## _wp03213

## Source details

**Canonical URL:** [_wp03213](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2003/_wp03213.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2003/_wp03213.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2003/_wp03213.pdf.json)

---

### I. Introduction: scope and motivation
- Focus: recent trends in bank credit to the private sector (BCPS) in Central and Eastern European and Balkan (CEB) countries, determinants, macroeconomic and financial-stability implications, and econometric estimates of structural determinants of the BCPS ratio.
- Country group (15 countries): Albania, Bosnia and Herzegovina (BiH), Bulgaria, Croatia, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, the Former Yugoslav Republic of Macedonia, Poland, Romania, Serbia and Montenegro (SM), the Slovak Republic, and Slovenia.
- Observed recent episodes:
  - Cumulative real BCPS growth, 1998–2002: Bulgaria 315 percent; Estonia 121 percent; Latvia 304 percent.
  - BCPS-to-GDP ratio increases, 1998–2002: Bulgaria 11 percentage points; Estonia 22 percentage points; Latvia 17 percentage points.
  - 2002 real BCPS growth: Bosnia and Herzegovina (BiH) over 50 percent; Serbia and Montenegro (SM) 23 percent.
- Anticipated drivers of continued strong BCPS growth:
  - Natural financial deepening.
  - Strong capital inflows through the banking system driven by capital/labor ratio differences with Western Europe, expectations of real exchange rate appreciation, and sizable nominal interest rate differentials.

### II. Financial structure and stylized facts
- Bank-centered systems: in most CEB countries, the share of bank assets over total assets held by financial institutions exceeds 85 percent (Table 1).
- Equity markets small; bank credit primary external financing channel to private firms.
- Selected bank assets and market-capitalization figures (Table 1, exact values and years preserved):
  - Albania: Bank Assets 98.1 (2002); Stock Market Capitalization 0.0
  - Bosnia and Herzegovina: Bank Assets 88.0 (2002); Stock Market Capitalization 4.8
  - Bulgaria: Bank Assets 86.6 (June 2001); Stock Market Capitalization 3.7
  - Croatia: Bank Assets 91.9 (June 2001); Stock Market Capitalization 16.8
  - Czech Republic: Bank Assets 85.3 (2001); Stock Market Capitalization 15.4
  - Estonia: Bank Assets 84.3 (1999); Stock Market Capitalization 27.0
  - Hungary: Bank Assets 75.1 (September 2001); Stock Market Capitalization 18.7
  - Latvia: Bank Assets 90.7 (2000); Stock Market Capitalization 9.2
  - Lithuania: Bank Assets 86.5 (2001); Stock Market Capitalization 10.0
  - Macedonia, FYR: Bank Assets n.a.; Stock Market Capitalization 2.0
  - Poland: Bank Assets 85.3 (2001); Stock Market Capitalization 14.0
  - Romania: Bank Assets 91.7 (2000); Stock Market Capitalization 6.0
  - Serbia and Montenegro: Bank Assets 87.5 (2002); Stock Market Capitalization 3.1
  - Slovak Republic: Bank Assets 87.7 (2001); Stock Market Capitalization 3.3
  - Slovenia: Bank Assets 74.3 (2001); Stock Market Capitalization 15.3
- Reference benchmark: stock market capitalization, on average, exceeded 70 percent in the euro area at the end of 2001.
- Data note: BCPS data refer to loans extended by commercial banks to households and nonbank private enterprises.

### III. Phases of banking-sector transformation
- Typical three-phase evolution:
  - (i) Recognition and resolution of nonperforming loans of public banks, with losses shifted to government.
  - (ii) Sale of banks, primarily to foreign investors.
  - (iii) Emergence of standard banking operations, including increased lending to private enterprises.
- Timing and speed of these phases varied across countries.

### IV. Patterns of BCPS expansion and financing
- Country groups and defining characteristics:
  - Early birds (Bulgaria, Croatia, Estonia, Hungary, Latvia, Poland, Slovenia):
    - BCPS ratio rising for at least five years at an annual average rate exceeding 1½ percentage point of GDP.
    - Peak annual average rate: almost 4½ percentage points in Estonia.
    - Financing: increases partly reflected rising deposit-to-GDP ratios; in all but one (Latvia) also reflected declines in bank credit to government-to-GDP ratio.
    - Bulgaria, Croatia, Hungary: increased deposits and declines in credit to government allowed banks to improve net asset positions with the rest of the world (domestic funds sufficient).
    - Estonia, Latvia, Poland, Slovenia: rise in BCPS ratio also supported by increased net borrowing from abroad.
  - Late risers (Bosnia and Herzegovina, Lithuania, Serbia and Montenegro):
    - Only recently experienced BCPS ratio rises exceeding 1½ percentage points of GDP per year.
    - Rise fueled by deposit surges; in BiH deposits surge allowed banks to improve net asset positions with the rest of the world.
    - In BiH and SM, deposits surge reflected response to introduction of the euro—holders deposited pre-euro currency notes to convert into euros.
  - Sleeping beauties (Albania, Czech Republic, Macedonia, Romania, Slovak Republic):
    - Showed initial declines and then stabilization with limited balance-sheet composition changes.
- Selected annual average changes (Table 4, percent of GDP per year; exact values preserved):
  - Early birds average: Bank Loans to Private Sector 2.4; Bank Credit to the Public Sector -0.4; Bank Deposits 2.2; Net Foreign Liabilities 0.1; Other Net Liabilities -0.3
  - Late risers average: Bank Loans to Private Sector 3.2; Bank Credit to the Public Sector -0.6; Bank Deposits 4.1; Net Foreign Liabilities -0.1; Other Net Liabilities -1.4
  - Sleeping beauties average: Bank Loans to Private Sector -1.6; Bank Credit to the Public Sector 0.2; Bank Deposits 0.7; Net Foreign Liabilities 0.6; Other Net Liabilities -2.7
- Notable country decompositions (Table 3 notes, selected exact totals):
  - Croatia 1995–2002: Change in Bank Loans to Private Sector 20.8; Bank Deposits change 30.0; Net Foreign Liabilities change -7.7; Other Net Liabilities change -12.9.
  - Estonia 1992–2002: Change in Bank Loans to Private Sector 35.2; Bank Deposits change 19.0; Net Foreign Liabilities change 9.9; Other Net Liabilities change 6.2.
  - Czech Republic 1998–2002: Change in Bank Loans to Private Sector -14.9; Bank Credit to the Public Sector 8.7; Bank Deposits -1.3; Net Foreign Liabilities 19.2; Other Net Liabilities -24.2 (large drop in Other Net Liabilities due to surge in central bank assets from sterilization).

### V. Sectoral composition and credit dynamics
- Early birds: both household and corporate credit increased relative to GDP between 1997 and 2002.
- Sleeping beauties: generally declines in corporate credit relative to GDP and lower household credit growth relative to early birds.
- Selected sectoral growth rates and cumulative figures (exact values preserved):
  - Latvia private sector credit (annual): 1998: 54.5; 1999: 26.6; 2000: 17.4; 2001: 30.8; 2002: 34.5; 2000–2002 cumulative: 106.5. Credit to households 2002: 61.7; 2000–2002 cumulative: 278.1.
  - Hungary private sector credit (annual): 1998: 10.1; 1999: 7.2; 2000: 17.4; 2001: 16.5; 2002: 7.6; 2000–2002 cumulative: 47.2. Credit to households 2002: 45.9; 2000–2002 cumulative: 185.7.
  - Estonia private sector credit (annual): 1998: 30.5; 1999: 4.6; 2000: 10.4; 2001: 20.4; 2002: 22.0; 2000–2002 cumulative: 62.2. Credit to households 2002: 34.0; 2000–2002 cumulative: 97.4.
  - Romania private sector credit (annual): 1998: 20.0; 1999: -2.0; 2000: -19.3; 2001: 5.4; 2002: 26.0; 2000–2002 cumulative: 7.2. Credit to households 2002: 68.2; 2000–2002 cumulative: 48.7.
  - Czech Republic private sector credit (annual): 1998: -5.1; 1999: -4.8; 2000: -5.4; 2001: -6.1; 2002: -0.3; 2000–2002 cumulative: -11.5. Credit to households 2002: 17.2; 2000–2002 cumulative: 33.3.
  - Slovak Republic private sector credit (annual): 1998: -1.2; 1999: -1.2; 2000: -3.8; 2001: -14.0; 2002: -12.3; 2000–2002 cumulative: -27.5. Credit to households 2002: 13.8; 2000–2002 cumulative: 56.9.
- Conclusion: institutional and banking-system factors, rather than demand-side factors alone, are responsible for group differences.

### VI. Drivers, assessment of future trends, and microeconomic outlook
- Micro drivers:
  - Privatization and foreign ownership: ongoing privatization expected to boost credit growth within observed lags (one-two years).
  - Bank profitability incentives: Early birds average ROA 1.47; ROE 13.98; ROER 7.17. Sleeping beauties average ROA 0.88; ROE 8.78; ROER -4.96. EU average (2001) ROA 0.60; ROE 10.60; ROER 7.85. Banks with low profitability likely to expand loans to raise returns.
  - Credit risk improvement: progress in accounting standards, creditor protection, and collateral quality expected to support credit expansion.
- Macro drivers:
  - Fiscal consolidation: average general government deficit during rising BCPS episodes—Early birds 2.2 percent of GDP; Sleeping beauties 4.5 percent of GDP (Table 6).
  - Capital inflows: expected to increase due to euro-convergence, lower risk premia, expectations of real exchange rate appreciation, and interest rate differentials.
  - Capital gap rationale: estimates based on equalizing capital-labor ratios with Germany (α assumed equal to 1/3; German capital-output ratio set at 1.74) imply substantial potential for inflows in some countries (Table 14, selected values preserved).
- Key microeconomic indicators and transition measures:
  - EBRD transition indexes and privatization: Early birds generally higher transition indexes; asset share of state-owned banks averaged 25 percent in early birds by end-1990s and fell below 16 percent by end-2001; sleeping beauties had over 41 percent in 1999 (Table 8).
  - Foreign ownership shares of banks exceeded 80 percent in a number of countries.

### VII. Econometric model and equilibrium BCPS findings
- Random effects panel regression (nontransition countries, 1973–76, Table 12 main specification, selected coefficients and significance preserved):
  - Public Debt / GDP: -0.136 (standard error 0.032, t-statistic -4.29) ***
  - Log (GDP per capita): 0.084 (0.022, 3.74) ***
  - High-inflation dummy*(1/Inflation rate - 1/Inflation threshold): 0.004 (0.001, 4.66) ***
  - Liberalization index: 0.190 (0.034, 5.64) ***
  - Accounting: 0.007 (0.003, 2.74) ***
  - German legal origin: 0.287 (0.086, 3.34) ***
  - Number of observations: 468; R-squared: 0.66
  - Inflation threshold used: 4 percent.
- Out-of-sample predicted BCPS (Panel B, Table 12, Actual 2002 versus Predicted and Absolute Deviation a - b, exact values preserved):
  - Albania: Actual 4.9 / Predicted 24.6 / Absolute Deviation -19.7
  - Bosnia: Actual 21.9 / Predicted 18.5 / Absolute Deviation 3.5
  - Bulgaria: Actual 15.6 / Predicted 52.6 / Absolute Deviation -37.0
  - Croatia: Actual 45.6 / Predicted 49.1 / Absolute Deviation -3.5
  - Czech Republic: Actual 42.6 / Predicted 69.3 / Absolute Deviation -26.7
  - Estonia: Actual 46.0 / Predicted 75.4 / Absolute Deviation -29.4
  - Hungary: Actual 29.3 / Predicted 70.5 / Absolute Deviation -41.2
  - Latvia: Actual 24.8 / Predicted 76.7 / Absolute Deviation -51.9
  - Lithuania: Actual 11.6 / Predicted 68.1 / Absolute Deviation -53.9
  - Macedonia, FYR: Actual 17.1 / Predicted 44.2 / Absolute Deviation -27.1
  - Poland: Actual 28.1 / Predicted 70.4 / Absolute Deviation -42.3
  - Romania: Actual 8.3 / Predicted 58.0 / Absolute Deviation -49.7
  - Serbia: Actual 14.7 / Predicted 21.0 / Absolute Deviation -6.3
  - Slovak Republic: Actual 31.5 / Predicted 59.9 / Absolute Deviation -28.4
  - Slovenia: Actual 38.4 / Predicted 63.8 / Absolute Deviation -25.4
- Interpretation:
  - BCPS ratios in almost all CEB countries are still significantly undersized relative to model-predicted equilibrium.
  - Average deviation from equilibrium is similar across sleeping beauties and early birds (about 33-30 percentage points of GDP).
  - BCPS ratio close to equilibrium only in Croatia and Serbia and Montenegro (SM); Bosnia somewhat above equilibrium following a 6¾ percentage points increase in 2002.

### VIII. Macroeconomic outlook, capital flows, and fiscal policy
- Fiscal consolidation and crowding-in (Table 13, selected exact values preserved):
  - Average (excluding Croatia): 2002 outturn 3.8 percent of GDP; Targets 2003 3.2 / 2004 2.9 / 2005 2.5.
  - Country examples (percent of GDP): Albania 2002 outturn 6.9 ; Targets 2003 6.6 / 2004 6.0 / 2005 5.8. Bosnia and Herzegovina 2002 outturn 4.3 ; Targets 2003 2.2 / 2004 2.0 / 2005 1.7. Bulgaria 2002 outturn 0.6 ; Targets 2003 0.7 / 2004 0.5 / 2005 0.0.
- Capital inflows and capital gap estimates (Table 14, selected values preserved):
  - Nominal interest rate differential (1999–2002 average), Parity Real Interest Rate, Estimated Capital Gap (percent of GDP): Albania 7.4 / 3.4 / 1131.1. Bulgaria 1.3 / -2.1 / 707.9. Croatia 0.2 / -0.8 / 502.2. Czech Republic 1.1 / -2.8 / 224.1. Hungary 7.8 / -6.8 / 282.6. Poland 11.4 / -2.7 / 427.5. Slovak Republic 5.5 / -4.3 / 347.7. Slovenia 6.3 / -0.2 / 157.7. (Romania reported nominal differential 42.4 / Parity Real Interest Rate -14.6 / Estimated Capital Gap --.)
  - Two-thirds of CEB countries were net lenders to the rest of the world in the average of 2002; only five countries were net borrowers (BiH, Czech Republic, Estonia, Latvia, and Poland).
  - Capital inflows expected to increase with euro-convergence, lower risk premia, expectations of real appreciation, and interest rate differentials.
- Interpretation of rising BCPS ratios:
  - A rising loans-to-GDP ratio is not per se evidence of overheating; overheating arises when loan-financed demand growth exceeds potential output growth and initial stock conditions satisfy inequality conditions (see formal discussion in the paper).
  - Empirical examples: 2002 credit surges in Bosnia, Serbia and Montenegro, and Croatia matched by sharp current account deteriorations; Bulgaria’s 2002 credit acceleration had no obvious macroeconomic implications.

### IX. Policy instruments, trade-offs, and practical responses
- Available instruments to address surges in BCPS and aggregate demand:
  - Standard monetary policy tools (open market operations, money market instruments).
  - Fiscal policy (fiscal tightening to avoid overheating); in 2002, 60 percent of sample countries had general government deficit over 4 percentage points of GDP.
  - Controls on capital inflows (limited by OECD/EU commitments and uncertain effectiveness).
  - Administrative credit ceilings (used in Croatia in early 2003; historically used in Greece in 1999).
  - Prudential measures and supervisory actions.
- Monetary policy and sterilized intervention:
  - Most CEB central banks (excluding Bulgaria, BiH, Estonia, Lithuania with currency boards) have money market instruments to control liquidity.
  - Empirical sterilization offset coefficients reported: Czech Republic earlier studies 1/3-2/3; Christensen (2003) below 0.15; OECD (1996) range 0.4-0.65. Hungary: Fernández-Ansola and Moghadam (1996) find 0.5 pre-March 1995; near unity after crawl but with large standard error.
  - Limits of sterilization: financial cost, perpetuation of capital inflows, buildup of short-term assets that can exit or fuel credit growth.
- Fiscal policy:
  - Fiscal tightening recommended where monetary policy constrained by exchange-rate objectives.
  - Practical experience: Croatia, BiH, and SM regarded fast credit growth and widening current-account deficits as reasons to tighten fiscal policy in 2003; fiscal responses were small in some cases.
- Capital controls:
  - Most CEB countries removed capital controls; Article 7 (OECD) allows suspension up to eighteen months under serious balance-of-payments distress.
  - Practical constraints: conflicts with OECD/EU commitments, mixed evidence on effectiveness, reputational costs.
- Credit ceilings and prudential instruments:
  - Credit ceilings can be effective if circumventing borrowing from abroad is limited; Croatia imposed credit ceilings in early 2003.
  - Prudential instruments suggested: pro-cyclical capital requirements, restrictive collateral rules, higher provisioning for banks with high credit growth.
  - Prudential actions strengthened in response to credit booms in several CEB countries, including BiH, Croatia, SM.

### X. Financial stability, lending booms detection, and risks
- Literature and evidence linking lending booms and crises:
  - Historical precedents: Argentina 1980; Chile 1982; Sweden, Norway, Finland 1992; Mexico 1994; Thailand, Indonesia, Korea 1997.
  - Empirical studies: Demirgüç-Kunt and Detragiache (1997) — 10 percent increase in lagged credit growth → increases crisis probability by 5½ percent for 1994 Mexican crisis example. Gourinchas et al. (1999, 2001), Kaminsky and Reinhart (1999), Hardy and Pazarbasioglu (1998) report lending-boom patterns preceding crises, though conditional crisis probability after boom remains below 20 percent in some studies.
  - Theoretical channels: financial accelerator / collateral-based amplification (Kiyotaki and Moore (1997); Aghion, Bacchetta, and Banerjee (2001); Dell’Ariccia and Marquez (2003)).
  - Alternative view: rapid credit growth can reflect financial deepening and permanent allocation improvements (Rajan and Zingales, 1998; Beck, Levine, and Loayza, 2000).
- Detecting lending booms in CEB countries (HP-filter rolling trend methodology, Gourinchas et al. (2001) approach):
  - Define a lending boom where BCPS ratio deviates from rolling backward-looking HP-trend (first trend constructed using first five years of data).
  - Caveats: short BCPS series make results sensitive; transition economies’ lack of tranquil pre-sample periods complicates trend proxies; write-offs in early years (e.g., Bulgaria) affect trend estimation.
  - Comparative findings:
    - Control-group crisis cases often saw annual BCPS ratio increases exceeding 5 percentage points, sometimes up to 10 percentage points.
    - CEB economies’ typical increases around 2 percentage points; only a few cases exceed 5 percentage points.
  - 2002 HP-deviation results (Table 16, exact values preserved):
    - Bulgaria: BCPS Ratio 15.6; HP-Estimated 11.8; Absolute Deviation 3.8; Relative Deviation 24.1 percent (noted as slightly below lower bound of threshold range).
    - Croatia: BCPS Ratio 45.6; HP-Estimated 43.5; Absolute Deviation 2.1; Relative Deviation 4.7
    - Estonia: BCPS Ratio 46.0; HP-Estimated 45.6; Absolute Deviation 0.4; Relative Deviation 1.0
    - Hungary: BCPS Ratio 29.3; HP-Estimated 26.0; Absolute Deviation 3.3; Relative Deviation 11.2
    - Latvia: BCPS Ratio 24.8; HP-Estimated 20.4; Absolute Deviation 4.4; Relative Deviation 17.6
    - Poland: BCPS Ratio 28.1; HP-Estimated 27.1; Absolute Deviation 1.0; Relative Deviation 3.5
    - Slovenia: BCPS Ratio 38.4; HP-Estimated 39.0; Absolute Deviation -0.6; Relative Deviation -1.7
    - Lithuania: BCPS Ratio 14.2; HP-Estimated 11.4; Absolute Deviation 2.8; Relative Deviation 19.7
    - Albania: BCPS Ratio 4.9; HP-Estimated 4.2; Absolute Deviation 0.7; Relative Deviation 14.3
    - Czech Republic: BCPS Ratio 42.6; HP-Estimated 45.2; Absolute Deviation -2.6; Relative Deviation -6.2
    - Macedonia, FYR: BCPS Ratio 17.1; HP-Estimated 15.9; Absolute Deviation 1.2; Relative Deviation 6.9
    - Romania: BCPS Ratio 8.3; HP-Estimated 8.8; Absolute Deviation -0.5; Relative Deviation -5.6
    - Slovak Republic: BCPS Ratio 31.5; HP-Estimated 36.6; Absolute Deviation -5.1; Relative Deviation -16.3
  - Conclusion: Using Gourinchas et al. (2001) reference thresholds (absolute deviations 4.8 percent to 6.4 percent; relative deviations 24.9 percent to 31.1 percent), only one CEB lending boom detected in 2002: Bulgaria’s relative deviation 24.1 percent is slightly below the lower bound; Latvia is a close case.
- Credit to households (BCH) deviations and household credit booms (Table 17 and Table 18 exact values preserved):
  - Hungary BCH 1999–2002 rolling deviations: 1999 Absolute Deviation 0.7; Relative Deviation 24.1; 2000 Absolute Deviation 1.1; Relative Deviation 32.1; 2001 Absolute Deviation 1.6; Relative Deviation 34.5; 2002 Absolute Deviation 2.4; Relative Deviation 34.9
  - 2002 BCH deviations (selected): Hungary BCH Ratio 6.8; HP-Estimated 4.4; Absolute Deviation 2.4; Relative Deviation 34.9. Romania BCH Ratio 0.7; HP-Estimated 0.5; Absolute Deviation 0.2; Relative Deviation 22.4.
  - Observation: household credit booms are present in more cases (e.g., Hungary exceeds 24 percent threshold), even when aggregate BCPS tests do not flag a boom due to HP-trend absorption.
- Financial-stability considerations:
  - IMF Board cautioned CEB countries about risks to bank portfolio quality from too rapid private credit growth and called for increased supervisory attention.
  - Historical and econometric evidence links lending booms to elevated probabilities of banking crises, but conditional crisis probability after boom episodes remains below 20 percent in some studies.
  - Policy implications depend on whether rapid credit growth is transitory (boom-bust) or a permanent financial-deepening shift.

### XI. Policy recommendations and practical guidance (synthesis of paper’s discussion)
- Priorities for authorities facing rapid BCPS growth:
  - Intensify prudential supervision and on-site inspection given rapid credit growth and rising complexity of banking systems.
  - Use fiscal policy to offset aggregate demand pressure when monetary policy constrained by exchange-rate objectives (fiscal consolidation consistent with medium-term targets and Stability and Growth Pact obligations).
  - Consider a mix of instruments rather than a single tool: monetary operations, fiscal tightening, prudential measures, and selective administrative measures where appropriate.
  - Prudential measures preferred to blunt distortions of credit ceilings: examples include higher provisioning for banks with high credit growth, pro-cyclical capital requirements, and restrictive collateral rules that increase bank solidity.
  - Reserve the use of capital-flow measures for exceptional circumstances given commitments to capital mobility and mixed empirical effectiveness; Article 7 (OECD) allows temporary suspension of capital mobility up to eighteen months under serious balance-of-payments distress.
  - Where necessary and feasible, use administrative measures (e.g., credit ceilings, raised reserve requirements) as temporary responses to rapid credit growth—recognizing these instruments may distort market efficiencies and are unlikely to be permanent features.
  - Address microeconomic distortions (e.g., remove housing subsidy schemes) that artificially boost sectoral credit demand.
- Monitoring and evaluation:
  - Monitor sectoral and currency composition of credit, maturity mismatches, and real-estate exposures that can amplify risks despite moderate aggregate BCPS growth.
  - Use rolling backward-looking trend tests for early warning but complement with balance-sheet decomposition, microprudential indicators, and qualitative assessment of structural reforms and banking privatization.

*Italic source attribution: Content derived from the IMF working paper (_wp03213 - References) provided in the source PDF.*

### References..............................................................................................................

### _wp03213 - References..............................................................................................................

### I. Introduction: scope and motivation
- Focus: recent trends in bank credit to the private sector (BCPS) in Central and Eastern European and Balkan (CEB) countries, determinants, macroeconomic and financial-stability implications, and econometric estimates of structural determinants of the BCPS ratio.
- Country group (15 countries): Albania, Bosnia and Herzegovina (BiH), Bulgaria, Croatia, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, the Former Yugoslav Republic of Macedonia, Poland, Romania, Serbia and Montenegro (SM), the Slovak Republic, and Slovenia.
- Observed recent episodes:
  - Cumulative real BCPS growth, 1998–2002: Bulgaria 315 percent; Estonia 121 percent; Latvia 304 percent.
  - BCPS-to-GDP ratio increases, same period: Bulgaria 11 percentage points; Estonia 22 percentage points; Latvia 17 percentage points.
  - 2002 real BCPS growth: Bosnia and Herzegovina (BiH) over 50 percent; Serbia and Montenegro (SM) 23 percent.
- Anticipated drivers of continued strong BCPS growth:
  - Natural financial deepening of these systems.
  - Likelihood of strong capital inflows through the banking system due to capital/labor ratio differences with Western Europe, expectations of real exchange rate appreciation, and sizable nominal interest rate differentials.

### II. Financial structure and stylized facts
- Financial system characteristics:
  - Bank-centered systems: in most CEB countries, the share of bank assets over total assets held by financial institutions exceeds 85 percent (Table 1).
  - Equity markets are small; bank credit is the primary external financing channel to private firms.
- Selected Bank Assets and Stock Market Capitalization (exact figures from Table 1):
  - Albania: Bank Assets 98.1 (2002); Stock Market Capitalization 0.0
  - Bosnia and Herzegovina: Bank Assets 88.0 (2002); Stock Market Capitalization 4.8
  - Bulgaria: Bank Assets 86.6 (June 2001); Stock Market Capitalization 3.7
  - Croatia: Bank Assets 91.9 (June 2001); Stock Market Capitalization 16.8
  - Czech Republic: Bank Assets 85.3 (2001); Stock Market Capitalization 15.4
  - Estonia: Bank Assets 84.3 (1999); Stock Market Capitalization 27.0
  - Hungary: Bank Assets 75.1 (September 2001); Stock Market Capitalization 18.7
  - Latvia: Bank Assets 90.7 (2000); Stock Market Capitalization 9.2
  - Lithuania: Bank Assets 86.5 (2001); Stock Market Capitalization 10.0
  - Macedonia, FYR: Bank Assets n.a.; Stock Market Capitalization 2.0
  - Poland: Bank Assets 85.3 (2001); Stock Market Capitalization 14.0
  - Romania: Bank Assets 91.7 (2000); Stock Market Capitalization 6.0
  - Serbia and Montenegro: Bank Assets 87.5 (2002); Stock Market Capitalization 3.1
  - Slovak Republic: Bank Assets 87.7 (2001); Stock Market Capitalization 3.3
  - Slovenia: Bank Assets 74.3 (2001); Stock Market Capitalization 15.3
  - Reference: stock market capitalization, on average, exceeded 70 percent in the euro area at the end of 2001.
- Data note: BCPS data refer to loans extended by commercial banks to households and nonbank private enterprises.

### III. Phases of banking sector transformation in CEB countries
- Typical three-phase evolution:
  - (i) Recognition and resolution of nonperforming loans of public banks, with losses shifted to government.
  - (ii) Sale of banks, primarily to foreign investors.
  - (iii) Emergence of standard banking operations, including increased lending to private enterprises.
- Timing of these phases varied across countries with differing speeds of institutional transformation.

### IV. Patterns of BCPS expansion and financing
- Heterogeneity across countries; three broad country groups identified:
  - The seven "early birds" (Bulgaria, Croatia, Estonia, Hungary, Latvia, Poland, Slovenia):
    - BCPS ratio has been rising for at least five years at an annual average rate exceeding 1½ percentage point of GDP.
    - Peak annual average rate: almost 4½ percentage points in Estonia (Table 2).
    - Financing patterns: increase partly reflected rising deposit-to-GDP ratios; in all but one of these countries, it also reflected a decline in bank credit to the government-to-GDP ratio.
    - In Bulgaria, Croatia, and Hungary: increased bank deposits and declines in credit to government allowed banks to improve net asset positions with the rest of the world (i.e., domestic funds were sufficient in these cases).
    - In Estonia, Latvia, Poland, and Slovenia: the rise in the BCPS ratio was also supported by increased net borrowing from abroad.
  - The three "late risers" (Bosnia and Herzegovina, Lithuania, Serbia and Montenegro):
    - Only recently experienced BCPS ratio rises exceeding 1½ percentage points of GDP per year.
    - Rise fueled by a surge in deposits; in BiH this allowed banks to improve net asset positions with the rest of the world.
    - In BiH and SM, deposits surge reflected response to the introduction of the euro—holders of pre-euro currency notes deposited them to convert into euros, and most of these resources remained in the banking system.

### V. Research agenda and paper structure (as presented)
- The paper:
  - Presents recent trends in BCPS in CEB countries.
  - Evaluates factors affecting trends, cross-country differences, and likely future trajectories; includes an econometric model estimating structural determinants of the BCPS ratio.
  - Assesses implications of fast credit growth (and rising BCPS ratio) for macroeconomic developments and policies, particularly where BCPS growth is coupled with strong capital inflows through the banking system.
  - Discusses effects of strong credit growth on financial stability, including sectoral composition of BCPS.
- Paper organization:
  - Sections II–V: the four substantive topics above.
  - Section VI: conclusions.

*Italic source attribution: Content derived from the IMF working paper (_wp03213 - References) provided in the source PDF.*

### introduction of the euro caused a permanent shift in the distribution of agents’ holdings of

### _wp03213 - introduction of the euro caused a permanent shift in the distribution of agents’ holdings of

### Summary of empirical findings
- The analysis groups Central and Eastern European and Balkan (CEB) countries into three categories based on Bank Credit to the Private Sector (BCPS)-to-GDP dynamics: Early birds, Late risers, and Sleeping beauties.
- Early birds experienced sustained rises in BCPS ratios after a turning-point year; sleeping beauties showed initial declines and then stabilization with limited changes in balance sheet composition, while late risers showed more recent upticks.
- The paper emphasizes growth-rate comparisons across countries rather than levels because of differences in coverage (for some countries BCPS series include public enterprises) and the effects of loan write-offs.

### Country groupings, turning points, and BCPS growth (selected figures from Table 2)
- Early birds (turning point year; Average Annual Real Growth Rate; Change in BCPS Ratio; Average Annual Change in BCPS Ratio):
  - Bulgaria: 1998; 33.0; 11.0; 2.2
  - Croatia: 1994; 12.3; 20.8; 2.6
  - Estonia: 1994; 25.9; 35.3; 4.4
  - Hungary: 1997; 11.1; 9.2; 1.5
  - Latvia: 1997; 31.4; 17.9; 3.0
  - Poland: 1996; 13.5; 12.9; 1.8
  - Slovenia: 1992; 13.0; 21.6; 2.2
- Late risers:
  - Bosnia and Herzegovina: 2001; 36.6; 9.2; 4.6
  - Serbia and Montenegro: 2002; 23.1; 2.3; 2.3
  - Lithuania: 2002; 23.7; 2.6; 2.6
- Sleeping beauties (no turning point year in the sample period):
  - Albania: --; 14.5; 1.5; 0.3
  - Czech Republic: --; -4.4; -15.0; -3.0
  - Macedonia, FYR: --; -6.7; -9.1; -1.8
  - Romania: --; 4.7; 1.3; 0.3
  - Slovak Republic: --; -6.7; -20.1; -4.0

### Financing the growth of BCPS: balance sheet changes (selected figures from Table 3 and Table 4)
- Table 3 shows changes in ratios between balance sheet items and GDP (percentage points of GDP) over reference periods; Table 4 shows annual average changes (percentage points of GDP per year).
- Early birds (annual average changes, Table 4):
  - Bulgaria 1998–2002: Bank Loans to Private Sector 2.2; Bank Credit to the Public Sector -1.2; Bank Deposits 2.8; Net Foreign Liabilities -1.3; Other Net Liabilities -0.1
  - Croatia 1995–2002: 2.6; -1.4; 3.8; -1.0; -1.6
  - Estonia 1992–2002: 3.2; -0.0; 1.7; 0.9; 0.6
  - Hungary 1997–2002: 1.5; -0.3; 0.4; -0.7; 1.5
  - Latvia 1998–2002: 3.3; 0.0; 2.2; 2.1; -1.0
  - Poland 1998–2002: 1.8; -0.2; 1.9; 0.5; -1.0
  - Slovenia 1995–2002: 2.3; -0.1; 2.2; 0.5; -0.3
  - Early birds average: 2.4; -0.4; 2.2; 0.1; -0.3
- Late risers (annual average changes, Table 4):
  - Bosnia and Herzegovina 2001: 4.6; -2.6; 6.1; -2.2; -1.9
  - Serbia and Montenegro 2002: 2.3; 0.1; 3.9; 0.6; -2.2
  - Lithuania 2002: 2.6; 0.6; 2.3; 1.2; -0.2
  - Late risers average: 3.2; -0.6; 4.1; -0.1; -1.4
- Sleeping beauties (annual average changes, Table 4):
  - Albania 1998–2002: 0.3; -1.9; 0.7; -0.0; -2.3
  - Czech Republic 1998–2002: -3.0; 1.7; -0.3; 3.8; -4.8
  - Macedonia, FYR 1998–2002: -1.8; 0.3; 3.1; -1.7; -2.9
  - Romania 1998–2002: 0.3; -0.9; -0.3; -0.0; -0.2
  - Slovak Republic 1998–2002: -4.0; 2.0; 0.3; 1.0; -3.3
  - Sleeping beauties average: -1.6; 0.2; 0.7; 0.6; -2.7

- Notable country-level decompositions and explanations from Table 3 notes:
  - Croatia 1995–2002 overall change in Bank Loans to Private Sector: 20.8; Bank Deposits change: 30.0; Net Foreign Liabilities change: -7.7; Other Net Liabilities change: -12.9.
  - Estonia 1992–2002 overall change in Bank Loans to Private Sector: 35.2; Bank Deposits change: 19.0; Net Foreign Liabilities change: 9.9; Other Net Liabilities change: 6.2.
  - Czech Republic 1998–2002 change in Bank Loans to Private Sector: -14.9; Bank Credit to the Public Sector: 8.7; Bank Deposits: -1.3; Net Foreign Liabilities: 19.2; Other Net Liabilities: -24.2 (large drop in Other Net Liabilities due to surge in assets at the central bank from sterilization operations).

### Sectoral composition and dynamics of credit (selected figures from Table 5 and Figure 2)
- Early birds saw both household and corporate credit increase relative to GDP between 1997 and 2002; sleeping beauties generally experienced declines in corporate credit relative to GDP and lower household credit growth relative to early birds.
- Selected annual real growth rates by sector (1998–2002 and cumulative 2000–2002 where reported):
  - Latvia: Credit to private sector (1998–2002 sample annual figures) include 1998: 54.5; 1999: 26.6; 2000: 17.4; 2001: 30.8; 2002: 34.5; 2000–2002 cumulative: 106.5. Credit to households 2002: 61.7; 2000–2002 cumulative: 278.1.
  - Hungary: Credit to private sector 1998: 10.1; 1999: 7.2; 2000: 17.4; 2001: 16.5; 2002: 7.6; 2000–2002 cumulative: 47.2. Credit to households 2002: 45.9; 2000–2002 cumulative: 185.7.
  - Estonia: Credit to private sector 1998: 30.5; 1999: 4.6; 2000: 10.4; 2001: 20.4; 2002: 22.0; 2000–2002 cumulative: 62.2. Credit to households 2002: 34.0; 2000–2002 cumulative: 97.4.
  - Romania: Credit to private sector 1998: 20.0; 1999: -2.0; 2000: -19.3; 2001: 5.4; 2002: 26.0; 2000–2002 cumulative: 7.2. Credit to households 2002: 68.2; 2000–2002 cumulative: 48.7.
  - Czech Republic: Credit to private sector 1998: -5.1; 1999: -4.8; 2000: -5.4; 2001: -6.1; 2002: -0.3; 2000–2002 cumulative: -11.5. Credit to households 2002: 17.2; 2000–2002 cumulative: 33.3.
  - Slovak Republic: Credit to private sector 1998: -1.2; 1999: -1.2; 2000: -3.8; 2001: -14.0; 2002: -12.3; 2000–2002 cumulative: -27.5. Credit to households 2002: 13.8; 2000–2002 cumulative: 56.9.
- The sectoral evidence supports the conclusion that institutional and banking-system factors, rather than demand-side (borrowing-sector) factors alone, are responsible for differences between groups.

### Reasons for trends and assessment of future trends (qualitative conclusions and methodology)
- The paper discusses multiple possible drivers of BCPS growth differences: institutional reforms, banking system development, foreign financing, deposit shifts, write-offs, and prudential operations (sterilization, reserve requirement changes).
- The authors regard a full econometric dynamic model encompassing equilibrium BCPS levels and transition speeds as too ambitious given the short time dimension of the country panel. Instead, they:
  - Examine available statistics informally to identify likely explanatory factors.
  - Assess how those factors might evolve in the future.
  - Estimate a model of equilibrium BCPS for a sample of non-transition countries and use it to project equilibrium BCPS for CEB countries.
- Key qualitative lessons:
  - No clear evidence that faster BCPS growth is primarily a catching-up from low initial bank intermediation: early birds started from a lower pre-turning-point average BCPS ratio (14.2 percent) but rose to levels averaging 29 percent in 2001 and 32.6 percent in 2002—above sleeping beauties.
  - No clear evidence that early birds’ BCPS expansion was mainly financed by foreign liabilities: the average annual increase in net foreign liabilities (percent of GDP) was almost zero in early birds.

### Prudential and policy implications highlighted
- Rapid BCPS growth in some countries (including some sleeping beauties where recent BCPS growth exceeded GDP growth) raises prudential supervision concerns: high growth rates, even from a low base, require close examination of supervisory capacity.
- The composition of balance sheet financing matters: bank deposits, net foreign liabilities, changes in other net liabilities (often driven by central bank sterilization operations or changes in required reserves), and shifts in government deposits at commercial banks can all materially affect BCPS dynamics and risks.

*Source: _wp03213 (selected tables and figures, 1992–2002 data as presented in the source content).*

### 0.6 percentage points of GDP in the sleeping beauties (Tables 3 and 4). Indeed, in some

### _wp03213 - 0.6 percentage points of GDP in the sleeping beauties (Tables 3 and 4). Indeed, in some

### Factors Affecting Bank Credit Growth: Cross-country patterns and statistics
- Early birds (selected values from Table 6):
  - Average BCPS Ratio Before Turning Point: 14.2
  - Average BCPS Ratio in 2001: 29.0
  - Average BCPS Ratio in 2002: 32.6
  - Average Fiscal Deficit (percent of GDP): 2.2
  - Country examples (BCPS Ratio Before / 2001 / 2002 / Average Fiscal Deficit):  
    - Bulgaria: 4.6 / 12.3 / 15.6 / 0.5  
    - Croatia: 24.8 / 38.8 / 45.6 / 3.2  
    - Estonia: 10.7 / 39.6 / 46.0 / 0.4  
    - Hungary: 20.1 / 28.1 / 29.3 / 5.0
- Late risers (selected values from Table 6):
  - Average BCPS Ratio in 2001: -- / Average BCPS Ratio in 2002: 16.9 / Average Fiscal Deficit: 3.5
  - Example: Serbia and Montenegro: -- / 12.4 / 14.7 / 5.0
- Sleeping beauties (selected values from Table 6):
  - Average BCPS Ratio Before Turning Point: 26.7
  - Average BCPS Ratio in 2001: 21.9
  - Average BCPS Ratio in 2002: 20.9
  - Average Fiscal Deficit (percent of GDP): 4.5
  - Country examples (BCPS Ratio Before / 2001 / 2002 / Average Fiscal Deficit):  
    - Albania: 3.3 / 4.2 / 4.9 / 9.2  
    - Czech Republic: 53.3 / 43.5 / 42.6 / 3.1  
    - Slovak Republic: 48.5 / 37.5 / 31.5 / 4.4

- Key empirical observations:
  - The growth in the BCPS ratio was essentially driven by domestic saving flows; external borrowing was sizable in some countries (notably Estonia and Latvia).
  - Deposit-to-GDP ratio increased in both early birds and sleeping beauties, but the BCPS ratio rose mainly in early birds—financial deepening alone was not sufficient.
  - Crowding-in occurred: in all but one early bird (Latvia) bank credit to the public sector ratio declined, while it increased on average for sleeping beauties.
  - Average general government deficit-to-GDP ratio during the period of rising BCPS ratio:
    - Early birds: 2¼ percent of GDP (reported as 2.2 percent in Table 6)
    - Sleeping beauties (last five years): 4½ percent of GDP (reported as 4.5 percent in Table 6)
  - Progress on structural reforms and banking privatization:
    - Early birds had higher average EBRD “transition index” and faster reduction in asset share of state-owned banks.
    - Asset share of state-owned banks averaged 25 percent in early birds by end-1990s (fell below 16 percent by end-2001, Table 8); sleeping beauties had over 41 percent in 1999.
  - Foreign ownership shares of banks exceeded 80 percent in a number of countries (Table 8).

### The role of transition, privatization, and legal framework (Tables 7 and 8)
- EBRD transition indexes (1989–2002 averages, Table 7):
  - Early birds average (selected indexes): Banking 2.65 / Sector and Interest Rate Liberalization 2.61 / Legal Transition Indicators 3.52
  - Sleeping beauties average: Banking 2.40 / Sector and Interest Rate Liberalization 2.26 / Legal Transition Indicators 3.20
  - Observations:
    - Early birds generally rank higher on transition indexes; differences larger for banking reform and interest rate liberalization.
    - Legal transition indicators (creditors’ rights) are stronger on average for early birds, though average differences are small and not significant at 10 percent level in some comparisons.
- Privatization and foreign ownership (Table 8, selected figures):
  - Early birds: Average asset share of state-owned banks (1999): 25.0; 2000: 14.6; 2001: 15.7.
  - Sleeping beauties: Average asset share of state-owned banks (1999): 41.3; 2000: 38.6; 2001: 22.9.
  - Selected country examples (Asset Share of State-Owned Banks 1999/2000/2001 ; Asset of Banks Controlled by Nonresidents (year)):
    - Bulgaria: 50.5 / 19.8 / 19.9 ; 74.0 (2001)
    - Czech Republic: 23.1 / 28.2 / 3.8 ; 84.0 (2001)
    - Slovak Republic: 50.7 / 49.1 / 4.9 ; 88.6 (2001)
- Subjective constraints on credit growth (Table 10):
  - Factors regarded as relevant (percentage of countries): Weak creditor rights 87.5; Dearth of trustworthy borrowers 50.0; Need to complete bank restructuring 50.0; Poor accounting standards 37.5; Direct financing from abroad 25.0.

### Microeconomic outlook for BCPS growth: bank ownership, profitability, and credit risk
- Ownership:
  - Ongoing privatization (including among sleeping beauties: Czech Republic and Slovakia privatized in 2001) is expected to boost credit growth within lags observed elsewhere (one-two years).
  - Privatization attempts in Albania and Romania in 2002 failed due to market conditions, but intentions remain.
- Profitability (Table 9, Bankscope 2001 data):
  - Early birds average: ROA 1.47 ; ROE 13.98 ; ROER 7.17
  - Sleeping beauties average (excluding Albania and Czech Republic as noted): ROA 0.88 ; ROE 8.78 ; ROER -4.96
  - EU average (2001): ROA 0.60 ; ROE 10.60 ; ROER 7.85
  - Country examples:
    - Estonia: ROA 2.70 / ROE 20.50 / ROER 13.94
    - Latvia: ROA 1.85 / ROE 21.52 / ROER 18.58
    - Czech Republic: ROA -3.66 / ROE -60.84 / ROER -62.60
    - Albania: ROA 2.99 / ROE 49.42 / ROER 44.90 (high ROER explained by market imperfections)
  - Implication: Low bank profitability in many CEB countries relative to EU averages suggests banks will seek to expand loans to boost profitability as loan market risk conditions improve.
- Credit risk:
  - Weak creditor protection and inefficient legal processes were identified as key constraints in many Financial Systems Stability Assessments.
  - Progress is occurring, and integration with Western Europe is expected to drive convergence in real estate prices, raising value of a key form of collateral.

### Econometric model and equilibrium BCPS ratios (Table 12)
- Random effects panel regression (nontransition countries, 1973–76) — selected coefficients and significance:
  - Public Debt / GDP: -0.136 (standard error 0.032, t-statistic -4.29) ***
  - Log (GDP per capita): 0.084 (0.022, 3.74) ***
  - High-inflation dummy*(1/Inflation rate - 1/Inflation threshold): 0.004 (0.001, 4.66) ***
  - Liberalization index: 0.190 (0.034, 5.64) ***
  - Accounting: 0.007 (0.003, 2.74) ***
  - German legal origin: 0.287 (0.086, 3.34) ***
  - Number of observations: 468; R-squared: 0.66
  - Inflation threshold used: 4 percent.
- Out-of-sample predicted values for BCPS ratio, CEB countries (Panel B, Table 12) — Actual BCPS ratio (2002) versus Predicted Value and Absolute Deviation (a - b):
  - Albania: Actual 4.9 / Predicted 24.6 / Absolute Deviation -19.7
  - Bosnia: Actual 21.9 / Predicted 18.5 / Absolute Deviation 3.5
  - Bulgaria: Actual 15.6 / Predicted 52.6 / Absolute Deviation -37.0
  - Croatia: Actual 45.6 / Predicted 49.1 / Absolute Deviation -3.5
  - Czech Republic: Actual 42.6 / Predicted 69.3 / Absolute Deviation -26.7
  - Estonia: Actual 46.0 / Predicted 75.4 / Absolute Deviation -29.4
  - Hungary: Actual 29.3 / Predicted 70.5 / Absolute Deviation -41.2
  - Latvia: Actual 24.8 / Predicted 76.7 / Absolute Deviation -51.9
  - Lithuania: Actual 11.6 / Predicted 68.1 / Absolute Deviation -53.9
  - Macedonia, FYR: Actual 17.1 / Predicted 44.2 / Absolute Deviation -27.1
  - Poland: Actual 28.1 / Predicted 70.4 / Absolute Deviation -42.3
  - Romania: Actual 8.3 / Predicted 58.0 / Absolute Deviation -49.7
  - Serbia: Actual 14.7 / Predicted 21.0 / Absolute Deviation -6.3
  - Slovak Republic: Actual 31.5 / Predicted 59.9 / Absolute Deviation -28.4
  - Slovenia: Actual 38.4 / Predicted 63.8 / Absolute Deviation -25.4
- Interpretation:
  - BCPS ratios in almost all CEB countries are still significantly undersized relative to the model’s predicted equilibrium.
  - Average deviation from equilibrium is similar across sleeping beauties and early birds (about 33-30 percentage points of GDP).
  - BCPS ratio close to equilibrium only in Croatia and Serbia and Montenegro (SM), with Bosnia somewhat above equilibrium following a 6¾ percentage points increase in 2002.

### Macroeconomic outlook and implications of fast BCPS growth (fiscal and capital flows)
- Crowding-in and fiscal consolidation (Table 13):
  - 2002 fiscal deficits and medium-term targets: On average, fiscal deficit ratio targeted to decline by 1¼ percentage point of GDP during 2003-05.
  - Table 13 2002 outturns and targets (selected examples, percent of GDP):
    - Albania: 2002 outturn 6.9 ; Targets 2003 6.6 / 2004 6.0 / 2005 5.8
    - Bosnia and Herzegovina: 2002 outturn 4.3 ; Targets 2003 2.2 / 2004 2.0 / 2005 1.7
    - Bulgaria: 2002 outturn 0.6 ; Targets 2003 0.7 / 2004 0.5 / 2005 0.0
    - Average (excluding Croatia): 2002 outturn 3.8 ; Targets 2003 3.2 / 2004 2.9 / 2005 2.5
- Capital inflows and capital gap estimates (Table 14, selected values):
  - Nominal interest rate differential (1999–2002 average), Parity Real Interest Rate, Estimated Capital Gap (percent of GDP):
    - Albania: 7.4 / 3.4 / 1131.1
    - Bulgaria: 1.3 / -2.1 / 707.9
    - Croatia: 0.2 / -0.8 / 502.2
    - Czech Republic: 1.1 / -2.8 / 224.1
    - Hungary: 7.8 / -6.8 / 282.6
    - Poland: 11.4 / -2.7 / 427.5
    - Romania: 42.4 / -14.6 / --
    - Slovak Republic: 5.5 / -4.3 / 347.7
    - Slovenia: 6.3 / -0.2 / 157.7
  - Interpretive notes:
    - Two-thirds of CEB countries were net lenders to the rest of the world in the average of 2002; only five countries were net borrowers (BiH, Czech Republic, Estonia, Latvia, and Poland).
    - Capital inflows are expected to increase due to euro-convergence, lower risk premia, expectations of real exchange rate appreciation, and interest rate differentials.
    - The capital gap estimates reflect potential for real investment from abroad needed to equalize capital-labor ratios with Germany (α assumed equal to ⅓; German capital-output ratio set at 1.74).

### Macroeconomic implications and policy responses
- On interpretation of rising BCPS ratios:
  - A rising loans-to-GDP ratio should not per se be taken as a sign of economic overheating; overheating arises when the growth of the demand component financed by loans exceeds potential output growth and initial loan stock conditions satisfy inequality conditions (see equation (1) and discussion).
  - Empirical example: In 2002, Bosnia, Serbia and Montenegro, and Croatia saw credit surges matched by sharp current account deteriorations; Bulgaria’s 2002 credit acceleration had no obvious macroeconomic implications.
- Transmission channels of fast credit growth:
  - (i) Increased propensity of banks to lend (given monetary base) — can be expansionary by lowering borrowing costs relative to bond market (IS rightward shift).
  - (ii) Increased capital inflows through banks — can increase base money via foreign exchange intervention, magnifying expansion.
  - Expansionary effects depend on elasticity of base money demand to interest rates and on banks’ secondary liquidity (holdings abroad and central bank paper).
- Historical parallel:
  - CEB countries’ experience may resemble that of countries "at the periphery of pre-euro Europe" in the late 1990s: convergence of interest rates matched by surges in bank loans and external deficits despite fiscal tightening.
- Policy instruments available to authorities in response to a surge in BCPS and aggregate demand:
  - Standard monetary policy tools.
  - Fiscal policy.
  - Controls on capital inflows.
  - Administrative credit ceilings.
  - Supervisory measures.
- Practical approach:
  - Recent episodes show authorities have used combinations of these instruments rather than a single response.

*Source: Appendix I.*

### introduction at the end of 1998 is included in Schadler and others (2003), from which Figures

### _wp03213 - introduction at the end of 1998 is included in Schadler and others (2003), from which Figures

### Monetary Policy
- Overheating pressures can be offset through open market operations or equivalent measures to contain base money growth, including offsetting a shift of banks out of their (base money-creating) secondary liquidity.
- With the exclusion of Bulgaria, Bosnia and Herzegovina, Estonia, and Lithuania (which have currency boards in place), almost all other CEB central banks now have one or more money market instruments to control liquidity.
- Feasibility of using money market instruments depends on the elasticity of foreign capital inflows to domestic interest rates (extent of capital mobility).
- Empirical evidence for sterilized intervention effectiveness in CEB countries (mainly Hungary and Czech Republic in the second half of the 1990s):
  - Studies generally found offset coefficients well below unity, signaling short-run effectiveness of sterilized intervention.
  - Specific offset coefficient findings referenced:
    - Czech Republic: earlier studies reported offset coefficients of ⅓-⅔; Christensen (2003) finds a coefficient below 0.15; OECD (1996) refers to a coefficient range of 0.4-0.65 estimated by the Czech National Bank.
    - Hungary: Fernández-Ansola and Moghadam (1996) find an offset coefficient of 0.5 for the period preceding March 1995; for the period following the introduction of the crawl they find a coefficient close to unity, but with a large standard error.
- Limitations and costs of sterilized intervention:
  - Financial cost over time.
  - If sterilized intervention keeps domestic interest rates higher than abroad, it perpetuates capital inflows and builds up a stock of short-term assets that can suddenly move abroad or be used to fuel domestic credit growth.
  - Sterilized intervention in some CEB countries has allowed banks to build up large secondary liquidity reserves that could potentially fuel credit growth or a speculative attack against the currency.
- Historical illustrations:
  - During 1996 sterilized intervention was particularly strong in Hungary—the net domestic assets of the central bank declined by some US$5 billion, mostly reflecting sterilization of inflows during that year.
  - During the speculative attack of January 16, 2003 the purchases of foreign exchange by the National Bank of Hungary (which were not sterilized at that time) amounted to several billions of U.S. dollars in a single day.
- Policy trade-offs:
  - High elasticity of capital inflows to interest rates would not prevent using monetary policy where authorities are willing to accept a sizable exchange rate appreciation.
  - Many CEB countries prefer exchange rate stability, making acceptance of strong appreciations unlikely.
  - Two reasons for reluctance to accept appreciations:
    - Underlying external current account positions are likely to be in deficit; countries with sizable gross external debt may be unwilling to see the external deficit widen further.
    - Risk of non-reversibility of an exchange rate appreciation for countries on the way to euro-adoption, since the market exchange rate may be used as the entry rate and could “freeze” an overvalued position.

### Fiscal Policy
- Fiscal policy is the obvious tool to avoid overheating when monetary policy is constrained by the desire to avoid exchange rate appreciation.
- In 2002, 60 percent of the countries in the sample had a general government deficit of over 4 percentage points of GDP.
- A fiscal tightening would be consistent with the medium-term fiscal consolidation path required by the current formulation of the Stability and Growth Pact.
- Contrast with Asia in the run-up to the late 1990s crises: those Asian countries had fiscal positions close to balance or in surplus, making fiscal tightening less available then.
- Practical experience and constraints:
  - Governments of Croatia, Bosnia and Herzegovina, and Serbia and Montenegro regarded fast credit growth and rapidly widening external current account deficits as sufficient reason to tighten fiscal policy in 2003.
  - In two of these countries, and in the run-up to general elections, the fiscal response was small.
  - In all of them authorities also relied on monetary policy instruments, including administrative measures.
  - Footnote: In the case of SM, eventually there was no tightening as the 2002 fiscal outturn was better than expected when the 2003 budget was formulated.

### Capital Controls
- Capital controls have been removed in most CEB countries.
- Reintroduction of capital controls conflicts with long-term commitments to freedom of capital mobility from OECD and EU membership, but:
  - OECD and EU rules do not seem to be inconsistent with temporary reintroduction of controls on capital inflows in principle.
- Practical considerations limiting use of capital controls:
  - Effectiveness is uncertain in a world of increasingly sophisticated financial transactions; evidence is mixed and effectiveness should not be assumed.
  - Reintroduction might be seen as a step back in the transition process, reducing attractiveness of capital controls for many CEB countries.

### Credit Ceilings
- Bank-by-bank credit ceilings (often with penalties such as mandated below-market investments for banks with high credit growth) were common in the 1960s and 1970s.
- Main rationale historically:
  - Ensure effective control of credit growth and aggregate demand in the absence of a smoothly functioning monetary policy transmission mechanism.
  - Affect credit growth and aggregate demand without causing a generalized increase in interest rates, particularly on government paper.
- Recent usage examples noted in the text:
  - Credit ceilings used in Croatia.
  - A rise in reserve requirements used in Bosnia and Herzegovina and Serbia and Montenegro.

### Specific Policy-relevant Observations and Examples
- Distortion removal to moderate unjustified credit growth:
  - Example: subsidy scheme on housing in Hungary—removal would moderate demand for housing and mortgage credit.
  - Existing subsidized credit schemes in Croatia similarly contribute to credit growth.
- Administrative and non-monetary measures have been used alongside fiscal and monetary responses in several CEB countries.

*Source: _wp03213 - introduction at the end of 1998 is included in Schadler and others (2003), from which Figures*

### introduction of controls on short-term capital inflows that are fueling an unsustainable

### _wp03213 - introduction of controls on short-term capital inflows that are fueling an unsustainable

### Controls on capital inflows (Table 15 highlights)
- Article 7 of the Capital Movements Code (OECD) allows OECD members to suspend capital mobility for up to eighteen months if “the balance of payments ... develops adversely at a rate and in circumstances, including the state of its monetary reserves, which is considered serious.”
- Summary of country practices (as reported in Table 15, excluding real estate purchases):
  - Albania: No controls
  - Bosnia and Herzegovina: No controls
  - Bulgaria: No controls, although there are some registration requirements
  - Croatia: Controls on purchases of short-term securities (to be removed four years after the coming into effect of the Stabilization and Association Agreement with the EU). The central bank can also introduce short-term “safety measures” if capital movements cause or threaten to cause “serious difficulties in the implementation of monetary or foreign exchange policies.”
  - Czech Rep.: No controls
  - Estonia: No controls
  - Hungary: No controls
  - Latvia: No controls
  - Lithuania: No controls
  - Macedonia, FYR: There are fees on the purchase of short-term securities by nonresidents
  - Poland: No controls
  - Romania: Nonresidents are not allowed to purchase t-bills and, in principle, other money market instruments
  - Serbia and Montenegro: Purchases of securities by nonresidents are restricted
  - Slovakia: No controls
  - Slovenia: No controls

### Credit ceilings and prudential instruments
- Credit ceilings
  - Used in some cases to constrain domestic credit growth when countries are unwilling to allow exchange rate appreciation and cannot tighten domestic monetary conditions in the presence of an open capital account. (Footnote: For credit ceilings to be effective in this case, it would, of course, be necessary that bank customers cannot borrow directly from abroad.)
  - Reintroduction/instances:
    - Greece: Credit ceilings were temporarily reintroduced in response to strong capital inflows, stimulated by Greece’s expected entry in the euro-zone in 1999.
    - Croatia: Imposed in early 2003 following a surge in BCPS, also in the context of strong capital inflows.
  - Effectiveness and costs:
    - Few studies (Cottarelli et al. (1986)) conclude that when effective (i.e., when not circumvented), credit ceilings significantly distort bank competition and damage banking market efficiency, and thus are inferior to controls on capital inflows.
    - To avoid forms of financial dirigisme reminiscent of planning, credit ceilings are unlikely to become a permanent feature of CEB countries, although they may be used occasionally in response to credit booms.
- Prudential instruments
  - Potential measures include:
    - (i) pro-cyclical capital requirements;
    - (ii) restrictive rules on collateral; and
    - (iii) higher provisioning for banks with high credit growth.
  - Distinction from credit ceilings:
    - Prudential tools aim at increasing the solidity of banks and address both macroeconomic and prudential aspects of credit booms, unlike credit ceilings which may force banks with excessive credit growth to invest in below-market-yield assets, potentially reducing bank solidity.
  - Implementation:
    - Prudential regulation has been strengthened—and on-site supervision intensified—in response to credit booms in a number of CEB countries (including Bosnia and Herzegovina, Croatia, Serbia and Montenegro).

### Fast credit growth and financial stability — concerns and evidence
- IMF board cautions:
  - The IMF board has on several occasions cautioned CEB countries against the risks to bank portfolio quality associated with too rapid growth in private sector credit, and called for increased supervisory attention.
- Literature linking lending booms to banking crises:
  - Historical episodes where strong credit growth preceded crises: Argentina 1980; Chile 1982; Sweden, Norway, and Finland 1992; Mexico 1994; Thailand, Indonesia, and Korea 1997.
  - Contributing factors in some cases: deregulation increasing bank competition, reduced reserve requirements, and capital account liberalization providing liquidity.
  - Econometric findings:
    - Demirgüç-Kunt and Detragiache (1997): Lending booms precede banking crises. Example: for the 1994 Mexican crisis, a 10 percent increase in the initial value of lagged credit growth would have increased the probability of a crisis by 5½ percent.
    - Hardy and Pazarbasioglu (1998): Robust evidence that credit to the private sector follows a boom-and-bust pattern ahead of banking crises.
    - Kaminsky and Reinhart (1999): Growth rate of the BCPS ratio accelerates markedly as banking crises or twin crises approach, remaining well above growth in tranquil times.
    - Gourinchas, Valdes, and Landerretche (1999): Probability of having a banking crisis significantly increases after lending boom episodes; the conditional incidence depends critically on the size of the boom. However, the conditional probability remains below 20 percent, indicating most lending booms are not followed by banking crises.
  - Theoretical models:
    - Financial accelerator models based on collateral: an exogenous shock raises collateral values, relaxes borrowers’ constraints, expands credit, pushes up asset prices further, and can lead to deterioration in bank portfolio quality and a reversal when the bubble bursts (references include Kiyotaki and Moore (1997), Aghion, Bacchetta, and Banerjee (2001), Dell’Ariccia and Marquez (2003)).
  - Alternative view:
    - Literature emphasizing financial development and growth (Rajan and Zingales, 1998; Beck, Levine, and Loayza, 2000; Levine, Beck, and Loayza, 2001) suggests rapid credit growth can reflect financial deepening and allocate savings to productive enterprises. Some fast lending phases may represent permanent shifts rather than transitory cycles; policy response should depend on whether the episode is expected to be transitory or a permanent shift.

### Detecting lending booms in CEB countries (methodology and findings)
- Methodology
  - Applied Gourinchas et al. (2001) approach: define a lending boom as an episode where the BCPS ratio deviates from a rolling, backward-looking, country-specific stochastic trend estimated by an Hodrick-Prescott (HP) filter.
  - Advantages: rolling backward-looking trend uses only information available up to the time evaluated, allowing examination of recent bank credit evolution.
  - Caveats:
    - Short range of BCPS time series makes results sensitive to individual observations.
    - For transition economies that have not experienced “tranquil times,” the HP filter of past BCPS realizations may be an imperfect proxy for the expected ratio, potentially leading to overestimation or underestimation of credit booms.
    - Specific data issues: early years’ observed decline in BCPS ratios may be affected by write-offs (example: Bulgaria), complicating trend identification.
  - Implementation detail:
    - For each country, construct the backward-looking stochastic trend for the BCPS series for each year by applying an HP filter using data from the beginning of the sample to that year. The first five years of data are used to construct the first trend.
- Comparison with crisis countries
  - In years preceding banking crises, control-group countries often experienced annual increases in the BCPS ratio often exceeding 5 percentage points, and sometimes as high as 10 percentage points.
  - In CEB economies, increases are generally around 2 percentage points and only in a few cases have exceeded 5 percentage points.
  - The initial level of the BCPS ratio was much higher in the control group than in CEB countries.
- Empirical classification (summary)
  - Using BCPS data and the HP trend for 2002:
    - All early birds, but Slovenia, had BCPS ratios in excess of their trend component in 2002.
    - Most sleeping beauties experienced negative or close to zero deviations in 2002, including relatively more developed financial systems such as the Czech Republic and the Slovak Republic.
    - Lack of consistent data prevented trend computation for Bosnia and Herzegovina and Serbia and Montenegro.
- Thresholds for classifying a lending boom
  - Gourinchas et al. (2001) thresholds used as reference:
    - Absolute deviation thresholds: 4.8 percent to 6.4 percent.
    - Relative deviation thresholds: 24.9 percent to 31.1 percent.
  - According to those boundaries, only one CEB lending boom could be detected in 2002: Bulgaria had a relative deviation of (text ends here in source).

### Key statistics and numeric points preserved from the source
- Article 7 suspension duration: up to eighteen months
- Croatia: controls to be removed four years after the coming into effect of the Stabilization and Association Agreement with the EU
- Greece expected euro-zone entry stimulated inflows in 1999
- Croatia imposed credit ceilings in early 2003
- Demirgüç-Kunt and Detragiache (1997) example: 10 percent increase in lagged credit growth → increases crisis probability by 5½ percent (Mexican 1994 case)
- Historical BCPS growth preceding crises: often exceeding 5 percentage points, sometimes as high as 10 percentage points
- Typical BCPS increases in CEB economies: around 2 percentage points, only a few cases exceed 5 percentage points
- Conditional incidence of banking crisis after lending booms: remains below 20 percent
- Gourinchas et al. (2001) suggested thresholds: absolute deviations 4.8 percent to 6.4 percent; relative deviations 24.9 percent to 31.1 percent

*Source: _wp03213 - introduction of controls on short-term capital inflows that are fueling an unsustainable (IMF PDF chapter/section).*

### 24.1 percent that is slightly below the lower bound of the threshold range. For all the other

### _wp03213 - 24.1 percent that is slightly below the lower bound of the threshold range. For all the other

### BCPS deviations from trend in 2002 (aggregate bank credit to private sector)
- Table 16 reported BCPS ratios, HP-estimated BCPS ratios, absolute deviations (a-b), and relative deviations (a-b)/a in Percent for 2002:
  - Bulgaria: BCPS Ratio (a) 15.6; HP-Estimated BCPS Ratio (b) 11.8; Absolute Deviation (a-b) 3.8; Relative Deviation (a-b)/a in Percent) 24.1
  - Croatia: BCPS Ratio 45.6; HP-Estimated BCPS Ratio 43.5; Absolute Deviation 2.1; Relative Deviation 4.7
  - Estonia: BCPS Ratio 46.0; HP-Estimated BCPS Ratio 45.6; Absolute Deviation 0.4; Relative Deviation 1.0
  - Hungary: BCPS Ratio 29.3; HP-Estimated BCPS Ratio 26.0; Absolute Deviation 3.3; Relative Deviation 11.2
  - Latvia: BCPS Ratio 24.8; HP-Estimated BCPS Ratio 20.4; Absolute Deviation 4.4; Relative Deviation 17.6
  - Poland: BCPS Ratio 28.1; HP-Estimated BCPS Ratio 27.1; Absolute Deviation 1.0; Relative Deviation 3.5
  - Slovenia: BCPS Ratio 38.4; HP-Estimated BCPS Ratio 39.0; Absolute Deviation -0.6; Relative Deviation -1.7
  - Lithuania: BCPS Ratio 14.2; HP-Estimated BCPS Ratio 11.4; Absolute Deviation 2.8; Relative Deviation 19.7
  - Albania: BCPS Ratio 4.9; HP-Estimated BCPS Ratio 4.2; Absolute Deviation 0.7; Relative Deviation 14.3
  - Czech Republic: BCPS Ratio 42.6; HP-Estimated BCPS Ratio 45.2; Absolute Deviation -2.6; Relative Deviation -6.2
  - Macedonia, FYR: BCPS Ratio 17.1; HP-Estimated BCPS Ratio 15.9; Absolute Deviation 1.2; Relative Deviation 6.9
  - Romania: BCPS Ratio 8.3; HP-Estimated BCPS Ratio 8.8; Absolute Deviation -0.5; Relative Deviation -5.6
  - Slovak Republic: BCPS Ratio 31.5; HP-Estimated BCPS Ratio 36.6; Absolute Deviation -5.1; Relative Deviation -16.3

- Interpretation:
  - Bulgaria’s relative deviation of 24.1 percent is “slightly below the lower bound of the threshold range” referenced in the text; most other countries’ deviations do not meet the definitions of a lending boom per Gourinchas et al. (2001), though Latvia is a close case.
  - Bulgaria’s BCPS ratio in 2002 remains well below its pre-1997 level and below its estimated “equilibrium” (Section III), implying part of the recent credit surge may reflect “catching-up” and medium-run financial deepening rather than a boom-bust episode.

### Credit to households (BCH): sectoral developments and risks
- Table 17 (2002 BCH deviations from trend) key figures:
  - Bulgaria: BCH Ratio 3.2; HP-Estimated BCH Ratio 3.1; Absolute Deviation 0.1; Relative Deviation 1.7
  - Croatia: BCH Ratio 20.5; HP-Estimated BCH Ratio 19.1; Absolute Deviation 1.4; Relative Deviation 6.7
  - Czech Republic: BCH Ratio 6.9; HP-Estimated BCH Ratio 5.7; Absolute Deviation 1.2; Relative Deviation 17.9
  - Estonia: BCH Ratio 10.4; HP-Estimated BCH Ratio 9.7; Absolute Deviation 0.6; Relative Deviation 6.2
  - Hungary: BCH Ratio 6.8; HP-Estimated BCH Ratio 4.4; Absolute Deviation 2.4; Relative Deviation 34.9
  - Latvia: BCH Ratio 6.2; HP-Estimated BCH Ratio 5.4; Absolute Deviation 0.8; Relative Deviation 12.3
  - Macedonia, FYR: BCH Ratio 2.0; HP-Estimated BCH Ratio 1.8; Absolute Deviation 0.2; Relative Deviation 10.8
  - Poland: BCH Ratio 11.2; HP-Estimated BCH Ratio 11.6; Absolute Deviation -0.4; Relative Deviation -3.3
  - Romania: BCH Ratio 0.7; HP-Estimated BCH Ratio 0.5; Absolute Deviation 0.2; Relative Deviation 22.4
  - Slovak Republic: BCH Ratio 5.3; HP-Estimated BCH Ratio 4.5; Absolute Deviation 0.8; Relative Deviation 15.6
  - Slovenia: BCH Ratio 10.9; HP-Estimated BCH Ratio 12.5; Absolute Deviation -1.6; Relative Deviation -15.0

- Hungary: detailed BCH rolling deviations from trend (Table 18):
  - 1999: BCH 2.7; HP-Estimated BCH 2.0; Absolute Deviation 0.7; Relative Deviation 24.1
  - 2000: BCH 3.5; HP-Estimated BCH 2.4; Absolute Deviation 1.1; Relative Deviation 32.1
  - 2001: BCH 4.8; HP-Estimated BCH 3.1; Absolute Deviation 1.6; Relative Deviation 34.5
  - 2002: BCH 6.8; HP-Estimated BCH 4.4; Absolute Deviation 2.4; Relative Deviation 34.9

- Observations and limits:
  - Credit to households is generally a small share of GDP in CEB countries (except Croatia), consistent with underdeveloped household credit in pre-transition economies.
  - Relative deviations can be sizable: Hungary’s BCH ratio exceeds the 24 percent threshold used for boom detection, indicating a household credit boom.
  - Short time-series and HP filter methodology can absorb sustained rapid growth into the trend, causing tests to miss booms; six countries (Bosnia, Bulgaria, Croatia, Estonia, Latvia, Macedonia) had high BCH growth in the past two years often above 20 percent but were not flagged by the test for this reason.

### Drivers of recent BCPS growth and projection
- Identified microeconomic drivers of BCPS increases:
  - Privatization in 2001–2003 and expected further privatization.
  - Attempts by CEB banks to raise profitability by expanding the share of bank loans in total bank assets (profitability low by EU standards).
  - Improvement in accounting standards, risk-assessment techniques, availability of collateral, overall creditworthiness, and protection of creditors’ rights.
- Identified macroeconomic drivers:
  - Fiscal consolidation aligning with EU fiscal constraints.
  - Capital inflows driven by declining risk premia as CEB countries enter the EU and approach euro participation.
  - Nominal interest rate differentials that remain sizable in many countries, expectations of real appreciation, and scarcity of capital relative to labor in the CEB area.
  - Growth of the banking market attracting capital inflows (example noted: Croatia).

- Outcome and equilibrium:
  - The econometric model linking BCPS to per capita income, public debt-to-GDP ratio, inflation, financial liberalization measures, and legal/regulatory/corporate culture predicts BCPS levels for CEB countries well above current levels, suggesting further increases toward estimated long-run equilibria.

### Macroeconomic and policy implications
- Macroeconomic effects:
  - Removal of microeconomic impediments to credit growth shifts the IS curve rightward (expansionary effect on aggregate demand).
  - With broad exchange rate stability, capital inflows reduce central banks’ ability to control money supply; sterilized intervention may be ineffective.
- Recommended policy responses:
  - Fiscal policy should play a critical role to avoid overheating; CEB countries are relatively well positioned to tighten fiscal policy given structural fiscal adjustment needs.
  - Other instruments discussed: administrative credit ceilings, controls on capital inflows, and enhanced prudential requirements.
  - Enhanced prudential supervision is important given rapid credit growth and the complexity of banking systems.
- Risks and caveats:
  - Up to 2002, BCPS growth in CEB countries was not quantitatively comparable to credit booms preceding banking crises elsewhere, but vigilance is warranted.
  - Tests are limited by short time series and transition dynamics, making it difficult to distinguish “abnormal” credit growth from catching-up.
  - Aggregate trends may understate risks from currency composition of balance sheets, maturity mismatches, and sectoral exposures (real estate, stock market).
  - Faster credit growth has been observed in subsectors, notably households, where more countries show abnormally high increases.

### Econometric model specification and estimation notes
- Model overview:
  - Random effects GLS panel regression estimated using annual data for 24 countries for 1973–1996.
  - Dependent variable: BCPSit = bank credit to the private sector as a ratio to GDP.
  - Key explanatory variables: public debt/GDP (PublicDebtYit), Log(GDP per capita) (Log(GDPPC)it), inflation with threshold/non-linear specification, liberalization index (LibIndexit), bank entry restrictions (EntryRestrictionsi), Accounting index, legal-origin dummies (including German legal origin), and break dummies for BCPS series.
- Functional form highlights:
  - Non-linear inflation effect modeled via threshold specification (following Khan, Senhadji, and Smith (2001)) with an inflation threshold set at 4 percent in the main specification; alternative specification used a 15 percent threshold with a high-inflation dummy and interaction terms.
- Fit and robustness:
  - Reported R-squared for the final specification: 0.66.
  - Alternative inflation specification (text table) reported R-squared 0.61 with 471 observations and significant coefficients:
    - Public debt / GDP coefficient: -0.164 (Standard Error 0.032; t-statistic -5.11; significance ***)
    - Log (GDP per capita) coefficient: 0.094 (Standard Error 0.024; t-statistic 4.02; significance ***)
    - Inflation rate coefficient: -0.843 (Standard Error 0.141; t-statistic -5.98; significance ***)
    - High-inflation dummy coefficient: -0.064 (Standard Error 0.017; t-statistic -3.71; significance ***)
    - High-inflation dummy * Inflation rate coefficient: 0.843 (Standard Error 0.141; t-statistic 5.98; significance ***)
    - Liberalization index coefficient: 0.189 (Standard Error 0.034; t-statistic 5.56; significance ***)
    - Bank entry requirements coefficient: -0.031 (Standard Error 0.027; t-statistic -1.15)
    - Accounting coefficient: 0.008 (Standard Error 0.003; t-statistic 2.68; significance ***)
    - German legal origin coefficient: 0.266 (Standard Error 0.105; t-statistic 2.54; significance **)
    - Trend coefficient: 0.003 (Standard Error 0.002; t-statistic 1.44)
    - Constant: -0.759 (Standard Error 0.290; t-statistic -2.62; significance ***)
  - Note on inflation dynamics in the alternative specification: at an inflation rate of 15 percent (the threshold in that specification), the negative contribution of inflation to BCPS rises up to 13 percentage points; once inflation exceeds the threshold, the negative contribution drops to 6 percentage points and remains constant thereafter — an implication the authors found awkward and thus preferred the main specification in Table 12.
- Data sources and construction (selected):
  - BCPS: geometric averages of end-of-year data. Source: International Financial Statistics, line 22d.
  - Public Debt ratio: Lane and Milesi-Ferretti (2002).
  - GDP per capita: U.S. dollars at PPP prices. Source: World Economic Outlook database.
  - Inflation Rate: CPI-based inflation rate. Source: World Economic Outlook database.
  - Liberalization Index: Abiad and Mody (2003) / constructed using 2002 EBRD Progress in Transition for out-of-sample prediction.
  - Accounting index: La Porta et al. (1998) (proxied for CEB out-of-sample predictions as described).
  - Bank entry restrictions index: Barth, Caprio, and Levine (2002) methodology.

*Source: Appendix I and II of the provided document.*

### REFERENCES

### REFERENCES

### Cited works

- Abiad, Abdul, and Ashoka Mody, 2003, “Financial Reform: What Shakes It? What Shapes It?,” IMF Working Paper 03/70.
- Barabás, Gyula, István Hamecz, and Judit Neményi, 1999, “Fiscal Consolidation, Public Debt Containment and Disinflation,” NBH Working Paper 1998/5 (May).
- Barth, R. James, Gerard Caprio, and Ross Levine, 2002, “Bank Regulation and Supervision: What Works Best?", Journal of Financial Intermediation, National Bureau of Economic Research, Working Papers Series (U.S.) No. 9323: pp. 1-46. (November).
- Beck, Thorsten, Aslí Demirgüç-Kunt, and Ross Levine, 2001, “Law, Politics, and Finance,” World Bank, Working Paper 2585
- Beck, Thorsten, Ross Levine, and Norman Loayza, 2000, “Finance and the Sources of Growth,” Journal of Financial Economics 58, pp. 261-300.
- Begg, David, 1998, “Pegging Out: Lessons From the Czech Exchange Rate Crisis,” Centre for Economic Policy Research, Discussions Papers Series No. 1956, International Macroeconomics and Transition Economies, (September), London.
- Boyd, John, Ross Levine, and Bruce Smith, 1996, “Inflation and Financial Market Performance,” Federal Reserve Bank of Minneapolis, Working Paper 573D.
- Calza, A., C. Gartner, and J. Sousa, 2003, “Modelling the Demand for Loans to the Private Sector in the Euro Area, Applied Economics, 35, pp. 107-117.
- Christensen, Jakob, 2003, Capital Inflows, Sterilization, and Private Bank Speculation: The Case of the Czech Republic, IMF unpublished.
- Cottarelli, Carlo, Giampaolo Galli, Paolo Marullo Reedtz and Giovanni Pittaluga, 1986, “Monetary Policy Through Ceilings on Bank Lending,” Economic Policy (October).
- Dell’Ariccia, Giovanni, and Robert Marquez, 2003, “Lending Standards and Banking Crises,” IMF unpublished.
- Demirgüç-Kunt, Aslí, and Enrica Detragiache, 1997, “The Determinants of Banking Crises: Evidence from Developing and Developed Countries”, World Bank Working Paper No. 1828.
- Drees, Burkhard, and Ceyla Pazarbaşioğlu, 1998, “The Nordic Banking Crises-Pitfalls in Financial Liberalization?” IMF Occasional Paper 161 (Washington, International Monetary Fund).
- European Bank for Reconstruction and Development (EBRD), 2002, Transition Report- Agriculture and Rural Transition, Economic Transition in Central and Eastern Europe and the CIS.
- Feldman, Robert A., and Nancy L. Wagner, 2002, “The Financial Sector in Central Europe: Interlinkages with Macroeconomic Policy and Performance on the Road to the EU,” paper presented at the 2002 EIB Conference on Economics & Finance, January 17, 2002, Luxemburg.
- Fernández-Ansola, Juan, and Reza Moghadam, Capital Flows, 1996, “Reverse Currency Substitution, and Sterilized Intervention,” IMF unpublished.
- Ghosh, Atish, Timothy Lane, Marianne Schulze-Ghattas, Aleš Buliř, Javier Hammann, and Alex Mourmouras, 2002, “ IMF-Supported Programs in Capital Account Crises,” IMF Occasional Paper 210, (Washington, International Monetary Fund).
- Goldfajn, Ilan, and Rodrigo O. Valdés, 1997, “Capital Flows and the Twin Crises: The Role of Liquidity,” IMF Working Paper /97/87 (Washington: International Monetary Fund).
- Gourinchas, Pierre-Olivier, Rodrigo Valdes, and Oscar Landerretche, 1999, “Lending Booms: Some Stylized Facts,” (October).
- Gourinchas, Pierre-Olivier, Rodrigo Valdes, and Oscar Landerretche, 2001, “Lending Booms: Latin America and the World,” Economia, pp. 47-99 (Spring). Italy.
- Hardy, Daniel C., and Ceyla Pazarbasioglu, 1991, “Leading Indicators of Banking Crises: Was Asia Different?” IMF Working Paper 98/91 (Washington, International Monetary Fund).
- Kaminsky, Graciela L., and Carmen M. Reinhart, 1999, “The Twin Crises: The Causes of Banking and Balance-of-Payments Problems,” American Economic Review, Vv89, N.3, pp. 473-500
- Khan, Mohsin, Abdelhak Senhadji, and Bruce Smith, 2001, “Inflation and Financial Depth,” IMF Working Paper 01/44 (Washington, International Monetary Fund).
- Kiss, Gergely, 2002, “The Housing Market and Financial Stability in the Light of EU Accession, National Bank of Hungary, Report on Financial Stability (December).
- Kiyotaki, Nobuhiro, and John Moore, 1997, “Credit Cycles,” Journal of Political Economy, Vol. 105, No. 2, The University of Chicago Press, pp. 211-248 (April).
- Klein, Michael and Giovanni Olivei, 1999, “Capital Account Liberalization, Financial Depth, and Economic Growth”, NBER Working Paper 7384.
- Köke, Jens, and Michael Schröder, 2002, “The Prospects of Capital Markets in Central and Eastern Europe,” Discussion paper No. 02-57, Center for European Economic Research.
- Lane, Philip R., and Gian Maria Milesi-Ferretti, 2002, “Long-Term Capital Movements,” NBER Macroeconomic Annual 2001, edited by B. Bernanke and K. Rogoff, Cambridge, MA: MIT Press.
- La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Schleifer, and Robert Vishny, 1998, “Law and Finance,” Journal of Political Economy, Vol. 106, p.1113.
- La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Schleifer, and Robert Vishny, 1999, “Quality of Government,” Journal of Law, Economics and Organization, Vol. 15 (1), pp 222-279
- Levine, Ross, Norman Loyaza, and Thorsten Beck (2000), “Financial Intermediation and Growth: Causality and Causes,” Journal of Monetary Economics, Vol 46 pp. 31-77.
- Lipschitz, Leslie, Timothy Lane, and Alex Mourmouras, 2002, “Capital Inflows to Transition Economies: Master or Servant?” IMF Working Paper 02/11 (Washington, International Monetary Fund).
- McKinnon, Ronald, 1973, Money and Capital in Economic Development, Washington, Brookings Institution
- Miller, Margaret, ed., 2003, Credit Reporting Systems and the International Economy, MIT Press, Cambridge, Mass.
- Neményi, J. 1996, “Capital Inflow, Macroeconomic Equilibrium, the Public Debt and the Profit and Loss of the National Bank of Hungary,” Acta Oeconomica, periodical of the Hungarian Academy of Sciences, Vol. 48 (3-4), pp. 241-270.
- Organization for Economic Cooperation and Development, 1996, OECD Economic Surveys, The Czech Republic, OECD Publications, Paris, France.
- Pissarides, Francesca, 2001, “Financial Structures to Promote Private Sector Development in South-Eastern Europe,” European Bank for Reconstruction and Development Working Paper No. 64 (June), London.
- Rajan, Raghuram G. and Luigi Zingales, (1998), “Financial Dependence and Growth,” American Economic Review, Vol. 88, Number 3, pp. 559-586 (June).
- Riess, Arman, Rien Wagenvoort, and Peter Zajc, 2002, “Practice Makes Perfect: A Review of Banking in Central and Eastern Europe,” EIB Papers, Vol 7. No. 1, pp. 31-53.
- Schadler, Susan, Paulo Drummond, Louis Kuijs, Suzana Murgasova, and Rachel van Elkan, 2003, “Adopting the Euro in Central Europe: an Examination of Policy Issues, IMF Occasional Paper, forthcoming.
- Shaw, Edward, 1973, Financial Deepening in Economic Development, New York, Oxford University Press.
- Thimann, Christian (ed), 2002, “Financial Sectors in EU Accession Countries,” European Central Bank, Frankfurt.
- Watson, C. Maxwell, 2003, “Financial Sector Efficiency in the EU Accession Countries of Central and Eastern Europe (CEE): The Impact of Policies and the Road Ahead,” paper presented at the 24th SUERF Colloquium, June 12-14, Tallinn.
- Williamson, John and Molly Mahar, 1998, “A Survey of Financial Liberalization,” Essays in International Finance, Nov. 1998, Princeton University.

*Source: _wp03213 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2003/_wp03213.pdf_