Bulgaria: Selected Issues and Statistical Appendix

This Selected Issues paper and Statistical Appendix analyzes the reasons behind the relatively low rates of savings in Bulgaria and prospects for their evolution over the medium term. The paper argues that low saving rates largely reflect the current stage of transition—characterized by still low income levels, incomplete structural reforms, the memories of the financial and banking crises of 1996–97, and an adverse demographic structure. An analysis of prospective saving rates indicates that as the transition process advances, saving rates may increase by 5 percentage points over the medium term.


This Selected Issues paper and Statistical Appendix analyzes the reasons behind the relatively low rates of savings in Bulgaria and prospects for their evolution over the medium term. The paper argues that low saving rates largely reflect the current stage of transition—characterized by still low income levels, incomplete structural reforms, the memories of the financial and banking crises of 1996–97, and an adverse demographic structure. An analysis of prospective saving rates indicates that as the transition process advances, saving rates may increase by 5 percentage points over the medium term.

III. Bulgaria’s Banking Sector 1

A. Introduction and Background

1. The banking sector has strengthened considerably since the 1996–97 financial and exchange rate crisis, but intermediation remains low. The present scope and quality of banking sector intermediation in Bulgaria still carries the legacy of the crisis, which was largely triggered by the unsound lending practices and weak governance and supervision of banks during the early transition years. Strong measures to restore financial stability, centered on the introduction of a currency board arrangement (CBA) and the strengthening of banking regulation and supervision, have dramatically improved the condition of the sector. The banking sector landscape has been transformed through substantial consolidation, including the resolution of 15 banks in 1996, successful privatization of the dominant state banks, and ongoing restructuring of the banks that survived the crisis. As a result, the Bulgarian banking sector is now generally well regulated and supervised, highly capitalized and liquid, and profitable. However, the level and composition of monetary aggregates suggest that the public still lacks full confidence in the banking system and that banks are still cautious in expanding lending to the private sector. This chapter discusses the current structure, performance, and resilience to shocks of the Bulgarian banking sector against the background of the 1996–97 crisis and key changes in banking sector behavior since the crisis.

2. Widespread insolvency and weak governance and supervision in the banking sector were major factors underlying the financial crisis of the mid-1990s. In the early transition years, the banking sector continued to be dominated by state-owned institutions which, on a massive scale, extended “soft loans” to loss-making state enterprises. These misguided lending policies resulted in a surge in bad loans—around 75 percent of loans were nonperforming by 1995—and widespread bank insolvency. The refinancing needs of insolvent banks intensified in the second half of 1995, prompting the Bulgarian National Bank (BNB) to provide large amounts of uncollateralized refinancing. A growing awareness of widespread insolvency in the banking system and doubts about the ability of the government to meet its debt service obligations weakened confidence in the currency. In the first half of 1996, currency outside the banks declined by more than 20 percent in real terms and lev deposits by more than 30 percent, respectively. Faced with bank runs, the BNB continued to extend unsecured loans to ailing banks.

3. In the period May to December 1996, the BNB made several unsuccessful attempts to restore confidence in the banking system and the currency. Successive stabilization packages tried to address weaknesses in the banking sector, including through support for viable institutions, and to raise interest rates to positive levels in real terms. The authorities were, however, unable to implement the policies as envisaged, the external financial position remained under pressure as foreign financing did not resume, and the depreciation of the currency accelerated. Following the resignation of the government in December 1996, a full-blown financial and exchange rate crisis erupted, underscoring the need for a fundamental shift in policy.

4. The new macroeconomic stabilization program, centered on a CBA, restricted the financial relations between the BNB and the banking system. Under the CBA, the BNB was reorganized into two principal financial departments. The Issue Department (or currency board) holds all the BNB’s monetary liabilities—to be covered at all times by foreign exchange assets and gold—and the deposit of the Banking Department, which enables the latter to function as a lender of last resort. The non-monetary liabilities of the BNB on behalf of the government were restructured, while the BNB ceased open-market operations and eliminated the repurchase facility. The conversion rate for the peg—leva 1,000 per DM—was set at a level that reconciled the objectives of preserving Bulgaria’s external competitiveness while more than fully covering the CBA’s monetary liabilities. At the selected rate, the excess cover—corresponding to the size of the deposit of the Banking Department with the Issue Department—was made available to support the banking system, under strict conditions (credit can be extended in case of systemic threat, but only to solvent banks experiencing illiquidity and against high grade collateral for a maximum of three months). These resources amounted to somewhat more than US$300 million, covering somewhat less than 20 percent of total deposits (including foreign exchange-denominated accounts), but around half of the lev-denominated deposits and all of the deposits at the State Savings Bank, which was formerly the monopoly in household deposit taking.

5. The introduction of the CBA in July 1997 resulted in a rapid return in confidence and initial remonetization. In the second half of 1997, interest rates fell back to low levels in nominal terms and turned positive again in real terms, credit to the non-government sector began to recover, and real money aggregates rebounded significantly. As a result of the rapid remonetization, the increase in the resources available to support the banking system—the excess cover at the Issue Department—could not keep pace with that in total deposits. The cover ratio was brought back to its initial level in September 1998, following the retention by the BNB of an IMF purchase and the accrual of net income at the BNB. The improvement in monetary conditions also allowed the BNB to ease access to reserve holdings. In April 1998, banks were given access to 100 percent of required reserves on any given day, up from 15 percent when the CBA was established.

6. Following the initial remonetization, monetary and credit expansion continued in 1998–99, but at a slower pace. As a pickup in real activity and continued low inflation boosted money demand, broad money rose by 8.5 percent in real terms in 1998, mainly on account of an increase in currency in circulation. It grew by a more moderate 4.9 percent in 1999 owing to a sizable increase in lev-denominated time and savings deposits. Credit to the non-government sector picked up modestly in this two-year period, by a combined 7.3 percent in real terms. However, this increase masked divergent developments. Net claims on nonfinancial state enterprises declined sharply, in part owing to the privatization of some enterprises. As banks redirected new lending to privatized and new private sector enterprises, claims on such enterprises showed steady growth in both years. Claims on the household sector, finally, grew very rapidly in 1998 as a result of a robust expansion in lending activities by the State Savings Bank, but increased only modestly the following year. A better management of disposable funds allowed banks to lower the level of excess reserves. Reflecting the still conservative lending policies, however, the reserves-to-deposit ratio remained well in excess of the minimum required 11 percent, at around 15 percent.

7. Money and credit growth accelerated considerably in 2001 from moderate levels in 2000, but hovered during the first half of 2002. M2 increased in real terms by more than 5 percent in 2000, having lev currency as the fastest growing component, a sign of a waxing confidence in the domestic currency. In 2001, M2 increased by more than 28 percent in real terms, and the growth in foreign currency deposits far outpaced that of lev money, largely reflecting the one-off effect of the deposit of cash holdings of former EMU currencies in the banking system. With this effect having run its course, and in line with usual seasonal patterns, monetary aggregates were broadly unchanged in real terms in the first half of 2002 relative to end-2001 levels. On the banking sector’s asset side, claims on private sector enterprises and households in real terms increased only marginally in 2000, as the additional resources freed by the reduction in the minimum reserve requirement from 11 to 8 percent in July 2000 were mainly channeled into additional placements abroad. Credit growth was, however, very robust in 2001 as claims on enterprises and household in real terms increased by more than 15 percent and more than 38 percent, respectively, with the solid growth rate in part still reflecting the low base. The recent credit expansion was supported both by banks’ increased focus on retail banking, a historically underdeveloped market segment, and by the strong deposit growth on the funding side and the fall in interest rates on alternative placements abroad. In part as a result of the slowdown in deposit growth, credit growth eased in the first half of 2002, but there continues to be an underlying trend of a steady expansion in credit to private sector enterprises and households.

B. Structure and ownership

8. The banking sector remains small relative to the size of the Bulgarian economy but dominates the rest of the financial sector. At the end of March 2002, the assets of the 35 licensed banks amounted to close to 40 percent of projected 2002 GDP, lending to the private sector to 12 percent of GDP, and deposits from non-financial institutions to 27 percent of GDP. These ratios are low by the standards of the other EU accession countries, but not out of line with those in other transition countries at a comparable stage of the transition. Moreover, banks are well developed relative to the other segments of the financial sector (agricultural credit unions, securities firms, finance companies, general and life insurance companies, pension funds, and other nonbank financial institutions) and account for more than 90 percent of the assets of the financial sector as a whole (Table 1). With only a fraction of the 35 banks serving domestic companies and the general public, there is no indication that the number of such banks is excessive relative to the population size.

Table 1.

Bulgaria: Structure of Financial Sector as of March 2002

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

Projected 2002 GDP

9. In recent years, Bulgaria has made major progress selling state-owned banks to foreign strategic investors. Since the closure of 15 banks in 1996.2 the total number of banks, at around 35, has been stable. Foreign ownership, however, has increased dramatically. In the immediate aftermath of the crisis, state-owned banks still dominated the sector, holding more than 80 percent of total bank assets. By end-1997, the share of state-owned banks in total bank assets had fallen to about two-thirds, while that of foreign-owned banks had risen to about 15 percent.3 A series of privatizations to foreign investors in 1998–2000—including the sale of the country’s largest financial institution, Bulbank—reduced the share of state-owned banks in total bank assets to less than 20 percent and raised that of foreign-owned banks to close to 80 percent.4 The privatization of Biochim Bank, expected to be completed by end-July 2002, will result in the transfer of another 5 percent of bank assets from the state sector to foreign investors. Following this transaction, out of 35 commercial banks, 26 will be subsidiaries or branches of foreign banks, 6 domestic private banks, 2 domestic state-owned banks, and 1 a municipally owned bank (Table 2).5

Table 2.

Bulgaria: Structure and Ownership of Commercial Banks, March 2002

article image
Source: Bulgarian National Bank.

10. The structure of bank assets changed substantially in response to the crisis.6 Banks responded to the crisis by cutting down on domestic lending while sharply increasing the placements with major banks abroad (Table 3). The share of claims on banks and other financial institutions, the majority of which are deposits in foreign banks abroad, by end-1997 had risen to above 30 percent of total assets, twice its pre-crisis level, and in March 2002 still amounted to 33 percent. Because of this shift to low-risk and liquid assets, the share of loans and advances to non-financial institutions dropped from around 40 percent at end-1995 to 15 percent at end-1997, but has since rebounded, reaching 34 percent in March 2002. As a result of the sharp reduction in the supply of new government securities associated with the shift to a tight fiscal policy, the ratio of such securities to total bank assets has fallen to about half the pre-crisis level. The share of vault cash and balances in current accounts with the BNB has remained relatively high at close to 10 percent of total bank assets and has changed little in recent years, as Bulgaria’s economy continues to be largely cash-based.

Table 3.

Bulgaria: Aggregate Balance Sheet of Commercial Banks, 1998-March 2002

(In percent of total, unless otherwise indicated, end of period)

article image
Source: Bulgarian National Bank.

Deposits in foreign banks abroad; FI stands for Financial Institutions

11. Underlying the overall rebound in lending activity, there has been a profound change in the composition of the lending portfolio. As a result of both the privatization process and banks’ more conservative approach toward lending to loss-making enterprises, the proportion of claims on nonfinancial public enterprises in the lending portfolio dropped from more than 45 percent at end-1995 to less than 4 percent in March 2002. In the same period, the share of private sector enterprises in total bank lending rose from around 50 percent to more than 75 percent and—reflecting intensifying competition in such retail market segments as consumer credit and residential mortgages—that of households rose from less than 2 percent to around 20 percent. In spite of this increase in the proportion of claims on private sector enterprises and households, in real terms, these claims in March 2002 were still almost 30 percent below their end-1995 level.7

12. The structure of liabilities has undergone major changes as well. The level of interbank deposits fell from more than 25 percent of total liabilities and capital at end-1995 to less than 10 percent at end-1997, as refinancing by the BNB was eliminated by the introduction of the CBA. Interbank deposits fell further to less than 7 percent of total liabilities in March 2002. The share of deposits from nonfinancial institutions in total liabilities, on the other hand, rose from around 53 percent at end-1995 to around 62 percent at end-1997 and further to around 71 percent in March 2002. The maturity structure of these deposits significantly shortened in the wake of the crisis, with the share of savings and time deposits falling from more than 80 percent before the crisis to around 57 percent at end-1997, and the share of demand deposits rising to the same extent. Marking a gradual return of confidence in the banking system, the proportion of time and savings deposits partially rebounded to 69 percent in March 2002. The contractual maturity of total deposits remains short, however, with over 90 percent of deposits bearing a maturity of less than one year.8 9

13. A large fraction of bank assets and liabilities continues to be denominated in foreign exchange,10 with some further increase in the household deposit segment since 1997. In part as a result of valuation effects following the sharp depreciation of the lev, the share of foreign currency-denominated assets increased from less than 40 percent of total assets before the crisis to close to 60 percent of assets at end-1997 and has remained around that level since. The currency composition of major asset categories has also been broadly stable in recent years, with around 95 percent of claims on financial institutions, 75 percent of the securities portfolio, and 35 percent of loans being denominated in foreign currency, respectively.11 Altogether, in March 2002, short-term foreign currency deposits in banks abroad still accounted for more than 30 percent of bank assets. On the liabilities side, the share of foreign exchange deposits rose from around 45 percent at end-1995 to around 60 percent at end-1997, with little change thereafter. This overall evolution masks, however, significant differences among depositors. While the share of foreign exchange in deposits attracted from other financial institutions has gradually decreased since 1997, that in household deposits has picked up from 53 percent in 1997 to 59 percent in March 2002, in part as interest rates on lev demand deposits continued to be lower than those on euro and U.S.-dollar deposits.

14. Concentration in the banking sector remains moderately high, but competition in the bank credit market is on the rise. Concentration—measured by such indicators as the Herfindahl-Hirschman Index or the aggregate share of the three largest banks—is not out of line with that in comparable transition countries.12 Concentration rose in the wake of the 1996-97 crisis as a number of insolvent banks were closed. With respect to total assets and loans to non-financial institutions, the indicators show a reversal in more recent years. The share of the three largest banks in total assets declined from around 55 percent at end-1997 to around 45 percent in March 2002.13 In part reflecting the erosion of the dominant position of DSK Bank in the credit market as a result of the penetration by foreign subsidiaries and branches, the combined share of the top three banks in loans to non-financial institutions in the same period fell from more than 45 percent to less than 34 percent. In the market for primary deposits, however, concentration has remained broadly stable since 1997, with the three largest banks consistently accounting for more than 50 percent of such deposits owing to their dominant branch network. Continuing to reflect the pattern of specialization under central planning, Bulbank and DSK Bank maintain the largest share of foreign exchange deposits and lev deposits, respectively.

C. Performance

15. Bulgarian commercial banks are highly capitalized and liquid, and have low sensitivity to interest rate and exchange rate risks. This low sensitivity reflects the limited mismatch in their portfolio as most assets and deposits have very short maturities with floating interest rates, while the net open foreign exchange position is marginal owing to a strict regulation on the position. The overall sound condition of the banking system is attributable to (i) conservative lending policies on the part of commercial banks, (ii) limited opportunities for domestic lending, (iii) tight prudential regulations and supervision, and (iv) the strong discipline imposed by the CBA, which limits central bank liquidity support to solvent but illiquid banks in a situation of systemic risk.


16. The level of capitalization of the banking system has gradually declined since end-1998, but remains relatively high compared with the minimum requirement as well as by international standards. Bank capital was restored through the removal of unviable banks from the system, an implicit recapitalization of the remaining banks with capital gains on foreign currency assets, the introduction of the capital adequacy regulation in July 1997 and the shift to a much more cautious lending behavior induced largely by the introduction of strict loan classification and provisioning regulations.14 The average capital adequacy ratio (CAR) excluding branches of foreign banks rose to more than 40 percent at end-1998. As banks have begun to cautiously expand lending again, the average CAR fell back to 29 percent in March 2002 (Table 4), although it is still much higher than the minimum CAR of 12 percent.15 The decline was most pronounced for the group of largest banks (group 1) and for subsidiaries of foreign banks, with the CAR of domestic banks hovering between 20–25 percent during 1999-early 2002.16 The on average relatively high CAR levels suggest that banks can still considerably expand lending and take on additional risk without having to raise capital, thereby achieving a better risk/return tradeoff on their invested funds. However, some medium-sized banks have a CAR closer to the 12 percent minimum requirement, while some small banks have a capital base that could be insufficient in case major risks were to materialize.

Table 4.

Bulgaria: Capital Adequacy Ratios of Commercial Banks, 1999-March 2002 1/

(End of period; in percent, unless otherwise indicated)

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

Data may differ from official BNB publication, as classification in this table is based on group and ownership as of December 2000 for comparative purpose. Data excludes branches of foreign branches in Group 5.

The classification reflects assets totals as of end-2000: Group I:> BGN 500 million (3 banks); Group II:>300.<500 million (5 banks); Group III:>100 <300 million (6 banks); Group IV:<100 million (13 banks).

Asset quality

17. The asset quality of the banking system has significantly improved following the 1996–97 banking crisis. The gross nonperforming exposure ratio (defined as watch, substandard, doubtful, and loss)—which includes not only domestic loans but also deposits in foreign banks in the denominator and numerator of the ratio—fell from more than 40 percent at end-1996 to 17 percent at end-1997, as a result of substantial write-offs and restructuring of bad loans. The gross nonperforming exposure ratio fell further to 6 percent in March 2002 of which 5 percent was provisioned (Table 5). As banks have continued to make provisions, the exposure ratio net of provisioning stood at around 1 percent in March 2002. Asset quality differs substantially from group to group. The ratio is very low in Group 1, in part because the largest bank, Bulbank, has a high proportion of its assets invested in foreign deposits, which are classified as standard. However, the total nonperforming exposure ratios for medium-size banks in groups 3 and 4 remain quite high, on the order of 15 percent. The nonperforming exposure ratios should be interpreted with caution in view of the inclusion of low-risk deposits in foreign banks among the assets. Not taking into account these deposits, the gross nonperforming loan ratio would be 13 percent and the net nonperforming loan ratio 6 percent, respectively, in March 2002.

Table 5.

Bulgaria: Quality of Domestic Loans of Commercial Banks, 1999-March 2002

(End of period; in percent of total, unless otherwise indicated)

article image
Sources: Bulgarian National Bank, and IMF staff estimates.

Nonperforming exposure or nonperforming loan is the total of assets or loans categorized in watch, substandard, doubtful, and loss.

Assumes all claims on financial institutions as being standard with no provisioning.


18. Banking sector profitability remains adequate. Bank profitability was boosted in 1997 by one-off foreign exchange revaluation gains stemming from the sharp lev depreciation, but deteriorated in 1998–99 under the impact of the Russian financial crisis and the Kosovo conflict (Table 6). Profitability, as measured by the return on equity, rebounded in 2000–01 as banks raised net interest income by expanding lending activity while maintaining wide interest rate spreads. Following the introduction of the CBA and the tight regulation on net open foreign exchange position, banks have come to rely less on earnings from foreign exchange open positions and trading. The contribution from fees and commissions also remains low, in part because of the limited range of services offered. Profitability has been dampened by operating expenses which, in spite of low wage costs, remain high relative to both total operating income and assets.

Table 6.

Bulgaria: Income Statement of Commercial Banks, 1999-First Quarter of 2002

(Million Leva)

article image
Source: Bulgarian National Bank

Annualized quarterly numbers.


19. Bank liquidity is high, although it is gradually declining.17 The ratio of total marketable assets over total assets fell from 59 percent at end-1999 to 48 percent in March 2002, in line with the increase in lending activities.18 As banks continue to invest an important fraction of their assets in liquid foreign currency assets abroad, the current level of the ratio is still high by international standards. The ratio of marketable assets to attracted funds has also come down, from 76 percent at end-1999 to 59 percent in March 2002, as a result of increased lending activity, but with a depositor base that is largely stable, this development does not raise immediate concerns.

D. Stress Tests

Summary of results

20. Stress tests show that the Bulgarian banking system as a whole is resilient to substantial shocks. Table 7 summarizes the results of the tests in terms of the impact of the hypothetical shocks on the capital position of the banks. The tests were conducted on banking sector data from March 2002, and assess how adverse changes in the quality of domestic loans, foreign exchange rates, and interest rates would affect the capital position of 27 banks (7 branches of foreign banks were excluded from the stress tests as they are not subject to the capital adequacy regulations).19 The impact of hypothetical shocks on the capital adequacy ratio (CAR) of each bank is estimated, and the individual results aggregated for (i) four groups of banks categorized by the BNB on the basis of asset volume, (ii) two groups of banks categorized by ownership (subsidiaries of foreign banks and domestic banks), and (iii) the entire system excluding branches of foreign banks. The results show a high degree of resilience of the Bulgarian banking system as a whole and of the various groups. Under the stress tests on credit risks, exchange rate risks, and interest rate risks separately, the CAR across groups as well as of the system remains well above the minimum required 12 percent. Even under the worst-case scenario, which combines adverse developments on all three fronts, the system-wide CAR remains as high as 19 percent, although this implies a reduction of the CAR to nearly two thirds of its level before the shocks.

Table 7.

Bulgaria: Summary of Stress Tests Assumptions and Results, March 20021

(In percent)

article image
Sources: BNB and IMF staff estimates.

Since the beginning of 2001, the BSD has categorized banks into five groups according to asset totals: Group 1: >BGN 800 million (3 banks); Group 2: >300,<800 million (6 banks); Group 3: > 100,<300 million (6 banks); Group 4: <100 million. (13 banks); Group 5: all branches of foreign banks (7 banks). Data exclude Group 5. Data may differ from those in Table 4 because classification of group in Table 4 was as of December 2000.

Exchange rate shocks applied to net open foreign exchange position excluding Euro position.

Based on the repricing-gap model. Shocks only applied to lev-denominated interest-rate sensitive assets and liabilities.

Main assumptions include: a) all deposits in foreign banks classified as standard with no specific provisioning; b) banks shift 50 percent of their deposits in foreign banks to local loans; c) distribution of quality of domestic loans remain unchanged.

Main assumptions include: a) 10 percent of guarantee and commitment called and classified as loss; b) no shift from deposits to bank lending; c) quality of existing domestic loans deteriorates and all doubtful loans become loss, 50 percent of substandard loans becomes doubtful, 5 percent of watch loans becomes substandard, and 1 percent of standard loans become watch.

The assumption on intensification of migration includes: 10 percent of watch loans become substandard, and 5 percent of standard loans become watch.

21. The strong resilience of the Bulgarian banking system mirrors the prevailing strong risk averseness of the system, and is accounted for by three main factors: (i) the high initial CAR percent, which reflects banks’ preference to deposit a large proportion of their assets in reputable foreign banks, (ii) the low currency mismatch partly due to the tight regulation on net open foreign exchange positions, and (iii) the low interest rate mismatch as a result of short maturities on both sides of the balance sheet.

Credit risks

22. The credit risk stress tests are conducted under two scenario. The first scenario envisages an acceleration in domestic lending, with banks assumed to increase domestic lending by shifting half of their deposits in foreign banks to domestic loans. Under the assumption that the ratio and distribution of nonperforming to total domestic loans is unchanged, this switch results in an increase in the overall level of non-performing loans (NPLs). The second scenario envisages a deterioration in the asset quality of outstanding domestic loans in response to adverse macroeconomic developments that affect the repayment ability of borrowers. This deterioration is reflected in a migration in NPLs to the next worst category. No quantitative value is attached to these shocks.

23. Both shocks result in an increase in NPLs, higher specific provisions, less income, and subsequently less capital. Overall, however, the Bulgarian banking system proves to be highly resilient to credit risks. Under the first scenario, the CAR of the system drops by 3 percentage points to 26 percent. There is little difference among the four groups of banks classified by asset volume in the extent to which their CARs are affected by the first shock. An acceleration in domestic lending has, however, a larger negative impact on the subsidiaries of foreign banks than on the domestic banks. Under the second scenario, the CAR of the system decreases by 2 percentage points to 27 percent. The stress is more strongly felt by the smaller banks, reflecting their higher NPL ratios prior to the shocks.

Exchange rate risks

24. The test assumes a 30 percent depreciation of the Lev vis-à-vis the U.S. dollar and estimates the losses/gains in the net open foreign exchange positions following the depreciation, and the subsequent reduction/increase in capital.20 Commercial banks in Bulgaria are hardly exposed to direct foreign exchange risks. The banking system is able to absorb this pronounced depreciation without a change in the system-wide CAR.21 The CAR is negatively affected in only Group 1 and 4 banks, but only marginally so. The resistance of the system to exchange rate risks is underpinned by (i) the public confidence in and the authorities’ commitment to the CBA, (ii) the strict open foreign exchange position regulations, and (iii) the preference on the part of commercial banks to place funds attracted in foreign currencies as U.S. dollar or Euro-denominated deposits in foreign banks abroad.

Interest rate risks

25. The stress tests on interest rate risks use the repricing gap model, which estimates losses/gains on the annual net interest income in each maturity bucket by subtracting interests paid on liabilities from interests earned on assets.22 Applying a 50 percentage points increase in Lev rates to interest rate sensitive Lev-denominated assets and liabilities, the system-wide CAR falls by only 2 percentage points over a 6 month period, a reduction that is easily absorbable by the high level of capital. This strong resilience to interest rate risks is accounted for by the short maturities of both assets and liabilities, which result in a limited interest rate mismatch. In addition, the dominance of floating interest rates enables banks to make rapid adjustments in response to any significant Lev-rate increase.

Worst-case scenario

26. The stress tests on credit risks, foreign exchange rate risks, and interest rate risks can be combined into a worst-case scenario, in which the three shocks occur simultaneously and the size of the credit and interest rate shocks is magnified (the migration of NPLs to the next worst category intensifies, and the Lev interest rates increase by 100 percentage points). Under these combined adverse effects, the system-wide CAR falls by 10 percentage points to 19 percent, still remaining well above the 12 percent minimum requirement. While the system would thus remain solvent even in the face of the large combined shock, a third of the tested banks, mainly of Groups 2 and 3 banks, would require considerable recapitalization. The total capital shortfall under this worst case scenario would, however, be limited to less than 1 percent of 2002 GDP.

E. Comparison with Bank Intermediation and Performance in Central and Eastern European Countries

27. Bank intermediation in Bulgaria is relatively low by the standards of other central and eastern European (CEE) transition countries (Table 8). With the exception of Romania, monetization, as measured by the M2-to-GDP ratio, is below that in other CEE countries. Domestic credit in percent of GDP (18 percent) falls short even more of the ratios in the Slovak Republic (63 percent), the Czech Republic, Hungary, and Slovenia (around 50 percent each), and Poland (38 percent). Bulgaria also scores relatively low in terms of the size of the spread between deposit and lending rates and the EBRD index of progress in banking sector reform. Also, measured in terms of net loans in percent of total assets, Bulgarian banks scored the lowest ratio (23 percent),23 less than half of those in Hungary, Poland, and Slovenia (48–53 percent) (Table 9). The relatively low level of intermediation in Bulgaria in part reflects the fact that the country’s transition overall is still lagging that of the comparator countries other than Romania. The differences in the scale and quality of intermediation can be expected to narrow once the transition process is completed and EU membership gained.

Table 8.

Bulgaria: Financial Development in Central and Eastern European Countries, 2001

article image
Sources: EBRD, and IMF staff estimates

2000 data.

The index ranks from 1 to 4, which is the benchmark for a fully functioning market economy.

Table 9.

Bulgaria: Indicators of Banks in Central and Eastern European Countries in 2000

(In percent)

article image
Sources: Bankscope, central banks in Central Eastern Europe, IMF and World Bank staff estimates.

ROAA=net income/total average assets.

ROAE=net income/total equity

Discrepancies with data in Table 3 are attributable to the fewer number of banks, which have higher proportion of assets deposited in foreign banks.

28. However, as a partially related phenomenon, banks in Bulgaria are very well capitalized,24 relatively well provisioned against losses, quite profitable, and highly liquid compared to peer banks in CEE countries. But, over the longer term, the high levels of capital and liquidity could be associated with relatively low profitability.

29. With regards to asset quality, banks in the Czech Republic and the Slovak Republic had the highest loan loss provisioning ratio (11 percent of total loans), followed by those in Bulgaria and Romania (7 percent) and Hungary, Poland and Slovenia (3–5 percent).25

30. Bulgarian banks received the highest return on average assets (ROAA) at 2.9 percent and the second highest return on average equity (ROAE) at 15.5 percent. Bulgarian banks were also characterized by their high liquidity. Liquid assets in percent of customer and short-term funding were at 84 percent in Bulgaria, much higher than those in the Czech Republic, Romania, and the Slovak Republic (38–51 percent), and Hungary, Poland, and Slovenia (7–25 percent).

F. Concluding Remarks

31. A transformed and healthier banking sector has emerged from the 1996–97 financial crisis. A shift in policies and measures to close insolvent banks and tighten regulation and supervision in the context of the CBA have successfully addressed the weaknesses that were at the root of the crisis. An ongoing process of consolidation, privatization involving foreign participation, and restructuring has helped to further improve banking sector soundness. The sector is currently generally well regulated and supervised, enjoys high capital adequacy and liquidity, has reduced considerably the scale of impaired lending, does not face major vulnerabilities associated with foreign currency and interest exposure, and is fairly profitable.

32. The level and key features of the bank intermediation process continue to reflect the legacy of the financial crisis. As their confidence in the banking sector has not yet been fully restored, households continue to maintain large cash holdings in both leva and foreign currency outside the sector and, when depositing, have a preference for instruments with short-term maturities and a foreign currency denomination. Banks only partially channel the attracted savings to borrowers as they keep in place conservative lending strategies. Low-risk deposits in foreign financial institutions still account for around one third of bank assets, about the same proportion as that of loans to domestic households and enterprises. In addition to the memory of the crisis, conservative lending strategies reflect such factors as limited capacity for credit risk assessment and weaknesses in the provision of corporate information and the enforcement of claims.

33. The banking sector is now facing the key challenge of moving beyond the phase of post-crisis return to soundness and caution and of embarking on a path of steady expansion in intermediation while avoiding an undue increase in risk. With an expansion in economic activity, intensifying competition and cost pressure on margins, banks have begun to reassess their earning strategies and increase lending to enterprises and households. The higher risk in banks’ portfolios associated with more lending will require an upgrade in the capacity to manage and monitor risk exposure on behalf of both banks and bank supervisors. The needed adjustments may be particularly challenging for some smaller banks which are operating with low profitability. More generally, measures to enhance risk management and monitoring need to be accompanied by steps to strengthen the legal and institutional arrangements covering credit rights and insolvency, financial transparency, and corporate governance.


Prepared by Mark De Broeck and Yuri Kawakami.


Although the insolvent banks ceased operations in 1996, it took until March 1998 before all closed banks were declared bankrupt.


Stephan Barisitz, “The Development of the Romanian and Bulgarian Banking Sectors since 1990,” Focus on Transition, 1/2001, pp. 79–118.


However, in some banks, direct shareholders are established in off-shore centers, and no further information is available on the ultimate indirect shareholders or beneficial owners.


The privatization of DSK Bank (the former State Savings Bank) is being prepared.


For a more detailed discussion with a historical perspective, see Jeffrey Miller and Stefan Petranov, “The Financial System in the Bulgarian Economy,” Bulgarian National Bank Discussion Paper DP/19/2001, December 2001.


For an analysis of the factors impeding the expansion of lending to the private sector, see Tarhan Feyzioğlu and Gaston Gelos, “Why is Private Sector Credit so Low in Bulgaria?” IMF Staff Country Report 00/54, April 2000 (Washington: International Monetary Fund).


The actual maturity is longer as 70–75 percent of deposits are stable.


The decline in the share of time and savings deposits in the wake of the crisis has been driven mainly by the weakening of household confidence in the banking sector. This lack of confidence is also reflected in the fact that in 2000 only 27 percent of Bulgarian households had bank accounts, 10 percent time deposit accounts and 11 percent saving deposit accounts, respectively; see Jeffrey Miller and Stefan Petranov, o.c.


While the BNB collects detailed information on the foreign currency structure of assets and liabilities, it has yet to collect information on the foreign currency structure of deposits and loans by the type of currencies. New reporting forms with each balance sheet item and each income and expense item presented by foreign currency, including the type of currency, have been developed and are expected to be used from 2003.


While around half of all bank loans to private enterprises are expressed in foreign exchange, only a very small fraction of loans to households is (4 percent in March 2002).


John P. Bonin, “Financial Intermediation in Southeast Europe: Banking on the Balkans,” October 2001 (Vienna: Vienna Institute for International Economic Studies).


These three banks are Bulbank, the former State Foreign Trade Bank, United Bulgarian Bank, and DSK Bank, the former State Savings Bank.


The capital adequacy regulation issued on July 15, 1997 established a risked-based measure of required minimum capital in line with the Basle Committee recommendations. The minimum CAR was gradually phased in to reach 12 percent by end-1999.


As discussed below, the CARs are boosted by the high proportion of zero percent risk-weighted assets such as government securities and low risk-weighted assets such as placements with banks abroad.


The Banking Supervision Department categorizes banks into five groups according to asset totals: Group 1: > BGN 800 million (3 banks); Group 2: > 300, <800 million (6 banks); Group 3: > 100, <300 million (6 banks); Group 4: < 100 million (12 banks); Group 5: all branches of foreign banks.


Branches and subsidiaries of foreign banks have a much lower liquidity ratio than other banks, as they are normally able to liquidate deposits placed in their head offices and parent banks, in case of need.


Marketable assets include cash, noninterest-bearing deposits, interest-bearing deposits with banks, and Bulgarian treasury bills and bonds, minus all interest-bearing deposits with banks classified as watch or worse and all assets pledged to third parties.


It also excludes the new bank which started to operate in October 2001.


Exchange rate shocks were applied to net open foreign exchange positions other than Euro positions.


Indirect exchange rate risks resulting from lending in foreign exchange to non-foreign exchange earners are not covered by the stress tests on exchange rate risks but by those on credit risks.


While the repricing gap model provides information on the maturity mismatches in the portfolio by estimating the effect of an interest rate change on the income position, it does not assume any effect of the changes in the market value of assets, thus effectively valuing assets and liabilities at book value.


This number (23 percent) is provided by Bankscope, which covers only 20 banks with a relatively high exposure in deposits in foreign banks.


Banks in Bulgaria have the highest capital adequacy among the comparator CEE countries. The ratio of equity in percent of total assets was 20 percent in 2000 in Bulgaria, followed by those in Romania (18 percent), Slovenia and Poland (10–11 percent), and the Czech Republic, Slovak Republic, and Hungary (6–8 percent).


A straightforward comparison of nonperforming loan ratio, could be misleading, owing to different definitions across countries.