A global reappraisal of credit risk and a repricing of financial assets commenced in the summer of 2007 against the background of mounting tensions in the U.S. subprime mortgage market. Reflecting the pervasive uncertainty about the magnitude of the risks faced by financial institutions, frictions spread across asset classes and throughout Europe. Despite the severity of the crisis, Europe’s financial systems have remained relatively resilient thus far. Yet financial system resilience is likely to be tested further, as loss recognition catches up and additional risks remain on the horizon. Although quantification is subject to considerable uncertainty and evidence so far is limited, financial turbulence is likely to take a significant toll on real activity.
How Financial Turbulence Spread to Europe
A reappraisal of credit risk and a repricing of financial assets began in the summer of 2007, stemming from mounting creditworthiness problems in the U.S. subprime mortgage market. After a period of high liquidity, aggressive credit expansion, growing complexity in mortgage securitization, and loosening in underwriting standards, credit markets abruptly changed direction. The pervasive uncertainty about the valuation of structured finance products—and the lack of clarity surrounding the extent of bank on-and off-balance-sheet exposures to these instruments—triggered a reversal in market sentiment and set the stage for a severe liquidity squeeze. The turmoil quickly spread to Europe, prompting bank rescues and capital injections. Despite the width and depth of the crisis, European mortgage markets have remained largely unscathed so far, owing to sounder household balance sheet positions, more cautious credit risk management, and stricter regulation.
Global Risk Repricing and Spillovers
Since end-July 2007, global financial markets have been witnessing a jarring repricing of risk from low historical standards, as investors are demanding more compensation for the credit risk they bear (Figure 14).6 While the correction has not, so far, been particularly notable with regard to the absolute level reached by risk premiums, the speed of the reversal in market sentiment has been remarkable. Besides, the complexity of financing structures and risk transfer strategies in the financial sector are prolonging the repricing process, causing risks to propagate in pervasive and unexpected ways.
Estimating Shifts in the Global Price of Risk, 2007–08 1/
(Basis points)
Sources: Bloomberg L.P.; and IMF staff calculations.1/ The fan chart plots, at each point in time, selected percentiles of the estimated probability distribution for the expected unit price of risk that is common across assets. There is a 50 percent chance that the global price of risk will be inside the blue-shaded range and 90 percent chance that the outturn will be inside either the blue- or the purple-shaded area. The central thick black line denotes the estimated median price of risk. See Lombardi and Sgherri (2008) for analytical underpinnings.With the increased presence of international investors—reacting to any given shock by rebalancing their portfolios in assets and markets that would otherwise seem to be unrelated—the abrupt shift in risk-taking attitude turned out to be the key channel through which financial turbulence spread from the U.S. mortgage market to other assets and countries, including advanced and emerging Europe.7
In advanced Europe, the rise in uncertainty and the drop in confidence among major banks have severely disrupted the interbank market, with money center banks becoming unable to finance large securities portfolios in wholesale markets (Table 3).8 Spreads between the three-month euro London interbank offered rate (LIBOR) and the overnight index swap (OIS) rate have remained wide, indicating some combination of greater preference for liquidity over unsecured lending to banks and widening counterparty risk premiums (Figure 15, panel 1).9 Owing to cash hoarding by international investors, sovereign bond markets have also come under strain (Figure 15, panel 2), with underbidding in some auctions for Italian notes and liquidity tightening in the Greek government debt market. Global risk repricing has also contributed more than 50 percent to the rise in volatility in corporate bond markets; the notable exception to this has been the financial sector, where uncertainty has been related mainly to institution-specific factors (Table 3). In a sign of discrimination among different classes of credit risk, the correction has primarily halted activity and widened the spreads of the high-yield segment of the European corporate bond market (Figure 15, panel 3).10
Reassessing Risks across Asset Classes and Borders
1. EU LIBOR Rate Minus Overnight Indexed Swaps, January 2007–March 2008
(Basis points)
Source: Bloomberg L.P.2. Selected Intra-Euro-Area 10-Year Government Bond Markets—Spread vis-à-vis Germany, January 2007–March 2008
Source: Datastream.3. iBoxx Euro-Corporate Bond Spread over 10-Year German Bond Rate, January 2007–March 2008
(Basis points)
Source: Datastream.Measuring the Impact of Risk Repricing on Selected Markets 1/
Estimation details are provided in Lombardi and Sgherri (2008).
Increase in spread (basis points) following a 100-basis-point increase in risk premium.
Percent of total volatility due to shifts in the global price of risk.
EMBI+ is the JPMorgan Emerging Markets Bond Index Plus.
Measuring the Impact of Risk Repricing on Selected Markets 1/
| Impact on Spread 2/ | Impact on Volatility 3/ | |
|---|---|---|
| U.S. asset-backed commercial paper | 69 | 65 |
| U.S. interbank | 31 | 75 |
| EU interbank | 29 | 52 |
| EU corporate bonds | 28 | 99 |
| Emerging Europe EMBI+ 4/ | 22 | 85 |
| U.S. corporate bonds | 20 | 46 |
| EU financial corporate bonds | 17 | 10 |
| Selected intra-euro-area sovereign bonds | 10 | 70 |
Estimation details are provided in Lombardi and Sgherri (2008).
Increase in spread (basis points) following a 100-basis-point increase in risk premium.
Percent of total volatility due to shifts in the global price of risk.
EMBI+ is the JPMorgan Emerging Markets Bond Index Plus.
Measuring the Impact of Risk Repricing on Selected Markets 1/
| Impact on Spread 2/ | Impact on Volatility 3/ | |
|---|---|---|
| U.S. asset-backed commercial paper | 69 | 65 |
| U.S. interbank | 31 | 75 |
| EU interbank | 29 | 52 |
| EU corporate bonds | 28 | 99 |
| Emerging Europe EMBI+ 4/ | 22 | 85 |
| U.S. corporate bonds | 20 | 46 |
| EU financial corporate bonds | 17 | 10 |
| Selected intra-euro-area sovereign bonds | 10 | 70 |
Estimation details are provided in Lombardi and Sgherri (2008).
Increase in spread (basis points) following a 100-basis-point increase in risk premium.
Percent of total volatility due to shifts in the global price of risk.
EMBI+ is the JPMorgan Emerging Markets Bond Index Plus.
Emerging markets remained initially unaffected by the financial turmoil, bolstered by improved fundamentals and financial buffers built up over recent years. Nonetheless, European emerging markets’ bond spreads rose significantly in the first quarter of 2008 (Figure 15, panel 4). Higher borrowing costs and more limited liquidity have delayed new member states’ issuance of Eurobonds originally scheduled for January. Since March, though, fixed-income markets in the region have also come under pressure (e.g., Hungary). Over the same period, syndicated and securitized bank lending—the main source of financing for companies in other emerging European markets—has also been slowing, raising concerns about prospective external borrowing in these economies.
Safer Homegrown Mortgages
Notwithstanding the troubles in U.S. credit markets, advanced Europe’s mortgage markets have remained relatively unscathed so far, showing no sign of a homegrown subprime crisis. Several features of the European housing finance system have contributed to an improved functioning of domestic loan markets, while enhancing financial stability in the region. Subprime loans—which account for about one-tenth of total mortgage lending in the United States—are virtually nonexistent in continental Europe and represent a limited share of the total mortgage stock in the United Kingdom.11 The adequate debt-servicing capacity of European borrowers provides a buffer during volatile periods in the real estate market. Most important, though, national boundaries, more cautious risk management, and stricter regulation make banks in most of Europe slow to adopt those risky financial innovations in secondary mortgage markets that have proved problematic in the United States. Despite their recent decline, retail deposits still constitute more than half of the funding source for housing finance in European Union (EU) member states. Securitization characterizes less than one-third of total mortgage financing and relies mainly on covered bonds—plain vanilla debt instruments regulated by EU legislation and secured against a pool of mortgages that remains on the balance sheet of the issuer. Loan-to-value ratios, prudent property valuation rules, and special investor protection reinforce the relative safety of European mortgage-covered bond markets, while allowing lenders to obtain comparatively cheap funds in capital markets and benefit from relatively low regulatory risk weightings.12
Emerging Europe’s mortgage market deserves, nonetheless, a note of caution. Housing loans have contributed significantly to the rapid credit growth experienced by the region, as well as to sustained increases in real asset prices. Given the importance of these assets as loan collateral, a downward adjustment in house prices might lead to increased provisioning requirements for banks and ultimately larger losses if delinquencies rise, adding to concerns about regional financial vulnerabilities. Moreover, the overall net asset position of emerging European households remains relatively low—both in absolute terms and compared with the euro area—contributing to the risks of a slowdown in the housing market.
Impact on Financial System Resilience
Resilient So Far…
Europe’s financial systems have thus far held up relatively well. With the subprime fallout translating into direct investment losses, higher funding costs, heightened uncertainty about funding availability, widespread reassessments of credit and counterparty risk, and large unanticipated assumptions of off-balance-sheet assets, financial systems have been tested. Yet, based on the information available through end-March 2008, the soundness of the major systemic players has been preserved, as shareholders and outside investors have injected fresh capital where needed. As a result, while capital ratios have been affected, those of the large and complex banking groups in the euro area have continued to exceed the regulatory minimums. However, liquidity remains seriously impaired despite aggressive responses by major central banks.
The continuing growth in lending in Europe thus far has also been suggestive of a relatively resilient banking system, even though this could be partly due to reintermediation (Figure 16).13 In advanced economies, where corporate lending has remained robust, household lending has continued its deceleration, which had begun well before the crisis broke out last summer. Credit standards have tightened across the board, but not as sharply as in the United States—in part because they had not been loosened by as much to begin with. Meanwhile, many emerging economies have continued to register rapid credit growth. While a cooling of credit growth has become apparent in some countries, this development has mostly preceded the onset of the current episode of financial turmoil and has in most cases also been considered helpful in dealing with overheating pressures.
Euro Area Lending to Nonfinancial Corporations and Households, 2000–08
(Trillions of euros)
Sources: European Central Bank; and IMF staff calculations.Individual financial institutions, however, have reported substantial write-offs. As markets reassessed risks, large money-center banks and major financial institutions booked extensive write-downs of structured securities, prompting credit downgrades and increases in default risk premiums. As the reassessment continued, European banks in several countries were forced to reintegrate into their balance sheets exposures to special investment conduits and structured investment vehicles. This resulted in significant changes in the market value of assets and the book value of short-term liabilities over the period January 2007–January 2008 (Figure 17).
… But Markets Remain Concerned
Market-based estimates of default frequencies have risen, but from a low base and to a lesser extent than in the United States (Figure 18). Considering a sample of 226 European and 666 U.S. banks, the time pattern of expected default frequencies (EDFs) derived using Moody’s KMV CreditEdge methodology suggests that, while the risk of bank default within the next year has risen since the onset of the subprime crisis, it remains relatively small.14 The resilience of European banks has also been less affected than that of the U.S. banks in the sample, particularly for the worst 10th percentile.
Expected Default Frequencies, 2007–08
(Percent)
Source: Moody's KMV.1/ Median of sample at each point in time.2/ Worst 10 percent of banks at each point in time.Implied spreads on credit default swaps suggest a similar picture (Figures 19 and 20). Banks have seen their credit default swap spreads widen significantly, suggesting a sharp rise in their risk profiles. The higher price for market risk (i.e., a higher Sharpe ratio) is one factor that has contributed to the widening of spreads. But, particularly for exposed banks, the market appears to have priced in a generalized increase in asset volatility due to concerns about asset quality, worsening liquidity conditions, and associated fire sales. Rising market leverage has been a further contributing factor, particularly for institutions that have had to expand their balance sheets unexpectedly due to exposures to special investment conduits and structured investment vehicles.
Implied Spreads, 2007–08
(Basis points)
Source: Moody's KMV.1/ EICDS = expected-default-frequency-implied (EDF-implied) credit default swap spread.2/ Median of sample at each point in time.3/ Worst 10 percent of banks at each point in time.Market Leverage versus Asset Volatility of Exposed EU Banks, 2007–08
(Percent)
Source: Moody's KMV.Note: Market leverage refers to the bank’s barrier relative to the market value of assets, where the distress barrier is an intermediate point between total and current liabilities.The market’s evaluation of assets under distress suggests that the quality of system assets has deteriorated to below investment grade. As shown in Figure 21, this deterioration in asset quality has produced a progressive shift in the curve linking an increasing share of total assets to higher expected default frequencies. At the beginning of 2007, market prices implied that, for banks with an investment grade, none of the assets were under distress, whereas by the end of January 2008 as much as 40 percent of these assets were below investment grade.15
European Banks: Market Evaluation of Assets in Distress
Source: Moody's KMV.Loss Recognition Needs to Catch Up
Direct exposures to the U.S. subprime mortgage markets have been a main driver of write-downs by European banks. With direct net exposures of global banks to U.S. subprime mortgage markets estimated at about $700 billion, European banks hold about 40 percent of this amount. Compared with U.S. banks, European banks are relatively less exposed to unsecuritized subprime mortgage loans and subprime mortgage-related collateralized debt obligations (CDOs); however, they are relatively more exposed to subprime mortgage-related asset-backed securities (ABS) (Table 4).
Net Bank Exposure to U.S. Subprime Mortgage Markets 1/
(Billions of U.S. dollars)
Calculations as of March 2008.
Percent.
ABSs = asset-backed securities.
CDOs = collateralized debt obligations.
Net Bank Exposure to U.S. Subprime Mortgage Markets 1/
(Billions of U.S. dollars)
| European Banks | U.S. Banks | |||
|---|---|---|---|---|
| Level | Share 2/ | Level | Share 2/ | |
| Subprime mortgage loans | 106 | 38 | 190 | 50 |
| Subprime mortgage-related ABSs 3/ | 85 | 30 | 40 | 10 |
| Subprime mortgage-related CDOs 4/ | 88 | 32 | 151 | 40 |
| Total exposure | 279 | 100 | 381 | 100 |
Calculations as of March 2008.
Percent.
ABSs = asset-backed securities.
CDOs = collateralized debt obligations.
Net Bank Exposure to U.S. Subprime Mortgage Markets 1/
(Billions of U.S. dollars)
| European Banks | U.S. Banks | |||
|---|---|---|---|---|
| Level | Share 2/ | Level | Share 2/ | |
| Subprime mortgage loans | 106 | 38 | 190 | 50 |
| Subprime mortgage-related ABSs 3/ | 85 | 30 | 40 | 10 |
| Subprime mortgage-related CDOs 4/ | 88 | 32 | 151 | 40 |
| Total exposure | 279 | 100 | 381 | 100 |
Calculations as of March 2008.
Percent.
ABSs = asset-backed securities.
CDOs = collateralized debt obligations.
The losses incurred by banks on these exposures are estimated to be significant. As of March 2008, expected losses on subprime mortgage-related exposures are estimated at $123 billion in Europe and $144 billion in the United States (Table 5). These estimates reflect the following scenario on a large sample of banks: average losses of 15 percent on unsecuritized mortgage loans, 30 percent on ABSs, 60 percent on CDOs, and 5 percent on liquidity lines to off-balance-sheet conduits and special investment vehicles (SIVs). For Europe, CDOs account for the largest share of these losses, followed by losses incurred on ABSs and off-balance-sheet liquidity lines. These estimates remain subject to a margin of uncertainty, as they are sensitive to assumptions and their underlying determinants continue to evolve.16
Estimated Losses on Mortgage-Related Subprime Bank Exposures 1/
(Billions of U.S. dollars)
Calculations as of March 2008.
SIV = structured investment vehicle.
Estimated Losses on Mortgage-Related Subprime Bank Exposures 1/
(Billions of U.S. dollars)
| Europe | United States | |||
|---|---|---|---|---|
| Expected losses | 123 | 144 | ||
| U.S. subprime loans | 16 | 29 | ||
| Asset-backed securities | 27 | 12 | ||
| Collateralized debt obligations | 53 | 90 | ||
| Conduits and SIVs 2/ | 27 | 13 | ||
| Reported losses | 80 | 95 | ||
| Remaining losses expected | 43 | 49 | ||
Calculations as of March 2008.
SIV = structured investment vehicle.
Estimated Losses on Mortgage-Related Subprime Bank Exposures 1/
(Billions of U.S. dollars)
| Europe | United States | |||
|---|---|---|---|---|
| Expected losses | 123 | 144 | ||
| U.S. subprime loans | 16 | 29 | ||
| Asset-backed securities | 27 | 12 | ||
| Collateralized debt obligations | 53 | 90 | ||
| Conduits and SIVs 2/ | 27 | 13 | ||
| Reported losses | 80 | 95 | ||
| Remaining losses expected | 43 | 49 | ||
Calculations as of March 2008.
SIV = structured investment vehicle.
These findings suggest that loss recognition will need to catch up. While several European banks have disclosed large losses, the aggregate number so far (roughly $80 billion as of March 2008) still falls short of the total estimate. The concern, therefore, is that financial institutions in Europe still need to own up to larger losses of approximately $43 billion.
The Test Is Not Over
Financial system resilience in advanced Europe is likely to be tested further.
-
As the crisis in U.S. subprime mortgage products deepens, this will inevitably negatively affect the direct exposures held by Europe’s financial institutions. With U.S. housing market conditions yet to stabilize, additional bouts of distress remain highly plausible in the mortgage markets. Such episodes would further lower the value of securitized and structured products and exacerbate existing losses.
-
The continuing financial distress is likely to accelerate the spread of the crisis to other forms of debt and debt default insurance. Asset quality is already deteriorating, or expected to deteriorate, for other forms of debt, such as other types of consumer debt (credit card, auto, and prime mortgage loans), commercial property loans, and corporate leveraged loans. Changes in the credit ratings of bond insurers could also translate into more credit losses for banks and investors. More generally, the market for debt default insurance, which has not yet been tested in a downturn, is also at risk. At this point, however, it is difficult to predict how an additional failure in one segment of the financial system would spread to other segments, and how this would play out precisely for Europe’s financial systems.
-
Equally significant, however, the slowing of global growth is likely to add to existing challenges. First, the deteriorating economic outlook could weaken European household and corporate balance sheets appreciably, even though they have from the outset generally been stronger than in the United States. Second, financial institutions’ attempts to preserve profitability, solvency, and liquidity are bound to off-load some of these evolving pressures to the real economy, an outcome that could not only amplify the growth slowdown but also feed back into the performance of financial systems. Third, if the crisis spilled over into emerging Europe, the profits of foreign banks that are significantly exposed to the region would be reduced, thereby weakening the banks’ soundness.
The impact of these risks on financial system resilience would manifest itself in several ways. First, deteriorating asset quality would amplify existing investment losses and reduce profitability and retained earnings. Second, if additional assets needed to be taken onto the balance sheet as a result of contingent credit lines, this would compound existing problems through an increase in leverage and a deterioration in asset quality. Third, as the need for additional capital injections intensifies, the availability of outside capital may decline. Fourth, liquidity risk could create further trouble if price discovery in previously well-established markets continues to be hampered and institutions are forced to sell at fire-sale prices to meet unexpected liquidity demands. Finally, the potential for extreme events to spill over from one bank to another—both domestically and across borders—has increased over time and represents an additional source of stress (Box 4).
Although emerging Europe’s financial systems have generally fared well thus far, they face increasing risks:
-
While the impact of the credit turmoil was relatively muted until early 2008, signs of spillovers have started to emerge. Limited direct exposures to complex structured products are a mitigating factor. But downside risks have risen, as tighter global liquidity could significantly curtail credit growth and problems in parent banks could easily be transmitted across borders. Evidence of such spillovers has surfaced. For example, sovereign debt markets have begun to be affected. Also, private bond issuance has fallen sharply—though from a low base—which could matter for banks that have relied on external financing to support rapid domestic credit growth. More generally, however, if credit growth were to slow considerably and standards tighten, there is the risk that asset quality would deteriorate where lending standards had previously been lax.
-
Foreign bank presence may be a boon, but foreign banks could also be a destabilizing factor by importing or exporting financial instability. While a strong foreign bank presence could help to contain domestic downturns, it would likely exacerbate financial instability in the case of more significant events, such as a parent bank’s experiencing liquidity shortages in the home country. If parent banks remain well capitalized, their financial support might lower the sensitivity of affiliates to local conditions and provide a measure of stability to local financial systems. Yet the concentration of international players in a number of countries and the similarity of their activities expose emerging economies to common-lender contagion risks (Box 5).
Recent stress tests in emerging Europe suggest that these vulnerabilities could seriously affect financial system resilience. First, capital buffers were not always found to be sufficient to accommodate shocks that are severe but still plausible, a particular concern for the smaller banks. Yet, on the positive side, the results also suggest that, when consolidated with parent banks, capital buffers are generally substantial. Second, credit risk is usually the key risk, particularly when the indirect effects of currency and interest rate risk are included. Foreign currency lending is still substantial, and the majority of loans are made with flexible rates. Third, in most countries, individual name and industry concentration (particularly in real estate and construction) is important. Finally, the sustainability of liquidity positions depends critically on the level of commitment of the parent banks. This risk is mitigated when parent banks have ample liquidity and strong profit and capital buffers. At the same time, the growing access of domestic banks to nonparent financing requires careful monitoring of the maturity structure of such debt.
Spillover Risks among Major EU Banks
The scope for cross-border spillovers among the major European banks can be examined using the extreme value theory framework. This framework analyzes comovements between extreme events (“co-exceedances”), specifically the comovement of extreme negative (left-tail) realizations of banks’ soundness measures. The soundness measure chosen in this analysis is the distance to default (DD), defined as the difference between the expected value of assets at maturity and the default threshold, which is a function of the value of the liabilities. A higher DD is associated with a lower probability of the bank’s default. It is generally a useful proxy for default risk if stocks are traded in liquid markets.
The DDs for 27 of the largest EU banking groups were computed for May 2000–April 2007 using daily stock prices and annual balance sheet data. A binomial logit model was used to estimate the probability of a bank experiencing a large negative DD change in response to large negative shocks to the DD changes of other banks. Large negative shocks were defined as those falling in the tenth percentile of the left tail of the common distribution of the changes in the DD across all banks. Four control variables—changes in the slope of the term structure, and the volatility of the domestic, regional, and world stock market indices—were also included in the model to account for common factors.
The results (table) suggest that, although spillovers within domestic banking systems generally remain more likely, the potential for extreme events to spill over from one bank to another appears to have increased, both among domestic banks and across borders. The number of significant cross-border links is already larger than the number of significant links among domestic banks, underscoring the need for greater cross-border supervisory cooperation in the European Union.
Significant Co-Exceedances among EU Banks, May 2000–April 2007
The number of bank pairs for which co-exceedances were found significant at the 5 percent level in the given period.
The number of significant links (in the previous line), in percent of all possible contagion channels (i.e., as percent of all possible domestic and cross-border pairings of banks, respectively).
Significant Co-Exceedances among EU Banks, May 2000–April 2007
| May 2000-Apr 2007 | May 2000–Nov 2003 | |||
|---|---|---|---|---|
| Domestic | Cross-border | Domestic | Cross-border | |
| Number of significant links 1/ | 19 | 57 | 14 | 50 |
| Percent of all possible links 2/ | 39.6 | 8.7 | 28.6 | 7.6 |
The number of bank pairs for which co-exceedances were found significant at the 5 percent level in the given period.
The number of significant links (in the previous line), in percent of all possible contagion channels (i.e., as percent of all possible domestic and cross-border pairings of banks, respectively).
Significant Co-Exceedances among EU Banks, May 2000–April 2007
| May 2000-Apr 2007 | May 2000–Nov 2003 | |||
|---|---|---|---|---|
| Domestic | Cross-border | Domestic | Cross-border | |
| Number of significant links 1/ | 19 | 57 | 14 | 50 |
| Percent of all possible links 2/ | 39.6 | 8.7 | 28.6 | 7.6 |
The number of bank pairs for which co-exceedances were found significant at the 5 percent level in the given period.
The number of significant links (in the previous line), in percent of all possible contagion channels (i.e., as percent of all possible domestic and cross-border pairings of banks, respectively).
Impact on the Real Economy
Empirically assessing the macroeconomic effects of the current financial turbulence is difficult. Indeed, existing macroeconomic models incorporate only a few elements of the many potential transmission channels through which the unfolding financial shocks may affect the real economy. Moreover, today’s turmoil is quite unprecedented, making estimates based on historical data less than fully reliable.
Regional Financial Interlinkages and Contagion Channels
With the pace of private sector credit growth remaining brisk in emerging Europe, dependence on nondeposit funding has increased in many countries. High and rising loan-to-deposit ratios (LTDs) mirror increasing reliance on foreign funding channeled primarily through the banking sector, given the relatively undeveloped domestic capital markets and easy access to cheap financing from the parent institutions of the mostly foreign-owned banks.1
LTDs have been rising in most countries in the region, particularly in the Baltic countries, where they have roughly doubled since the early 2000s, and in Ukraine, Hungary, and Russia, where they ranged from 120 to 150 percent in 2007. Except in a few cases (Moldova; Serbia; Macedonia, FYR; and Bosnia and Herzegovina), the changes in bank credit–to-GDP ratios significantly exceeded those in the ratio of bank deposits to GDP, suggesting that credit growth has significantly outpaced deposit growth in recent years (first figure).
The sizable cross-border financial linkages across Europe highlight the vulnerabilities arising from reliance on concentrated foreign funding. International banking statistics from the Bank for International Settlements provide consolidated foreign claims of reporting banks on individual countries (through both direct lending and local banking systems) and give a sense of the magnitude and distribution of exposures of emerging European countries to western European banking systems. Similarly, they provide exposures of western European countries to emerging European economies by the nationality of the reporting banks.
Data suggest that most emerging European economies are heavily exposed to—and dependent on—western European banks (either directly or through the local banking systems). Exposures are significant, in relation to both the recipient countries’ GDP and the size of their banking system assets (second figure).
Outstanding foreign claims owed to reporting banks in all western European banks are larger for emerging European countries whose banking systems are largely foreign owned: although risks in these countries should not be exaggerated where local bank lending is financed mainly by local deposits, vulnerabilities could increase in the event of problems in parent banks. Exposures are also significant in terms of the host country’s banking sector assets for many countries; however, there are several notable exceptions, such as Russia, Turkey, and Ukraine.
Change in Deposit and Credit to GDP, 2003–07
(Percentage points)
Sources: IMF, International Financial Statistics; and IMF staff calculations.The concentration of funding exposure is contributing to the vulnerabilities associated with heavy reliance on foreign funding, particularly when it concerns the funding of banks.2 Most countries in the region have concentrated exposures to banks in Austria, Italy, and Germany, and the Baltic countries have large exposures to Sweden (third figure). Among those most dependent on foreign funding, some are more diversified (e.g., the Czech Republic and Poland), while several depend on funding from a very few countries. Sweden provides 85 percent of the external funding for the Baltic countries, while Austria holds at least 40 percent of the claims on Bosnia and Herzegovina, Croatia, Serbia, and the Slovak Republic, and more than 30 percent on Romania. Similarly, Italy provides more than 30 percent of foreign funding for Bulgaria, Croatia, and Bosnia and Herzegovina.
Such concentration of funding renders emerging Europe vulnerable to a credit squeeze or crunch originating from advanced economies. Banking systems that are heavily dependent on foreign borrowing to support credit growth could face a sudden shortfall of funding or a sharp increase in its cost if there were a sudden reassessment of parent bank exposure to a host country (e.g., due to concerns about growing vulnerabilities in that country or region). While reputational risks could make it unlikely that parent banks would abandon their subsidiaries, unconditional support is unlikely. Liquidity or solvency problems in a parent bank could also be easily transmitted to local banks in a concentrated banking system.
Relative Magnitudes of Funding Exposure for Emerging and Western Europe, June 30, 2007 1/
(Percent)
Sources: Bank for International Settlements, Quarterly Banking Statistics, December 2007; and IMF staff calculations.1/ Funding exposure concerns bank and nonbank sectors in recipient countries. In countries with foreign bank ownership, consolidated claims include claims by foreign-owned banks on residents; net claims are lower to the extent that banks hold domestic deposits.The impact of a retrenchment of credit would be amplified if funding from other (nonbank) sources were also limited. Some emerging European economies that turned to international capital markets for funding in recent years have seen demand decline significantly since August 2007 (e.g., Eurobond issuance by the Russian and Ukrainian financial sectors), and some banks in central and eastern Europe reportedly postponed their planned bond issues as a result of wider spreads. International bond issuance has been a negligible source of funding for most other countries in the region.
The magnitude of western European bank exposures to emerging Europe is far smaller than the exposures of the latter, with the exception of Austria. For Austria, the claims of the reporting banks on emerging Europe amount to 23 percent of its banking system assets, though exposures seem well diversified across several countries and small with respect to each individual country. The exposures of banks in Belgium and Sweden to emerging Europe are also significant but remain below 10 percent of banking system assets. For the remaining countries, the exposures are negligible, including those of France, Germany, and Italy with the most active banks in central and southeatern Europe.
However, even where the exposures seem well diversified, the ultimate impact of possible adverse developments in one country may be much more significant, as troubles in one country may spill over to others. Such regional spillovers could make even Austria’s exposures substantial.
Concentration of Emerging Europe Exposure to Western Europe, June 30, 2007 1/
(Percent)
Source: Bank for International Settlements, Quarterly Review, December 2007.Note: Country names are abbreviated according to the ISO standard codes.1/ Emerging Europe exposure to western European banks is defined as the share of the reporting banks in each western European country in the total outstanding claims on a given emerging European country (both bank and nonbank sectors). For example, about 42 percent of Croatia's exposures to western European reporting banks is owed to Austrian banks, 38 percent to Italian banks, 13 percent to French banks, etc. For the Baltic countries, 85 percent or more of exposures to the reporting banks is owed to Swedish banks.Subject to these caveats, our analysis indicates that the macroeconomic impact of the turmoil is likely to be significant, especially in advanced economies in Europe, while uncertainty remains about the probability of a more severe credit squeeze that could result in further output losses. In Europe’s emerging economies, the output slowdown stemming from the direct impact of financial shocks, as well as the spillover effects from lower demand from advanced economies, are expected to curb the robust growth rates experienced by these economies in recent years. The output loss would be bigger in the event of a significant credit squeeze, especially in countries where the recent expansion has been largely driven by foreign borrowing and very rapid credit growth.
Channels of Transmission to the Real Economy
The current financial turmoil will continue to affect the real economy through several channels. First, borrowing costs for firms and households are rising, in line with banks’ increasing funding costs, thereby adversely affecting investment and consumption decisions. Second, firms also face higher costs in issuing bonds and stocks, as corporate spreads and equity premiums—the excess return of equity over the risk-free rate—increase to compensate investors in the face of declining risk appetite.
On top of the effect on the price of credit, the financial turmoil is expected to have an adverse impact on the supply of credit, which, in turn, would dampen consumption and investment. First, banks’ higher funding costs reduce their liquidity and ability to extend loans.17 More important, subprime-market-related losses incurred by banks in Europe’s advanced economies impair their capital adequacy and, hence, their lending capacity. Furthermore, the significant contraction in the issuance of structured products reduces financing opportunities for banks in these countries and their ability to extend credit.18 While not directly exposed to the subprime market, banks in emerging Europe also face the risk of a credit squeeze, as parent banks in advanced Europe could cut back lending to their subsidiaries and branches.
The turnaround of the housing market cycle in a number of countries and the recent weakness of equity markets will also have macroeconomic consequences. Asset prices affect consumption through wealth effects. Property prices also have an impact on construction activity.19 Furthermore, declining equity and house prices, as well as increasing borrowing rates, distress firms’ and households’ balance sheets and cash flows, thus reducing their creditworthiness and ability to borrow.20
Quantifying the Impact
The current financial turmoil is likely to have a significant impact on output in advanced European economies. Widening corporate spreads are expected to reduce euro area output by 0.3 and 0.4 percent in 2008 and in 2009, respectively, relative to a hypothetical scenario without financial shocks. The tightening of lending standards—associated with a rise in the spread between lending and deposit rates—together with the decline in equity and house prices, will contribute to a further output slowdown of at least 0.4 and 0.5 percent in 2008 and 2009, respectively, compared with a scenario without financial shocks (Table 6).21 In the United Kingdom, the GDP loss stemming from the rise in corporate spreads is estimated to be 0.3 percent in 2008 and 0.4 percent in 2009, relative to a no-shock scenario. The assumed fall in house prices in the United Kingdom will, however, produce a larger output slowdown than in the euro area. In the other advanced European economies, financial shocks are expected to have a milder impact on GDP.
Output Response to Financial Shocks: Simulation Results
(Real GDP percent deviations compared with a no-shock scenario)
Excluding Cyprus and Malta.
Denmark, Norway, Sweden, and Switzerland.
Bulgaria, Estonia, the Czech Republic, Hungary, Latvia, Lithuania, Poland, Romania, and the Slovak Republic.
Output Response to Financial Shocks: Simulation Results
(Real GDP percent deviations compared with a no-shock scenario)
| Euro Area 1/ | United Kingdom | Other Advanced European Economies 2/ | New EU Member States 3/ | United States | |
|---|---|---|---|---|---|
| 100-basis-point increase in corporate spreads for two years in all advanced economies | |||||
| 2008 | −0.3 | −0.3 | −0.1 | −0.1 | −0.2 |
| 2009 | −0.4 | −0.4 | −0.2 | −0.2 | −0.3 |
| 100-basis-point increase in the spread between lending and deposit rate for two years in all advanced economies | |||||
| 2008 | −0.1 | −0.1 | −0.1 | −0.1 | −0.1 |
| 2009 | −0.1 | −0.1 | 0.0 | 0.0 | −0.1 |
| 15 percent decline in equity prices for one year in advanced and emerging economies | |||||
| 2008 | −0.2 | −0.2 | −0.2 | −0.4 | −0.4 |
| 2009 | −0.1 | 0.0 | 0.0 | −0.1 | 0.0 |
| House prices decline in United States, United Kingdom, Spain, and Ireland | |||||
| 2008 | −0.1 | −0.5 | −0.1 | −0.2 | −0.5 |
| 2009 | −0.3 | −0.8 | −0.2 | −0.2 | −1.1 |
Excluding Cyprus and Malta.
Denmark, Norway, Sweden, and Switzerland.
Bulgaria, Estonia, the Czech Republic, Hungary, Latvia, Lithuania, Poland, Romania, and the Slovak Republic.
Output Response to Financial Shocks: Simulation Results
(Real GDP percent deviations compared with a no-shock scenario)
| Euro Area 1/ | United Kingdom | Other Advanced European Economies 2/ | New EU Member States 3/ | United States | |
|---|---|---|---|---|---|
| 100-basis-point increase in corporate spreads for two years in all advanced economies | |||||
| 2008 | −0.3 | −0.3 | −0.1 | −0.1 | −0.2 |
| 2009 | −0.4 | −0.4 | −0.2 | −0.2 | −0.3 |
| 100-basis-point increase in the spread between lending and deposit rate for two years in all advanced economies | |||||
| 2008 | −0.1 | −0.1 | −0.1 | −0.1 | −0.1 |
| 2009 | −0.1 | −0.1 | 0.0 | 0.0 | −0.1 |
| 15 percent decline in equity prices for one year in advanced and emerging economies | |||||
| 2008 | −0.2 | −0.2 | −0.2 | −0.4 | −0.4 |
| 2009 | −0.1 | 0.0 | 0.0 | −0.1 | 0.0 |
| House prices decline in United States, United Kingdom, Spain, and Ireland | |||||
| 2008 | −0.1 | −0.5 | −0.1 | −0.2 | −0.5 |
| 2009 | −0.3 | −0.8 | −0.2 | −0.2 | −1.1 |
Excluding Cyprus and Malta.
Denmark, Norway, Sweden, and Switzerland.
Bulgaria, Estonia, the Czech Republic, Hungary, Latvia, Lithuania, Poland, Romania, and the Slovak Republic.
In all advanced economies, further output losses would materialize if banks were to reduce credit supply significantly. Our estimates indicate that European banks’ current increased risk profile is likely to affect the supply of loans only moderately, and that the resulting impact on GDP growth will be limited (Box 6). Nevertheless, a sharper-than-expected credit squeeze cannot be ruled out. This could trigger a downward spiral, whereby lower credit expansion could result in a sharp economic slowdown, which, in turn, would curtail the demand for credit.
The financial turmoil is expected to have negative real effects also in emerging Europe. In the new member states,22 declining equity prices, together with the total spillover effects from lower demand from advanced economies, are estimated to result in output losses of at least 0.8 and 0.5 percent in 2008 and 2009, respectively, relative to a hypothetical scenario without financial shocks (Table 6). Given the brisk rate of growth experienced by these economies in the past years, such estimates point to a relatively mild slowdown. However, rising domestic and foreign borrowing premiums could take a further toll on growth in emerging Europe. Specifically, the concurrent rise in foreign and domestic risk premiums by 100 basis points (bps), sustained for two years, could reduce output by 0.8–1 percent in 2008, and by 0.3–0.8 percent in 2009, compared with a scenario without shocks (Box 7).23
The output slowdown in emerging Europe would be sharper if a significant credit squeeze were to materialize. Estimates of vector autoregressions (VARs) for selected emerging economies suggest a considerable impact. If credit growth were to fall temporarily by 10 percentage points, real GDP growth would decline significantly in the first year in Estonia and Romania—two economies where credit growth has been among the strongest within emerging Europe since the mid-1990s—before rebounding in the second year (Table 7).24 In the Czech Republic, the negative effect on output would be smaller on impact but more prolonged, while in Hungary the growth reduction would be more contained.
Response of GDP Growth to a 10 Percentage Point Decline in Credit Growth
(Percentage points, year-on-year)
Response of GDP Growth to a 10 Percentage Point Decline in Credit Growth
(Percentage points, year-on-year)
| Czech | ||||
|---|---|---|---|---|
| Republic | Estonia | Hungary | Romania | |
| Year 1 | −0.7 | −2.0 | −0.7 | −1.3 |
| Year 2 | −0.8 | 0.7 | 0.1 | 0.7 |
Response of GDP Growth to a 10 Percentage Point Decline in Credit Growth
(Percentage points, year-on-year)
| Czech | ||||
|---|---|---|---|---|
| Republic | Estonia | Hungary | Romania | |
| Year 1 | −0.7 | −2.0 | −0.7 | −1.3 |
| Year 2 | −0.8 | 0.7 | 0.1 | 0.7 |
Banks’ Risk Profile, Credit Growth, and the Real Economy
To what extent is the current financial turmoil, and the associated increase in banks’ risk profile, likely to affect banks’ supply of credit? And what is the potential impact on output? To address these questions we estimate the relationships between banks’ risk conditions, loan growth, and real GDP growth on a sample of 42 major banks from seven advanced European economies (France, Germany, Italy, Spain, Sweden, Switzerland, and the United Kingdom) over 1991–2007.
The model specification consists of two simultaneous equations. In the first, loan growth in 42 banks is regressed over a measure of banks’ risk profile (lagged), real GDP growth, and a number of control variables (including the interbank rate, house and stock price indices, and a measure of banks’ size). The second equation explores the relationship between loan growth and output growth.1 Banks’ risk profile is proxied by Merton-type credit risk indicators, including different versions of the distance-to-default (DD) and Moody’s KMV expected default frequency (EDF).2
Our model estimates indicate that a rise in banks’ risk profile results in (statistically significant) lower credit growth. Specifically, on average, a 1 percent increase in a bank’s EDF can result in up to a 0.05 percentage point decrease in its annual loan growth, with some variations across countries.3 Given the change in banks’ risk profile observed since August 2007 (Figure 18 in main text), our model predicts a reduction in annual credit growth of up to 4 percentage points. In our model, such a slowdown in credit expansion implies a decline in GDP growth of less than 0.1 percentage point.4 We obtained similar results when estimating the model on macroeconomic rather than individual bank data.
Note: The main authors of this box are Andrea M. Maechler and Alexander Tieman. The analytical underpinnings of this box are presented in Maechler and Tieman (forthcoming). 1 The model is estimated at quarterly frequency. 2 Both risk indicators are based on a Merton-type option pricing structural model. The DD represents the number of standard deviations a market-based estimate of a bank’s equity is away from default. It has a negative relationship with financial risk (a lower DD means more risk). Moody’s KMV EDFs, which measure probabilities of default, use a database of historic defaults to translate the model-implied risk-neutral probabilities to real-world (risk averse) expected default frequencies. EDFs exhibit a positive relationship to financial risk (a higher EDF means more risk). 3 Relative to French banks, the top German, Spanish, and Swedish banks contract their credit expansion more sharply in response to a deterioration in their risk profiles. 4 For longer-term estimates of the potential impact of the financial turmoil on credit expansion and output in the United States, see IMF (2008a).Real Sector Implications of Financial Turbulence for an Emerging Market Economy
To assess the potential macroeconomic consequences of the current financial turmoil on emerging Europe, we simulated a simple dynamic general equilibrium model for a stylized emerging market economy, characterized by (1) a high degree of openness; (2) a fixed/tightly managed exchange rate; (3) high levels of financial dollarization (with most loans and deposits linked to foreign currency); (4) rapid credit growth to the private sector, facilitated by banks’ foreign borrowing; and (5) corporate sector direct access to international financial markets.1
We consider the following two-year shock: a 100-basis-point increase in the foreign risk premium, which affects the rate at which agents borrow from abroad, together with a 100-basis-point rise in the domestic risk premium, reflecting banks’ smaller risk appetite. The shock results in an immediate rise in lending rates and a drop in bank foreign borrowing and corporate credits. Direct corporate foreign borrowing increases, however, as it is cheaper than bank borrowing, given the rise in the domestic risk premium. Bank household credits fall, but less than bank corporate credits, as households cannot resort to direct foreign borrowing (figure).
Real output declines by 0.8–1 percent in 2008, and by 0.3–0.8 percent in 2009, compared with a scenario without shocks (figure). Consumption also falls, although less than production, as households reduce their stock of deposits in smoothing consumption. The output slowdown reduces inflation in 2008 and 2009. As consumption and the relative price of domestic to imported goods both decline, imports fall and the trade balance improves. When the premium shocks have subsided, bank foreign borrowing and bank corporate credits rebound, contributing to output and consumption recovery, which is associated with inflationary pressure and trade balance deterioration.
Response to an Increase in Domestic and Foreign Premiums
(Percent deviations compared with a no-shock scenario)
Source: IMF staff simulations.Note: The main authors of this chapter are Philip Schellekens, Silvia Sgherri, and Edda Zoli. The underpinning analytical work is presented in Lombardi and Sgherri (2008).
The price of risk—i.e., the expected return that, in equilibrium, international investors require to hold an extra unit of risk—is not directly observable. The approach adopted in this chapter—to estimate shifts in the international price of risk—relies on the notion that risk premiums embedded in different asset prices are determined jointly in the market and are influenced by the riskiness of the specific asset in question and global risk factors, such as the willingness and the ability of international investors to bear that risk. To account for more realistic data-generating processes, estimates are conducted within a multivariate generalized autoregressive conditional heteroscedasticity framework, featuring fat-tail shock distributions. For further details, see Lombardi and Sgherri (2008).
Recent work analyzing the role of risk repricing as a channel of transmission during periods of financial distress includes, for example, Kumar and Persaud (2002), Gai and Vause (2006), Coudert and Gex (2007), and Gonzalez-Hermosillo (2008).
Theoretically, a number of recent studies have explained that an asset’s market liquidity—the ease with which an asset is traded—and traders’ funding liquidity—the ease with which traders can obtain funding—(1) are inherently related to uncertainty, (2) comove with the market, and (3) have commonalities across securities. In particular, see Brunnermeier and Pedersen (2007), Caballero and Krishnamurthy (2007), and Adrian and Shin (2008).
Interbank lending involves payment up front, whereas OIS contracts are settled on a net basis at maturity. The LIBOR spread, hence, reflects both a credit and a liquidity premium. For a decomposition of LIBOR spreads, see Bank of England (2007).
On this point, see also the IMF’s World Economic Outlook (2008b).
Bank for International Settlements (2006); and European Commission (2006).
Avesani, García Pascual, and Ribakova (2007).
Moody’s KMV CreditEdge provides daily updates and changes in the probability of default on publicly listed firms.
Investment grade (BBB– or higher) corresponds to a mean expected default frequency of approximately 0.07 percent in the year ahead.
This is the so-called lending channel, first identified by Bernanke and Blinder (1988).
Since 1999, securitization has contributed significantly to credit growth in the euro area (Altumbas, Gambacorta, and Marqués, 2007).
Several studies have estimated the link between asset prices and the business cycle in advanced economies. See, for example, IMF (2000), Catte and others (2004), and IMF (2008b).
The role of balance sheet effects and collateral in credit cycles was first pointed out by Bernanke and Gertler (1989), who developed the so-called financial accelerator theory. See also Kiyotaki and Moore (1997); and Bernanke, Gertler, and Gilchrist (1999).
These estimates are based on simulations performed with the National Institute Global Economic Model (NiGEM). NiGEM is a multicountry model that allows an assessment of the impact of shocks hitting individual economies directly, as well as spillover effects from the slowdown in other countries. Specifically, the following scenarios were considered:
(1) a 100-basis-point (bp) increase in the spread between corporate and government bonds in all advanced economies (including non-European), lasting for two years, in line with the observed increase in spreads, and consistent with the assumption of a sustained repricing of risk;
(2) a 100 bp rise in the spread between lending and deposit rates sustained for two years, in advanced economies (including non-European), reflecting banks’ attempt to maximize profitability to rebuild their capital, as well as the tightening of lending standards;
(3) a 15 percent decline in equity prices in advanced and emerging economies (including non-European), lasting for one year, in line with equity market developments since the beginning of 2008; and
(4) a decline in nominal house prices in the United States, United Kingdom, Spain, and Ireland of 10–16 percent over two years, consistent with expected developments in the respective housing market.
This group comprises the following countries: Bulgaria, Estonia, the Czech Republic, Hungary, Latvia, Lithuania, Poland, Romania, and the Slovak Republic.
These estimates are based on simulations of a general equilibrium model for a stylized emerging economy, where GDP growth is largely driven by consumption and investment, which are financed by foreign borrowing and very rapid credit growth (see Box 7 for details). We resort to this alternative model because NiGEM does not capture the important relationship between foreign and domestic risk premiums, credit growth, and the real sector, which is a key part of the transmission of the current financial shocks to emerging Europe.
The VAR includes the following variables: credit growth, real GDP growth, inflation (all annualized), the lending rate, and, as an exogenous variable, the LIBOR. Credit growth is expressed in nominal terms following Bernanke and Lown (1991), who argue that this measure is the most appropriate in proxying the real value of credit extensions in the context of bank-borrower relationships, where the effective maturity of loans is very long. The model is estimated on quarterly data between the mid-1990s and the third quarter of 2007. The number of lags is determined through standard information criteria. Shock responses are computed using the generalized impulse response function.
