of 25 credit default swaps on commercial mortgages.

There are notable differences, though, that may prevent the risks to the commercial real estate sector from intensifying to the same extent as in the residential mortgage sector. First, only about one-quarter of the commercial real estate sector is securitized, substantially lower than the 80 to 90 percent securitization rates observed in the subprime residential market at its peak, and there is less repackaging into structured products. This should increase the “skin in the game” for the sector as a whole. Second, commercial mortgage borrowers are less likely to face payment shocks associated with resetting mortgage rates, since most commercial mortgages are standard, 7- to 10-year fixed-rate loans. Third, borrowers in the commercial sector typically have audited financial statements, which should help keep the incidence of fraud well below that observed in the residential sub-prime sector.

Concerns about the economic outlook and tighter lending conditions are also starting to weigh on U.S. consumer credit markets.

Despite the weakening in mortgage markets, credit quality in the $2.5 trillion U.S. consumer debt market has remained fairly strong, suggesting that some borrowers have made it a priority to stay current on credit card and auto debts.15 Delinquency and charge-off rates have picked up slightly since late 2005 across the various consumer credit markets, but remain low relative to levels observed during the last U.S. economic downturn in 2001 (Figure 1.8).16 This may reflect the fact that consumer loans have not grown at the same pace as mortgages over the last few years and that declaring bankruptcy to avoid paying consumer debt has become a less attractive option for some borrowers following bankruptcy reforms in 2005.17,18 However, consumer credit performance is expected to weaken as the rate of personal bankruptcies rebounds and unemployment increases. Econometric work used to estimate consumer loan losses indicates that rising unemployment rates have made the most significant contribution to increases in consumer loan charge-offs.

Figure 1.8.Charge-Off Rates for U.S. Consumer Loans

(In percent)

Sources: Federal Reserve; and IMF staff estimates.

Reflecting concerns about the deteriorating outlook, spreads on consumer-related asset-backed securities (ABS) have widened to record levels. However, a simple comparison of credit card charge-off rates to discounts on consumer credit ABS suggests that spreads are implying an extreme high in charge-off rates relative to the historical trend (Figure 1.9).19 As in some other credit markets, the repricing in risk premia appears to be more reflective of the broader credit market stress than of the underlying collateral quality.

Figure 1.9.Credit Card Charge-Off Rates versus Credit Card Asset-Backed Spreads on Securities

Sources: Federal Reserve; JPMorgan Chase & Co.; Standard & Poor’s; and IMF staff estimates.

Note: Data are based on monthly observations. ABS = asset-backed security.

The corporate debt market appears vulnerable as default rates are set to rise, owing to both macroeconomic and structural factors.

Financial innovation and low policy rates have helped keep corporate default rates at historically low levels long after they had been forecast to rise. The October 2007 GFSR warned that highly leveraged firms were vulnerable to business and economic shocks (IMF, 2007a). Experience is already bearing out this view. U.S. corporate defaults on high-yield debt in January 2008 alone roughly equaled defaults for the whole of 2007, and January’s leveraged loan defaults were twice those seen in all of 2007. Meanwhile, the ratio of downgrades to upgrades on U.S. debt has already risen back to the level of May 2005, when General Motors and Ford were downgraded to subinvestment grade. Downgrades occurred across a range of assets, not just structured finance, and rating agencies appear to be ready to change ratings more promptly than in the past. At the same time, supply factors continue to weigh on the market. The pipeline of leveraged loans and related high-yield bonds has shrunk only modestly, as banks have preferred to take loans onto their balance sheets rather than sell them at deep discounts. Nevertheless, loan prices have fallen (Figure 1.10) in secondary markets and some collateralized loan obligations (CLOs) used to repackage leveraged loans are unwinding, forcing banks to take loans back onto their balance sheets.

Figure 1.10.LCDX Prices and Spreads

Source: JPMorgan Chase & Co.

Note: LCDX = an index comprised of 100 credit default swaps referencing first-lien loans.

Looking ahead, high-yield default rates may rise to 4 to 12 percent if the economy goes into recession (see Box 1.1). The higher side of that range would be comparable to the last recession in 2001 and come close to the peak in defaults during the 1990–91 recession. The unprecedented issuance of low-tier corporate debt over 2003–07, combined with the increase in leverage, may exacerbate corporate distress during the credit downturn (Figure 1.11).20,21 Refinancing risk could further pressure defaults in the near term as $650 billion of leveraged loans are set to mature starting in 2008 over the next three years.22

Figure 1.11.U.S. Leveraged Buyout Loans: Credit Quality Indicators

Sources: Standard & Poor’s LCD; and IMF staff estimates.

Note: EBITDA = earnings before interest, tax, depreciation, amortization.

Systemic Risks Have Risen Sharply

The previous section detailed the deepening and the broadening of the crisis to other market segments. This section attempts to quantify the potential losses that can be expected from the crisis, while tracing the potential systemic effects.

Broader credit deterioration, a weakening economy, and falling credit prices combine into a substantial hit to the capital of systemically important financial institutions.

We estimate aggregate potential writedowns and losses to be approximately $945 billion as of March 2008 (see Table 1.1 and Annex 1.2 for details on the methodology).23,24 Aggregate losses are on the order of $565 billion for U.S. residential loans (nonprime and prime) and securities and $240 billion on commercial real estate securities. Corporate loans (including leveraged loans and CLOs) are expected to account for $120 billion of losses, while consumer loan losses are likely to add an additional $20 billion. Most of the nonprime losses are in securities rather than unsecuritized loans. At present, pricing of mortgage-related derivative indices suggests higher losses than do calculations based on projected cash flows for the underlying loans.25 Since the October 2007 GFSR, ABS prices have declined between 20 and 40 percent across tranches rated AAA to BBB–, and as much as 50 percent on ABS collateralized debt obligations (ABS CDOs) across all ratings categories, reflecting market expectations of future deterioration and illiquidity of the underlying securities. (See Box 2.2 to 2.4 in Chapter 2 for more details on the fragility of structured product ratings and their valuations.) Market prices continue to adjust on an almost daily basis, pressuring mark-to-market losses higher.

Table 1.1.Estimates of Financial Sector Potential Losses as of March 2008(In billions of U.S. dollars)
Estimates of Losses on Unsecuritized U.S. LoansBreakdown of Losses on Unsecuritized Loans
OutstandingEstimated lossBanksInsurancePensions/SavingsGSEs and governmentOther (hedge funds, etc.)
Subprime3004520-30<5<510-155-10
Alt-A6003015-20<5<55-10<5
Prime3,8004015-20<5<515-20<5
Commercial real estate2,4003015-20<5<5<5<5
Consumer loans1,4002010-15<5<5<5
Corporate loans3,7005025-30<5<515-20
Leveraged loans170105-10<5<5<5
Total for loans12,370225100-13010-2010-2030-5040-50
Estimates of Mark-to-Market Losses on Related SecuritiesBreakdown of Losses on Securities
OutstandingEstimated mark-to-market lossBanksInsurancePensions/SavingsGSEs and governmentOther (hedge funds, etc.)
ABS1,10021085-10020-3535-4520-3520-45
ABS CDOs400240145-16035-5015-250-2515-50
Prime MBS3,8000
CMBS94021085-9520-3530-4520-3520-45
Consumer ABS6500
High-grade corporate debt3,0000
High-yield corporate debt6003010-15<55-10<5
CLOs3503015-20<5<50-10
Total for securities10,840720340-38095-11070-12040-9070-150
Total for loans and securities23,210945440-510105-13090-16070-140110-200
Sources: Goldman Sachs; JPMorgan Chase & Co.; Lehman Brothers; Markit.com; Merrill Lynch; and IMF staff estimates.Note: ABS = asset-backed security; CDO = collateralized debt obligation; CLO = collateralized loan obligation; CMBS = commercial mortgage-backed security; GSE = government-sponsored enterprise; MBS = mortgage-backed security.

Potential credit losses would lower aggregate capital adequacy ratios at U.S. banks by about 250 basis points, and at European banks by about 150 basis points. Although aggregate ratios remain above regulatory norms, a bottom-up analysis of losses indicates that some banks and regions will suffer disproportionately. Put in historical perspective, this crisis is of similar dollar magnitude to the Japanese banking crisis of the 1990s (Figure 1.12).26

Figure 1.12.Comparison of Financial Crises

Sources: World Bank; and IMF staff estimates.

Note: U.S. subprime costs represent staff estimates of losses on banks and other financial institutions from Table 1.1 All costs are in real 2007 dollars. Asia includes Indonesia, Korea, the Philippines, and Thailand.

Uncertainty over the size and spread of losses further elevates systemic risks, even as markets price in losses for banks and insurance companies.

Global banks are likely to shoulder roughly half of aggregate potential losses, totaling from $440 billion to $510 billion, with insurance companies, pension funds, money market funds, hedge funds, and other institutional investors accounting for the balance.27 Banks generally hold the most senior tranches of these products, but even these are now likely to incur substantial losses (see Box 2.3 and 2.4 in Chapter 2). European banks hold sizable amounts of complex structured products such as MBS and CDOs and have been exposed to losses related to structured investment vehicles (SIVs) (Figure 1.13).

Figure 1.13.Expected Bank Losses as of March 2008

(In billions of U.S. dollars)

Sources: Goldman Sachs; UBS; and IMF staff estimates.

Note: ABS = asset-backed security; CDO = collateralized debt obligation; SIV = structured investment vehicle.

By mid-March 2008, U.S. banks had reported most of their estimated losses, with European banks’ disclosures catching up owing partly to the longer reporting lags of European banks (see Annex 1.2). In addition, nonbank financial institutions, including insurance companies, may yet also report sizable additional writedowns.

Bank equity and debt capital markets appear to have taken into account the effect of credit-market-related losses. The market capitalization of banks globally declined by some $720 billion through March 2008. Insurance companies have also experienced a decline in market value that appears to be commensurate with the top-down loss estimate of $105 billion to $130 billion.

Strains are compounded by pressures on financial guarantors…

Additional bank losses may originate from the knock-on effects of rating downgrades on financial guarantors, as the ratings on insured bonds would decline and certain hedges would become less effective. IMF staff estimate the total losses to banks from potential downgrades of financial guarantors to be $60 billion to $90 billion, depending on whether the downgrade is one grade (from AAA to AA) or two (to A).28,29 Since 1998, most financial guarantors (such as AMBAC, MBIA, and FGIC) have expanded their traditional business of insuring bonds issued by U.S. municipalities to include structured credit (i.e., ABS and ABS CDOs) and, to a lesser extent, corporate bonds. Losses on ABS protection have now eaten into the capital of a number of financial guarantors, threatening both their own credit ratings and those of the debt they insure (Figure 1.14).30 Additional downgrades of financial guarantors would cause the value of the $800 billion of structured credit they have insured to fall further, imposing additional losses on banks.

Figure 1.14.Financial Guarantors

Source: Bloomberg L.P.

Box 1.1Outlook for U.S. High-Yield Corporate Debt Markets and Default Rates1

The corporate debt market is set to weaken and default rates are expected to rise from historic lows due to both macroeconomic and structural factors. Macroeconomic variables, credit and financial conditions, and market perception of risk are typically used to model and forecast default rates. All of these indicators and models have predicted rising corporate debt defaults since 2007. However, increased financing flexibility extended by lenders may have deferred realized defaults. As well, structural changes in the composition of the corporate debt market may add to market distress in a downturn.

Macroeconomic Indicators and Default Rates

Sources: Bureau of Economic Analysis; Federal Reserve; JPMorgan Chase & Co.; Merrill Lynch; Moody’s; National Bureau of Economic Research; and IMF staff estimates.

1Year-on-year changes; standardized over 1983–present; inverted scale.

2Net survey balances; standardized over 1990–present.

3Issuer-weighted.

Three empirical approaches discussed below all point to a rise in defaults in 2008, with macroeconomic and credit market conditions being the key drivers.

Financial and Corporate Indicators and Default Rates

Sources: Bloomberg L.P.; Bureau of Economic Analysis; JPMorgan Chase & Co.; Merrill Lynch; Moody’s; National Bureau of Economic Research; and IMF staff estimates.

1Year-on-year changes; standardized over the sample periods; inverted scale.

2Standardized over 1986–present.

3Issuer-weighted.

Macroeconomic and credit conditions. Historically, default rates are inversely related to the level of economic activity (see first figure). Both GDP and industrial production closely track the contemporaneous level of default rates. Bank lending standards tend to lead a rise in default rates and are considered a reliable forecasting indicator. Both macroeconomic and credit variables have been signaling a pickup in the default rate over the last year, with expected defaults far exceeding actual defaults.

Financial and corporate indicators (see second figure). Another way to project default rates examines corporate profits (to proxy corporate debt market performance), the implied volatility of the S&P 500 (to capture uncertainty over the future earnings stream), and the debt-to-earnings ratio for high-yield companies (to capture the degree of debt burden relative to revenue). After posting strong growth during 2002–06, corporate profits contracted 1.9 percent year-on-year in 2007, and are expected by the market to remain flat in 2008. Implied equity volatility (VIX) rose from 11 percent in January 2007 to 25 percent as of February 2008, and futures markets expect volatility to remain elevated during 2008. The debt-to-earnings ratio for high-yield corporates has been growing since 2005, and is likely to increase further in 2008. In short, financial indicators also point in the direction of increasing default rates.

Extraction of default probabilities from credit risk transfer markets. Observed prices or yields on corporate bonds and credit default swaps can also be used to derive the implied probability of default. The corporate debt and credit default swap markets have already partly priced in a heightened probability of default (see third figure and Annex 1.2).

Weakening credit discipline may have both delayed and masked the rise in defaults. Loosening credit standards, especially in the leveraged buyout market, resulted in the growth of “covenant-lite” loans, whose holders are not obliged to meet quarterly maintenance criteria. This increased financing flexibility from the lender’s side may help to explain the unusually low number of defaults in the last two years.

Valuation of Financial Instruments Based on Implied Probability of Default

Sources: Bloomberg L.P.; Dow Jones; JPMorgan Chase & Co.; Merrill Lynch; Moody’s; National Bureau of Economic Research; and IMF staff estimates.

1Index based on a blend of realized collateralized debt swaps (CDS) and high-yield bond indices.

2Issuer-weighted.

As the credit cycle turns, the rise in default rates may magnify stress in bond markets owing to several factors. First, there was an unprecedented issuance of low-rated debt over 2003–07, which has raised the share of CCC-rated bonds in total high-yield debt above the end-2000 level. Second, increased leverage on corporate debt—amid deterioration in overall debt quality—may have aggravated vulnerability to external financial conditions, affecting asset quality and earnings streams. In addition, the increase in the share of secured corporate debt from 5 to 11 percent of total high-yield debt over the last seven years may lower recovery rates and prices of unsecured bonds. Third, the maturity profile of leveraged loans is fairly short, subjecting them to near-term refinancing risk as well as raising default risk.

Share of “Stressed Debt”

(Percent of high-yield market trading 1000+ basis points above U.S. treasuries)

Source: Bloomberg L.P.

Forecasts of U.S. High-Yield Default Rates in 2008(In percent)
Assumptions1Forecasts2 (In percent of U.S. high-yield corporate debt)
Industrial productionLending standardsStructural delay3No structural delay4
Best case scenario (No deterioration of economic conditions)1.7204.09.3
Baseline scenario (Moderate deterioration of economic conditions)-2.0304.710.4
Worst case scenario (Economic recession)-5.5505.812.3
Sources: Bureau of Economic Analysis; Federal Reserve; and IMF staff estimates.

When realized default rates diverge from fundamentals, some analysts rely on proxies for distressed debt, such as the share of “stressed debt” (trading 1000 basis points or more above U.S. treasuries). As the fourth figure illustrates, the pendulum has swung dramatically, presaging rising defaults, with the share of stressed debt rising from 9 percent in December 2007 to 21 percent in February 2008. Other measures of debt distress attempt to estimate the number of companies that are able to raise additional debt in the absence of cash to pay interest on existing debt. Liquidity ratings compiled by the major rating agencies suggest that liquidity positions of leveraged borrowers weakened dramatically during 2007.

The different scenarios for the default rate in 2008 are outlined using econometric modeling based on macroeconomic and credit variables and taking into account the possibility of a delay in a full realization of defaults (see table). If the loosening financing standards from lenders continue to delay realized default rates, the default rate is projected in the range of 4 to 6 percent, depending on the extent of the U.S. economic slowdown. If default rates are set to revert to the levels implied by economic fundamentals that were observed before 2007, defaults could rise more sharply, in the range of 9 to 12 percent, based on our estimates.2

Note: Sergei Antoshin prepared this box.1 While this box relates exclusively to U.S. credits, it is recognized that losses related to European-issued securities could be substantial. Indeed, European leveraged buyout deals saw a similar, albeit less pronounced, rise in leverage. In addition, the European high-yield market has also become riskier (as reflected by the higher share of low-tier debt issuance), although it still only represents 15 to 20 percent of the global high-yield debt market.2 These forecast ranges are in line with the 2 to 10 percent array of forecasts produced by credit agencies and market analysts.

…raising concerns about counterparty risks and spill-overs in the credit default swap market…

In view of the weakened capital position of financial guarantors—and because guarantors are not required to post maintenance margins on credit default swap (CDS) contracts that they have sold—many banks have begun to write down the value of the protection they have bought from financial guarantors. For the CDS market overall, losses incurred by protection sellers should equal the gains of protection buyers, but specific sectors may be heavily positioned one way, leading to an increase in counterparty credit risk in the event of a rise in corporate defaults. The concentration of counterparty risk in the CDS market could further compound the risk of multiple failures, for instance, if an individual protection seller is unable to fulfill its payment obligations.31,32

Weaknesses in infrastructural arrangements for CDS markets may further exacerbate risks. Despite earlier attempts to address back-office processing delays, recent slippage in the time-liness of confirmations and affirmations in over-the-counter markets—including corporate CDS—means that many market participants cannot assess in real time changes in their CDS exposures. Moreover, the absence of a central counterparty and multilateral netting of contracts leaves the system dependent on potentially long exposure chains that are vulnerable to a default at any one point. In addition, CDS contracts often require delivery of the underlying bond, and since the volume of contracts often exceeds the volume of underlying instruments, large-scale defaults could result in settlement problems. Since the corporate CDS market may be tested over the coming months, these potential problems need to be monitored closely by policymakers.

…and stability at the core of the global financial system…

Measures of default risk for large complex financial institutions and the potential for contagion within the financial system derived from market prices point to heightened concern about system risk (Figure 1.15).33 The highest likelihood of a single default and the likely number of defaults in the event of a single default in the group—a measure of contagion risk within the global banking system—have both risen significantly.

Figure 1.15.Systemic Bank Default Risk

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

1Among 15 selected large and complex financial institutions (LCFIs).

2Measures the largest probability of default among the sampled 15 LCFIs each day.

…despite sizable injections of bank capital from sovereign wealth funds and elsewhere.

Sovereign wealth funds have contributed about $41 billion of the $105 billion of capital injected into major financial institutions since November 2007. This compares with total reported losses among global banks of some $193 billion (see Box 1.2). Such injections are welcome and critical to restoring bank balance sheets. However, despite these injections, market indicators suggest that many investors believe that some banks still need to raise additional capital.

Bank funding strains are symptomatic of a broad deleveraging of the global financial system and systemic stress.

Some banks have rapidly expanded their balance sheets in recent years, largely by increasing their holdings of highly rated securities that carry low risk weightings for regulatory capital purposes (see Box 1.3 on page 31). Part of the increase in assets reflects banks’ trading and investment activities. Investments grew as a share of total assets, and wholesale markets, including securitizations used to finance such assets, grew as a share of total funding (Figure 1.16). Banks that adopted this strategy aggressively became more vulnerable to illiquidity in the wholesale money markets, earnings volatility from marked-to-market assets, and illiquidity in structured finance markets. Equity markets appear to be penalizing those banks that adopted this strategy most aggressively (Figure 1.17).

Figure 1.16.Securitization Volume in the European Union (EU-15)

(In billions of euros)

Sources: Bloomberg L.P.; Citibank; and Dealogic.

Figure 1.17.Bank Equity Price Changes and Balance Sheet Leverage

(In percent)

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

The forced deleveraging has impacted other leveraged institutions, especially hedge funds.

Until recently, one of the remarkable features of the current crisis was how few large hedge funds had failed. Among the funds that have folded, most appear to have unwound their positions without undue difficulty, suggesting that collateral was liquidated at close to the pledge value. Even as they shrank their balance sheets elsewhere, large banks tried to maintain their prime brokerage lending to hedge funds, on the basis that it enhanced the bank’s long-run franchise value. This situation is changing with the intensification of the crisis as margin locks roll off and pressure on bank balance sheets increases.34 “Haircuts” and margins have increased, and fewer hedge funds are able to secure the leverage required to meet return targets on low-yielding assets. A forced deleveraging of the type outlined in the October 2007 GFSR may therefore be under way, further reducing demand for AAA-rated assets. The example illustrated in Table 1.3 in the October 2007 GFSR shows that, even with no change in value or redemptions by investors, an increase in margin to 10 percent, from an initial 3 percent, would force a fund to sell nearly 70 percent of its holdings (IMF, 2007a). Table 1.2 shows that such increases in margins have been far from unprecedented. Some hedge fund indices already suggest cumulative hedge fund returns have been zero for the last 12 months, even before taking account of the survivorship and reporting biases that tend to overstate returns. It would therefore be unsurprising if there were more hedge fund failures in coming months.

Table 1.2.Typical “Haircut” or Initial Margin(In percent)
January–May 2007April 2008
U.S. treasuries0.253
Investment-grade bonds0–38–12
High-yield bonds10–1525–40
Equities1520
Investment grade CDS15
Synthetic super senior12
Senior leveraged loans10–1215–20
2nd lien leveraged loans15–2025–35
Mezzanine level loans18–2535+
ABS CDOs:
AAA2–415
AA4–720
A8–1530–50
BBB10–2040–70
Equity50100
AAA CLO410–20
AAA RMBS2–410–20
Alt-a MBS3–520–50
Sources: Citigroup; and IMF staff estimates.Note: ABS = Asset-backed security; CDO = collateralized debt obligation; CDS = credit default swap; CLO = collateralized loan obligation; RMBS = residential mortgage-backed security.
Table 1.3.Macro and Financial Indicators in Selected Emerging Market Countries(Estimates for 2007)
Current Account (percent of GDP)Growth in Private Credit (percent year-on-year)Change in Private Credit as Share of GDP (percentage points)External Position vis-á-vis BIS Reporting Banks (percent of GDP)
Europe, the Middle East, and Africa
Bulgaria-21.462.519.7-11.9
Croatia-8.817.83.4-50.8
Estonia-16.041.815.1-68.7
Hungary-5.616.81.6-42.5
Kazakhstan-6.755.212.5-9.5
Latvia-22.945.010.7-53.9
Lithuania-13.345.310.9-34.7
Poland-3.739.68.0-12.7
Romania-14.560.410.7-25.7
Russia5.951.07.18.3
Serbia-16.540.16.0-7.6
South Africa-7.422.05.49.6
Turkey-7.626.54.1-13.9
Asia
China11.119.52.10.8
India-1.421.72.6-3.0
Indonesia2.322.42.0-7.9
Korea0.613.58.7-13.9
Malaysia13.711.83.40.5
Philippines4.43.3-1.5-0.4
Thailand5.63.9-1.45.1
Latin America
Argentina0.737.01.4-7.1
Brazil0.328.55.1-7.8
Chile4.720.85.9-8.0
Colombia-3.823.54.7-7.3
Mexico-0.819.02.2-5.8
Peru1.622.36.2-0.5
Venezuela9.272.54.92.9
Sources: Bank for International Settlements (BIS); European Central Bank; IMF, International Financial Statistics and World Economic Outlook; and IMF staff estimates (preliminary data as of March 3, 2008).Note: The gray boxes of the table point to areas of potential concern. Cutoff values are as follows: current account balance below –5 percent of GDP; private sector credit growth greater than 20 percent year-on-year; growth in the ratio of private sector credit to GDP of more than 10 percent year-on-year; and net external position to BIS banks less than –10 percent of GDP.

Box 1.2.Do Sovereign Wealth Funds Have a Volatility-Absorbing Market Impact?

Between November 2007 and February 2008, sovereign wealth funds (SWFs) were frequently in the news, as major mature market financial institutions required additional capital. This box examines the impact that SWF-provided capital may have had in current volatile market conditions. It may be premature to draw strong conclusions in the absence of a broader set of data and the need for a better understanding of the diverse investment policies and risk management practices of the SWFs. However, given their typically long time horizon and limited liquidity needs, SWFs can have a shock-absorbing role, at least in terms of abating short-term market volatility.1

Credit Default Swap Spreads on Selected Financial Institutions

(In basis points)

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

Note: Vertical lines indicate capital injections to each institution on that date.

SWFs as investors. There are several factors that facilitate the ability of the SWFs to act as a countervailing force in times of market stress.

  • Most SWFs have a long-term investment horizon and limited liquidity needs (with the notable exception of stabilization funds), as they are commonly established to meet long-term macroeconomic objectives;

  • Many SWFs aim to meet long-term real return objectives, and accept short-term volatility in return for expected higher long-term returns and the diversification benefits from a less-constrained strategic asset allocation;

  • Compared with other institutional investors, SWFs also have a stable funding base and no capital adequacy or prudential regulatory requirements;

  • The below-average valuations of stocks in crisis-hit financial markets may have provided a window for SWFs to accumulate significant exposure in the global financial sector.

The table provides a summary of the transactions in which SWFs have injected capital into mature market financial institutions. Common features of these transactions are that they were (1) significant in size, while remaining minority stakes in companies; (2) privately negotiated rather than executed in public markets; and (3) often in convertible bonds, high-yielding bonds that are to be converted to equity stakes in the future. While many SWFs execute their strategic asset allocation decisions in public markets, historically, some of the major SWFs have also used privately negotiated transactions. Increasingly, some of the SWFs are broadening the set of eligible asset classes, including through private equity, in order to implement their long-term investment and strategic asset allocation decisions.2

Sovereign Wealth Fund (SWF) Capital Injections into Financial Institutions and Market Response
Date of AnnouncementFinancial InstitutionsWritedown (of financial institution)SWFs and Other Investor(s)Amount (percent of total stakes) from SWFs and Other Investor(s)Immediate Market Response (change after announcement compared to previous transaction day)
SWFsOther investor(s)SWFsOther investor(s)Stock price (%)CDS (%)
Nov. 26, 2007Citigroup$6 billion in 2007: Q3Abu Dhabi Investment Authority$7.5 billion (4.9%)-1.2-6
Dec. 10, 2007UBS$18 billion in 2007Government of Singapore Investment CorporationUnknown Middle Eastern investor$9.7 billion (10%)$1.8 billion (2%)1.4-9
Dec. 19, 2007Morgan Stanley$9.4 billion in 2007: Q4China Investment Corporation$5 billion (9.9%)4.20
Dec. 21, 2007Merrill Lynch$8.4 billion in 2007: Q3Temasek HoldingsDavis selected Advisors, L.P.$4.4 billion (9.4%)$1.2 billion (2.6%)1.90
Jan. 15, 2008Citigroup$18.1 billion in 2007: Q4Government of Singapore Investment Corporation, Kuwait Investment AuthoritySanford Weill, Saudi Prince Alwaleed bin Talal, Capital Research global Investors, Capital World Investors, New Jersey Investment Division$6.8 billion from Government of Singapore Investment Corporation (3.7%) and $3 billion from Kuwait Investment Authority (1.6%)$2.7 billion (1.5%)-7.3-5
Jan. 15, 2008Merrill Lynch$14.1 billion in 2007: Q4Korea Investment Corporation, Kuwait Investment AuthorityMizuho Financial Group Inc.$2 billion (3.2%) from Korea Investment Corporation and Kuwait Investment Authority, respectively$2.6 billion (4.1%)-5.3-12
Feb. 18, 2008Credit Suisse$2.85 billionQatar Investment AuthorityApproximately $500 million (1% to 2%); the purchase was on the open market3.22
Sources: Bloomberg L.P.; Citigroup; and IMF staff estimates.Note: The stock price of Citigroup rose 6.5 percent on November 28, 2007, the third day after the announcement of the first capital injection. The stock price declines of Citigroup and Merrill Lynch on January 15, 2008 were confounded owing to the simultaneous announcement of huge writedowns and dilution of the claims of existing shareholders.

Recent capital injections. The capital injections by SWFs have augmented the involved financial institutions’ capital buffers and have been helpful in reducing their risk premium, at least in the short term, as the injection curtailed the need to reduce bank assets to preserve capital. The figure and table suggest that the announcements of capital injections from SWFs have assisted in stabilizing share prices and the elevated CDS spreads, at least over the short run.

In most cases, after the announcement of new capital injections, the initial share price reactions to the SWF investments were positive, since announcements of asset writedown went hand-in-hand with a solution based on the capital injection from investor groups in which the SWF had a significant role (see table). Also, share price volatility declined somewhat following the capital injections, which supports the view that SWFs could have a volatility-reducing impact on markets. However, the long-term impact and the potentially stabilizing role of SWFs as major institutional investors will require a broader set of data and assessment.

Next steps. The IMF is currently working across a broad range of issues relating to SWFs. Recognizing the growing importance and relevance for its surveillance activities, the Finance Committee has encouraged the IMF to analyze SWF issues and engage in a dialogue with SWFs to identify best practices. The IMF Executive Board has endorsed the call and asked the staff to prepare a set of commonly agreed best practices for SWFs, which will be a voluntary framework developed in close partnership with SWFs during 2008.

Note: Kristian Flyvholm, Heiko Hesse, and Tao Sun prepared this box.1 It is not the first time that SWFs have invested in financial firms. For instance, China recapitalized its banking sector in 2003 (via Central Huijin Investment Company Limited, which was later merged into the China Investment Corporation as a wholly-owned subsidiary), and Temasek owns stakes in banks in the United Kingdom and in Asia.2 For example, there are recent investments by the China Investment Corporation in Blackstone, and a prospective investment by the Government of Singapore Investment Corporation in the Texas Pacific Group.

Central banks have worked to contain the crisis, giving direct support to term funding markets…

Central banks have adopted a novel and pivotal role in interbank funding markets, different from previous periods of market stress. As private banks retrenched from interbank markets and nonbanks backed away from term funding markets, major central banks became key counterparties in those markets (Figure 1.18 and 1.19).35 They accepted collateral—including some structured products—that many private banks would not. For example, the European Central Bank has accepted as collateral highly rated ABS and MBS, allowing banks to continue to securitize some high-quality assets to use as collateral (see Chapter 3 for more detail).

Figure 1.18.U.S. Funding Market Liquidity

(In billions of U.S. dollars)

Source: Federal Reserve.

Note: ABCP = asset-backed commercial paper.

Figure 1.19.Euro Area Funding Market Liquidity

(In billions of euros)

Source: European Central Bank.

…but while liquidity strains have eased, bank counter-party credit risks remain elevated, making a central bank exit difficult.

Central bank operations had relieved some of the liquidity strains, especially during the turn of the year, but term interbank rates picked up again, possibly reflecting a significant counterparty credit risk component Figure 1.20).36 Thus, it is difficult for central bank operations to target liquidity concerns in term funding markets without distorting (lowering artificially) the market pricing of credit risk. This makes other private and official measures to restore counterparty confidence and reduce risks in the financial system vital to diminish the need for central banks to interpose themselves as counterparties in term funding markets.

Figure 1.20.Decomposing Interbank Spreads

(In basis points)

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

Note: Credit strains are derived by averaging the one-year credit default swap spreads of the banks that determine dollar LIBOR and euro LIBOR rates. These results are then subtracted from the spread between LIBOR and overnight index swaps (OIS) to determine noncredit strains, which are likely to be liquidity related.

Will Emerging Markets Remain Resilient?

Emerging markets have so far proved broadly resilient to the financial turmoil. Improved fundamentals, abundant reserves, and strong growth have all helped to sustain flows into emerging market assets. However, as noted in the October 2007 GFSR, there are macroeconomic vulnerabilities in a number of countries that make them susceptible to deterioration in the external environment (Table 1.3). Eastern Europe, in particular, has a cluster of countries with current account deficits financed by private debt or portfolio flows, where domestic credit has grown rapidly. A global slowdown, or a sharp drop in capital flows to emerging markets, could force painful adjustment.

There are several distinct risks to emerging markets arising from the current turmoil.

First, mature market banks may pare back funding to their local subsidiaries, particularly in circumstances where external imbalances are large.

Second, balance sheet contraction by global financial institutions may reduce funding for investments by hedge funds and other institutions, raising their dollar funding costs, and inducing financial stress within some emerging markets.

Third, emerging market corporate credit risks may continue to increase. Emerging market corporate debt spreads have already moved out about as much as those of similarly rated credits in mature markets.

Fourth, emerging market financial institutions may yet prove vulnerable to financial contagion through exposure to subprime or other structured credit products.

Fifth, a spike in exchange rate volatility could slow or reverse flows into emerging market fixed-income assets, leading to higher funding costs. Negative terms-of-trade shocks could raise difficulties for emerging markets in Latin America and elsewhere that have benefited from the commodity price boom. More broadly, a global slowdown could affect flows into emerging market assets.

For some emerging markets there remains a risk of overheating. Countries whose monetary policy is tied to the U.S. dollar may experience a buildup of domestic liquidity.

Potential funding pressures on foreign banks active in emerging Europe pose risks to a soft landing.

Domestic banks in Eastern Europe have built up large negative net foreign positions vis-à-vis parent banks and international lenders, as credit growth has far outpaced growth in domestic deposits (Figure 1.21). Most European parent banks have plans to sustain cross-border financing of their subsidiaries in the Baltics and southeastern Europe, while gradually slowing credit to cool the economies. Swedish, Austrian, and Italian banks take a long-term view of the growth opportunities in the Baltics and southeastern Europe, and seek to protect their franchise values.

Figure 1.21.External Position of Emerging Markets by Region vis-à-vis BIS Reporting Banks

(In billions of U.S. dollars)

Source: Bank for International Settlements (BIS).

The main parent banks are vulnerable to continued financial turbulence because they obtain a substantial part of their funding on international wholesale markets, as do many mid-sized European banks (Figure 1.22). A soft landing in the Baltics and southeastern Europe could be jeopardized if external financing conditions force parent banks to contract credit to the region. For example, with about half of their funding denominated in foreign currencies, Swedish banks—the main suppliers of external financing to the Baltics—could come under pressure.37

Figure 1.22.Selected European Banks: Dependence on Wholesale Financing as of March 2008

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

Locally owned banks make up one-third of the banking sector in Latvia. These banks are under substantial external funding pressure, which could force them to curtail lending. As with other banks that rely heavily on external bond markets, liquidity for these banks has all but dried up, and spreads have widened 500 basis points. In response, local banks are seeking alternative sources of financing and have worked to increase local deposits.

In Bulgaria and Romania, tighter credit risk controls by parent banks have not been effective in slowing aggregate credit growth, as new entrants, notably Greek and Portuguese banks, have sought to expand market share. Since Bulgaria and Romania only recently joined the European Union, they are still seen by many banks as offering attractive growth opportunities. However, there is a danger that local banks may underestimate the deterioration in the quality of loan portfolios that often accompanies rapid credit growth.

A credit crunch could create pressures for asset quality deterioration in many of the central and southeast European countries.

Banks active in the region also face risks on the asset side of the balance sheet. House prices have soared in tandem with domestic credit growth, and the credit portfolios of banks in emerging Europe have increasingly become exposed to the real estate sector (Figure 1.23 and 1.24). In Estonia and Latvia, house prices have now started to fall, which has led banks to curtail lending to many construction projects, while more developers have resorted to pre-selling apartments in order to receive financing for them. Banks have not experienced a significant increase in loan losses so far, but they have centralized and strengthened risk management in a manner similar to mature market banks. Internal risk controls could force a sharp reduction in credit to protect bank capital, if asset quality deteriorates sharply.

Figure 1.23.Central and Eastern Europe: Growth in Private Credit and House Prices, 2002–06

(In percent)

Sources: Égert and Mihaljek (2007); and IMF staff estimates.

Note: The speed of credit growth is defined as the annual percentage point increase in the private credit-to-GDP ratio, averaged over 2002–06.

Figure 1.24.Baltic States, Bulgaria, and Romania: Credit to Households by Type

(In percent of GDP)

Sources: European Central Bank; and IMF staff estimates.

Note: The figure aggregates credit and GDP across countries. The ratio of household credit to GDP is considerably higher in Estonia and Latvia (above 40 percent in 2007), and lower in Romania (18 percent in 2007).

Perceptions of higher risks are reflected in bank stocks exposed to the region, in CDS, and in the Romanian leu (Figure 1.25).38 The stocks of Swedish banks exposed to the Baltics have underperformed other Nordic bank shares partly owing to significant short-selling. CDS spreads on sovereign debt have surged since August 2007, as investor demand for credit protection has pushed up prices.

Figure 1.25.Baltic States’ 5-Year Credit Default Swap Spreads and Romanian Leu

(In basis points, left scale, unless indicated)

Sources: Bloomberg L.P.; and Datastream.

Reduced access to international funding is having an impact across regions, with some risks to domestic credit markets.

External funding difficulties have arisen in a number of emerging markets and have been particularly acute among some emerging market economies. In Kazakhstan, banks that relied heavily on bond and syndicated loan markets, and where investors are now more concerned about credit risks and weak disclosure practices, have run into funding difficulties, as evidenced by the recent sharp widening in bank CDS spreads. Some private Russian banks have encountered similar problems. In Hungary, tightening credit conditions have pushed up swap and interbank rates, prompting some leveraged investors funding at the swap rate to sell off holdings of government bonds. While pressures on Turkish banks are not as strong, there has been a shift in funding sources away from external bond markets and back toward syndicated loan markets. At the same time, spreads in the cross-currency swap market—used to transform currency exposure and maturities—have moved against domestic Turkish banks.

Despite generally strong external positions, some concerns about dollar funding have arisen in Asia, particularly in Korea, Taiwan Province of China, and, to an extent, in India. Korea’s large stock of external dollar-denominated banking debt—about $95 billion as of September 2007—presents some potential rollover risk, although much of it reflects currency hedging by exporters (notably shipbuilders) enjoying record order flows. In India, some corporations have borrowed dollars and swapped the resulting debt into yen, increasing the difference between borrowing and lending rates, but leaving a large open exposure.39 Nevertheless, the risk to the Indian financial sector arising from these transactions currently appears manageable.

External funding pressures in Latin America remain modest by the standards of past episodes of financial turmoil, due in part to a decline in regional dependence on foreign capital flows. In many countries in the region, much of the financing for domestic credit growth in recent years has come from an expanding domestic deposit base. In Brazil, the development of this credit channel is evident in domestic currency interbank spreads that have remained stable despite the global turmoil. Nevertheless, dollar spreads in Brazil have widened somewhat, particularly at longer maturities. Elsewhere in the region, external funding costs, as indicated by corporate global bond spreads, have also risen.

The widening in corporate spreads could point to future funding issues.

Emerging market corporate spreads have widened substantially since the beginning of the turmoil, signifying that the concerns about funding and credit risks in mature markets have spilled over to emerging market credit. Corporate credit has been more highly correlated with similarly rated mature market credit than it has with other types of emerging market assets, particularly sovereign bonds. In contrast to corporate spreads, the widening in sovereign bond spreads has so far been quite moderate by the standards of previous financial crises, due in part to debt repurchases that have reduced outstanding supply.

With the expansion of emerging market corporate debt as an asset class and the development of CDS and index-based contracts that facilitate the trading of that debt, investors have drawn fewer distinctions between mature and emerging corporate bonds. That perspective, while positive for the asset class, has opened a new potential channel of contagion. Should mature market credit spreads widen further, emerging market corporate funding costs would probably increase, pushing credit demand into domestic banking systems, and increasing domestic funding pressures (Figure 1.26).

Figure 1.26.Emerging Markets: Private Sector External Bond Issuance

(In billions of U.S. dollars)

Source: Dealogic.

The degree of exposure to mortgage-related credit is not yet fully known.

Thus far, exposure to subprime instruments appears to be quite limited in most emerging markets. Some emerging Asian financial institutions have revealed subprime exposures, but writedowns have been less than $1 billion. There has also been rapid growth in Asian-originated structured credit products—most of which are not related to real estate—but the growth has been from a low base, and the total outstanding is likely still below $100 billion.40 Purchases of subprime and structured credit products in Latin American markets appear to have been quite limited, as yield-seeking domestic investors have regarded high domestic nominal interest rates as an attractive alternative to offshore instruments, while tight banking regulations have helped limit exposure to riskier assets. In the emerging Europe region, banks have typically focused on expanding domestic lending, often at high expected real rates of return, rather than acquiring foreign assets. Nevertheless, experience in mature markets suggests that subprime exposure often turns out to be larger than initially indicated.

Exchange rate volatility could prompt outflows.

Cross-border carry trades into emerging market currencies that have flourished during the past half-decade may still be vulnerable to bouts of volatility (Figure 1.27).41,42 Popular carry trade destinations have included Brazil, Colombia, Iceland, Indonesia, New Zealand, Turkey, and South Africa, with funding most often from the Japanese yen or Swiss franc, as well as, now, the U.S. dollar. Since July 2007, risk repricing and yen appreciation have prompted the unwinding of a substantial proportion of yen carry trades, but cross-border interest rate differentials have persisted, and lower U.S. interest rates have increased the use of the dollar as a carry trade funding currency. The continued strength of a number of emerging market currencies—including the Brazilian real and the Indian rupee—suggests that some carry trades have persisted. This could present a channel of vulnerability in the event of future volatility spikes.

Figure 1.27.Carry-Trade Index and Currency Volatility

(In billions of U.S. dollars)

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

Note: Implied currency volatility obtained from 1-month U.S. dollar–Japanese yen options.

A generalized slowdown could still prompt a broad retreat from emerging market assets.

A global slowdown, in turn, could lead to a decline in most types of capital inflows to emerging markets. While there have been some signs of slowing, inflows to emerging equity markets have generally remained positive. Some supply-side factors continue to favor emerging markets, with institutional investors in Europe and North America still seeking portfolio diversification, retail investors in Japan continuing to look for higher returns abroad, and institutional or sovereign investors in the Middle East recycling oil-based surpluses. High commodities prices are also supportive. Nevertheless, the experience of previous bouts of global risk reduction in the midst of slowing growth suggests that the possibility of a reversal in equity flows remains considerable, particularly if other factors are unfavorable.

For certain emerging markets there may be a risk of overheating as investors shift away from mature market assets.

For countries with strong balance of payments positions and tight links to the dollar, the possibility of overheating remains.43 A number of Middle East oil exporters have currencies that are closely linked to the dollar, and many of these have already experienced strong inflationary pressures. In some Asian economies, steps taken to limit the pace of appreciation against the dollar may lead to monetary policy settings that are looser than would otherwise be optimal. Despite the financial turmoil, some “Asia play” flows into currencies such as the Chinese renminbi and Indian rupee have continued.44 In contrast to the predominant view in prior crises, a few investors have even taken the position that emerging market assets could provide a form of safe haven from mature market upheavals. Under such circumstances, further downward pressure on the dollar, particularly if it emanates from subprime or similar shocks, could boost liquidity and lead to an intensification of inflationary pressures in some emerging markets.

Box 1.3.The Rise in Balance Sheet Leverage of Global Banks

For the past decade, high levels of liquidity and low volatility supported significant asset growth among the largest banks, while asset growth that contributed to holdings of regulatory capital was more moderate. This trend is evident in the 10 largest publicly listed banks from Europe and the United States, which doubled in aggregate assets in the last five years to 15 trillion euros, while risk-weighted assets, which drive the capital requirement, grew more moderately to reach about 5 trillion euros (see figure). While considerable differences are present among individual institutions, the widening gap between risk-weighted assets and total assets reflects an expanding share of assets that for regulatory capital purposes carried a lower risk weighting. Two key factors are responsible for the difference.

  • The adoption of international financial reporting standards (IFRS) in Europe caused the re-recognition on the balance sheet of substantial activity associated with the originate-to-distribute business model. Activities that were earlier transferred under national accounting standards to special-purpose vehicles (SPV) were brought onto bank balance sheets. Under Basel I, which used a different measure for risk transfer, the banks were able to record a lower or no risk weight for the associated assets (and for backup credit lines extended to SPV).

  • The increase in trading and investment activities (e.g., asset-backed securities, and hedging). The associated risk weights on these instruments were substantially less than loans because they were generally highly rated, showed relatively stable prices, or were used for hedging.

Regulatory capital requirements did not constrain asset growth. The banks continued to meet the Basel I capital requirement with relative ease. The banks showed on average a Tier 1 capital-to-risk-weighted-asset ratio of between 7 and 9 percent—well above the 4 percent minimum. With the high capital ratios, many of the large banks were able to engage in stock repurchases through the third quarter of 2007.

Balance Sheet Profiles for 10 Large Publicly Listed Banks

Sources: Thomson Financial; and IMF staff estimates.

The composition of bank balance sheets for large banks moved away from loans funded by deposits. Loans declined as a share of total assets, and investments (securities holdings and trading activities) grew (see figure). A companion to the loan decline was a falloff in the importance of retail deposits as a source of stable funding, which is most significant among the banks that grew the quickest.

Banks became more reliant on liquidity from money markets (i.e., interbank borrowing and other forms of short- and long-term debt, including securitized funding) or from the sale of marketable securities. These funding sources, however, entailed higher market-sensitive interest costs (compared to slower growing consumer deposits), which increased and became more difficult to obtain with the tightening of market liquidity starting in the third quarter of 2007. Moreover, the ability to sell marketable securities at close to book values proved increasingly more difficult, as fears of underlying credit quality tainted market valuations.

Note: Michael Moore prepared this box.

Credit Squeeze or Credit Crunch?

What began as a fairly contained deterioration in portions of the U.S. subprime market has metastasized into severe dislocations in broader credit and funding markets that now pose risks to the macroeconomic outlook in the United States and globally. This is best illustrated by Figure 1.28, which documents how the deterioration that first emerged in nonprime mortgage markets spread to leveraged finance and mortgage-related structured credit markets, global money markets, and then moved up the credit spectrum from low- to high-grade corporate credit markets, and to prime residential and commercial mortgage markets, finally threatening to broaden to emerging market assets. Spreads have widened across the full range of credits—not only subprime but high-grade—and around the globe to Europe as well as the United States and to emerging as well as mature markets (Figure 1.29).

Figure 1.28.Heat Map: Developments in Systemic Asset Classes

Source: IMF staff estimates.

Note: The heat map measures both the level and 1-month volatility of the spreads, prices, and total returns of each asset class in terms of deviation relative to the average during 2004–06 (i.e., wider spreads, lower prices and total returns, and higher volatility). That deviation is expressed in terms of standard deviations. Green signifies a standard deviation under 1, yellow signifies 1 to 4 standard deviations, and black signifies greater than 4 standard deviations. ABS = asset-backed security; MBS = mortgage-backed security; RMBS = residential mortgage-backed security.

Figure 1.29.Spreads Across Credit: Historical Highs, Lows, and Current Levels

(In basis points)

Sources: JPMorgan Chase & Co.; Merrill Lynch; and IMF staff estimates.

Note: Yellow lines indicate period ranges. Black squares are as of March 2008. Data inception in parentheses. CMBS = commercial mortgage-backed security.

Off-balance-sheet structures and leveraged entities are being forced to unwind leverage, adding supply to the market from distressed debt sales and a downward spiral of credit prices.

Rising funding costs and low valuations are forcing off-balance-sheet credit vehicles, some hedge funds, and some investment funds to sell assets to raise liquidity and reduce leverage. SIVs are under rising pressure to sell assets as they struggle to roll over much of their medium-term financing. Falling prices on leveraged loans have triggered unwinds of some of the $300 billion of market-value CLOs, requiring their managers to sell the underlying loans onto the market, depressing prices further.45 These sales added to the pressure from the estimated $230 billion overhang of debt sitting on bank balance sheets from buyout deals completed in 2007.46 Financial guarantor concerns have spilled over to municipal markets and guaranteed bonds, as funding pressure is now being felt across markets wherever AAA-rated paper was issued to finance assets with lower ratings. Markets for other types of short-term securities have also come under pressure, suggesting some contagion effects.47 Spreads on the municipal bonds backed by the financial guarantors have widened, and corporates are also finding it more expensive to issue.

Both engines of credit creation are sputtering.

Against this backdrop, the environment for new issuance in some securities markets is more challenging. This year, private sector net debt issuance is expected to contract markedly. Investment-grade corporate issuance is thought likely to hold up relatively well, and highly rated firms should still be able to borrow on reasonable terms, but mortgage issuance and high-yield corporate loan issuance are likely to fall sharply (Figure 1.30). Many of the structures created over recent years are struggling, as the traditional buyer base of the high-rated securities has shifted to more liquid and less risky assets. Confidence in the architecture, ratings, and process of structured finance will require reform and time to be restored.

Figure 1.30.U.S. Private Sector Net Debt Issuance by Sector

(In billions of U.S. dollars)

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

Note: ABS = asset-backed security; CDO = collateralized debt obligation; CLO = collateralized loan obligation; MBS = mortgage-backed security.

1Only gross debt issuance data are available.

2Subtotal is based on data for which net figures are available.

Bank balance sheet adjustment could crimp or bind credit.

The possible immediate credit impact of the aggregate loss estimates on banks is that credit growth could be substantially squeezed. Estimating the impact on credit to the private sector is difficult. One gauge is to assume that banks will cut back lending to offset part, but not all, of the worsening of their key ratios that would result from the losses they will incur and involuntary balance sheet expansion. Using this approach, and spreading the credit withdrawal over three quarters, the pace of credit growth in a squeeze would be reduced to a little over 4 percent of the outstanding private sector debt stock in the United States. It is worth noting that credit had grown on average by nearly 9 percent in the United States in the post-war period. A credit squeeze might therefore feel roughly like the normal constriction of credit seen at the bottom of the business cycle in mature markets.

A supply shock to credit would result in a more painful credit crunch. In a negative scenario, funding markets remain restricted, forcing banks to de-lever and hold more capital in support of their balance sheets, banks’ profits fall and fee-earning sources shrink, and raising fresh capital is more difficult. Banks may not only limit exposure to lower-quality loans, but curtail credit across the board—central bank surveys show a remarkably consistent picture of tightening of credit standards, including across categories of lending (Figure 1.31). In this case, credit growth could be reduced to 1 percent of the outstanding private sector debt in the United States. The resulting slowing of credit growth would be similar to that experienced during the 1990–91 recession, and worse than those in previous recessions (Figure 1.32).48

Figure 1.31.G-3 Bank Lending Conditions

(Net percentage of domestic respondents reporting tightening standards for loans)

Sources: Bank of Japan; European Central Bank; Federal Reserve; and IMF staff estimates.

Note: Monthly interpolated GDP-weighted average. Euro area 1999:Q1 to 2002:Q4 based on values implied by credit growth.

Figure 1.32.U.S. Private Sector Borrowing

(Borrowing by households and nonfinancial corporations as a percent of debt outstanding)

Sources: Federal Reserve; National Bureau of Economic Research; and IMF staff estimates.

Note: Yellow bars represent recession periods

Simulations suggest that a supply shock to credit is likely to have a significant impact on economic growth.

We develop a simple vector autoregression model to get some feel for how credit growth and other economic variables affect one another. The model includes real GDP growth, inflation, private sector borrowing, and the prime loan rate on quarterly data for the United States between the first quarter of 1952 and the third quarter of 2007.49 Private sector borrowing is measured as a percentage of the outstanding stock of private sector debt.50

The model detects a statistically significant impact of a negative shock to credit growth on GDP growth.51 A credit squeeze and a credit crunch spread evenly over three quarters will reduce GDP growth about 0.8 and 1.4 percentage points year-on-year, respectively, assuming no other shocks to the system (Figure 1.33). This suggests that the adjustment process is likely to be long lasting, and would continue to dampen growth well into 2009.

Figure 1.33.Impulse Response of U.S. GDP to Credit Shocks

(In percent, year-on-year)

Source: IMF staff estimates.

Note: Credit withdrawal spread over three quarters.

A great deal of uncertainty surrounds such an exercise. The model does not account for the unusually aggressive monetary policy easing being undertaken by the Federal Reserve, which is likely to mitigate some of the predicted impact on growth. At the same time, however, the effect on GDP could get substantially larger if market dislocations were to affect the issuance of nonfinancial corporate debt more significantly. Furthermore, the fact that this credit shock is taking place in the heart of the banking system, where securitization and structured credit products have been used to shift credit risks to other holders, not simply in smaller banks where such risks were retained, means that the impact could be more profound than suggested by historical patterns in the data. Finally, although not modeled here, the slowing of credit growth in Europe would be substantial, and the greater role of banks in credit intermediation in many European economies than in the United States means that the impact on European economies could be significant.

Immediate Policy Challenges

Against a backdrop of continuing weakness in global credit markets, threats to systemic stability have intensified. Despite some reductions in policy rates in the United States, United Kingdom, Canada, and a few other economies, as well as a sizable U.S. fiscal package, global growth is likely to slow significantly in 2008. The risks of a credit crunch are heightened by spreading dislocations in securities markets, significant bank balance sheet adjustment, and growing concerns about counterparty credit risks. This more negative scenario, however, is not a forgone conclusion. Banks are seeking capital injections and private participants, including banks, financial guarantors, and credit rating agencies are taking steps to rebuild market confidence and stem systemic risks.52 Nevertheless, a range of financial policies—in addition to macroeconomic policies—will be needed to mitigate downside risks. These policies aim to foster counterparty confidence, and set the stage for more medium-term reforms discussed in Chapter 2 and 3.

Restoring counterparty confidence is an immediate priority to reduce systemic threats and spillovers.

Lack of reliable information about exposures and risks has led to misunderstandings and misperceptions that have amplified systemic risks. More rapid and informative disclosure by financial institutions is needed, including how complex structured credit securities are valued and the extent of losses. However, some financial institutions may lack incentives to do this, and addressing such shortcomings will take time and require international agreements. More immediately, national authorities should seek to remove misperceptions about the vulnerabilities of national financial institutions and markets. One approach would be to issue special financial stability reports drawing information from supervisory authorities that assesses risks, provides information and analysis relevant to financial stability, and highlights plans to restore financial soundness as needed. Such reports would complement other policy measures aimed at containing systemic risks.

Systemically important financial institutions need to continue to raise capital and funding to support balance sheets.

To strengthen confidence and avoid capital reductions that could constrain lending, banks with weak capital positions should be strongly encouraged to raise capital. In some instances, supervisors may need to direct banks to strengthen capital ratios and fortify funding positions, even in the more costly current environment. To improve confidence in reported information in Europe, consideration could be given to making nonconfidential information from supervisory prudential reports public, as is the practice in a few other countries. Financial guarantors along with others will need to continue to explore avenues for shoring up capital to back up commitments to structured credit products and protect or restore ratings, while reinforcing risk management and governance. Regulators will need to develop a capital adequacy framework for financial guarantors that is less dependent on rating agency ratings and models.

A strengthening of supervisory oversight should reduce the incidence of unsuspected risk exposure and contribute to the rebuilding of counterparty confidence.

Repeatedly during the crisis, banks have revealed unexpectedly large risk exposures. This risk came through many channels—purchases of securities based on loans that had initially been sold on by banks, implicit guarantees provided to off-balance-sheet vehicles, and large lines of credit extended to hedge funds and other high-risk clients, among others. At the same time, the degree of leverage undertaken by hedge funds and other market participants has often turned out to be much higher than expected. The revelation of such high and previously unsuspected levels of systemic risk underlines the important role that supervisory oversight should play in ensuring that institutions’ risks are well managed. Confidence in financial institutions can be enhanced through supervisory oversight that examines more broadly the risks banks are taking, with closer coordination among supervisors when they are international. There is an urgent need to review the regulatory framework and effectiveness of supervision. In particular:

  • Banks must be able to show sufficient capital to absorb shocks from the reduction in mark-to-market valuations or losses on asset sales. They need to demonstrate that they have sufficient capital and liquidity resources to reassure counterparties that good access to funding and money market liquidity, including during periods of severe turbulence, can be maintained. Pillar 2 of Basel II—supervisory review—can be used to ensure that banks hold additional capital beyond the minimum requirement identified by risk weights or by internal models under Pillar 1, when the supervisors identify deficiencies (see Chapter 2).

  • Bank supervisors need to take more account of balance sheet leverage as they assess capital adequacy. The risks (particularly market and liquidity risks) that have accompanied balance sheet growth need to be properly considered for capital adequacy purposes. While banks continue to meet the minimum regulatory capital requirements, the low absolute capital levels for many large banks at present and the prospect of further losses are adding to concerns about whether capital is sufficient. Banks that must be particularly vigilant are those that hold high levels of assets subject to mark-to-market valuations, that are highly reliant on wholesale funding markets, and that employ high leverage.

  • Banks need to improve their management of liquidity risk. This may include improvements in measurement, evaluation of the backup contingency lines, severe stress tests, and contingency plans for long periods when wholesale markets are unavailable. Supervisors need to be more proactive in countering signs that banks have inadequately protected against liquidity risks (see Chapter 3).

  • Stricter rules are needed on the use of off-balance-sheet entities by banks, and disclosure should be improved so that investors can assess the sponsor’s risk to the entity. Supervisors may need to strengthen guidelines regarding the circumstances under which risk transfers to off-balance-sheet entities warrant capital relief (see Chapter 2).

Public measures can help alleviate some stress in the U.S. mortgage markets, but longer-term policy repercussions need to be considered carefully.

Public measures to alleviate mortgage-related stress should help cushion some of the fallout from the crisis. In addition to a sharp easing in monetary policy and broader tax relief, measures adopted in the United States include a moratorium on interest rate resets for subprime borrowers; an increase in the limit on the size of loans that conform to packaging requirements at the GSEs; a removal of the cap on the GSEs’ retained portfolios; and an expansion in the Federal Housing Administration lending program. These steps, though helpful, are not a panacea. The planned moratorium, for example, seeks to limit foreclosures, but may also redistribute the cost from borrowers to lenders, servicers, and investors. Other measures will need to be weighed carefully to ensure that a balance is struck between (legitimate) issues of consumer protection and protection of legal contracts that underpin modern finance, as some of these measures may undermine existing contracts.

If systemic risks significantly increase, remedial measures may be warranted.

Public policy should seek to safeguard financial stability and market functioning. However, care should be taken to avoid creating adverse incentives or moral hazard that undermines discipline imposed on private players by such events. At the same time, the public resources should be kept as small as possible. Supervisors need to ensure prompt recognition of mark-to-market losses but should recognize that prices in illiquid markets can overshoot their new equilibrium (see Chapter 2). In a case of depleted capital, the preferred approach would be to take remedial measures and resolve the institution if it is no longer viable. Shareholders should bear the brunt of the adjustment, and the resources raised by the liquidation of the institution should be shared with creditors. When the failure of the institution poses a systemic threat, the case for public assistance may need to be considered, but only after shareholders have borne the full brunt, with clear mechanisms in place to ensure that operations continue on a commercial basis, and with an unambiguous plan for exit by the public sector.

Resolution should avoid adding to pressures of distressed debt sales. Under extreme scenarios, sales of structured finance assets from off-balance-sheet entities and banks under resolution could place further pressure on credit and may force other banks to become under-capitalized, leading to potentially disruptive and costly strains on insured depository institutions. Accordingly, disposition of assets should be managed in an orderly fashion.

Resolving institutions should go hand-in-hand with reforms to strengthen the financial system.

An important lesson from the crisis has been the role that underlying vulnerabilities and weakness in the financial system architecture has played in amplifying problems and raising costs to both private and public parties. Although a rush to regulate should be avoided, supervisors need to be able to respond proactively to address mis-aligned incentive structures—such as in the “originate-to-distribute” model—that together with an overall resolution strategy should reduce future risks. For example, some German Landesbanken were particularly exposed to subprime instruments, and IMF missions have called for a restructuring of these state-sponsored banks—a process that may gain new impetus. In the United Kingdom, a review of financial stability arrangements is under way—following the events at Northern Rock. This anticipates the establishment of a stronger system for the detection of banking sector problems, and associated with this a special resolution regime. An addition reform of the payment system oversight arrangements is being considered. In the United States, the experience of the financial guarantors argues for reforms to U.S. insurance regulation. Responsibility currently resides with the states, which has impeded coordination of regulatory efforts across states and with federal bank and securities regulators where spillovers are now evident. A new strategy for regulation of the financial guarantor sector needs to be implemented, including a coherent approach to capital adequacy and new limits on financial guarantors’ activities.

Restoring counterparty confidence in funding markets should support an exit by central banks as conditions stabilize.

Central bank operations in the term funding markets pose challenges for monetary operations in the presence of counterparty credit concerns. Term premiums reflect, in part, market perceptions and pricing of credit risk. Therefore, determining the size, tenor, and vigor of such operations needs to balance the desire to stabilize market conditions without unduly distorting the market pricing of credit risk. Importantly, central banks will find exiting the role of term funding support difficult without the implementation of the above policy measures, because central bank operations can address liquidity but not credit problems. Once counterparty confidence is restored and banks have strengthened their liquidity and funding positions, central banks should seek to gradually exit from significant support to term funding markets.

Emerging markets need to strengthen their resilience to global turmoil.

Policy improvements have contributed to the resilience of many emerging markets in the face of the global turmoil. In many countries, macroeconomic stabilization programs have helped to eliminate distortions and reduce external imbalances, making domestic markets less vulnerable to external shocks. Countries vulnerable to external financing shocks and higher inflation need to adjust to the new tighter external financing conditions and adopt policies to reduce domestic repercussions of sustained financial turmoil. These policies may include a tightening of limits on external borrowing by banks and other financial institutions. In addition, to prepare for the possibility of a deeper global liquidity shock, policymakers should map out contingency plans with potential responses to short-term funding problems. The importance of transparency in bolstering investor confidence has also become more apparent. The limited exposure to subprime and other impaired instruments in emerging markets should not lead to complacency, as the same benign conditions have underpinned higher risk-taking in some countries. As well, the lessons from the turmoil underscore the need to make further progress on fine-tuning the design and strengthening the implementation of accounting and disclosure standards for financial institutions.

The IMF is developing new methods to examine various types of risk and is seeking to strengthen its assessments of macro-financial linkages (see Box 1.4). These efforts will be intensifying given the now more urgent task of limiting the knock-on effects of the current crisis to the IMF’s broader membership.

Annex 1.1. Global Financial Stability Map: Construction and Methodology53

This annex outlines our choice of indicators for each of the broad risks and conditions in the stability map. To complete the map, these indicators are supplemented by market intelligence and judgment that cannot be adequately represented with available indicators.

To begin construction of the stability map, we determine the percentile rank of the current level of each indicator relative to its history to guide the assessment of current conditions, relative both to the October 2007 GFSR and over a longer horizon. Where possible, we have therefore favored indicators with a reasonable time series history. However, the final choice of positioning on the map is not mechanical and represents the best judgment of IMF staff. Table 1.4 shows how each indicator has changed since the October 2007 GFSR and the overall assessment of the movement in each risk and condition.

Table 1.4.Changes in Risks and Conditions Since the October 2007Global Financial Stability Report
Conditions and RisksChange since October 2007 GFSR
Monetary and Financial Conditions
G-7 real short rates
G-3 excess liquidity
Financial conditions index
Growth in official reserves
G-3 lending conditions
Risk Appetite↓↓
Investor survey of risk appetite
Investor confidence index
Emerging market fund flows
Risk aversion index
Macroeconomic Risks↑↑↑
World Economic Outlook global growth risks
G-3 confidence indices
Economic surprise index
OECD leading indicator
Implied global trade growth
Emerging Market Risks↑↑
Fundamentals EMBIG spread
Sovereign credit quality
Credit growth
Median inflation volatility
Corporate spreads
Credit Risks↑↑↑
Global corporate bond index spread
Credit quality composition of high-yield corporate bond index
Speculative-grade corporate default rate forecast
Banking stability index
G-3 loan delinquencies
Market Risks
Hedge fund estimated leverage
Speculative positions in futures markets
Common component of asset returns
World implied equity risk premia
Composite volatility measure
Financial market liquidity index
Source: IMF staff estimates.Note: Changes are defined for each risk/condition such that ↑ signifies more risk or easier conditions and ↓ signifies the converse; ↔ indicates no appreciable change. The number of arrows for the six overall conditions and risks corresponds to moves on the global financial stability map.

Monetary and Financial Conditions

The availability and cost of funding linked to global monetary and financial conditions (Figure 1.34). To capture movements in general monetary conditions in mature markets, we begin by examining the cost of short-term liquidity, measured as the average level of real short rates across the G-7. From there, we take a broad measure of excess liquidity, defined as the difference between broad money growth and estimates for money demand. Realizing that the channels through which the setting of monetary policy is transmitted to financial markets are complex, some researchers have found that including capital market measures more fully captures the effect of financial prices and wealth on the economy. We therefore also use a financial conditions index that incorporates movements in real exchange rates, real short- and long-term interest rates, credit spreads, equity returns, and market capitalization. Rapid increases in official reserves held by the central bank create central bank liquidity in the domestic currency and in global markets. To measure this, we look at the growth of official international reserves held at the Federal Reserve. While the above measures capture the price effects of monetary and financial conditions, to examine the quantity effects, we incorporate changes in lending conditions based on senior loan officer surveys in mature markets.

Box 1.4.Quantitative Financial Stability Modeling

In the wake of the U.S. subprime crisis, the IMF has expanded its research agenda in quantitative financial stability modeling to strengthen the analysis of macro-financial linkages.

The IMF is developing new applications for stress tests and other risk assessment models to help identify and address financial system vulnerabilities in member countries. This work aims at enhancing the quality of quantitative analyses performed in the context of the Financial Sector Assessment Program, supporting technical cooperation on risk-based supervision and Basel II implementation, and facilitating offsite surveillance of national and global financial systems, and hence IMF surveillance more broadly.

Among the specific areas in which the IMF has been active are the further development of credit risk modeling; analysis of the “second-round effects” of shocks—both interactions within the financial sector and feedback between the financial sector and the real economy; and expansion of existing approaches to liquidity risk modeling.

Credit Risk Modeling

Work in this area revolves around three methodologies. One application models portfolio credit risk based on CreditRisk+, a tool used by financial institutions and supervisors to compute credit portfolio loss distributions (Avesani and others, 2006). This application can be useful for scenario stress testing when complemented with models of the probability of default and loss given default. Other recent work includes macro stress testing in the presence of data constraints, an approach that seeks to quantify the impact of macroeconomic shocks on banks’ economic capital in the presence of short time series of default probabilities (Segoviano Basurto, 2006). It simultaneously accounts for changes in the correlation among banks’ assets through the economic cycle. The contingent claims approach (CCA)—a method that combines balance sheet and market information with widely used finance techniques to construct risk-adjusted balance sheets—is also being used to conduct scenario analysis and can be applied to financial institutions that issue securities in sufficiently deep markets (Gray, Merton, and Bodie, 2007.

Measurement of Second-Round Effects

This includes a measure of financial fragility at the system level—a banking stability index—based on banks’ joint probability of default (see Box 1.5), which represents the expected number). This approach can also be applied at the global level by looking at joint probabilities of default (or other measures of stability) for key large complex financial institutions. Another approach to modeling contagion uses the extreme value theory framework to capture the possibility that large, extreme shocks are transmitted across financial systems differently than small shocks (Chan-Lau, Mitra, and Ong, 2007). A third approach is to develop a CCA-based framework that provides risk indictors and can be linked to macroeconomic models of varying degrees of complexity.

Liquidity Risk Modeling

Work is under way to enhance the range of tools and methods available to stress test exposures to liquidity risk—a risk area that the current turmoil has made more apparent. The three main directions of work in this area are (1) building on existing methodologies to identify funding liquidity risk (including non-traditional sources, such as securitization) and expanding them to incorporate market liquidity risk (including the effects of asset fire sales and crowded trades); (2) capturing off-balance-sheet concentration risk—for example, excessive committed and uncommitted credit lines to a single counterparty; and (3) extending the CCA-based framework using information from equity option prices to capture the effects of increased uncertainty of asset values, market illiquidity, potential for fire sales, and funding liquidity risk.

Note: The main author of this box is Marina Moretti.

Figure 1.34.Global Financial Stability Map: Monetary and Financial Conditions

Sources: Bloomberg L.P.; Goldman Sachs; OECD; lending surveys by Bank of Japan, European Central Bank, and Federal Reserve Board for households and corporates; and IMF staff estimates.

Note: Dashed lines are period averages. Vertical lines represent data as of the October 2007 GFSR.

1Only G-3 subindicators are shown.

2A GDP-weighted average of China, euro area, Japan, and the United States. Each country index represents a weighted average of variables, including interest rates, credit spreads, exchange rates, and financial wealth.

3Monthly interpolated GDP-weighted average. Euro area 1999:Q1 to 2002:Q4 based on values implied by credit growth.

Risk Appetite

The willingness of investors to take on additional risk by increasing exposure to riskier asset classes, and the consequent potential for increased losses (Figure 1.35). We aim to measure the extent to which investors are actively taking on more risk. A direct approach to this exploits survey data. The Merrill Lynch Fund Manager Survey asks about 200 fund managers what level of risk they are currently taking relative to their benchmark. We then track the net percentage of investors reporting higher-than-benchmark risk-taking. An alternative approach is to examine institutional holdings and flows into risky assets. The State Street Investor Confidence Index uses changes in equity holdings by institutional investors relative to domestic investors to measure relative risk tolerance.54 The index extracts relative risk tolerance by netting out wealth effects and assuming that changes in fundamentals symmetrically affect all kinds of investors. We also take account of flows into emerging market equity and bond funds as these represent another risky asset class. Risk appetite may also be inferred indirectly by examining price or return data. As an example of this approach, the Goldman-Sachs Risk Aversion Index measures investors’ willingness to invest in risky assets as opposed to risk-free securities, building on the premises of the capital asset pricing model.55 By comparing returns between government bills and equities, the model allows the level of risk aversion to move over time. Taken together, these measures provide a broad indicator of risk appetite.

Figure 1.35.Global Financial Stability Map: Risk Appetite Conditions

Sources: Emerging Portfolio Fund Research, Inc.; Goldman Sachs; Merrill Lynch; State Street Global Markets; and IMF staff estimates.

Note: Dashed lines are period averages. Vertical lines represent data as of the October 2007 GFSR.

1The estimated changes in relative risk tolerance of institutional investors from Froot and O’Connell (2003) are integrated to a level, scaled, and rebased so that 100 corresponds to the average level of the index in the year 2000.

Macroeconomic Risks

Macroeconomic shocks with the potential to trigger a sharp market correction, given existing conditions in capital markets (Figure 1.36). Our principal assessment of the macroeconomic risks is based on the analysis contained in the WEO and is consistent with the overall conclusion reached in that report on the outlook and risks for global growth (see, in particular, Figure 1.12 of the April 2008 WEO). We complement that analysis by examining various economic confidence measures. The first of these is a GDP-weighted sum of confidence indices across the major mature markets to determine whether businesses and consumers are optimistic or pessimistic about the economic outlook. A second component is a “surprise” index that shows whether data releases are consistently surprising financial markets on the upside or downside. The aim is to capture the extent to which informed participants are likely to have to revise their outlook for economic growth. Third, recognizing the importance of turning points between expansions and slowdowns of economic activity, we incorporate changes in the Organization for Economic Cooperation and Development’s composite leading indicator. Finally, to gauge inflection points in global trade, we include global trade growth estimates implied by the Baltic Dry Index, a high-frequency indicator based on the freight rates of bulk raw materials that is commonly used as a leading indicator for global trade.

Figure 1.36.Global Financial Stability Map: Macroeconomic Risks

Sources: IMF, World Economic Outlook; Bloomberg L.P.; Dresdner Kleinwort; OECD; The Baltic Exchange; and IMF staff estimates.

Note: Dashed lines are period averages. Vertical lines represent data as of the October 2007 GFSR.

1The 2008 revised datapoint accounts for skewness in the distribution of risks to the baseline forecast.

2Amplitude adjustment is carried out by adjusting mean to unity and the amplitude of the raw index to agree with that of the reference series by means of a scaling factor.

3The Baltic Dry Index is a shipping and trade index measuring changes in the cost of transporting raw materials such as metals, grains, and fuels by sea.

Emerging Market Risks

Underlying fundamentals in emerging markets and vulnerabilities to external risks (Figure 1.37). These risks are conceptually separate from, though closely linked to, macroeconomic risks insofar as they focus only on emerging markets. Using an econometric model of emerging market sovereign spreads, we identify the movement in Emerging Market Bond Index Global (EMBIG) spreads accounted for by changes in fundamentals, as opposed to the movement in spreads attributable to other factors. Included in the fundamental factors are changes in economic, political, and financial risks within the country.56 This is complemented with a measure of the trend in actions by sovereign rating agencies, such as Moody’s and Standard & Poor’s, to gauge changes in the macroeconomic environment and progress in reducing vulnerabilities arising from external financing needs. We also measure fundamental conditions in emerging market countries that are separate from those related to sovereign debt, particularly given the reduced need for such financing in many emerging market countries, by including an indicator of growth in private sector credit. Other components of the subindex include a measure of the volatility of inflation rates, and a measure of corporate credit spreads relative to sovereign counterparts.

Figure 1.37.Global Financial Stability Map: Emerging Market Risks

Sources: Bloomberg L.P.; JPMorgan Chase & Co.; The PRS Group; IMF, International Financial Statistics; Credit Suisse; and IMF staff estimates.

Note: Dashed lines are period averages. Vertical lines represent data as of the October 2007 GFSR.

1EMBIG = Emerging Market Bond Index Global. The model excludes Argentina because of breaks in the data series related to debt restructuring. Owing to the short data series, the model also excludes Indonesia and several smaller countries. The analysis thus includes 32 countries.

2Net actions of upgrades (+1 for each notch), downgrades (–1 for each notch), changes in outlooks (+/– 0.25), reviews and creditwatches (+/–0.5).

344 countries.

4Average of 12-month rolling standard deviations of consumer price changes in 25 emerging markets.

Credit Risks

Changes in and perceptions of credit quality that have the potential for creating losses resulting in stress to systemically important financial institutions (Figure 1.38). Spreads on a global corporate bond index provide a market-price-based measure of investors’ assessment of corporate credit risk. We also examine the credit-quality composition of the high-yield index to identify whether it is increasingly made up of higher- or lower-quality issues, calculating the percentage of the index comprised of CCC or lower-rated issues. We also incorporate forecasts of the global speculative default rate produced by Moody’s. Another important component of the subindex is a Banking Stability Index (see Box 1.5), which represents the expected number of defaults among large complex financial institutions (LCFIs), given that at least one LCFI defaults. This index is intended to highlight market perceptions of systemic default risk in the financial sector. Finally, to capture broader credit risks, we include delinquency rates on a wide range of noncorporate credit, including residential and commercial mortgages and credit card loans.

Figure 1.38.Global Financial Stability Map: Credit Risks

Sources: Merrill Lynch; Moody’s; Bloomberg L.P.; Mortgage Bankers Association; Federal Reserve; and IMF staff estimates.

Note: Dashed lines are period averages. Vertical lines represent data as of the October 2007 GFSR.

130-, 60-, and 90-day delinquencies for residential and commercial mortgages, and credit card loans in the United States..

Market and Liquidity Risks

The potential for instability in pricing risks that could result in broader spillovers and/or mark-to-market losses (Figure 1.39). An indicator attempting to capture the extent of market sensitivity of hedge fund returns provides an indirect measure of institutional susceptibility to price changes. The subindex also includes a speculative positions index, constructed from the noncommercial average absolute net positions relative to open interest of a range of futures contracts as reported to the Commodity Futures Trading Commission. These typically rise when speculators are taking relatively large positional bets on futures markets, relative to commercial traders. Also included is an estimation of the proportion of return variance across a range of asset classes that can be explained by a common factor. The higher the correlations across asset classes, the greater the risk of a disorderly correction in the face of a shock. An additional indicator is an estimate of equity risk premia in mature markets using a three-stage dividend discount model. Low ex ante equity risk premia may suggest that investors are underestimating the risk attached to equity holdings and so increasing potential market risks. There is also a measure of implied volatility across a range of assets. Finally, to capture perceptions of funding, secondary market trading, and counterparty risks, we incorporate the spread between major mature market government securities yields and interbank rates, the spread between interbank rates and expected overnight interest rates, bid-ask spreads on major mature market currencies, and daily return-to-volume ratios of equity markets.

Figure 1.39.Global Financial Stability Map: Market and Liquidity Risks

Sources: Credit Suisse Tremont Index LLC; Bloomberg L.P.; JPMorgan Chase & Co; IBES; Morgan Stanley Capital International; and IMF staff estimates.

Note: Dashed lines are period averages. Vertical lines represent data as of the October 2007 GFSR.

136-month rolling regressions of hedge fund performance versus real asset returns.

2Data represent the absolute value of the net position taken by noncommercial traders in 17 select U.S. futures markets. High values are indicative of heavy speculative positioning across markets, either net-long or net-short.

3Represents an average z-score of the implied volatility derived from options from stock market indices, interest, and exchange rates. A value of 0 indicates the average implied volatility across asset classes is in line with the period average (from 12/31/98 where data are available). Values of +/–1 indicate average implied volatility is one standard deviation above or below the period average.

4Based on the spread between yields on government securities and interbank rates, term and overnight interbank rates, currency bid-ask spreads, and daily return-to-volume ratios of equity markets. A higher value indicates tighter market liquidity conditions.

Annex 1.2. Methodology for Calculating Global Losses and Bank Exposures57

This annex describes the methodology for estimating losses on holdings of U.S. residential and commercial mortgages, consumer credit, and corporate debt.

Loss estimates vary widely depending on the methodology employed. Our estimates are based on potential loan losses that have occurred since the subprime crisis began and over the next two years, consistent with the period of expected slowing of the U.S. economy and mark-to-market losses on related securities over the course of the past year reflecting the credit deterioration that has occurred and is anticipated to occur. The objective of the analysis is to identify the scale of losses that market participants have already recognized and could potentially recognize in the period ahead. Losses on loans are based on projections of cash flow shortfalls, while losses on securities are based on changes in the market pricing of cash and derivative indices.

The loans captured in the exercise include subprime, alt-A, prime residential and commercial real estate mortgages, consumer, corporate, and leveraged loans. Securities include ABS and ABS CDOs based on subprime and alt-A residential mortgage loans, prime MBS, CMBS, auto loan and credit card ABS, CLOs, and high-yield and investment-grade corporate debt.

Losses on different types of loans were estimated from regression analysis using various relevant factors, such as changes in unemployment, lending standards, and housing and commercial real estate pricing, as relevant. In each case, the outstanding stock of the type of loan was multiplied with the change in the forecasted loss (charge-off) rate. The underlying historical data on loan loss rates and changes in lending standards were obtained from the Federal Reserve. Although the loan loss data are for banks only, it was assumed that loans held by other lenders would exhibit similar performance.

Losses on residential and commercial mortgages were also estimated by a second procedure. This one involved a three-step process. We first estimated the percentage of loans that would become delinquent, then the percentage of delinquent loans that would default, and finally losses on defaulted loans after completion of the foreclosure or recovery process. Each of these steps is detailed below.

In the first step, we projected delinquencies on residential and commercial loans over a multi-year period using historical patterns and the current trajectory of recent vintage loans. An average delinquency for each loan type (prime, alt-A, subprime, and commercial) was computed by weighting the maximum projected delinquency on loans issued each year by the size of issuance. In the second step, 70 percent of prime, alt-A, and commercial real estate loans were assumed to convert from late stage (60-day) delinquency into default. One hundred percent of 60-day delinquent subprime loans were assumed to default. These figures are broadly consistent with market estimates.

Box 1.5.Banking Stability Index

Simultaneous large losses in several banks can affect a banking system’s financial stability, and so the likelihood of such an event needs to be monitored and measured. This box describes the banking stability index and additional indicators.

The proper estimation of default dependence among banks is vital for financial stability surveillance because banks are usually linked—either directly, through the interbank deposit market, or indirectly, through lending to common sectors. This default dependence varies across the economic cycle, rising in times of distress so that the fortunes of banks decline concurrently. Thus, simultaneous large losses in several banks could affect stability in the overall banking system. Supervisors should assess both the risk of large losses and possible default of a specific bank, and the impact that this would have on other banks in the system.

To model the stability of the banking system, we follow Goodhart and Segoviano (forthcoming) in treating the banking system as a portfolio of banks. Then, using market-based probabilities of default (PoDs) of individual banks, and employing a novel nonparametric copula approach, we derive the joint probability of default (JPoD) of the banking system.1 The JPoD represents the probability of all the banks in the portfolio going into default, that is, the tail risk of the system. In periods of financial distress, the banking system’s JPoD may experience larger and nonlinear increases than those experienced by the PoDs of individual banks. Based on the JPoD, we estimate a Banking Stability Index (BSI), which reflects the expected number of bank defaults given that at least one bank defaults. A higher number signifies greater instability. This framework allows for the estimation of additional measures of stability, including the probability that each bank in the system will default, given that another bank in the system defaults. Such pair-wise conditional probabilities provide insights into the likelihood of contagion and can be presented in a default contagion matrix (DCo).

Tail Risk and Average Probability of Default1

(In percent)

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

1From January 1, 2007 to March 10, 2008, the average probability of default increased by a factor of 14.8, while the JPoD, measure of tail risk, increased by a factor of 203.6.

2Joint probability of 15 simultaneous defaults.

3Unweighted average of individual banks’ probabilities of default.

To examine the effects of the current credit turmoil on the banking system, the average PoD for a portfolio of 15 systemically important large and complex financial institutions (LCFIs) is compared with changes in the system’s JPoD.2 As stress grew from mid-2007 to the present, the JPoD increased more than 10 times than the average PoD. The difference is mainly explained by an increased default dependence among the banks in the system, which has significantly augmented the tail risk in the system (see first figure) and sharply increased the BSI.3 This increased instability was driven by banks under greater stress, which can be seen when grouping the 15 LCFIs into two categories; that is, lesser-stressed banks (L) and higher-stressed banks (H).4 As the credit woes worsened, the JPoD for each group increased significantly, though more severely for H banks, indicating that tail risk within the H banks increased more sharply (see second figure). Also, contagion among the H banks is higher, as indicated by the mean-DCo (see third figure). These estimations provide evidence that a bank’s resilience to shocks is affected by the overall resilience of the other banks within the financial system. Thus, unless banks’ default dependence is taken into account, supervisors may not accurately estimate the banking system’s stability.

Joint Probability of Default (JPoD) for Banks with Higher and Lower Losses1

(In percent)

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

1On March 5, 2008, the JPoD for H banks was higher than that for L banks by a factor of 1.3.

Mean Default Contagion (DCo) for Banks with Higher and Lower Losses1

(In percent)

Source: IMF staff estimates.

1Unweighted average of pair-wise conditional probabilities of default, which indicate that H banks or L banks default given that any other bank (B) defaults. In order to keep the confidentiality of the analyzed institutions, we report the mean-DCo, rather than the institution-specific DCo. From January 1, 2007 to March 5, 2008, the mean-DCo rose by a factor of 1.5 among L banks, and by a factor of 1.7 among H banks.

Note: The main author of this box is Miguel Segoviano.1 The structure of linear and nonlinear dependencies among banks in a system can be represented by copula functions. Our approach infers copulas from the joint movement of individual banks’ PoDs. This is in comparison with traditional approaches, in which parametric copulas have to be chosen and calibrated explicitly—usually a difficult task, especially under data constraints.2 ABN Amro, Bank of America, Bear Stearns, BNP, Citigroup, Credit Suisse, Deutsche Bank, Goldman Sachs, HSBC, JPMorgan, Lehman Brothers, Merrill Lynch, Morgan Stanley, UBS, and Société Générale.3 The BSI is used to construct the credit risk component of the global financial stability map.4 This classification was based on the expected size of banks’ losses due to subprime mortgage exposures relative to Tier 1 capital. The methodology used for this classification is further explained in Annex 1.2.

For the final step, the recovery rate of principal from the foreclosure process was assumed to be 60 percent for prime and commercial real estate loans, and 50 percent for both alt-A and subprime loans. The loss on each category of residential and commercial loans was computed as the vintage-weighted delinquency times the conversion-to-default rate multiplied by the loss given default (or one minus the recovery rate). Average projected cash flow losses were estimated to be 15 percent of principal for subprime, 5 percent for alt-A, 1 percent for prime, and 1 percent for commercial loans.

Losses for securities were next estimated by multiplying the outstanding stock of each type of security by the change in the market price of the relevant index over the course of a year. The average price change was obtained by weighting price changes for constituent indices comprised of different vintages and ratings by the issuance in each of these categories.

Beginning with the residential mortgage market, subprime-related ABS and CDO securities were priced using ABX and TABX derivative indices, respectively. Average losses on securities were estimated as 30 percent of principal for ABS and 60 percent for ABS CDOs since last year. The corresponding dollar loss estimates for subprime and alt-A securities were adjusted for any overlap of losses on ABS with those on CDOs. For prime-mortgage-related securities, conforming and nonconforming issues were treated separately and weighted appropriately. The prices of on-the-run agency pass-through securities were used as reference for conforming securities, while quotes on pools of jumbo loans were used to represent the pricing of nonconforming securities. Spreads on agency pass-throughs have widened versus U.S. Treasury securities, as have spreads on pools of jumbo loans versus agency securities. However, the absolute change in market prices of these prime securities has been positive over the course of the past year because of falling yields on U.S. treasuries. No losses were therefore estimated on holdings of prime securities.

Appropriately weighted indices were also used for other types of securities: CMBS, consumer ABS, and corporate debt. The CMBX derivative indices were used to estimate losses on CMBS, while cash indices were used for consumer ABS (autos and credit cards), investment-grade corporate debt, high-yield debt, and for the LCDX for CLOs. No losses were estimated for holdings of consumer ABS or investment-grade U.S. corporate debt, as corresponding indices have been positive over the last year.

The loss estimates are subject to the following caveats and uncertainties:

  • The fall in market prices may be overshooting potential declines in cash flows over the lifetime of underlying loans.

  • Projected delinquency patterns may not fully account for recent structural changes in markets, including a rise in the proportion of adjustable-rate mortgages likely to experience rate resets in the near term.

  • Falling U.S. house prices and further deterioration in the macroeconomic environment could increase rates of delinquency, default, and loss. Conversely, fiscal stimulus, monetary easing, and loan modification measures could lower these rates.

Based on this approach, we estimate total losses from broad credit market deterioration of $945 billion globally, $565 billion of which is due to losses on residential mortgage debt, $240 billion on commercial real estate debt, $120 billion on corporate debt, and $20 billion on consumer credit debt.58 Securitized debt (rather than whole loans) accounts for the bulk of losses (Table 1.5).

Table 1.5.Losses by Asset Class as of March 2008(In billions of U.S. dollars)
Base Case Estimates of Losses on Unsecuritized U.S. Loans
OutstandingEstimated loss October 2007 GFSREstimated loss March 2008
Subprime3003045
Alt-A6001030
Prime3,800Not estimated40
Commercial real estate2,400Not estimated30
Consumer loans1,400Not estimated20
Corporate loans3,700Not estimated50
Leveraged loans170Not estimated10
Total for loans12,37040225
Base Case Estimates of Mark-to-Market Losses on Related Securities
OutstandingEstimated mark-to-market loss October 2007 GFSREstimated mark-to-market loss March 2008
ABS1,10070210
ABS CDOs400130240
Prime MBS3,800Not estimated0
CMBS940Not estimated210
Consumer ABS650Not estimated0
High-grade corporate debt3,000Not estimated0
High-yield corporate debt600Not estimated30
CLOs350Not estimated30
Total for securities10,840200720
Total for loans and securities23,210240945
Sources: Goldman Sachs; JPMorgan Chase & Co.; Lehman Brothers; Markit.com; Merrill Lynch; and IMF staff estimates.Note: ABS = asset-backed security; CDO = collateralized debt obligation; CLO = collateralized loan obligation; CMBS = commercial mortgage-backed security; MBS = mortgage-backed security.

Banks globally are expected to shoulder roughly half of the subprime mortgage-related losses, based on bottom-up analysis using publicly disclosed exposures. Specifically, banks are estimated to have $740 billion of net subprime exposure, mostly held by U.S. banks (53 percent), with the remainder held by European (41 percent), Asian (5 percent), and Canadian (1 percent) banks. In terms of composition, U.S. banks (together with government-sponsored enterprises) hold a greater proportion of overall exposure to the subprime market through unsecuritized subprime loans and ABS CDOs compared with European banks. On the other hand, European banks hold a greater proportion of their exposure to the subprime market via ABS. Banks are assumed to hold the most senior tranches.

Based on average loss estimates of 15 percent for unsecuritized mortgage loans, 30 percent on ABS, and 60 percent on ABS CDOs as described above, potential losses of U.S. banks ($144 billion) are likely to be similar to those borne by European banks ($121 billion). Losses of Asian banks are likely to be less than one-tenth of losses in Europe. More than half of the aggregate subprime-related loss would likely come from exposure to CDOs, while the remainder is expected to come from ABS, unsecuritized subprime loans, and losses on off-balance-sheet liquidity lines. In particular, potential losses on off-balance-sheet conduit and SIV liquidity lines could result in $40 billion of losses globally ($27 billion for European banks and $13 billion of losses for U.S. banks). These estimates are based on the assumption of an average loss of 5 percent on liquidity lines to off-balance-sheet conduits and SIVs. The 5 percent loss assumption is based on losses on a typical asset composition for conduits and SIVs. Losses on conduit assets are assumed to pass directly to the liquidity line, but losses on SIV assets are assumed to be mostly absorbed by the junior notes, given their funding structures (see Box 2.5 in Chapter 2). Conduits and SIVs are weighted by their market proportions—90 percent and 10 percent of the total, respectively—and it is assumed that all liquidity lines eventually get called.

Through mid-March 2008, banks had reported $190 billion in losses on U.S. mortgage market exposure. Much of that, however, represents mark-to-market losses, and some could yet be recoverable going forward. Most of subprime-related losses appear to have been reported already. U.S. banks and government-sponsored enterprises could report a further $49 billion in additional writedowns, while European banks could report as much as $43 billion in additional writedowns (Table 1.6). These loss estimates should be regarded with caution for the following reasons:

Table 1.6.Global Bank Losses as of March 2008(In billions of U.S. dollars)
Country/RegionTotal Reported LossesEstimated Losses on U.S. Subprime/Alt-A LoansEstimated Losses on ABSEstimated Losses on CDOsEstimated Losses on Conduits/SIVsTotal Estimated Subprime-Related LossesRemaining Subprime-Related Losses Expected
Europe801627532712343
Of which:
United Kingdom1916112114022
Switzerland2307151230
Scandinavia0000111
Euro area3301020154512
Unallocated50960149
United States952912901314449
Asia excluding Japan1030043
Of which: China1030032
Japan100550100
Asia110950133
Canada7025070
Gulf Cooperation Council1011010
Total19344501534028895
Sources: Goldman Sachs; UBS; and IMF staff estimates.Note: Bank allocation to asset-backed securities (ABS) in Table 1.1 includes estimated losses on ABS and conduits/SIVs. CDO = collateralized debt obligation; SIV = structured investment vehicles.
  • Loss estimates ultimately depend on the quality of disclosure about holdings. Where data have not been available, we have used estimates of exposure to subprime loans, ABS, and CDOs.

  • Because the loss ratio on CDOs differs from that on unsecuritized loans, the aggregate loss estimate is highly sensitive to the estimated proportions of bank exposure accounted for by unsecuritized loans, ABS, and CDOs.

  • The timing of loss recognition is uncertain. UK banks, in particular, appear to have significant exposure to unsecuritized loans, for which it may take some time to recognize losses relative to holdings of securities. There are also differences in methodology across countries regarding recognition of losses.

  • Estimates are also sensitive to the breakdown of exposure to different tranches of securities, as there is substantial variation in the pricing on which the mark-to-market estimates are based. For instance, a recent vintage AAA-rated ABX is quoted at 75 cents on the dollar, while a subordinated A-rated tranche of a different vintage is quoted at 16 cents. Lack of information appears to be an even bigger problem in Asia, including in Japan, where the breakdown of bank holdings of ABS and CDOs is largely unavailable.

  • Estimates of bank exposure to ABS and CDOs rely upon market indices, which may not represent the secondary market prices of actual bank holdings, as individual ABS and CDO tranches held by banks could have significantly different collateral and cash flow characteristics.

  • Implementation of remedial measures, including modification of mortgage loan terms, could lower loss estimates.

References

    Avesani, R., K.Liu, A.Mirestean, and J.Salvati, 2006, “Review and Implementation of Credit Risk Models of the Financial Sector Assessment Program,IMF Working Paper 06/134 (Washington: International Monetary Fund).

    Bank of England, 2007, “Markets and Operations,Bank of England Quarterly Bulletin—Q4, Vol. 47, No. 4, pp. 490510.

    Chan-Lau, J., S.Mitra, and L.Ong, 2007, “Contagion Risk in the International Banking System and Implications for London as a Global Financial Center,IMF Working Paper 07/74 (Washington: International Monetary Fund).

    Égert, Balázs, and DubravkoMihaljek, 2007, “Determinants of House Price Dynamics in Central and Eastern Europe,CESifo Working Paper No. 2152 (November). Available via the Internet: http://www.cesifo.de/DocCIDL/cesifo1_wp2152.pdf.

    Froot, Kenneth, and Paul G.J.O’Connell, 2003, “The Risk Tolerance of International Investors,NBER Working Paper No. 10157 (Cambridge, Massachusetts: National Bureau of Economic Research).

    Goodhart, Charles, and MiguelSegoviano, forthcoming,Banking Stability Index,IMF Working Paper (Washington: International Monetary Fund).

    Gray, D., R.Merton, and Z.Bodie, 2007, “New Framework for Measuring and Managing Macrofinancial Risk and Financial Stability,NBER Working Paper No. 13607 (Cambridge, Massachusetts: National Bureau of Economic Research).

    International Monetary Fund (IMF), 2007a, Global Financial Stability Report, World Economic and Financial Surveys (Washington, October).

    International Monetary Fund (IMF), 2007b, Global Financial Stability Report,World Economic and Financial Surveys (Washington, April).

    International Monetary Fund (IMF), 2008, World Economic Outlook,World Economic and Financial Surveys (Washington, April).

    SegovianoBasurto, MiguelA., 2006, “Portfolio Credit Risk and Macroeconomic Shocks: Applications to Stress Testing under Data Restricted Environments,IMF Working Paper 06/283 (Washington: International Monetary Fund).

Note: This chapter was written by a team led by Peter Dattels and comprised of Sergei Antoshin, Sean Craig, Martin Edmonds, Kristian Hartelius, Phil de Imus, Rebecca McCaughrin, Ken Miyajima, Michael Moore, Chris Morris, Mustafa Saiyid, Ian Tower, and Chris Walker.

Annex 1.1 details how indicators that compose the rays of the map are measured and interpreted. The map provides a schematic presentation that incorporates a degree of judgment, serving as a starting point for further analysis.

Credit risks measure changes in credit quality that have the potential for creating losses resulting in stress to systemically important financial institutions.

Indicators on market and liquidity risks measure the potential for instability in funding and pricing risks that could result in broader spillovers and/or mark-to-market losses.

Monetary and financial conditions represent a broader measure than that presented in the WEO, in that they incorporate both quantity and price aspects, whereas the WEO metric only captures price effects. See Annex 1.1 for further details and Figure 1.4 in the April 2008 WEO (IMF, 2008).

Nonprime refers primarily to subprime and alt-A mortgages. Subprime loans are typically made to borrowers that display one or more of the following characteristics at the time of origination: weakened credit histories that include payment delinquencies and bankruptcies; reduced repayment capacity as measured by credit scores or debt-to-income ratios; or incomplete credit histories. Alt-A mortgages, though of higher quality than subprime mortgages, are considered lower credit quality than prime mortgages due to one or more nonstandard features related to the borrower, property, or loan.

“Risk layering” refers to the practice whereby mortgage lenders combine nontraditional mortgages with weaker credit controls, for instance, by accepting high combined loan-to-value ratios, reduced documentation, and little or no downpayment.

As of the third quarter of 2007, 43 percent of foreclosures were on subprime ARMs, 19 percent on prime ARMs, 18 percent on prime fixed-rate mortgages, 12 percent on subprime fixed-rate mortgages, and 9 percent on loans with insurance protection from the Federal Housing Administration. That foreclosures have been dominated by ARMs likely reflects the shift in the mortgage landscape from fixed to floating rates over the last few years. Indeed, anecdotal evidence suggests that foreclosures have primarily occurred well ahead of the reset period, suggesting that the deterioration thus far has been a function of fraud, speculation, over-extension by borrowers, and the effects of weak underwriting standards.

In 2008, $250 billion of subprime mortgages are scheduled to reset, versus $82 billion in prime mortgages and $29 billion in alt-A mortgages. Easier monetary policy (and hence lower six-month LIBOR rates to which ARMs are traditionally indexed) helps to alleviate, but not fully eliminate, some payment shock as ARMs reset. IMF staff estimates suggest that average monthly payments on subprime mortgages will still rise by roughly 15 percent upon reset, while alt-A and jumbo interest-only ARMs will be subject to even higher payment shock, as borrowers are required to amortize their principal at the initial reset. Moreover, it will be difficult for borrowers to benefit fully from any further monetary policy easing, since most ARMs have floors and caps. Refinancing would be difficult in the current environment of tighter lending conditions or just as costly, since fixed rates on mortgages are still elevated.

The prime mortgage market is comprised of loans, which conform to the standards of government-sponsored entities (GSEs), and jumbo loans extended to creditworthy borrowers who do not conform to the GSEs’ criteria for securitization.

Econometric work suggests that the deterioration in lending standards typically contributes only partially to the deterioration in prime mortgage performance, with other factors, especially the unemployment rate, proving to be a more important determinant.

Many UK borrowers coming off fixed rates will face rate increases of 100 to 200 basis points.

As mentioned in the October 2007 GFSR, UK non-conforming loans have some features in common with U.S. nonprime loans (IMF, 2007a). Lending criteria for UK nonconforming loans were tightened in late 2007 and early 2008.

For instance, an increasing proportion of new loans were full-term, interest-only loans. Such loans do not amortize until the final payment, and thus offer less amortization over the life of the loan than other types of mortgages. In addition, subordination levels in securitized products declined, typical of the countercyclical pattern observed in rating cycles. Only in early 2007 did the major rating agencies begin to require higher subordination levels on new deals, leading to some improvement in credit quality later in the year.

Technical factors may have played a role in the spread widening, as speculative and hedging activity shifted from the ABX to the CMBX, indices of credit default swaps linked to a subset of underlying subprime and commercial mortgage-backed securities, respectively.

As of 2007, U.S. households held $2.5 trillion in consumer debt in the form of revolving ($900 billion), primarily credit card debt, and nonrevolving debt ($1.6 trillion), most of which is auto loans. The securitized market represents roughly $780 billion, spanning a wide range of assets, including credit cards ($343 billion), auto leases ($199 billion), student loans ($236 billion), and other miscellaneous securitized loans.

A charge-off occurs when payments are no longer collectible, due either to bankruptcy or default.

Consumer debt grew at an average annual rate of 5 percent during 2002–06 compared with the 12 percent growth rate of secured mortgage debt, which included home equity loans.

Consumer charge-off rates dropped significantly after a spate of accelerated personal bankruptcies in late 2005 before the implementation of a stricter bankruptcy law.

An alternative explanation could be that markets are anticipating a deeper downturn and retrenchment of credit card debt, which would increase the correlation among the underlying individual risks, and would have an impact on valuation and capital requirements.

Over the last five years, low-tier bonds accounted for an average of 21 percent of total high-yield debt issuance (peaking at 37 percent in 2007), compared with an average of 15 percent in 1998, which preceded escalating defaults over 1999–2002. Typically, 60 percent of CCC-rated bonds default before they mature, and 36 percent default within three years of issuance.

Leverage was needed to boost returns over the last few years, owing to a lack of distressed debt. This led to 7 times (and sometimes as much as 10 times) leverage on U.S. leveraged buyouts. In Europe, debt multiples also were stretched, with leverage of 5.5 times in 2007, versus 4.7 times in 1998.

The increase in “covenant-lite” loans may hinder early intervention by lenders, possibly delaying some defaults until later in the cycle, but potentially increasing the probability of default.

Loss estimates vary considerably, given different assumptions about inputs and valuation methods, so IMF staff estimates should be regarded as merely an exercise to help gauge the indicative magnitude of risks to the financial system. We estimate losses in two parts as indicated in Table 1.1, which is a composite of market-implied accumulated losses in the securitized markets and potential loan losses associated with the slowdown in economic activity. The top panel estimates projected losses on unsecuritized loans, net of recoveries, on real estate, consumer, and corporate loans, based on projected shortfalls in cash flows in the near term. Underpinning cash flow estimates is an expected deterioration in the U.S. economy, consistent with increasing macroeconomic risks highlighted in the global financial stability map and detailed in the April 2008 WEO.

Note the term “losses” used in this context refers to potential writedowns, as opposed to negative net profits.

ABS prices are based on the ABX, an index of credit default swaps linked to 20 underlying subprime mortgages. ABS CDO prices are based on the TABX, an index that tranches synthetic CDOs based on the BBB– and BBB ABX indices.

It should be noted that the current scenario is not directly comparable to prior crises, since the subprime crisis reflects potential estimated losses to financial institutions, some of which have yet to occur.

The exposure of market participants to losses is uncertain partly because placement data for various types of securities are imprecise.

These estimates are subject to considerable uncertainty given the limited information on individual banks’ exposures, especially to credit default swap (CDS) contracts written by financial guarantors and used by banks to hedge CDOs.

Initiatives to resolve the problems affecting some financial guarantors are continuing. The New York state insurance regulator has been working with banks on plans to recapitalize and potentially restructure those companies most affected by losses on structured finance business. Some of the companies have now raised new capital, enabling them to retain AAA ratings for the time being. But it remains unclear whether there will be further ratings downgrades of financial guarantors in the future. The New York regulator has committed to a review of its regulatory approach to financial guaranty business.

Several financial guarantors have already been downgraded.

The requirement to post margins mitigates this risk. A protection seller posts an initial margin (2 to 3 percent) and from then on daily margin equal to changes in the market value of the underlying security. Therefore, unless defaults increase abruptly and are largely unanticipated, most market participants will not experience substantial margin calls over a short period.

The 10 largest market makers account for close to 90 percent of the $45 trillion outstanding notional value of CDS.

This GFSR enhances the use of credit-derivatives-based credit risk indicators used in prior GFSRs to monitor the evolution of market perceptions of default risk in mature market financial systems. The mature market credit risk indicators measure the expected number of bank defaults given at least one bank default for 15 financial institutions, implied from the prices of CDS. See Box 1.5 for details.

Many hedge funds had negotiated “margin locks” that prevented their prime brokers from increasing the margins they pay when borrowing securities, or the “haircuts” they pay when pledging securities as collateral with their brokers for a fixed period of time.

In the United States, Federal Home Loan Banks have also stepped in to re-intermediate the credit market.

Figure 1.20 subtracts the average CDS spread referencing U.S. banks from the 1-year LIBOR overnight index swap spread to give an indicative decomposition into a credit and other component, the residual of which likely represents liquidity. See Bank of England (2007, pp. 499–500) for more detail.

So far, Swedish banks have been able to access euro funding through private placements with European investors, and the Swedish covered bond market has continued to function even when the European market has shut down.

The Romanian leu is the only floating currency with a liquid forward market among the group of eastern European countries with large external imbalances. It has depreciated substantially since July 2007, as some investors have expressed negative views on the region as a whole.

Indian corporations had net cross-border obligations of $31 billion as of September 2007, while Indian banks had very limited net exposure as of January 2008, according to the Bank for International Settlements. The October 2007 GFSR cited estimates that up to one-half of Indian firms’ short dollar positions had been swapped into yen (IMF, 2007a). Market sources suggest that the ratio of yen borrowing has likely diminished since then.

In fact there have already been some writedowns. For example, one Korean bank has written down $440 million in mortgage-backed CDO exposure and $20 million in nonmortgage-backed CDO exposure.

Currency volatilities have risen across the board, in both actual and implied terms, for mature and emerging market currencies.

A cross-border carry trade is normally defined as the combination of a short position in a lower-yielding currency with a long position in a higher-yielding currency, with the aim of collecting the interest rate differential between the two. Such trades can be highly leveraged and entail exposure to currency risk.

See the April 2008 WEO for other sources of over-heating, including high energy and food prices in some emerging market economies (IMF, 2008).

The “Asia play” can be loosely defined as the purchase of Asian-currency-denominated assets on the view that the local currency will likely appreciate against the dollar, especially if authorities are expected to reduce the scope of interventions.

CLOs are securitized packages of leveraged loans. A market-value CLO is one in which the manager has latitude to trade assets within the portfolio. Payments to investors come from both cash flows from the underlying assets and sales of some assets. Payments to tranches are not contingent on the adequacy of the underlying assets’ cash flows (as in a “cash-flow CLO”), but rather on whether the market value of the CLO exceeds certain thresholds. If those thresholds are breached, an automatic unwind of the structure is triggered to protect the position of the senior creditors.

The $175 billion or so of leveraged loans include the $17 billion issued by Bell Canada Enterprises, $15 billion by Clear Channel Communications, $10.5 billion by Alltel, $6 billion by Harrah’s Entertainment, and $8.8 billion by the Texas Utility Corporation. The remainder is high-yield bonds.

For example, demand for auction rate securities issued by student loan lenders and some U.S. municipalities have fallen dramatically. Similar dislocations are observed in the tender option bond (TOB) sector, primarily reflecting concerns that a downgrade of a financial guarantor will lead to a downgrade of the municipal bonds that serve as collateral for TOB products.

The shock will be mitigated to the extent banks can raise fresh capital, either from existing shareholders or from new ones (see Box 1.2). Other important factors include the rate at which losses are recognized, the amount of profits insulated from the credit crunch, and the extent to which some banks (and rating agencies) tolerate a temporary dip in capital ratios.

The model includes two lags, which is what the Schwarz information criterion prescribes for this particular sample. Parameters are stable according to Quandt-Andrews tests.

The data on borrowing and debt are from the Federal Reserve’s Flow of Funds Accounts. Borrowing is defined as the increase in credit market liabilities for households and nonfarm, nonfinancial corporations. It includes mortgages, consumer credit, bank loans, and issuance of commercial paper and corporate bonds. Over the sample period, private sector borrowing has averaged 8.8 percent of outstanding private sector debt, quarterly annualized, with a standard deviation of 2.9 percent.

The impulse response function is based on Cholesky decomposition, with the variables ordered as above. One caveat is that this simple model cannot distinguish between demand and supply shocks to credit. Figure 1.33 introduces three sequential shocks to borrowing, which bring borrowing growth down to 4 and 1 percent in a credit squeeze and a credit crunch, respectively. The simulation takes into account the model’s endogenous path for borrowing, as well as the dynamic effects of previous shocks.

In response to the crisis of confidence, market participants have already begun to strengthen their due diligence. With less support from rating agencies, financial guarantors, and traditional prepayment and cash flow models, though, credit analysis is now more operationally intensive. For instance, in the mortgage sector, each loan in a pool must be analyzed to determine equity build-up, prepayment history, triggers, and other credit attributes to forecast borrower behavior. Typically, each pool has 7,000 loans, with 70 different credit attributes across each pool that must be analyzed against several different home price scenarios.

The main author of this annex is Ken Miyajima.

The estimated changes in relative risk tolerance of institutional investors from Froot and O’Connell (2003) are aggregated using a slow, exponentially weighted moving average in order to account for slow-moving secular changes in the data. The index is scaled and rebased so that 100 corresponds to the year 2000.

The index represents the value of the coefficient of risk aversion, constrained to values between 0 and 10.

The model uses three fundamental variables to fit EMBIG spreads: economic, financial, and political risk ratings. The economic risk rating is the sum of risk points for annual inflation, real GDP growth, the government budget balance as a percentage of GDP, the current account as a percentage of GDP, and GDP per capita as a percentage of the world average GDP per capita. The financial risk rating includes foreign debt as a percentage of GDP, debt service as a percentage of GDP, net international reserves as months of import cover, exports of goods and services as a percentage of GDP, and exchange rate depreciation over the last year. The political risk rating is calculated accounting for 12 indicators representing government stability and social conditions.

The main author of this annex is Mustafa Saiyid.

58

Losses on the residential mortgage market were estimated as the sum of losses on subprime, alt-A, and prime loans, as well as on ABS, ABS CDOs, and prime mortgage securities. Losses on corporate debt were estimated as the sum of losses on corporate and leveraged loans, as well as on related securities, including investment-grade debt, high-yield debt, and CLOs.

    Other Resources Citing This Publication