6 Risk Profiles of Financial Institutions

Jörg Decressin, Wim Fonteyne, and Hamid Faruqee
Published Date:
September 2007
  • ShareShare
Show Summary Details

This chapter1 explores the impact of financial integration on system-wide risk profiles of publicly traded European financial institutions. The chapter provides an overview of indicators of banking market penetration, and then assesses whether the benefits of risk diversification arising from integration are reflected in convergence of financial institutions’ risk profiles to lower risk levels. It documents the evolution and convergence of risk profiles of publicly traded banks and insurance companies, offering insights into the role of financial integration as a driver of risk profile dynamics. A key finding is that the risk profiles of financial institutions have indeed converged, but not to lower risk levels. Convergence has likely been driven by increased exposures to common financial shocks. Enhanced links stemming from integration of European capital markets may have played a role, as increased exposures have occurred despite the lagging integration of the relevant retail markets. The lack of clear improvement in risk profiles suggests that diversification benefits have been offset by higher risk taking.

Integration of Bank Credit Markets

National barriers to cross-border banking in the European Union appear to be breaking down only slowly. Foreign bank penetration has proceeded most rapidly in Central and Eastern Europe, as the less developed financial systems of transition countries offered significant growth opportunities and a high return on direct investment (Focarelli and Pozzolo, 2005; ECB, 2004b). As a result, western European banks expanded rapidly into Central and Eastern Europe well before the recent eastward enlargement of the European Union and are now important players in the new member states (Box 6.1). Cross-border banking has been less visible within the EU-15. It has been comparatively intense in some countries, where language and cultural factors hastened cross-border ties and where banks expanded and consolidated to take advantage of economies of scale. Because large European multinational corporations operate across borders, they can bank where they can get the most favorable credit terms and the services they need—including outside Europe. They also have easier access to various forms of nonbank financing. However, as noted in Degryse and Ongena (2004), small and medium-sized enterprises (SMEs) and households still tend to bank with their local banks.

Box 6.1.Foreign Bank Penetration in Central and Eastern Europe

The European Union’s 10 new member states from Central and Eastern Europe (the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, the Slovak Republic, and Slovenia, all of which joined in 2004, and Bulgaria and Romania, which joined in 2007) and EU accession candidate Croatia all opened up their markets for financial services in the 1990s as they liberalized their economies and prepared for EU membership. This resulted in rapid and extensive foreign bank penetration as western European banks jumped on the opportunities for expansion these countries offered.

In almost all these countries, well over half of banking sector assets (and in some cases practically all of them) are now foreign-owned. German, Dutch, and Austrian banks were especially quick to enter. By 1999, the banking systems from each of these three countries accounted for 10 percent or more of total EU-15 exposures to the Central and Eastern European countries. Swedish and Finnish banks rapidly established a large presence in the Baltic markets. During 2000–05, further expansion followed, as Italian banks built up substantial exposures in Hungary, Poland, the Slovak and Czech Republics, Bulgaria, Romania, and Croatia.

Most western European banks make substantial profits in these markets. For example, in 2005, Central and Eastern European activities accounted for roughly 16 percent of total assets but about 35 percent of total profits of the six largest Austrian banks. Nonetheless, for most EU-15 banking systems, exposures in Central and Eastern Europe remain a limited share of their total foreign exposures, the Austrian (41 percent) and Italian (17 percent) systems being the exceptions.

Nonetheless, indicators of cross-border banking activity point to a steady increase in banking integration in the EU-15 in recent years. Volume-type proxies for integration can be constructed from data on cross-border holdings of credit institutions and foreign exposures from the ECB and the Bank for International Settlements (BIS).2 They show the following:

  • Cross-border activity by euro area banks progressed unevenly across different types of activities. It progressed most rapidly in securities holdings, less in interbank loans, and least in loans to nonbanks (ECB, 2004b).

  • Assets of branches and subsidiaries of credit institutions from other EEA countries have increased significantly in the banking systems of most EU-15 countries since 1997 (Table 6.1).3 The average amount of nonhost banking sector assets has risen from the equivalent of less than 30 percent of GDP in 1997 to over 47 percent of GDP in 2004, with most of the expansion occurring through subsidiaries. An increase took place in all countries, with the exceptions of Belgium and Luxembourg where banks from other European countries were already well established. The assets of other EU banks and subsidiaries now amount to a sizable share of domestic GDP in most smaller countries and the United Kingdom, but remain relatively small in the large continental countries.

  • Foreign branches and subsidiaries are increasingly likely to come—in some cases almost exclusively—from other EU-15 countries. The share of total assets of foreign branches and subsidiaries held by European institutions has risen from about 75 percent to close to 90 percent, on average. The main exception is the United Kingdom, where the rapid expansion of EEA bank branches and subsidiaries has been matched by expansion from other localities, keeping London a large, geographically diverse international financial center and setting it up to be the primary euro area financial center.

Table 6.1.Assets of Branches and Subsidiaries of Foreign Credit Institutions from European Economic Area (EEA) Countries(in percent of GDP or total foreign-held assets in the banking system)








Large countries111.165.612.072.913.077.514.877.714.980.716.384.916.586.618.188.0
Small countries1382.279.2388.780.7390.681.4371.484.9414.090.2369.990.3354.991.3367.494.0
EMU-11 (ex LX)126.272.630.076.430.678.835.681.240.986.038.287.739.789.050.591.4
Total EMU-1221.374.023.179.324.882.528.383.331.383.731.285.331.385.635.386.9
United Kingdom77.745.282.050.981.649.685.448.589.648.880.848.688.545.7107.251.5
Total non-EMU-360.645.865.551.864.950.369.149.376.050.569.750.775.147.890.353.1
Total EU-1529.658.532.264.433.764.837.864.441.466.439.967.940.865.947.5100.0
Source: European Central Bank.

Unweighted average.

Source: European Central Bank.

Unweighted average.

  • Cross-border exposures of EU-15 banks to euro area countries have risen sharply, with the rise driven in large part by banks in London. As a result, exposures from the United Kingdom drove total euro area exposures of banks in the three non-euro EU-15 countries (Denmark, Sweden, and United Kingdom) up by over 370 percent in nominal terms from 1999 to 2006. This raised the share of euro area country exposures in total foreign exposures of the three non-euro countries from 22 percent to 27 percent (Figure 6.1). Within the euro area, growth of cross-border exposures of banks was also rapid, rising some 150 percent in nominal terms during this period. However, this was sufficient to raise the share of total foreign exposures to other euro area countries only modestly.

  • European banks outside the euro area have increased their share of euro-denominated assets, while holdings of euro-denominated assets by euro area banks recently have tended to stabilize at a high level (Figure 6.2). In the United Kingdom, the share remains much lower, at around 25 percent, but this represents an increase from just over 18 percent in the first quarter of 2000. Taken together with the data on foreign exposures, it seems that banks outside the euro area have been diversifying into the euro area, whereas euro area banks’ expansion has been primarily domestic or within the region.

Figure 6.1.Consolidated Foreign Exposure

Source: Bank for International Settlements.

Note: EMU-12 includes Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, and Spain. Non-EMU-3 includes Denmark, Sweden, and the United Kingdom.

Figure 6.2.Euro-Denominated Assets of the Banking System

(In percent of total assets)

Sources: IMF, International Financial Statistics; and Bank of England.

Large banks in particular have increased their balance sheets, in part through cross-border activities, and have expanded their links to financial markets (Table 6.2). In recent years, similarities in business strategies can be detected for large banks in all the major EU-15 countries. Most have pursued rapid growth, either organically or through mergers and acquisitions, and have significantly raised their share of noninterest income. Through this process, the share of banks’ income that is generated by financial market activities has trended upward. However, whereas the direction of change has been consistent, there are substantial differences across countries in the structure of balance sheets, as well as in profitability. For example, the average share of noninterest income in 2005 ranged from below 40 percent in Germany, the Netherlands, and Ireland, to around 50 percent in the United Kingdom, France, and Italy, and returns on equity varied from 11 percent in Germany to over 20 percent in the Netherlands and Ireland.

Table 6.2.Balance Sheet Data for Individual Banks
Interest (percent of total)75.370.468.961.353.0
Operating (percent of total)24.729.631.138.747.0
Asset growth in 2 years (percent)27.55.864.840.4
Liabilities growth in 2 years (percent)27.25.664.241.4
Return on equity9.510.413.210.216.6
Total capital ratio10.510.910.912.411.11
Interest (percent of total)75.174.572.662.063.4
Operating (percent of total)24.925.527.438.036.6
Asset growth in 2 years (percent)3.918.111.76.62
Liabilities growth in 2 years (percent)3.718.011.77.12
Return on equity7.06.01.5–6.210.9
Total capital ratio9.39.910.912.312.5
Interest (percent of total)61.558.859.855.261.8
Operating (percent of total)38.541.240.244.838.2
Asset growth in 2 years (percent)17.820.254.064.5
Liabilities growth in 2 years (percent)
Return on equity22.822.415.116.520.6
Total capital ratio11.211.510.810.911.13
Interest (percent of total)68.759.260.452.451.0
Operating (percent of total)31.340.839.647.649.0
Asset growth in 2 years (percent)8.00.736.854.8
Liabilities growth in 2 years (percent)
Return on equity–2.913.310.48.011.6
Total capital ratio10.08.98.810.39.34
Interest (percent of total)66.465.465.857.661.45
Operating (percent of total)33.634.634.242.438.65
Asset growth in 2 years (percent)25.519.451.630.2
Liabilities growth in 2 years (percent)25.422.351.930.6
Return on equity12.816.217.718.520.7
Total capital ratio10.710.911.011.813.15
United Kingdom
Interest (percent of total)58.856.454.947.851.16
Operating (percent of total)41.243.645.152.248.96
Asset growth in 2 years (percent)13.574.253.559.76
Liabilities growth in 2 years (percent)13.072.954.261.56
Return on equity22.720.813.316.816.3
Total capital ratio12.013.011.912.012.0

Excluding Credit Agricole because of missing data.

Excluding Deutche Bank for the comparison between 1999 and 2001 and between 1997 and 1999 because of missing data.

Excluding Anglo Irish Bank Coprporation because of missing data.

Excluding Banca Intensa and UniCredito because of missing data.

Excluding ING because of missing data.

Excluding HBOS because of missing data.

Excluding Credit Agricole because of missing data.

Excluding Deutche Bank for the comparison between 1999 and 2001 and between 1997 and 1999 because of missing data.

Excluding Anglo Irish Bank Coprporation because of missing data.

Excluding Banca Intensa and UniCredito because of missing data.

Excluding ING because of missing data.

Excluding HBOS because of missing data.

In addition, during the last few years, the EU-15 countries have seen an increase in large cross-border banking mergers. In 2004, Spain’s Banco Santander took over Abbey National to become a major player in the U.K. market. In 2005, Unicredit took over Germany’s HVB (which itself had bought Bank Austria Creditanstalt in 2000), thus acquiring major operations in Central and Eastern Europe. Still in 2005, the Italian market opened up after a prolonged takeover fight and an institutional crisis. Banca Antonveneta and Banca Nazionale del Lavoro, two relatively large Italian banks, were acquired by ABN Amro and BNP Paribas, respectively. All the acquiring banks listed a desire to have a major presence in a new, large market (to create a second, third, or fourth “home market”) as a main rationale for their acquisitions. However, all these transactions could be dwarfed by a merger between Barclays and ABN Amro, on which both banks reached agreement in April 2007.4

In summary, the data paint a picture of growing cross-border penetration indicative of increasing integration of European banking. Branches and subsidiaries from the EU-15 are a rising presence, and cross-border exposures continue to grow rapidly. Outsiders are diversifying into euro assets and raising their exposures to euro area countries, reflecting in part the development of London as the euro area’s main financial center. That said, and despite the recent wave of large acquisitions, it continues to be the typical case that few of the banks found on High Streets across Europe come from other countries, as foreign banks tend to be concentrated in the country’s financial center.

Evolution and Convergence of Risk Profiles

As it expands investment opportunities, financial integration may have different effects on the risk profiles of individual financial intermediaries and financial systems as a whole. On the one hand, financial integration may enhance diversification opportunities for individual intermediaries, which may rely on an enlarged set of investments across activities and borders to enhance expected returns for the same or a lower level of risk. On the other hand, a financial system can become less diversified as a whole if the constituent intermediaries choose greater exposure to the same risks or if the risks to which they are exposed become more similar. If so, the probability increases that a large number of financial institutions would adjust in a similar way to an adverse shock, thereby amplifying the overall impact on the economy or financial markets.

The extent to which financial institutions’ business strategies tilt toward specialization or diversification has different implications for the evolution and convergence of their risk profiles. If specialization strategies dominate, then intermediaries’ risk profiles should exhibit heterogeneity and a lack of convergence. Conversely, if diversification strategies dominate, then their risk profiles should exhibit more similarity and more convergence. Whether convergence is toward lower- or higher-risk profiles will be determined by the desired risk-return combination embedded in their business strategies.

The dynamics of system-wide risk profiles and their convergence are explored here through construction of the distance-to-default measure, a market-based soundness indicator. The distance to default (DD) is derived from the finance literature and based on the structural valuation models of Black and Scholes (1973) and Merton (1974).5 The DD is a composite measure computed as the sum of the return on the estimated market value of assets and the capital-to-assets ratio at market prices, divided by the volatility of assets. It thus combines measures of profitability, balance sheet strength, and market uncertainty. The level of the DD should not be seen as an absolute measure of soundness, but changes in the DD over time can be interpreted as changes in financial soundness. An increase in the DD indicates an improvement in financial soundness, caused by improved profitability, higher capitalization, reduced volatility, or a combination of these three factors.6 Although DD measures are sensitive to the underlying assumptions, they have been shown to predict supervisory ratings, bond spreads, and rating agencies’ downgrades.7 Of course, being an equity-based measure, the DD depends in part on general stock market trends.

The analysis in this chapter focuses specifically on the “portfolio” DD for a set of publicly traded European financial institutions during 1991–2005. The portfolio DD measure is constructed for a portfolio of banks and insurance companies belonging to each of the available Datastream stock indices of 13 of the EU-15 countries (Box 6.2).8 As with the standard DD, an increase (decrease) in portfolio DD indicates a lower (higher) risk profile for the set of institutions in the sample, which can result from higher expected profitability, better capitalization, lower asset volatility, or a combination of these factors. Cross-country convergence (or divergence) in system-wide risk profiles is measured by a decrease (or increase) in the cross-sectional standard deviation of DDs.9


In general, the risk profiles for the sample of banks considered do not appear to have improved over the past 15 years. In none of the countries do bank DDs exhibit a systematic upward trend (Figure 6.3). Indeed, in 11 of the 13 countries analyzed, DDs have tended to narrow between 1991 and 2003, before recovering in 2003–05. This suggests that, in most countries, risk reductions achieved through diversification have likely been offset by higher risk taking. This finding does not appear to be unusual, as similar patterns can be detected for large U.S. banks in this period (see De Nicolò and others, 2004).

Figure 6.3.EU Bank Distance to Default and Trend Component

(Standard deviations)
(Standard deviations)

Sources: Datastream; and IMF staff calculations.

Reflecting wide differences in size, business focus, and market penetration, the evolution over time of the risk profiles of publicly traded European banks is far from homogeneous. In Belgium, Germany, and the Netherlands, and to some extent in Spain and the United Kingdom, the DDs exhibit a downward trend between 1991 and 2003. In most other countries, the trend has typically been downward since the mid-1990s, often reversing a preceding upswing.10 Since 2003, the DDs have improved in almost all countries in the sample, largely reflecting benign global financial conditions.

However, the largest banks exhibit more similar patterns, as increases in asset return volatility between 1991 and 2003 have not necessarily been offset by increases in capitalization and improvements in asset returns (Figure 6.4). Large banks in France, Italy, Spain, the Netherlands, and the United Kingdom have experienced significant increases in asset return volatility in the period between the beginning of the 1990s and 2003, a trend not found in Germany. Substantial increases in capitalization and improvements in returns occurred simultaneously, but they were not sufficient to offset increases in risk taking captured by the rise in asset return volatility. Put differently, large European banks may have supported higher-risk/higher-return investments with larger capital buffers. Yet, risk-adjusted asset returns and overall risk profiles have not improved.

Figure 6.4.Large Banks: Return, Volatility, and Capitalization

(Percent, normalized January 1991)

Sources: Datastream; and IMF staff calculations.

Box 6.2.The “Portfolio” Distance-to-Default Measure

The basic structural valuation model by Black and Scholes (1973) and Merton (1974)—hereafter BSM—underpins the “portfolio” distance-to-default (DD) measure used in this paper. In the BSM model, the portfolio’s equity is viewed as a call option on the portfolio’s assets, with strike price equal to the current book value of total liabilities. When the value of the portfolio’s assets is less than the strike price, its equity value is zero. The market value of assets is not observable, but can be estimated using equity values and accounting measures of liabilities. The monthly DD measures used here are estimated with the methodology described in Vassalou and Xing (2004) using daily equity data and annual accounting data.

Under BSM assumptions, the distance to default of a portfolio of N firms is given by:

where Vp=ΣiVtiandLip=ΣiLti are the total value of assets and liabilities, respectively. The mean and variance of the portfolio are respectively given by μp=Σiwtiμiandσp=ΣiΣjwtiwtjσij,wherewti=Vti/ΣiVtiandσij is the asset return covariance of firm i and j. Thus, the “portfolio” DD embeds the structure of risk interdependencies among firms. “Default” at date t + 1 occurs when Vtp<Ltp. Thus, the DD indicates how many standard deviations Ln(Vtp/Ltp) has to deviate from its mean in order for default to occur. Since Vtp=Ltp+Etp,whereEtp is the value of equity, declines in Vtp/Ltp are equivalent to declines in capitalization (Etp/Ltp).

The portfolio DD can be viewed as a risk profile measure tracking the evolution of the joint risks of failure of the firms composing a portfolio. Lower (higher) levels of the DD imply a higher (lower) probability of firms’ joint failure. Since positive and negative variations in the individual firms’ DD are allowed to offset each other owing to firms’ return correlation, the DDofaportfolioisalwayshigher than the (weighted) sum of the DDs of the individual firms. As a result, the probabilityof“failure”associatedwiththe“portfolio”DDisalwayslower than that associated with the actual probability of joint failures of sets of firms in the portfolio. Thus, the portfolio DD can be viewed as tracking the evolution of a lowerbound to the joint probabilities of failure.

Despite the strong underlying assumptions, the dynamics of the portfolio DD provide useful information regarding the market valuation of systemic risk potential. The basic DD measures are constructed assuming that asset values follow a lognormal process, which does not capture extreme events adequately, and that the liability structure is composed of only equity and debt with fixed maturity for all firms and no rollover of debt. As a result, the implied estimates of probability of failure at a point in time may be imprecise. Moreover, without additional assumptions, the measures do not allow an identification of supply and demand factors that may drive their components, including overall stock market valuations. However, their dynamics have high informational content in signaling (forward-looking) market valuations of financial distress, as their predictive content for financial distress has been found to be significant. They have been shown to predict supervisory ratings, bond spreads, and rating agencies’ downgrades in both developed and developing economies (see Krainer and Lopez, 2001; Gropp, Vesala, and Vulpes, 2002, 2006; and Chan-Lau, Jobert, and Kong, 2004). Importantly, they have been found to have significant predictive power for actual defaults, even superior to measures based on “reduced form” statistical models of default intensities (see Arora, Bohn and Zhu, 2005). As a result, DD dynamics have become a standard tool of surveillance kits for financial as well as nonfinancial sectors.

Similar patterns among the large banks also show during the sharp recovery in DDs between 2003 and 2005, in which benign global financial conditions seem to have played a particularly large role. During this period, financial markets generally showed a steady upward trend, reflected in a uniform decrease in volatility at the large European banks (see Figure 6.4). As there are no indications that these banks have made a strategic choice to shed risks, these benign global financial conditions seem to be a main driver behind the recent simultaneous improvements in the banks’ risk profiles. At the same time, therefore, a reversal of these benign conditions could in turn lead to a rapid deterioration in the risk profiles.

Still, the improvements in capitalization and risk management undertaken by European banks in the past few years have also played a large role in the DD improvements since 2003. This conclusion is supported by, for instance, ECB (2004d), which also reports recent improvements in DD measures for 37 large EU banks (chart S43). Such strengthening may have supported increased risk taking in many instances, but it also helped the banks to weather a sequence of adverse financial shocks during 2000–03 (ECB, 2004a, 2005a).

Insurance Companies

Qualitatively, the dynamics of system-wide risk profiles for insurance companies in most European countries present a similar picture to that of banks (Figure 6.5). The DD also tended to decline up until 2003, after which it increased until 2005. Compared to banks, the dynamics of the DD for insurance companies are more heterogeneous across countries. However, the dynamics of risk profiles for banks and insurance companies have become more similar both within countries, and, as documented below, between countries, in part as a result of ongoing or increased conglomeration.11

Figure 6.5.Insurance Distance to Default and Trend Component

Large EMU countries

Sources: Datastream; and IMF staff calculations.

Driving Forces

Real business cycle developments do not appear to provide a systematic explanation of the dynamics of bank or insurance company risk profiles across countries. The correlation of the cyclical component of the bank DD with the cyclical component of GDP growth varies widely in magnitude. It is significantly negative in seven countries (Belgium, Germany, Greece, Ireland, the Netherlands, Portugal, and the United Kingdom), while it is positive in Austria, and not significantly different from zero in the remaining countries.12 This correlation also varies widely for the insurance sectors.

Notwithstanding cross-country heterogeneity, the risk profiles of banks and insurance companies converged markedly during 1991–2005 (see Figure 6.6). Convergence of banks’ risk profiles has occurred steadily, with the trend of the standard deviation of DDs dropping by over 50 percent (from 2.9 in 1991 to 1.4 in 2005). Insurance sectors exhibit a similar pattern, with the trend of the standard deviation of DDs dropping by 17 percent (from 3.0 in 1991 to 2.5 in 2005). Together with the lack of a systematic correlation across countries between DDs and the business cycle, this suggests that convergence in risk profiles across countries is unlikely to have been driven to an important extent by increased synchronicity in real business cycles.13

Figure 6.6.Convergence of Bank and Insurance Distances to Default

Sources: Datastream; and IMF staff calculations.

Increased exposure to financial cycles would appear to be a significant, although by no means unique, driver of convergence in risk profiles at large listed banks. Integration of money, bond, and equity markets may have played a role, as it has likely favored the diversification of institutions’ securities portfolios, which in turn may have increased their exposures to common financial shocks.14 European banks have exhibited a substantial increase in noninterest income (ECB, 2004a). This increase has been accompanied by a volatility of noninterest income growth significantly higher than that of interest income growth at large banks since 1997 (Table 6.3). These facts, as well as the dominance of convergence of the trend components of DDs, suggest that intermediaries’ business strategies, albeit different in many dimensions, may have produced the same outcome: a heightened exposure to common financial shocks.15 Again, similar results are found for U.S. banks (see Stiroh, 2004). The much less pronounced convergence of the cyclical component suggests that exposures of European banks to common sources of real shocks have not played a critical role yet, consistent with the current segmentation of the European retail bank markets. Thus, common exposures to financial shocks, as opposed to real shocks, seems to be an important factor explaining the convergence of risk profiles.

Table 6.3.Volatility of Income Growth Rates for Large Banks, 1997–2005
Number of

Growth of

Net Interest

Growth of



of Profits

before Tax
United Kingdom1123.527.327.4
Source: BankScope.Note: Volatility is measured by the sample standard deviation.
Source: BankScope.Note: Volatility is measured by the sample standard deviation.


There has been a convergence of risk profiles of listed banks and insurance companies despite the lagging integration of the relevant retail markets. Convergence appears to be the result not of more synchronized business cycles but of increased similarity in income sources and in exposures to financial shocks, because financial institutions have pursued similar growth strategies. In turn, such strategies may have been favored by the integration of money, bond, and equity markets and the ensuing diversification of securities portfolios. The similarity of the U.S. and European trends suggests that other drivers not necessarily related to integration may have played a role, such as technological developments.16

This chapter was prepared with research assistance from Marianne El-Khoury and is based on the authors’ work published in De Nicolò and others (2005).

For a discussion of quantity-type measures of integration, see Manna (2004). Annual data from the ECB on assets of EU branches and subsidiaries of foreign credit institutions from countries inside the European Economic Area (EEA; the EEA comprises the EU countries plus Iceland, Norway, and Liechtenstein) are available from 1997 to 2004. Quarterly data from the BIS on foreign exposures are available from 1999:Q2 to 2006:Q1 for most countries in the euro area and the United Kingdom.

EEA countries are selected for consistency with the relevant statistics produced by the ECB for earlier years.

At the time of writing, a rival offer for ABN Amro was proposed by a consortium of three banks.

The estimation is done using the procedure described in Vassalou and Xing (2004) on daily market and annual accounting data.

Two caveats apply. First, the DD as employed in this paper assumes a constant risk-free interest rate, and hence does not take into account the stochastic interest rate risk stemming from the correlation between the risk-free rate and the value of a company’s assets. As this is potentially another important source of risk that banks face, risks might be underestimated in the analysis (see Liu, Papakirykos, and Yuan, 2004, for an extension incorporating interest rate risk). Second, being based on market data, DDs can be subject to large fluctuations, which tend to be associated with the business cycle and “expectation cycles” regarding future earnings prospects.

The number of banks in the available index of each country is: Austria (8), Belgium (7), Denmark (9), France (8), Germany (16), Greece (9), Ireland (3), Italy (29), the Netherlands (4), Portugal (7), Spain (15), Sweden (5), and the United Kingdom (11). The number of insurance companies in the available index of each country is: Austria (3), Denmark (3), France (6), Germany (15), Greece (1), Ireland (1), Italy (10), the Netherlands (2), Spain (2), and the United Kingdom (15).

As illustrated in Solnik and Roulet (2000), the evolution of the cross-sectional standard deviation for a set of variables captures the degree to which correlations among these variables changes through time.

The initial upward trend in Sweden is essentially the outcome of the banking crisis of the early 1990s, when financial institutions experienced a large drop in the DD measure.

See Deutsche Bundesbank (2005). On the relationship and evidence between conglomeration and risk in an international context, see De Nicolò and others (2004).

Trend and cyclical components are constructed by applying the Hodrick-Prescott filter to monthly frequency data, adopting the value of the smoothing parameter used in Ravn and Uhlig (2002). The filter is applied to interpolated quarterly GDP growth data for the 1985:Q1–2003:Q4 period.

Other factors potentially at work may be a clustering of traders’ strategies that increased comovements in volatility, as well as converging market valuations of risk management practices.

Potential increases in direct risk interdependencies in the form of heightened exposures to potential contagion have been documented in ECB (2004d).

Some caution is warranted about generalizing the conclusions. Because BankScope primarily includes data on larger banks, in some countries with low banking concentration (such as Germany), a large share of banking assets is not covered. That said, the ECB (2000) documented the similarity in interest and noninterest income volatility for all banks with data up to 1998.

On the other hand, as argued in Chapter 4, there has been some parallelism in banking integration in the United States and the European Union.

    Other Resources Citing This Publication