Bulgaria: Selected Issues and Statistical Appendix
Author: International Monetary Fund
Publication Date:
08 Aug 2006
eISBN:
9781451804553
Language:
English
Keywords:
ISCR; CR; Bulgaria; bank; net; EU budget; solvency risk; bank lending; credit ceiling; management corporation; bank soundness; Credit; Bank credit; Credit booms; Foreign banks
Download PDF (533.6 KB)

This paper discusses Bulgaria’s prospects for converging to the living standards of the more advanced members of the European Union (EU). The unfavorable economic environment of the early 1990s and the economic crisis in 1996–97 hurt Bulgaria’s output, employment, and investment. Following the crisis, structural reforms and a sound macroeconomic framework set the stage for a sustained recovery. The structure of the Bulgaria economy has shifted markedly over the last decade, and investment has become the main engine of growth.

Abstract

This paper discusses Bulgaria’s prospects for converging to the living standards of the more advanced members of the European Union (EU). The unfavorable economic environment of the early 1990s and the economic crisis in 1996–97 hurt Bulgaria’s output, employment, and investment. Following the crisis, structural reforms and a sound macroeconomic framework set the stage for a sustained recovery. The structure of the Bulgaria economy has shifted markedly over the last decade, and investment has become the main engine of growth.

IV. Bulgaria—The Implications of Bank Behavior and Credit Measures for Solvency Risk51

A. Introduction

82. Bulgaria’s banking sector began a new phase following the introduction of the currency board in mid-1997. Government borrowing from banks was largely prohibited, and capital controls were abolished. A formal deposit insurance scheme was introduced, and prudential regulations were brought in line with international standards and strictly enforced by the BNB. Confidence in the banking system returned quickly, and bank deposits grew rapidly, although bank lending to the non-government sector took longer to recover. In 2001, the flow of bank lending to the non-government sector reached 3.6 percent of GDP, increasing rapidly to 12 percent of GDP in 2004.

83. The growth of Bulgaria’s banking sector has brought significant benefits, but spillover risks may be rising. The literature is unequivocal about the benefits of financial development for economic growth and development.52 In Bulgaria, financial intermediation has been dominated by the banking sector, which accounts for 90 percent of financial sector assets. Bulgaria’s banking sector has developed faster than the overall economy, and banking assets increased from 35 percent of GDP in 1999 to 78 percent by end 2005. That said, the rapid growth of credit has generated concerns over the stability of the banking sector, a potential overheating of the economy, and a more general underpricing of risk.

84. In 2005, the Bulgarian National Bank (BNB) introduced quantitative limits on credit growth on the back of concerns surrounding the prudential risks associated with the credit boom. The BNB reluctantly introduced these credit ceilings on banks at the end of Q1 2005, after a series of liquidity-draining measures failed to curb the very strong growth in lending (Figure 1). The BNB’s main reason for taking these measures was its concern that too much risk was concentrated in the banking system. This move was supported by the IMF, as concerns mounted about the impact of the credit boom on macroeconomic performance.53 At the time, it was acknowledged that the administrative measures would have only a temporary impact as circumvention would grow over time.

Figure 1.
Figure 1.

Bulgaria: Credit Growth

(In percent, year-on-year)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Source: Bulgarian National Bank and IMF staff calculations.

85. A little over a year after the implementation of the measure, the demand restraining impact of the credit measures has diminished. During the months following the introduction of the measures, there was anecdotal evidence that some companies had difficulty obtaining credit as banks reached their quarterly ceiling. However, since then, non-bank intermediation—mainly leasing and corporate borrowing outside the banking system—as well as cross-border lending have offset the impact on demand from the credit measures. Thus, while the data suggest that lending by banks has slowed, several predictable “distortions” in intermediation have arisen, which are disruptive to the competitiveness and efficiency of financial intermediation.

86. The objective of this paper is to determine the impact of bank- and policy-specific factors on the soundness of the banking sector. Specifically, we examine the extent to which variables such as the nature of bank ownership, banks’ growth strategy and credit measures imposed by the authorities affect the soundness of banks, that is, their solvency risk. Solvency risk is measured in terms of the “distance to default.” The method itself is not original, but this is the first time that it has been applied specifically to the Bulgarian banking sector. The BNB’s policy on the transparency of disclosure in the banking sector, and the availability of frequent (quarterly) data allow for such analysis.

87. The empirical results show that banks’ structure, growth strategy and credit policies have significant impact on their solvency. Specifically, banks with positive-growth strategies tend to face greater solvency risk concerns. Over time, locally-owned private banks have experienced greater solvency risk than their foreign-owned counterparts. This finding is consistent with the BNB’s stress test results. Meanwhile, the introduction of credit ceilings appears to have slowed the general decline in the bank soundness indicator—at least for the time being—although there are insufficient data at this stage for a definitive conclusion on their longer-term effects.

88. The paper is structured as follows. An overview of developments in Bulgaria’s banking sector is presented in section B, highlighting deposit and credit growth, and the macroeconomic benefits and risks associated with these developments; the opening of the banking sector to foreign competition is also discussed. Section C discusses the BNB’s response to the rapid credit growth to the economy, while section D provides an empirical analysis of the impact of behavioral, structural and policy factors in determining banks’ solvency risk. Section E concludes.

B. Banking Sector Developments since 1997

89. Reforms of the banking sector since the crisis has improved transparency and enhanced banks’ decision-making process. Information on outstanding loans of individuals improved in mid 2004 when the threshold of loans that have to be reported to the credit registry was eliminated and it became cheaper for banks to obtain information from the registry. The legal framework for enforcement of creditor rights has also been established. Over time, the economy recovered and confidence in the banking sector returned. Money and quasi-money increased as share of GDP (Figure 2), and remonetization gradually returned to levels similar to other countries in the region (Figures 2 and 3).

Figure 2.
Figure 2.

Bulgaria: Money and Quasi-Money as a Share of GDP, 1995-2005

(In percent)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Sources: Bulgarian National Bank, International Financial Statistics and IMF.
Figure 3.
Figure 3.

Cross-Country Comparison: Money and Quasi-Money as a Share of GDP, 1995 and 2005

(In percent)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Sources: Bulgarian National Bank, International Financial Statistics and IMF.

90. The liberalization of the banking sector and the entry of foreign banks have played an important role in its development. There is currently no restriction on foreign ownership of banks in Bulgaria. Consistent with the open capital account policy, non-residents are free to invest in Bulgarian banks provided they adhere to the prudential provisions that are enforced by the BNB. However, the BNB has not issued new banking licenses since 2003, as it considers the 34 banks that are active in Bulgaria as at end-2005 to be more than adequate to serve the local economy. That said, upon Bulgaria’s accession to the EU, banks from EU-member states will be entitled to open branches in Bulgaria subject only to notification, not the approval, of the BNB.

Participants in the Banking Sector

91. Over the years, state-owned banks’ share of total banking assets has been sharply reduced through restructuring and privatization. Banks privatized during the 1997-2003 period had a total market share of some 70 percent in 1999. Most privatized banks were sold to foreign investors in a key step towards the opening of the local banking sector to international participants (Table 1).54

Table 1.

Bulgaria: Key Privatizations of Commercial Banks since 1997

article image
Sources: Barisitz (2002), Bulgarian National Bank data and IMF staff calculations.

92. Foreign banks dominate the Bulgarian banking landscape. The share of banking assets held by foreign-owned subsidiaries increased from 38 percent in 1999 to 72 percent in 2005 (Figure 4). That said, these foreign investors had revamped the management of the purchased banks, and many experienced a loss in market share in the process.55 In contrast, locally-owned private banks have been growing faster than the banking sector as a whole since 2002. Their share of total bank lending to the non-government sector increased to 21 percent in 2005, from 15 percent in 1999, while their share of total deposits from non-financial institutions rose to 13.4 percent in 2005, up from 5.4 percent in 1999. As of end 2005, 15 foreign-owned subsidiaries managed 67 percent of bank assets; 12 locally-owned private banks managed 26 percent bank assets; foreign branches managed just over 5 percent of bank assets; and two small state owned banks managed less than 2 percent of bank assets.

Figure 4.
Figure 4.

Bulgaria: Share of Banking Assets by Bank Ownership Category, 1999-2005

(In percent of total assets)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Source: Bulgarian National Bank.

93. Ownership of banks in Bulgaria may be classified into four different categories, namely, state-owned, Bulgarian privately-owned, foreign-owned subsidiaries and foreign bank branches. Foreign-owned subsidiaries and Bulgarian privately-owned private banks compete directly in the same market, but the foreign owned banks have better access to best-practice business and risk management, banking infrastructure and capital, compared to locally-owned banks. Branches of foreign banks and state-owned banks serve specific market segments. Foreign branches have different legal status relative to foreign-owned subsidiaries, and tend to operate more as investment banks, while state-owned banks do not operate fully on a commercial basis. Meanwhile, state-owned banks slated for privatization are temporarily managed by state-owned bank management corporations.

94. As a whole, the banking sector has been reaping the benefits of earlier restructuring efforts. Despite solid economic growth during the 1998-2001 period, the flow of bank loans to the economy remained at around 2-4 percent of GDP during this time. However, bank lending started to grow rapidly in 2002, and over the 2003-04 period, credit increased by some 50 percent per year. Indeed, bank lending grew from 6 percent of GDP in 2002 to 12 percent of GDP in 2004.

The Credit Boom

95. Prior to the credit boom, banks had accumulated significant deposits, as confidence within the banking sector slowly recovered. These deposits were subsequently used to finance lending to non-financial institutions in Bulgaria, at the onset of the “boom” in 2003.56 This was particularly so for the locally-owned private banks, which are more dependent on deposits to finance lending, compared to their foreign-owned counterparts. Indeed, Bulgarian-owned banks have made significant efforts to attract deposits from the public, by offering higher deposit interest rates. Their higher costs of borrowing, as well as their more limited sources of funding has seen Bulgarian-owned banks consistently report lower pre-tax profits as share of assets since the start of the credit boom in 2003 (Figure 5).

Figure 5.
Figure 5.

Bulgaria: Average Annual Return on Assets by Bank Category, 1999-2005

(In percent)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Sources: Bulgarian National Bank and IMF staff calculations.

96. The drawdown of net foreign assets began in 2002 (Figure 6). The stock of commercial banks’ net foreign assets dropped from a peak of 15 percent of GDP in September 2000 to become negative (-1 percent) as at end-2004. By the end of 2005, commercial banks were contributing more than a quarter to Bulgaria’s private foreign debt.

Figure 6.
Figure 6.

Bulgaria: Financial Flows as a Share of GDP, 2000-05

(In percent)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Sources: Bulgarian National Bank.

97. Banks adopted different growth strategies during the credit boom period. During this time, several trends emerged (Table A.1). The share of bank capital in total assets declined much faster for Bulgarian-owned banks than for subsidiaries of foreign banks. Branches of foreign banks and state-owned banks actually increased capital as a share of assets. Not surprisingly, the capital-to-asset ratio declined particularly rapidly for banks which were growing market share, compared to competitors with no growth in market share.

Table A.1.

Bulgaria: Bank Capital as a Share of Total Assets, 1999-2005

(In percent)

article image
Sources: Bulgarian National Bank and IMF staff calculations.

98. The deposit-to-asset ratio remained relatively stable for locally-owned banks, but declined for the others (Table A.2). Deposits have become less important for the subsidiaries of foreign banks as these banks have access to alternative sources of financing. Locally-owned private banks are more constrained in accessing alternative sources of financing and hence are forced to expend greater effort to attract deposits locally.

Table A.2.

Deposits by Non-Financial Institutions and Other Clients, 1999-2005

(In percent of total assets)

article image
Sources: Bulgarian National Bank and IMF staff calculations.

99. The share of credit to the non-government sector in total assets has increased over time, but varies by bank category (Table A.3). This is not unusual, given that the rising share of credit to the non-government sector is common during a credit boom period when banks reallocate their assets. The loan-to-asset ratio is somewhat lower for locally-owned private banks than for foreign-owned banks. It suggests that Bulgarian-owned banks may keep more of their assets liquid so as to be able to absorb any shock to the banking system, while foreign owned banks may have less need to do so, given that they be have access to the resources of parent institutions.

Table A.3.

Bulgaria: Credit to the Non-Financial Non-Government Sector as a Share of Total Assets, 1999-2005

(In percent)

article image
Sources: Bulgarian National Bank and IMF staff calculations.

100. Initially, loans to the corporate sector represented the most dynamic growth segment, but the focus has switched to household lending during the credit boom years. Consumer loans started growing rapidly in 2003, but mortgage loans have become the fastest growing category since 2004. Total credit increased by 32 percent during 2005—household credit grew by almost 60 percent, while credit to the non-financial corporate sector increased by 23 percent (Figure 7). Hence, the share of household credit in the total has expanded rapidly.

Figure 7.
Figure 7.

Bulgaria: Total Bank Credit Flow to the Non-Government Sector, 2000-2005

(In percent of total credit flows)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Source: Bulgarian National Bank.

101. Credit quality had remained relatively high up until recently, but appears to have deteriorated (Table 2). Credit quality indicators tend to lag actual developments, and classified loans tend to increase faster than loan growth during a period of deceleration. However, even adjusted for the slowdown, the recent pickup is a reason for vigilance as classified loans grew by 56 percent, higher than the loan growth rate in recent years. The rise in classified loans has clearly accelerated—during the same period a year ago, classified loans rose by 30 percent while credit grew by 50 percent. On the other hand, it should also be noted that the changes in classification and provisioning rules, introduced in 2004, likely accounted for part of this increase.

Table 2.

Bulgaria: Changes in Asset Quality, 1997-2005

(In percent of total loans)

article image
Source: Bulgarian National Bank Annual Reports.Note: Standard exposure is in effect when principal and interest are 30 days or less overdue; watch exposure is in effect when principal and interest are past due 31-60 days,. Substandard exposure is in effect when principal and interest have been past due 61-90 days. Loss or nonperforming exposure is in effect when principal and interest have been past due over 90.

C. The Implementation of Credit Measures

102. Initially, the BNB welcomed the recovery in bank lending, but then became increasingly concerned with the associated risks of a credit boom. While the BNB was mainly concerned about the prudential risk, Fund staff highlighted both prudential and macro risks.57 In the course of 2004, the BNB implemented various measures to withdraw liquidity from the banking system, including raising the reserve requirement, moving government deposits and funds of the Deposit Insurance Fund from commercial banks to the BNB (Table 3). However, the measures were largely ineffective as banks were able to freely borrow abroad, given the open capital account. Individual banks were keen to maintain or increase their market share, and were loathe to take the lead in curbing credit to the private sector.

Table 3.

Bulgaria: Impact of Liquidity Reducing Measures, 2004

(In millions of Bulgarian leva)

article image
Sources: Bulgarian National Bank and IMF staff calculations.

103. In early-2005, the BNB attempted to reduce the aggregate credit expansion to the non-government sector in a further effort to contain risks to the stability of the financial sector. The aim was to limit credit growth to 30 percent, from 49 percent in 2004.58 In February 2005, the BNB announced the implementation of credit ceilings to limit the growth of credit to the non-government sector. Banks were allowed to expand credit by 6 percent per quarter, taking end-March 2005 as the base period.59 Bank credit in excess of this limit would be subject to a marginal reserve requirement by the BNB, of 200 percent of the excess.

104. The introduction of the credit ceilings in March 2005 created an incentive for all banks to expand their credit portfolio. The timing of the announcement provided banks with an opportunity to increase credit prior to end-March so that future credit growth would be calculated from a higher base. That said, the extent of banks’ reaction differed. For instance, credit flows were much more volatile for the small foreign bank branches that operated as investment banks than for the much larger foreign-owned subsidiaries around this time. Indeed, foreign-owned subsidiaries and locally-owned private banks also grew their credit base, but less as a percentage of outstanding loans. (Figure 8).

Figure 8.
Figure 8.

Bulgaria: Quarterly Credit Growth, 2003-05

(In percent)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Sources: Bulgarian National Bank data and IMF staff calculations.

105. While some of the credit flows could be directly attributable to increased advertising activity by commercial banks, part of the credits were “fictitious.” One strategy was to “recycle” funds—banks extended low interest rate loans to companies that deposited the money with them. Another form “window dressing” was the extension of credit to non-residents by way of taking over loans extended by overseas parent banks. The monetary data show that the flow of credit during the month of March alone was some 3.3 billion leva (7.8 percent of 2005 GDP).

106. The BNB modified its credit expansion limits on April 21, 2005, taking into account the strategies adopted by banks to increase lending. It was decreed that quarterly credit growth would be measured by comparing the daily average stock of credit with a base that excluded the artificial end-March credit boom. As banks had already made significant loans during March 2005, it was decided that during Q2 2005, a 10 percent rate of growth would be applied instead of the “artificial” end-March 2005 base. Following these new measures, much of the credit flow recorded during the March quarter was reversed in the following quarter.

107. The credit measures have been credited with slowing the credit expansion to 32.3 percent (11 percent of GDP) in 2005, from 49 percent (12 percent of GDP) in 2004. In October 2005, the BNB extended the existing credit expansion limits of 6 percent of the adjusted end-March 2005 base per quarterly, with the objective of further slowing banks’ credit expansion to 17.5 percent (7.1 percent of GDP) in 2006. Other additional measures have also been implemented in an attempt to prevent banks from circumventing the credit restraint measures (see Table 4).

Table 4.

Additional Credit Measures Taken by the BNB Following the Introduction of Credit Limits in March 2005

article image
Source: Bulgarian National Bank.

108. Despite the decline in credit flows from the banking sector, total domestic financial flows to the non-government sector has continued to increase modestly (Table 5). Non-bank financial institutions, notably leasing companies, have expanded rapidly. Companies have also been able to issue bonds in the local market and to borrow from abroad, frequently through foreign branches and foreign-owned subsidiaries in Bulgaria. According to BNB data, corporates were the main drivers of the growth in the leasing market during 2005. In 2006, the non-bank financial sector continued to grow rapidly and banks increasingly moved loans off the balance sheet to locally-registered special purpose vehicles.

Table 5.

Bulgaria: Financial Flows, 2004-05

(In percent of GDP)

article image
Source: Bulgarian authorities.

IMF staff estimates.

D. Determining the Soundness of the Banking Sector

109. In line with our earlier discussion, we quantify the extent to which bank-specific factors and credit ceilings affect banks’ solvency risk in this section. Specifically, the experience during the credit boom period suggests that: (i) the type of ownership affects banks’ credit policies, and their ability to access funding differ; (ii) within each bank category, some—but not all—banks pursue positive growth policies; (iii) the imposition of credit ceilings induced banks to behave differently to circumvent these measures. The results could provide some insights into the risks for the financial sector, and potentially provide quantitative support for future banking policies.

Data and Method

110. The effects of various structural, behavioral and policy factors on bank soundness is assessed using pooled OLS.60 Our study adds to existing work, which examines at the vulnerability of financial institutions and financial systems to a variety of macroeconomic and prudential variables.61 Maechler, Mitra and Worrell (2005) test for the impact of bank-specific and market risk factors, risk-mitigating (supervisory framework) indicators and macroeconomic variables on banking sector stability in the new EU member states and surrounding countries. In turn, we focus on the impact of bank ownership, credit growth strategy and administrative credit measures on banks’ distance-to-default, or insolvency risk.

111 A commonly used measure of bank soundness is the z-score. This statistic shows a bank’s risk of insolvency or distance to default, that is, the probability that losses (negative profits) exceed equity. The generic form of the z-score is defined as follows:

(1)zi≡μ^i+kiσ^i,
A04lev3sec3

where μ is the average return on assets (ROA) for bank i, k is the time average equity to asset ratio and σ is the sample estimate standard deviation of the ROA (which proxies for the volatility in returns). In other words, the Ζ statistic measures the number of standard deviations a return realization would have to fall in order to deplete equity.62 Thus, a higher Ζ-score corresponds to a lower probability of insolvency risk.

112. In this paper, we calculate the Ζ-scores for individual banks at quarterly intervals. These quarterly Ζ-scores would enable an assessment of the changes in the solvency risk of banks over the sample period. The total assets and Ζ-score components are derived from quarterly financial statements of individual commercial banks in the Bulgarian banking system.63 The data are publicly available from the BNB on a quarterly basis. The sample period covers Q4 1999 to Q4 2005.64

113 A visual observation indicates that, in the aggregate, foreign-owned subsidiaries have higher Ζ-scores than locally-owned private banks (Figure 9). This is not surprising given that locally-owned private banks have to pay higher interest rates to attract deposits, compared to foreign-owned subsidiaries which are able to access cheaper overseas funding more easily.65 Locally-owned private banks also pay particularly high interest rates on inter-bank loans.

Figure 9.
Figure 9.

Bulgaria: Solvency Risk by Bank Ownership Category, 1999-2005

(In z-scores)

Citation: IMF Staff Country Reports 2006, 299; 10.5089/9781451804553.002.A004

Source: IMF staff calculations.

114. We incorporate the total assets variable into our model, to capture the role of bank size in influencing its solvency risk. Existing studies show that bank size matters in determining the bank soundness. As an example, De Nicolo (2000) finds that banks’ solvency risk increases with size (defined as total assets), which he attributes to higher risk-taking by medium-to-large sized banks, which more than offset diversification benefits.

115. We also include three separate clusters of dummy variables in our model. These dummies represent aspects of bank behavior and banking policies, discussed previously:

  • bank ownership category;66

  • banks’ growth strategy (increased market share versus unchanged or decreased market share); and

  • banks’ lending behavior during Q1 2005, in the lead-up to the introduction of credit ceilings at the end of the quarter, and thereafter.

We also include seasonal dummies, given the quarterly nature of the data.

116. The pooled OLS model is defined as follows:

1n(zi,t)=βASSET1n(ASSETSi,t)+Σk=14βSEAS,kSEASi,t+Σk=14βOWN,kOWNi,t(3)+βGROW,kGROWi,t+Σk=12βCC,kCEILi,t,
A04lev3sec3

where ASSETSi,t represents the total assets, denominated in leva, of bank i at time t; and SEASi,t, OWNi,t, GROWi,t and CEILi,t represent dummy clusters, as defined in Table 6, corresponding to bank i at time t67 The Ζ-scores and asset variables are transformed using natural logarithms, to reduce the effects of outliers.68

Table 6.

Definitions of Dummy Variables

article image
Notes:

The control variable for this cluster is foreign subsidiaries.

The control variable for this cluster is the group of banks with decreased share of the loan market, or which show no growth in market share. A bank is defined as having a positive growth strategy if its share of the loan market increased over the four quarters to period t.

The control variable for this cluster is the pre-credit ceilings period, up to Q4 2004.

117. As a next step, we focus our test on the risks for individual bank categories, such that:

1n(zi,t)=βASSETS1n(ASSETSi,t)+Σk=14βSEAS,kSEASi,t+βGROW,kGROWi,t(4)+Σk=12βCC,kCEILi,t.
A04lev3sec3

We run both models (3) and (4) over the entire sample period of Q4 1999 to Q4 2005; we also test the models over the credit boom period of Q1 2003 to Q4 2005, for robustness.

Results

118. The regression results show clear trends within the different variable clusters being considered.69 We test models (3) and (4) for the full sample period of Q4 1999-Q4 2005.70 The results are presented in Tables 7 and 8.

Table 7.

Equation (3): Results for Pooled OLS Regression over the Sample Period, Q4 1999 to Q4 2005

article image
Source: IMF staff calculations.Note: The significance of the t-statistics are assessed against the 0.1, 1, 5 and 10 percent levels of significance. A notation of “--” means that the t-statistic is not significant at the 10 percent level.
Table 8:

Equation (4): Results for Pooled OLS Regression over Sample Period, Q1 2003 to Q4 2005

article image
Source: IMF staff calculations.Note: The significance of the t-statistics are assessed against the 0.1, 1, 5 and 10 percent levels of significance. A notation of “--” means that the t-statistic is not significant at the 10 percent level.
Seasonality

119. The seasonal dummy coefficients are very significant for all four quarters. As we would expect, the quarterly nature of the data supports the existence of strong seasonal effects.

Bank Size

120. Banks’ size appears to have little influence on their solvency risk. In other words, banks’ solvency risk within the Bulgarian banking sector appears to be independent of their ability to achieve greater diversification, or economies of scale, of operations. This is in contrast to existing evidence that bank size significantly affects solvency risk.

Type of Bank Ownership

121. Banks in different categories exhibit significantly different levels of solvency risk, on average, relative to the foreign-owned subsidiaries (control variable). State-owned banks have significantly higher Ζ-scores; foreign branches and locally-owned private banks are significantly more negative, that is, the solvency risk for these institutions is higher. The results are robust when tested over the credit boom period. The results are also consistent with the BNB’s stress tests, which show that locally-owned private banks have consistently been more vulnerable to shocks than foreign-owned subsidiaries.

Growth strategy

122. Banks with positive loan growth strategies have higher solvency risk (lower Ζ-scores), on average. The results are consistent over both, the entire sample period and the credit boom period only. Interesting differences in the relationship between growth strategy and solvency risk show up when the tests are run on an ownership basis, per equation (4). Positive-growth locally-owned private banks and foreign-owned subsidiaries tend to exhibit higher solvency risk (lower Ζ-scores) relative to the negative- and no-growth banks in their respective categories (Table 8).

Credit ceilings

123. Solvency risk for the banking sector as a whole remained statistically unchanged in Q1 2005 and in subsequent quarters, relative to the pre-2005 period. The observed decline in the Ζ-scores of locally-owned private banks slowed post-ceiling (Figure 9), and was only significantly different from the pre-ceiling period at the 10 percent level. However, these results would be more reliable once more data becomes available, as time progresses and the ability of borrowers to continue to service their loans becomes clearer.

E. Conclusion

124. Following a period of recovery from the crisis in 1996-97, the rapid growth of credit in recent years has generated concerns over the stability of the banking sector in Bulgaria. In this context, this paper analyzes the soundness of banks within the financial sector, using a “distance to default” model to determine the impact of behavioral, structural and policy factors on the solvency risks of banks. Our empirical findings suggest that:

  • Banks with positive loan growth strategies tend to face greater solvency risk concerns, as a matter of course. This particular feature holds true irrespective of whether the strategy is implemented in a credit boom environment, or otherwise. This illustrates the legitimate concerns of the BNB at the time the credit measures were introduced.

  • Locally-owned private banks tend to exhibit higher solvency risk than their foreign-owned counterparts. This has proved to be a consistent trend since the start of the credit boom phase in 2003.

  • The introduction of credit ceilings appears to have slowed the decline in bank soundness. This is especially true for the locally-owned private banks. However, given the limited length of the business cycle and the relative immaturity of the banking sector, the “equilibrium” Ζ-scores for the banking sector remains unclear at this stage.71

125. These findings support the importance of prudential measures, to enhance the soundness of the financial sector. Market participants are reportedly resorting to a myriad of (legal) circumvention techniques around the credit ceilings. For instance, banks are said to assign loans to their foreign parents, packaging loans through off-balance-sheet special purpose vehicles. Also of some concern is that the credit measures may have catalyzed the shift in credit from the banking sector into less well-supervised segments of the financial sector, with domestic non-bank financing growing rapidly. As such, prudential measures help to ensure that risk-taking by banks are “correctly” priced and appropriately managed.

126. The banking system is well-capitalized and profitable, but inherent risks exist in an environment of rapid credit growth. While it could be argued that foreign bank branches and subsidiaries may be able to rely on the support of their parents in the event of financial sector turbulence, this is not the case for locally-owned private banks. Furthermore, although the systemic risk posed by the rapidly growing non-bank financial sector remains small at this stage, the issue of reputation risk remains. In other words, the failure of any financial institution could result in contagion across the sector, especially given the increasingly close links between banks and their non-bank financial counterparts in Bulgaria.

References

  • Barisitz, Stephan, 2002, “Les Systèmes Bancaires Bulgare et Roumain Défis de la Transformation”, Le Courrier de Pays de l’Est, Vol. 37, No. 1021, January, pp. 42–58.

    • Search Google Scholar
    • Export Citation

  • Bulgarian National Bank (various years), Annual Report (Sofia: Bulgarian National Bank).

  • Cihak, Martin (2004), “The Determinants of Lending Rates and Domestic Spreads in Croatia,” IMF Country Report No. 04/251 (Washington, International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • De Nicolo, Gianni, 2000, “Size, Charter Value and Risk in Banking: An International Perspective,” Federal Reserve International Finance Discussion Paper No. 689 (Washington, December).

    • Search Google Scholar
    • Export Citation

  • Detragiache, Enrica and Poonam Gupta (2004), “Foreign Bank in Emerging Market Crises: Evidence from Malaysia,” IMF Working Paper No. 04/129 (Washington, International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Detragiache, Enrica, Poonam Gupta and Thierry Tressel, “Foreign Bank in Poor Countries: Theory and Evidence,” IMF Working Paper No. 04/129 (Washington, International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Duenwald, Christoph, Nikolay Gueorguiev and Andrea Schaechter, 2005, “Too much of a Good Thing? Credit Booms in Transition Economies: The Cases of Bulgaria, Romania and Ukraine,” IMF Working Paper No. 05/128 (Washington, International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • European Commission, 2003, The Financial Sector in Bulgaria, European Economy Enlargement Papers No.18 (Brussels, September).

  • Frankel, Jeffrey A. and Andrew K. Rose, 1995, “A Panel Project on Purchasing Power Parity: Mean Reversion within and between Countries,” Journal of International Economics, Vol. 40, Nos. 1-2, pp. 209–24.

    • Search Google Scholar
    • Export Citation

  • Goldsmith, Raymond, 1969, Financial Structure and Development (New Haven, CT: Yale University Press).

  • Hakkio, Craig S., 1984, “A Re-examination of Purchasing Power Parity: A Multi-Country and Multi-Period Study, Journal of International Economics, Vol. 17, Nos. 3-4, pp. 265–77.

    • Search Google Scholar
    • Export Citation

  • International Monetary Fund, 2005, Bulgaria Country Report, No. 05/169 (Washington: IMF).

  • Kennedy, Peter, 1998, A Guide to Econometrics, Fourth Edition (Cambridge, Massachusetts: The MIT Press).

  • King, Robert and Ross Levine, 1993, “Finance and Growth: Schumpeter Might be Right,” Quarterly Journal of Economics, Vol. 108, No. 3, pp. 717–37.

    • Search Google Scholar
    • Export Citation

  • Miller, Jeffrey and Stefan Petranov, 2001, “The Financial System in the Bulgarian Economy”, Bulgarian National Bank Discussion Paper No. 19/2001 (Sophia: Bulgarian National Bank).

    • Search Google Scholar
    • Export Citation

  • Maechler, Andrea, Srobona Mitra and DeLisle Worrell, 2006, “Financial Stability and the Convergence Process in Europe,” IMF Working Paper, forthcoming (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Rajan, Raghuram G. and Luigi Zingales, 1998, “Financial Dependence and Growth,” The American Economic Review, Vol. 88, No. 3, pp. 559–86.

    • Search Google Scholar
    • Export Citation

  • White, Halbert, 1980, “A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity,” Econometrica, Vol. 48 (May), pp. 817–38.

    • Search Google Scholar
    • Export Citation

  • Worrell, DeLisle, 2004, “Quantitative Assessment of the Financial Sector: An Integrated Approach,” IMF Working Paper No. 04/153 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
51

Prepared by Johannes Herderschee and Li Lian Ong.

53

See empirical evidence in section D.

54

The Bulgarian Bank Restructuring Company owned the banks prior to privatization and managed the sales.

55

For example, Bulbank’s share of total banking assets declined from over 25 percent 1999 to under 10 percent in 2003.

56

Although the strengthening in credit growth started during the course of 2002, we define the credit boom period is as having begun from Q1 2003, as the quality of published financial statements improved with the presentation of more detailed data.

58

Assuming Fund staff GDP projection at the time, this would be equivalent to a credit flow of 10 percent of GDP, down from 12 percent of GDP in 2004.

59

However, this restriction would not apply to banks where the ratio of credits (including risk-weighted off-balance-sheet items) minus capital to total deposits—other than those by other financial institutions—was below 60 per cent. Eight banks fell into this category as at end-September 2005.

60

Standard errors are adjusted for heteroskedasticity, using the White (1980) correction technique.

61

Worrell (2004) provides a survey of existing studies.

62

Normality of returns is assumed.

63

Ideally, market values of equity, assets and liabilities should be used in the calculations. In the absence of reliable market data, however, accounting values are used in this instance. The ROA is calculated as profit before foreign exchange revaluation, extraordinary and tax items. Typically, foreign exchange revaluation should be included in the profit item, as it is usually part of a bank’s normal operations. In this case, however, the breakdown of the components was not available prior to Q1 2003.

64

Existing studies generally apply annual data, while Cihak (2004) uses unpublished data of listed banks in Croatia. The panel regression approach is appropriate in that it reduces the amount of time-series data required, but still provides sufficient data for powerful tests; moreover, it exploits any cross-sectional variation in the data (see, for example, Hakkio, 1984; Frenkel and Rose, 1995)

65

The sharp rise in the Ζ-scores of foreign bank branches in 2005 is attributable to the substantial funding received by one of the 6 branches in this group from its parent.

66

The inclusion of state-owned banks, foreign branches and bank management corporations are for completeness; the caveat for the reliability of associated results is their relatively small sample sizes.

67

To enable a clearer interpretation of the resulting dummy coefficients, we assign a dummy variable to all four quarters (SEAS), while constraining the intercept term to zero.

68

The original Ζ-scores are all transformed by adding a constant to ensure that they are all positive, prior to the application of the natural logarithm. Interestingly, the majority of negative Ζ-scores tend to correspond with the branches of foreign banks. Not surprisingly, non-branch banks with negative Ζ-scores do not “survive” over time.

69

The results for equation (3) over the period up to Q4 2004 show that the banks increasing their share of the loan market up to that point had experienced a significant increase in solvency, perhaps justifying the implementation of credit ceilings by the authorities in an attempt to improve the soundness of the banking sector.

70

The pooled OLS results show very high adjusted R-squared coefficients, of greater than 90 percent. This is common for models containing dummy variables designed to capture structural sifts or seasonal factors, as these dummies may play a key role in generating the high R-squared figures (see Kennedy, 1998).

71

The authorities specifically note that the decline in the Ζ-score levels are not of particular concern at this stage, given that the banking sector as a whole is largely considered to have been over-capitalized.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Bulgaria: Selected Issues and Statistical Appendix
Author: International Monetary Fund
Volume/Issue:
Volume 2006: Issue 299
Publisher:
International Monetary Fund
ISBN:
9781451804553
ISSN:
1934-7685
Pages:
148
DOI:
https://doi.org/10.5089/9781451804553.002
  • View in gallery

    Bulgaria: Credit Growth

    (In percent, year-on-year)

  • View in gallery

    Bulgaria: Money and Quasi-Money as a Share of GDP, 1995-2005

    (In percent)

  • View in gallery

    Cross-Country Comparison: Money and Quasi-Money as a Share of GDP, 1995 and 2005

    (In percent)

  • View in gallery

    Bulgaria: Share of Banking Assets by Bank Ownership Category, 1999-2005

    (In percent of total assets)

  • View in gallery

    Bulgaria: Average Annual Return on Assets by Bank Category, 1999-2005

    (In percent)

  • View in gallery

    Bulgaria: Financial Flows as a Share of GDP, 2000-05

    (In percent)

  • View in gallery

    Bulgaria: Total Bank Credit Flow to the Non-Government Sector, 2000-2005

    (In percent of total credit flows)

  • View in gallery

    Bulgaria: Quarterly Credit Growth, 2003-05

    (In percent)

  • View in gallery

    Bulgaria: Solvency Risk by Bank Ownership Category, 1999-2005

    (In z-scores)

Bulgaria: Credit Growth

(In percent, year-on-year)

Bulgaria: Money and Quasi-Money as a Share of GDP, 1995-2005

(In percent)

Cross-Country Comparison: Money and Quasi-Money as a Share of GDP, 1995 and 2005

(In percent)

Bulgaria: Share of Banking Assets by Bank Ownership Category, 1999-2005

(In percent of total assets)

Bulgaria: Average Annual Return on Assets by Bank Category, 1999-2005

(In percent)

Bulgaria: Financial Flows as a Share of GDP, 2000-05

(In percent)

Bulgaria: Total Bank Credit Flow to the Non-Government Sector, 2000-2005

(In percent of total credit flows)

Bulgaria: Quarterly Credit Growth, 2003-05

(In percent)

Bulgaria: Solvency Risk by Bank Ownership Category, 1999-2005

(In z-scores)