Decomposing Financial Risks and Vulnerabilities in Eastern Europe
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Contributor Notes

Authors’ E-Mail Addresses: amaechler@imf.org; smitra@imf.org; dworrell@imf.org

This paper assesses how various types of financial risk such as credit risk, market risk, and liquidity risk affect banking stability in the ten countries that joined the European Union most recently, and eight neighboring countries. It also examines how the quality of supervisory standards may have mitigated the vulnerabilities arising from these risk factors. Using panel data, the study finds substantial variation in the impacts of financial risks, the macroeconomic environment, and supervisory standards on banks' risk profile across different country clusters. Credit quality is of general concern especially in circumstances where credit growth is accelerating.

Abstract

This paper assesses how various types of financial risk such as credit risk, market risk, and liquidity risk affect banking stability in the ten countries that joined the European Union most recently, and eight neighboring countries. It also examines how the quality of supervisory standards may have mitigated the vulnerabilities arising from these risk factors. Using panel data, the study finds substantial variation in the impacts of financial risks, the macroeconomic environment, and supervisory standards on banks' risk profile across different country clusters. Credit quality is of general concern especially in circumstances where credit growth is accelerating.

I. Introduction

This paper tests for the impact of various financial risks on bank stability in eastern Europe. Risks include credit, liquidity, and market risks; and risks from the macroeconomic environment. Furthermore, the paper investigates the extent to which vulnerabilities might be mitigated by good supervisory and regulatory policies and practices. Financial sector assessment programs (FSAPs) undertaken by the IMF and World Bank in most of the countries in the region in recent years have generally reported remarkable success in financial reforms after a period of financial turbulence in the early 1990s, as reflected in rapidly improving financial stability indicators and greater resilience to financial risk exposure. However, based on experience in other parts of the world, there remains a concern that financial supervision and regulation needs to be further upgraded especially since risks from rapid credit growth and potentially unsustainable macroeconomic imbalances could materialize in the future.

The study covers data on banks of the 10 countries that joined the European Union (EU) in 2004 (EU10), and 8 countries in the surrounding region (S8).2 The S8 share many financial characteristics with the EU10, including, in many cases, the large presence of EU-based foreign banks and financial institutions. They have also witnessed the rapid credit growth typical of much of our sample, and they share a concern about the financial sector implications of exchange rate policy. Also included in the study are three non-core EU countries (EU3) to act as a control group within the sample.3

In the next section, we present a literature review, which discusses the results of earlier studies and the variables used. A description of the underlying data, including regional variations by country clusters, is provided in Section III. The methodology and results are presented in Section IV, followed by conclusions in Section V.

II. Literature Survey

The empirical test in our study explores risk factors which are most often discussed in the recent literature and in policy circles, using an existing risk measure, and incorporating information on the quality of regulation and supervision. Our discussion includes rapid growth in bank credit, exchange rate regime and volatility, the extent of foreign ownership, the size of financial institutions, macroeconomic stability, and the quality of the regulatory and supervisory framework. This section provides a background for the study, drawn from the literature, and explains the choice of variables included in the empirical test.

A measure of risk

An increasingly used measure of bank soundness is the risk of insolvency or distance to default, also referred to as the z-index. This index which is directly related to the probability of loss exceeding equity capital can be summarized by:

zμ+kσ

where μ is average return on assets (in percent), k is equity capital in percent of assets, and σ is the standard deviation of return on assets as a proxy for return volatility.4 Statistically speaking, z measures the number of standard deviations a return realization has to fall in order to deplete equity, under the assumption of normality of banks’ returns. A higher level of z corresponds to a greater distance to equity depletion and therefore higher banks stability.

Credit growth

The risk of a credit boom-and-bust is the subject that has attracted most attention, among possible financial risks in European countries. At end-December 2006, 16 European countries experienced an annual private sector credit growth exceeding 20 percent, 5 of which had rates over 50 percent (Figure 1). All of the countries with a credit growth rate above 20 percent were Central and Eastern European (CEE) countries, except for Ireland (23 percent), Spain (24 percent), and Luxembourg (34 percent). At end-December 2006, the level of financial intermediation in CEE countries remains low, with a ratio of private sector credit to GDP ranging between almost 15 percent (Albania) and 64 percent (Latvia), in contrast to an average of almost 130 percent for the Euro area. While credit growth is largely perceived as part of a welcome catch-up process after many years of limited financial intermediation, some policymakers are increasingly concerned about its negative implications on macroeconomic and financial sector soundness.

Figure 1.
Figure 1.

Credit Growth in Europe, End-2006 1/

Citation: IMF Working Papers 2007, 248; 10.5089/9781451868111.001.A001

1/ 2005-2006 y-o-y growth, except for Albania and Poland (2004-2005 y-o-y growth), Credit/GDP For 2006; 2005 for Albania, Poland and Slovenia.

A very swift rise in credit may be the outcome of rapid income growth or the development of new credit markets such as housing and mortgage credit.5 In such circumstances, credit expansion may coexist for some time with low and declining inflation. Credit may also increase rapidly in cases of successful stabilization and significant economic reform, with credible economic policies. In practice, however, credit boom-and-bust cycles have often been associated with the absence of close financial surveillance. Thus, despite seemingly sound fundamentals, most studies generally agree that financial soundness indicators should be carefully monitored for early warnings of distress, that standards of prudential regulation and supervision should be strengthened and their implementation intensified, and that materialization of excess demand pressures should be closely analyzed.6

In terms of policy responses, there is also widespread agreement that, should signs of financial instability appear, tightening fiscal policy can be an effective response to slow down credit growth, whereas monetary policy measures, especially in countries with closely managed exchange rates and open capital accounts, have generally proved largely ineffective. Administrative measures and direct policy tools—such as reserve requirements, credit controls, etc—are sometimes seen to encourage excessive risk-taking by diverting local currency-denominated credit demand to foreign currency sources. To maintain the quality of banks’ loan portfolios, prudential tightening is the typically recommended policy response, although there is little evidence that such measures help reduce the speed of credit growth. If prudential measures are used to ration credit, there is an incentive to satisfy the excess credit demand through nonbank financial institutions, transferring the risk to nonblank financial institutions and/or non-financial borrowers.7

Exchange rate strategy

The literature has focused on the sustainability of exchange rate strategies rather than the implications for financial stability. But stable and sustainable exchange rate regimes are a necessary, though not sufficient, condition for financial stability. Backé and others (2004) distinguish between countries that have given up their monetary policy through the adoption of a currency board or a fixed exchange rate regime (Cyprus, Estonia, Latvia, Lithuania, Malta, and Slovenia), and all other countries, which can use the exchange rate as a stabilizing tool. No change in strategy is expected for the former group, although, in some cases, there may need to be some adjustments in the parities. Other studies seem to confirm this result. Burgess, Fabrizio, and Xiao (2003) concluded that the strategy of fixed exchange rates leading up to Euro adoption was viable for all Baltic countries, provided fiscal policy was sufficiently tight to counteract capital inflow surges. Other conditions include sound export performance and competitive (albeit appreciating) real exchange rates, owing to on-going productivity growth and economic reforms. Gulde, Kähkönen, and Keller (2000) concluded that, in general, countries with currency board arrangements (CBA) have experienced lower inflation and higher growth than countries with floating rates and simple pegs. They suggest that it may be possible to go directly from CBA to the European monetary union, given a conservative fiscal stance, a healthy financial system, cautious external debt management, and flexible labor markets.

IMF (2005) found no signs of flagrant exchange rate over- or undervaluation in accession countries, even though there was a wide range of relative competitiveness positions, suggesting that a range of parities may be manageable. Overvaluation may diminish over time, and the change will be more rapid and less costly if it is achieved by prices rather than quantities. Fiscal policy adjustment may help to reduce costs. However, not all studies concurred with this point of view. Egert and Lahrèche-Révil (2003) concluded that the Polish, Czech and Slovenian exchange rates were out of equilibrium, and De Haan, Berger, and van Fraasen (2001) argued that, while Estonia’s D-Mark based currency board was very much in line with the criteria for an optimal monetary regime, Lithuania’s initial choice of a U.S.-dollar based currency board was not.8 For the Czech Republic, Hungary, Poland, the Slovak Republic, and Slovenia, Borghijs and Kuijs (2004) found that the exchange rate responded little to shocks that affected output.

Foreign ownership

Foreign direct investment in financial institutions may have helped to integrate countries’ financial markets into the global financial system, bringing significant benefits of efficiency and stability, but it may also have highlighted country risk and financial vulnerability.9 Naaborg and others (2003) concluded that foreign bank entry was among the most striking features of European transition countries, with foreign banks accounting for over half the number and two-thirds the assets of their banking systems within less than ten years (Table 1). These foreign banks, most of which are owned by reputable western European bank groups, have increased stability and efficiency by revamping the banking sector in many CEE countries and re-establishing public confidence in their financial system. However, the presence of these banks has also introduced new challenges for host country supervisors, who must assess the risks that may arise from a change in the parent institution’s strategy or risk appetite and that are managed by the parent’s centralized risk management on a group-wide basis.10

Table 1.

Bank Ownership in Selected CEE Countries, 2003

article image

Macroeconomic stability

Financial vulnerability and resilience depend largely on the soundness of macroeconomic policy, as reflected in stable non-inflationary GDP growth rates, with sustainable debt and fiscal and external balances.11 Since the emerging market crises of the 1990s, many studies have confirmed a strong correlation between rising macroeconomic vulnerabilities, including large external deficits and debt and financial risks (Kaminsky, Lizondo, and Reinhart 1998). In particular, large current account deficits make emerging countries vulnerable to sudden capital flow reversals, as these deficits tend to be driven by financial market imperfections (such as liability euroization, and limited access to longer-term capital and equity finance) and financed through foreign bank funding rather than domestic saving and investment decisions, as is the case in rich countries (Blanchard, 2007; and Calvo,1998).

When foreign investors stop rolling over domestic debt, the resulting financing gaps have required a drawdown of reserves or higher interest rates. This has typically led to pressures on the exchange rate, which in turn affected bank portfolios, as holders of foreign currency or variable interest rate debt found it difficult to make repayments (Roubini and Setser, 2004). The impact of such shocks can be amplified by balance sheet mismatches and the extent to which the inflows have been channeled into productive vs. nonproductive sectors. If inflows have been absorbed primarily by nontradables, concerns about a country’s debt sustainability further raise financing costs and thereby, banks’ liquidity risks, market risks and credit risks (Sorsa et al., forthcoming).

The quality of regulation and supervision

Since 2001, the International Monetary Fund and the World Bank conduct joint FSAPs, in which they assess a country’s ability to withstand shocks and develop in a sustainable way. An important aspect of these assessments is the capacity of regulatory systems to reduce risks and increase the system’s resilience in case of a disturbance.12 The reports include assessments of country performance in relation to a variety of internationally agreed standards and codes, which typically include the Basel Core Principles of Effective Banking Supervision (BCP) and other codes of good practices, such as the Core Principles for Supervision of Systemically Important Payments Systems (CPSS) and guidelines issued by the International Association of Insurance Supervisors (IAIS) and the International Organization of Securities Commissioners (IOSCO).

In line with earlier studies, we use some of these international standard assessments to compile scores of the overall quality of supervision and regulation.13 Podpiera (2004) found some evidence of a positive impact of compliance with the BCP on banking sector performance. In our paper, we use a similar methodology to calculate a compliance index based on various elements of the BCP and CPSS assessments.14 In particular, from the we use the principles that relate to prudent credit policies and loan loss provisioning (CPs 7–8); limits on large exposures (CP 9) and connected lending (CP 10); market risk management (CPs12–13); quality of financial information (CPs 14, 19, and 21); and consolidated supervision (CPs 23–24). From the CPSS, we focus on the principles related to payment systems risk management (CPSS 2–7).

FSAPs and reports on standards and codes (ROSCs) for a majority of the countries in our sample, between June 2001 and present, reveal that the regulatory frameworks of almost all countries were adequately supervised, and many were described as well supervised and regulated. In all cases where supervision was only adequate, the FSSAs reported that a process of further strengthening was already underway. Compliance with BCP was generally good, even though there remained a few areas of weaknesses, with respect to lack of transparency of bank ownership, weak governance, and inadequate credit and other risk management policies in some countries.

III. Methodology

The financial risk variables used in this paper are common to those found in similar studies, except for the data on compliance with certain financial supervisory standards, which have rarely been applied in the literature on financial risk.15

The model

Our model follows in the tradition of studies that focus on the joint effect of a variety of macroeconomic and prudential variables on the vulnerability of financial institutions or the financial system as a whole.16 However, rather than test for financial institution failure, as is typical in these studies, our dependent variable is a measure of insolvency risk, or distance-to-default, of an individual bank—logz_rol, based on the z-index described above in Section II.

We estimate the following model to test for different risk factors that affect logz_rol:

logz_rolit=α+β1(Sizeit)+β1f(fodi*Sizeit)+s=13βBR,s(BRits)+s=13βBRC,s(BRits*CPiBRs)+βMR(MRit)+βMRC(MRit*CPiMR)+s=13βMs(Macit)+εit

The subscript i stands for bank; subscript t for year. Our dependent variable, logz_rol, is a variation of De Nicolo’s (2000) indicator of banking stability. In particular, logz_rol is computed as the sum of the average return on assets (in percent) and equity capital (as percent of assets) over the standard deviation of return on assets. To take advantage of as much year-on-year variation as possible, we use a three-year rolling z-index, which is computed by using the three-year moving average of return on assets (profitability) plus the three-year moving average of equity to assets (capitalization) over the three-year standard deviation (of return on assets). All variables, including the dependent variable, are transformed into natural logarithms.

The list of explanatory variables aims to incorporate a wide variety of possible risks, from those discussed in the literature and found in FSAP reports. The right-hand side variables are grouped into those that describe bank size (Size), including an interaction term with foreign-owned banks (fod*Size); bank-specific risks factors (BRs), country-specific market risk factors (MR), and interaction of each of bank risk factors and market risk factors with the countries’ compliance level with certain core principles of effective banking supervision and payment systems (CP)17; and variables describing the macroeconomic environment (Mac) that vary with country and year.18 The bank-specific factors included are credit growth, loan loss provisions, liquidity, bank size, and foreign ownership. Market risk is measured by exchange rate volatility, while macroeconomic risks include the ratio of credit to GDP, trade openness, and the inflation rate.

The effect of various risks and risk mitigating factors on bank stability is estimated by means of pooled OLS with heteroskedasticity-corrected (White) standard errors. A log-log specification is chosen so that the estimated coefficients can be interpreted as elasticities.19 Furthermore, to see which component of logz_rol is influenced by the various risk factors, we also run the same model with the individual components of logz_rol—profitability, equity over assets, and return-volatility—as dependent variables. Finally, to test for robustness, given substantial regional variation in the indicators, we run a simplified version of the model over pooled regional sub-samples.

Two caveats are in order. First, the specification of our model is not designed to infer causal relationships between bank stability and the various risk factors. Rather, the purpose of this paper is to identify statistically significant conditional correlations between these variables. In other words, the aim of our study is to investigate whether the presence of stronger banks is associated with, say, a stricter prudential and regulatory framework. Our results do not allow us to infer whether this stricter prudential framework has caused banks to become stronger or whether stronger banks prefer to operate in an environment with a stricter prudential and regulatory framework.

Second, in many countries of our sample, a significant portion of total loans is either denominated in foreign currency (dollars or euros) or indexed to the euro. As a result, it would be important to control for the impact of dollarization and euroization on financial stability, and to examine how exchange rate volatility may affect credit or liquidity risk directly (through banks’ balance sheets) or indirectly (through banks’ exposures to borrowers that may not be able to repay their debts denominated in foreign exchange). Unfortunately, neither the currency breakdowns of banks’ loan portfolios nor information on borrowers’ ability to withstand an exchange rate shock are readily available, making it very difficult to analyze this type of risk.

Data coverage

The data is based on annual data from Bankscope over the period 1997–2004. For the 21 countries included in our three groups (EU3, EU10 and S8), we selected all banks available in Bankscope for which data was available up to (at least) 2003. This yielded a total of 334 banks. Branches and subsidiaries of multinational banks are consolidated on a national basis—that is, various subsidiaries of a foreign bank in different countries are reported as separate entities.

Explanatory variables

Bank-specific risks

Banks’ risks are captured by credit risk and liquidity risk, and their interactions with the countries’ compliance with certain supervisory standards.20 A summary of various risks and risk-mitigating factors is given in Appendix Table 6; the discussion below draws on this table. See Table 2 for details on the variables used in the econometric exercise.

Table 2.

Variable Description 1/

article image

IFS = International Financial Statistics, IMF; WEO = World Economic Outlook; BS = Bankscope; assessment codes refer to the following: 4 = observed (in TFP and CPSS) or compliant (in BCP); 3 = broadly observed (in TFP and CPSS) or largely compliant (in BCP); 2 =partly observed (in TFP and CPSS) or materially non-compliant (in BCP); and 1 = non-observed (in TFP and CPSS) or non-complaint (in BCP). An ‘l’ in front of a variable denotes its natural log—lx=ln(x). An ‘s’ after the name of the variable denotes its square—xs=x*x or lxs=ln(x)*ln(x).

Credit risk

Credit risk from banks’ loan portfolios (in both local and foreign currency) is the main vulnerability of banks in EU10+S8 region, as identified in several FSAP reports.21 This is especially true in the case of a credit boom, which may hide the potential for future nonperforming loans (NPLs). There may also be indirect exchange rate-related credit risk on loans made in foreign currency (fx) to unhedged borrowers, even though banks keep foreign exchange open positions within the regulatory limit. We capture the risks associated with a credit boom-and-bust by including bank-by-bank credit growth (cg) and its square term (cgs) in the model. As a proxy for the riskiness of banks’ lending portfolio, we include loan-loss provisions in percent of net interest revenue (prov).22 We do not have any prior as to the sign on the coefficient of this variable. High provisioning may reflect high non-performing loans, and may be associated with a lower distance-to-default. Conversely, high provisioning could indicate prudence if a sound and profitable bank decides to boost precautionary reserves rather than distribute profits.23

Strong bank supervisory practices could mitigate some of the credit risk in so far as prudential guidelines encourage prudent risk management practices by banks. Assessment of these policies is made using the BCP. We used some of these assessments to see to what extent the countries and regions in EU10+S8 that have a high compliance with best practices, are better able to withstand shocks (higher logz_rol). For this, we interact the two principles that assess the quality of credit and provisioning policies (CPs 7–8) with prov. We also interact credit growth (cg) with an aggregated index that combines the four principles (CPs 7–10) that assess the overall quality of banks’ credit risk management practices (including policies on connected lending and large exposures).

Liquidity risk

Liquidity risk is modeled by taking the ratio of liquid asset to deposits and short-term funding (liq). Although rising liq is a positive influence on stability at low levels of liquidity, excessive liquidity could be a structural problem for the bank, reducing the value of our stability indicator. Thus, a bank could be highly liquid by not lending enough and holding large quantities of government securities, often in the absence of liquid secondary markets in such securities.

A key to avoiding systemic liquidity problems is the smooth functioning of, and management of risk in, payments and settlement systems. We make use of CPSS 2–7 to judge the level of country compliance on these policies. For a bank-specific effect, we combine CPSS 2–7 with liq.

Bank size and foreign ownership

We include total assets (ta) to capture the size of banks. A priori the sign on the coefficient of this variable is indeterminate, because the presence of very large banks could either be stabilizing or risky for the financial system, depending on the importance of economies of scale in each banking system (See, for example, De Nicolo, 2000).

Foreign bank ownership, which is very high in Central and Eastern Europe, introduces the risk that parent banks may fund credit expansion in the region in order to relieve tightening profit margins at home, generating rapid credit growth in the EU10+S8 countries. As a result, the foreign branches and subsidiaries may have contributed to a disproportionately large portion of the bank group profits compared to their risk exposures. Moreover, as parent banks tend to own subsidiaries in more than one country in the region, the resulting cross-border networks of bank groups introduces the risk that problems in one bank belonging to the regional network may spread to others, and that macroeconomic deterioration may be transmitted across borders. We capture risks of foreign ownership by interacting ta with a dummy variable that takes the value of 1 if the bank is foreign-owned (fod).

Country-specific market risk

The standard deviation of monthly exchange rate changes is used as our proxy for market risk (sd_exchg). High exchange rate volatility is a source of potential vulnerability, but good risk management policies to monitor market risks could mitigate the balance sheet effects of such fluctuations. We capture the latter by interacting sd_exchg with (BCP) CPs 12–13— supervisors should be satisfied that banks have in place systems that accurately measure, monitor, control market and other risks, and (supervisors) have the power to impose prudential limits or capital charges against such risks.

Macroeconomic environment

As country experience reported in the literature survey suggests, the macroeconomic environment could show some broad variations in stability trends across countries and country-clusters. We chose private sector credit to GDP (credgdp) as an indicator for overall financial development; trade openness (topen) to indicate susceptibility to real foreign shocks; and the inflation rate (infl) to indicate overall success of monetary policy.

IV. Regional Variation in the Data

Before turning to our empirical results, we present key regional variations found in the data. For purposes of comparison, we created seven clusters—Total (the total pooled sample), EU3 (Spain, Portugal, Greece), Surroundings (also referred to as S8—Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Macedonia, Moldova, Romania, and Serbia and Montenegro), High Credit Growth (Albania, Bulgaria, Estonia, Latvia, Lithuania, Moldova, and Romania) based on the classification in Hilbers, Otker-Robe, Pazarbasioglu, and Johnsen (2005),24 Baltics (Estonia, Latvia, and Lithuania), New Member States (also referred to as EU10— Cyprus, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovak Republic, and Slovenia), and Foreign Owned Banks. The variables and their sources are described in Table 2. Differences across clusters are depicted in Figures 25, which show the pooled means of various macroeconomic variables, banking characteristics, and regulatory compliance.

Figure 2.
Figure 2.

Mean of Z-Index and its Components

Citation: IMF Working Papers 2007, 248; 10.5089/9781451868111.001.A001

Figure 3.
Figure 3.

Mean of Macroeconomic Background

Citation: IMF Working Papers 2007, 248; 10.5089/9781451868111.001.A001

Figure 4.
Figure 4.

Mean of Banking Characteristics

Citation: IMF Working Papers 2007, 248; 10.5089/9781451868111.001.A001

Figure 5.
Figure 5.

Mean of Regulatory Indices

Citation: IMF Working Papers 2007, 248; 10.5089/9781451868111.001.A001

Figure 2 suggests a number of regional variations across country clusters in the overall stability indicator (z-index):

  • Compared to EU3 banks, banks in the S8 region are highly capitalized. In part owing to high interest margins, the banks in this region are also the most profitable, although they have the highest returns-volatility (measured by the standard deviation of returns on assets).

  • In spite of comparatively low capitalization and average profitability levels, EU3 banks appear to enjoy a lower insolvency risk (a higher z-index) than other banks in the sample, primarily because they experience much lower returns-volatility.

  • Countries experiencing high credit growth do not appear to be more vulnerable than other banks in the region, because their equity levels are high. Also, even though their rates of return are modest, they do not vary greatly.

The differences in macroeconomic characteristics as shown in Figure 3 are:

  • The EU3 countries have the highest ratio of financial assets to GDP but the lowest trade openness, whereas the reverse is true for the New Member States.

  • There seems to be a positive association between the inflation rate and bank insolvency risk (see Figures 2 and 3).

As far as liquidity and other bank characteristics are concerned:

  • The S8 banks exhibit ample liquidity and the highest average credit growth rate (see Figure 4).25 Despite higher profitability and capitalization, they are not more stable than the High Credit Growth group of banks (see Figure 2), mainly due to their higher return-volatility.

  • In the EU3 and S8 countries the loans to deposit ratios are higher than for the Baltic countries and the High Credit Growth countries, which could reflect higher indebtedness.

Bank sizes differ considerably among groups—average EU3 banks are nearly twice as large as the average for the entire pool, and new member country banks almost five times as big as the S8 ones. However, there does not seem to be systemic association between size and the stability (logz_rol) (see Figures 2 and 4).

As in some other studies (Podpiera, 2004), we have converted qualitative indicators of supervisory standards to quantitative scores (see Table 2 for details). The computed scores are shown in Figure 5 and the standards are elaborated in Appendix Table 7:

  • The regulatory regime shows less variation across regions, except for the S8 countries, which stand out with the lowest BCP scores.

  • Overall, countries seem to benefit from fairly strong payment systems infrastructure and oversight and there is not much difference in CPSS scores between regions.

V. Empirical Results

The main results are shown in Table 3. Columns 1–3 present the results with all the risks discussed in the previous section. Columns 4–6 focus on credit risk, as this has been consistently outlined as the main stability risk for banks. For each specification, we ran the regression controlling for banks in EU3 countries (columns 1 and 4), for all banks (columns 2 and 5), and banks in EU10+S8 region (columns 3 and 6). Table 4 provides estimates for the same model run on, respectively, profitability (columns 1–2), equity-to-assets (columns 3–4), and returns-volatility (columns 5–6) as dependent variables. We then run the credit risk part of the model on sub-sections of regional banks (Table 5). Overall, we find broadly robust results across specifications.

Table 3.

Risks and Stability 1/

article image

Absolute values of the t-statistics are in parentheses; significance at 1 percent level is shown by **, at 5 percent by *, and at 10 percent by +.

Table 4.

Components of logz_rol 1/

article image
Absolute values of the t-statistics are in parentheses; significance at 1 percent level is shown by **, at 5 percent by *, and at 10 percent by +. Columns 1–4 drop observations for which either equity/assets or roaa or both are negative.
Table 5.

Regional Credit Risk 1/

article image

Absolute values of the t-statistics are in parentheses; significance at 1 percent level is shown by **, at 5 percent by *, and at 10 percent by +.

Credit risk

In our model, several variables capture various aspects of credit risk and empirical tests yield the following results for each of them:

Credit growth

  • Higher bank-by-bank credit growth is associated with greater stability (positive sign on cg). Regressions on the components of logz_rol suggest that this result is driven by the association of faster credit growth with higher profits, higher equity, and lower volatility, all of which raise the stability indicator logz_rol (see Table 4).

  • However, banks become more vulnerable as credit growth accelerates (the quadratic effect of credit growth, lcgs, is strongly negative). This appears to be the case because returns become more volatile as credit growth accelerates (column 6 of Table 4), particularly in the case of the EU10+S8 group of banks.

  • Two results on the credit policy regime are puzzling. First, the returns to a bank with higher credit growth are lower where the supervisory regime is stronger (there is a negative coefficient of lcg_cp710s in Table 4, columns 1-4). Second, banks operating under a stricter credit policy regime (including limits on large exposures and connected lending, CP 7–10), have lower stability, indicators when credit growth is higher (there is a negative coefficient of lcg_cp710s in Table 3). In future work it would be useful to explore these results further with a dynamic model. One hypothesis is that in the short run, tougher supervisory standards may adversely affect banks’ profitability (through higher provisioning) and therefore reduce their apparent stability. Overtime, however, the higher costs associated with a stricter regulatory framework should translate into a greater ability to withstand shocks and therefore a lower returns-volatility and a higher value of the stability indicator. Our current model specification does not allow us to test this hypothesis.

Loan-loss provisioning

  • Higher provisioning for loan-losses is associated with a lower logz_rol (i.e. greater vulnerability). Evidence from the components of logz_rol indicates that banks with higher provisioning tend to be less profitable (see columns 1–2 of Table 4) and exhibit higher returns-volatility (see columns 5–6 of Table 4).

  • A higher score on the BCP that address credit and provisioning policies (CP 7–8) mitigates the negative effect of provisioning on stability. (The coefficient of lprov_cp78s is positive and statistically significant in Table 2) Higher compliance with CP 7–8 is associated with higher profitability (see columns 1–2 of Table 4).

Liquidity risk

Liquidity risks (lliq) have mixed effects on stability as defined by logz_rol.

  • Overall, there appears to be a negative association between liquidity and stability. Individual component estimations indicate that highly liquid banks tend to exhibit significantly higher returns-volatility. This is particularly true for the EU10+S8 group of banks (see column 3 of Table 3).

  • However, banks operating in countries with a good payment systems infrastructure and oversight (CPSS 2–7) experience lower returns-volatility, and hence, a lower insolvency risk.

Market risk

Country-wide exchange rate volatility has a somewhat counter-intuitive effect on stability.

  • Exchange rate volatility (sd_exchg) is associated with a higher logz_rol (higher stability) (see columns 1–3 of Table 3), mostly through higher capitalization and, in the case of EU10+S8, reduced return-volatility (see columns 3–4 and 6 of Table 4). This is plausible if banks anticipate the impact of possible exchange rate fluctuations on their balance sheets and allow for higher capital buffers.

  • However, the positive effect of exchange rate volatility on bank stability is somewhat mitigated when bank supervisors enforce strict market risk management practices (CP 12–13). This suggests that a strict regulatory framework may induce banks to better match their capitalization levels with the underlying risks, leaving less need for extra capital buffers (sde_cp1213s).

Macroeconomic performance and structure

  • Banks in countries with greater financial depth—a higher private sector credit (as a percent of GDP), lcredgdp—are more stable, which is the expected result.

  • Trade openness (ltopen) has a negative effect on stability, especially through its negative impact on capitalization. This result may reflect the greater inherent riskiness of foreign exposures.

  • Higher inflation is associated with higher profitability but has no significant effect on the stability indicator, logz_rol. This is a plausible result in a period of moderate inflation.

Bank structure and ownership

  • Larger foreign-owned banks are less stable (negative coefficient of fodlta), mainly due to lower profitability, lower capitalization and a mildly higher volatility (mainly in EU10+S8). While this result is counter-intuitive, it probably reflects the inherent weakness of our stability measure. Because foreign banks typically have access to a very large pool of equity funds abroad, they may safely operate with much lower levels of capitalization of local operations, than would be the case for local banks.

  • Profitability increases with size (except for EU10+S8 banks), whereas both capitalization and returns-volatility decrease in size. The effects appear to cancel each other, and the size variable is not significant for the overall stability of banks included in our sample.

Regional Credit Risk

Table 5 examines credit risk variation among the different groups defined earlier, with a simplified model containing only credit risk. Table 5 shows that the basic conclusions of the fuller model remain intact.

  • For all the regions, credit acceleration is associated with greater vulnerability, although financial depth (credgdp) does not matter for the S8 region.

  • Higher provisioning is associated with lower bank stability (negative impact on logz_rol) in EU10 and High Credit Growth countries. According to the regressions on the individual logz_rol components, this result is driven by the negative effect of higher provisioning on profitability and, in the case of EU10&S8 banks, higher returns-volatility.

  • For banks operating in S8 countries, a high score in the quality of supervision of credit policies (CPs 7–8) is associated with lower insolvency risk, higher profitability and lower returns-volatility.

VI. Concluding Remarks

The results indicate that while a focus on credit quality is justified, it is the acceleration of credit, rather than its rate of growth, that warrants extra vigilance. The observed rates of growth of credit are associated with greater bank stability for our sample, and it is only when credit growth speeds up that banks appear more vulnerable. When credit growth accelerates it is important to ensure sound supervisory practices, in order to minimize risk exposure.

Two results on the credit policy regime may need to be further explored, using a dynamic model. First, the returns to a bank with higher credit growth fall with the strength of the supervisory regime. Second, banks experiencing rapid credit expansion in a context of stricter credit policy regime exhibit lower stability. These phenomena may result from the adjustments that banks were required to make in response to supervisory tightening (i.e., higher provisioning), but we were unable to investigate that possibility.

Higher loan-loss provisioning is associated with lower stability, mainly through lower profitability and higher returns volatility. Procyclical provisioning practices—that is, provisioning more when returns are low—could increase profit volatility. However, improved supervisory policies on provisioning help to sustain profits and reduce volatility.

Foreign banks tend to have a higher risk profile than domestic banks because of their relatively lower capitalization, which is a reflection of their ability to rely on extra funding from their parent institutions when needed. There is no significant difference between the risk profiles of larger and smaller banks, although the returns-volatility of larger banks tends to be lower, suggesting a positive diversification effect

This paper is a first attempt at identifying the role of selected risk factors in affecting banking stability and how they may be mitigated by a strong prudential and regulatory framework.

Over time, with the availability of a wider dataset, the research may be extended to a wider sample of countries, a broader range of exchange rate regimes and macroeconomic diverse profiles. Longer data series will permit the investigation of dynamic effects such as the impact of costly risk mitigation regulations on (future) financial stability benefits. Access to a currency breakdown of banks’ balance sheet information and financial income statements will permit exploration of banks’ exposure to credit risk induced by potential exchange rate volatility. There is a need to refine the bank instability indicator, to ensure that it more faithfully reflects market perceptions of bank risk exposure. Finally, much work remains to be done on refining the computation of financial regulation indices, from the impact of using different weighting and scoring systems to documenting changes over time.

Appendix Table 6.

Sources of Risk and Risk-Mitigation Practices

article image
article image
Appendix Table 7.

Standards and Codes

article image
article image

References

  • Altman, Edward, and Anthony Saunders, 1998, “Credit Risk Measurement: Developments over the Last 20 Years,” Journal of Banking and Finance, Vol. 21, pp. 172142.

    • Search Google Scholar
    • Export Citation
  • Backé, Peter, and others, 2004, “The Acceding Countries’ Strategies Towards ERM II and the Adoption of the Euro: An Analytical Review,” ECB Occasional Paper No. 10, February.

    • Search Google Scholar
    • Export Citation
  • Bank for International Settlements, 2005, “Foreign Direct Investment in the Financial Sector – Experiences in Asia, Central and Eastern Europe, and Latin America,” Committee on the Global Financial System, June.

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier, (2007), “Current Account Deficits in Rich Countries,” Staff Papers, International Monetary Fund, Vol. 54, No. 2, pp.(191-219), June.

    • Search Google Scholar
    • Export Citation
  • Bonin, John, and Paul Wachtel, 2003, “Financial Sector Development in Transition Economies: Lessons from the First Decade,” Financial Markets, Institutions & Instruments, Vol. 12, No. 1, February.

    • Search Google Scholar
    • Export Citation
  • Borghijs, Alain, and Loius Kuijs, 2004Exchange Rates in Central Europe: A Blessing or a Curse?IMF Working Paper 04/02 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Borio, Claudio, and Philip Lowe, 2002, “Asset Prices, Financial and Monetary Stability: Exploring the Nexus,” BIS Working Paper No. 114, July.

    • Search Google Scholar
    • Export Citation
  • Burgess, Robert, Stefania Fabrizio, and Yuan Xiao, 2003, “Competitiveness in the Baltics in the Run-up to EU Accession,” IMF Staff Country Report No. 03/114 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Calvo, G., (1998), “Capital Flows and Capital-Market Crises: The Simple Economics of Sudden Stops,” Journal of Applied Economics, Vol. 1, No. 1.

    • Search Google Scholar
    • Export Citation
  • Corker, Robert, Gianni de Nicolo, Alexander Tieman, and J-W van der Vossen, 2005, “European Financial Integration, Stability and Supervision,” European Commission ECFIN Conference, October 6–7.

    • Search Google Scholar
    • Export Citation
  • Coricelli, Fabrizio, and Igor Masten, 2004, “Growth and Volatility in Transition Countries: the Role of Credit,” Festschrift in honor of Guillermo Calvo, IMF, April 15–16, available at http://www.imf.org/external/np/res/seminars/2004/calvo/pdf/corice.pdf.

    • Search Google Scholar
    • Export Citation
  • Cottarelli, Carlo, Giovanni Dell’Ariccia, and Ivann Vladkova-Hollar, 2003, “Early Birds, Late Risers, and Sleeping Beauties: Bank Credit Growth to the Private Sector in Central and eastern Europe and the Balkans,” IMF Working Paper 03/213 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Das, Udaibir, Plamen Iossifov, Richard Podpiera, and Dmitriy Rozhkov, Quality of Financial Policies and Financial System Stress,” IMF Working Paper 05/173 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • De Haan, Jakob, Helge Berger, and Erik Van Frassen, 2001, “How to Reduce Inflation: An Independent Central Bank or a Currency Board? The Experience of the Baltic Countries,” Center for Economic Studies and Ifo Institute for Economic Research Licos Working Paper No. 96/2001, January.

    • Search Google Scholar
    • Export Citation
  • De Nederlandsche Bank, 2001, “Economic Convergence and Monetary Policy in Accession Countries,” Quarterly Bulletin, March.

  • De Nicoló, Gianni, 2000, “Size, Charter Value and Risk in Banking: an International Perspective,” Federal Reserve International Finance Discussion Paper No. 689, December.

    • Search Google Scholar
    • Export Citation
  • De Nicoló, Gianni, Peter Hayward, and Ashok Vir Bhatia, 2004, “U.S. Large Complex Banking Groups: Business Strategies, Risks, and Surveillance Issues,” in IMF Staff Country Report No. 04/228 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Egert, Balázs, and Amina Lahrèche-Révil, 2003, “Estimating the Equilibrium Exchange Rate of the Central and Eastern European Acceding Countries: the Challenge of Euro Adoption,” Weltwirtschaftliches Archiv, Vol. 139, No. 4, pp. 683708.

    • Search Google Scholar
    • Export Citation
  • European Commission, 2004, “Financial Integration Monitor,” Available via internet: http://europa.eu.int/comm/internal_market/en/finances/actionplan/index.htm, 10th Progress Report Press Release link, June.

    • Search Google Scholar
    • Export Citation
  • European Forecasting Network, EFN, 2004, “The Euro Area and the Acceding Countries,” EFN Report, Spring.

  • FitchRatings, 2005, “Assessing Bank Systemic Risk: A New Product,” July 26, Available via Internet: www.Fitchratings.com.

  • Gulde, Anne-Marie, Juha Kähkönen, and Peter Keller, 2000, “Pros and Cons of Currency Board Arrangements in the Lead-up to EU Accession and Participation in the Euro Zone,” IMF Policy Discussion Paper 00/1 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Hilbers, Paul, Inci Otker-Robe, Ceyla Pazarbasioglu, and Gudrun Johnsen, 2005, “Assessing and Managing Rapid Credit Growth and the Role of Supervisory and Prudential Policies,” IMF Working Paper 05/151 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2007, Global Financial Stability Report: Market Developments and Issues, April, Washington, DC.

  • International Monetary Fund, 2005, Adopting the Euro in Central Europe - Challenges of Next Steps in European Integration, IMF Occasional Paper No. 234 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2004a, World Economic Outlook April 2004: Advancing Structural Reforms.

  • International Monetary Fund, 2003, “Euro Area Policies - Selected Issues,” Country Report No. 03/298, August 12.

  • IMF Economic Forum, 2004b, “Adopting the Euro in the New Member States: The Next Step in European Integration,” May 4.

  • Kaminsky, Graciela, Saul Lizondo, and Carmen Reinhart, 1998, “Leading Indicators of Currency Crisis”, Staff Papers, International Monetary Fund, Vol. 45, No. 1, pp.148, March.

    • Search Google Scholar
    • Export Citation
  • Kóbor Kobor, Adám, and István Székely, 2004, “Foreign Exchange Market Volatility in EU Accession Countries in the Run-up to Euro Adoption: Weathering Uncharted Waters,” IMF Working Paper 04/16 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Lin, Shu Ling, Jack Penm, Shang-Chi Gong, and Ching-Shan Chang, 2005, “Risk-based Capital Adequacy in Assessing on Insolvency Risk and Financial Performances in Taiwan’s Banking Industry,” Research in International Business and Finance, Vol. 19, No. 1, March, pp. 11153.

    • Search Google Scholar
    • Export Citation
  • Manna, Michele, 2004, “Developing Statistical Indicators of the Integration of the Euro Area Banking System,” ECB Working Paper No. 300, January.

    • Search Google Scholar
    • Export Citation
  • Naaborg, Ilko, Bert Scholtens, Jakob de Hann, Nanneke Bol, and Ralph de Haas, 2003, “How Important Are Foreign Banks in the Financial Development of European Transition Countries?CESifo Working Paper No. 1100, December.

    • Search Google Scholar
    • Export Citation
  • Podpiera, Richard, 2004, “Does Compliance with Basel Core Principles Bring Any Measurable Benefits?IMF Working Paper 04/204 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Portes, Richard, and Hélène Rey, 1999, “The Determinants of Cross-Border Equity Flows,” London Business School and LSE, August.

  • Roubini, N., and B. Setser, (2004), Bailouts or Bail-ins? Responding to Financial Crises in Emerging Economies, (Washington, DC: Institute for International Economics).

    • Search Google Scholar
    • Export Citation
  • Schadler, Susan, Paulo Drummond, Louis Kuijs, Zuzana Murgasova, and Rachel van Elkan, 2004, “Euro Adoption in the Accession Countries: Vulnerabilities and Strategies,” Paper for Conference on Euro Adoption in the Accession Countries—Opportunities and Challenges, Czech National Bank, February 2–3.

    • Search Google Scholar
    • Export Citation
  • Schipke, Alfred, Christian Beddies, Susan George, and Niamh Sheridan, 2004, Capital Markets and Financial Intermediation in the Baltics, IMF Occasional Paper No. 228 (Washington: International Monetary Fund)

    • Search Google Scholar
    • Export Citation
  • Schinasi, Garry, 2006, Safeguarding Financial Stability: Theory and Practice, (Washington: International Monetary Fund).

  • Sorsa, Piritta, Bas B. Bakker, Christoph Duenwald, Andrea M. Maechler, and Andrew Tiffin, 2007, “Vulnerabilities in Emerging Southeastern Europe--How Much Cause for Concern?IMF Working Paper 07/236, October.

    • Search Google Scholar
    • Export Citation
  • Tornell, Aaron, and Frank Westermann, 2002, “Boom-Bust Cycles: Facts and Explanation,” Staff Papers, International Monetary Fund, Vol. 49, Special Issue, (pp.111-155).

    • Search Google Scholar
    • Export Citation
  • Vincze, János, 2001, “Financial Stability, Monetary Policy and Integration: Policy Choices for Transition Economies,” National Bank of Hungary, WP 2001/4, December.

    • Search Google Scholar
    • Export Citation
  • Wagner, Nancy, and Dora Iakova, 2001, “Financial Sector Evolution in the Central European Economies: Challenges in Supporting Macroeconomic Stability and Sustainable Growth,” IMF Working Paper 01/141 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Worrell, DeLisle, 2004, “Quantitative Assessment of the Financial Sector: An Integrated Approach,” IMF Working Paper 04/153 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
1

The authors especially thank Pamela Madrid for helping to inspire and conceptualize the study, and for many useful references and comments, during its gestation; Kiran Sastry and Nada Oulidi, for their invaluable data assistance; Jochen Andritzky, Martin Cihak, Vidhi Chhaochharia, Gianni De Nicolo, Alain Ize, Inutu Lukonga, Kathleen McDill, Franziska Ohnsorge, David Parker, Mark Swinburne, Jan-Willem van der Vossen, and Francesco Vasquez; and participants at the Second Annual DG ECFIN Research Conference, for helpful suggestions. We are also indebted to Richard Podpiera, for providing material for the computation of the BCP indices. The authors are responsible for any remaining errors.

2

The EU10 comprise Cyprus, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovak Republic, and Slovenia. The S8 are Albania, Bosnia and Herzegovina, Bulgaria, Croatia, Macedonia, Moldova, Romania, and Serbia and Montenegro.

3

Adopting the terminology introduced by Schadler et al. (2004), the three “non-core” countries refer to Greece, Portugal, and Spain, as these countries joined the EU much later than the rest of their western European counterparts.

4

A higher z—and higher log(z)—implies a lower upper bound of insolvency risk and hence a lower probability of bank insolvency. Ideally, the z-score should be computed based on the market values of shareholders’ equity and assets rather than the book value of banks’ balance sheets, as is done here to overcome the lack of data on market capitalization of most of the banks in our sample. Since book values are invariably lower than market values, our measure of z gives a more conservative (but less volatile) measure of risk than would in fact exist.. For other studies using a similar methodology, see, for example, De Nicolo (2000), Altman and Saunders (1998), and Lin, Penm, Gong, and Chang (2005).

5

Studies of rapid credit growth include IMF (2004a)—which also considers this phenomenon in East Asia— Schadler, Drummond, Kuijs, Murgasova, and van Elkan (2004); Coricelli, and Masten (2004); Cottarelli, Dell’Ariccia, and Vladkova-Hollar (2003); IMF (2005); IMF (2004b); and Borio and Lowe (2002), who deal with this issue in a more general context.

6

A few studies, however, have found limited evidence of credit boom-induced banking crises (IMF, 2004a, and Tornell and Westermann, 2001).

7

See, for example, Hilbers et al. (2005).

8

Both countries are now pegged to the euro.

9

See the Bank for International Settlements (2005) for an analysis of the experiences of Asia, Central and Eastern Europe, and Latin America.

10

A discussion on the role of foreign banks as a risk transmission mechanism in emerging Europe can be found in Sorsa et al., (forthcoming). The risk implications of the centralization of operational functions in cross-border bank groups are discussed in IMF (2007).

11

See, for example, Schinasi (2006). For an in-depth discussion on the impact of rising vulnerabilities on the macroeconomic and financial sector stability of Emerging Southeastern European countries, see Sorsa et al., forthcoming,

13

See Podpiera (2004); and Das, Iossifov, Podpiera, and Rozhkov (2005).

14

There are other sources of risk which we were unable to explore for lack of data, including issues of financial integration among European countries (Manna, 2004; the European Commission, 2004; and Corker, de Nicolo, Tieman, and van der Vossen, 2005); capital flows, including spillovers and sudden large scale reversals (IMF, 2005; Kóbor and Székely, 2004; Vincze, 2001; and Portes and Rey, 2001); and direct and indirect euroization risks.

15

Podpiera (2004); and Das, Iossifov, Podpiera, and Rozhkov (2005) are two exceptions.

16

They are surveyed in Worrell (2004).

18

All variables are taken as natural logarithms, except for the dummy variables. For variables that can take 0 or negative values, we have used a transformation when taking logs as follows: ln(1+x), for small x (expressed as fraction).

19

The presence of time-invariant and country-specific supervisory variables (the CPs) makes it difficult to use a (fixed-effect) panel estimation model, which would drop a number of relevant bank-specific variables. A pooled OLS, however, enables us to exploit both variations within and between banks as well as regional variations.

20

We interact the supervisory scores with a bank-specific variable to avoid losing too many degrees of freedom.

21

See http://www.imf.org/external/np/fsap/fsap.asp for published FSSAs by country.

22

NPLs would have been a good indicator, but using this would have led to a sharp decline in our sample size due to missing observations on most banks.

23

As Fitch (2005) notes, prudential behavior of banks could be a risk factor if banks’ risk behavior is procyclical—excessively optimistic or pessimistic prudential behavior could amplify the business cycle and result in higher risk of bank failure.

24

Countries with real credit growth exceeding 16.8 percent (y-o-y) on an average between 2000 and 2004. However, the banks included in the High Credit Growth countries are not necessarily the ones with the highest average bank-by-bank nominal loan growth because of differences in their inflation rates.

25

Because of inflation, some countries in High Credit Growth and Surroundings overlap.

Decomposing Financial Risks and Vulnerabilities in Eastern Europe
Author: Mr. Rupert D Worrell, Andrea M. Maechler, and Ms. Srobona Mitra