## _wp0557 - Appendix I

## Source details

**Canonical URL:** [_wp0557 - Appendix I](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp0557.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp0557.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp0557.pdf.json)

---

### I. INTRODUCTION — Policy question and mechanism
- Policy question: why policymakers treat debt devaluation (default or inflation-induced debt devaluation) as a last resort despite academic literature suggesting state-contingent capital levies can reduce distortionary taxation costs.
- Practical constraints emphasized:
  - Bonds with explicit state-contingent returns are rarely available.
  - Default or debt devaluation is typically seen as a policy of last resort and avoided in fiscal practice.
- Hypothesis: asymmetric information and legal/institutional imperfections help explain policymakers’ reluctance to use debt devaluation.
- Key mechanism:
  - Government nominal debt can serve a collateral-like function for financial intermediation.
  - Producers borrow from intermediaries that require a minimum fraction of loans to be covered by government debt as collateral.
  - Fiscal policy creates two distortions:
    - Positive labor tax → suboptimal work effort.
    - Collateral requirement → suboptimal level of capital.
  - Trade-off: volatile inflation (state-contingent real returns) improves the labor-effort distortion but erodes government debt used as collateral, undermining intermediation.
- Empirical motivation: concerns about damage to domestic banks have driven delayed restructurings in practice (International Monetary Fund (2002) evidence).

### II. GOVERNMENT DEBT IN DEVELOPING COUNTRIES — empirical findings and implications
- Domestically held government debt
  - Domestically held government debt is of greater and growing importance than external debt for almost every emerging market country in the sample.
  - The domestic share of marketable debt has risen and now represents in excess of 70 percent of overall debt in almost all major developing countries (conservative domestic classification).
  - Exceptions where domestic share stayed below 50 percent due to severe crises and devaluations: Russia and Argentina; Mexico (1995) and Thailand (1997) experienced such episodes but returned to domestic debt shares of over 70 percent afterwards.
  - Time coverage for referenced figures: quarterly data from 1994Q1 through 2003Q4.
  - Marketable debt-to-GDP ratios: developing countries’ marketable debt-to-GDP ratios were around 20 percent in the mid-1990s and exhibited a positive growth trend.
- Holders of government debt
  - Banks and pension funds are the major holders of government debt; banks are particularly important.
  - Bank exposure metric (average ratio 1998-2002 of financial institutions’ net credit to the government to their total assets):
    - Industrialized countries: ratio typically around 10 percent.
    - Developing countries: ratios of 20-40 percent are common; ratios above 50 percent observed in some large developing countries.
  - Typical bank capitalization ratio cited as less than 10 percent of assets, implying government debt losses can render banks insolvent (Argentina example referenced).
- Reserve requirements and financial repression
  - Reserve requirements "no longer exist in several key developing countries, and have been falling in others, to the point that they are actually no longer binding."
  - Binding reserve requirements tend to affect pension funds in some countries.
  - Conclusion: large holdings of government debt by banks must therefore be largely voluntary in many countries.
- Legal and institutional constraints on lending
  - Problems: difficulty, cost, and uncertainty in creating security interests over movable property; imperfect perfection and registration; slow and expensive enforcement.
  - United States empirical contrast:
    - 70 percent of all loans secured by collateral.
    - Almost 50 percent of all credit secured by movable collateral.
    - Movable capital represents 67 percent of U.S. industrial capital stock and 75 percent of gross business investment.
  - Argentina empirical contrast:
    - Around 10 percent of all bank credit is secured at all; practically none secured by movable collateral.
  - When acceptable collateral is available, private creditors offer six to eight times more credit, two to ten times longer maturity, and 30 to 50 percent lower interest rates (Asian Development Bank (2000)).
  - Example: loans for agricultural machinery in Argentina — lending rates around 60 percent; decomposition suggests:
    - about one sixth of the rate difference explained by macroeconomic risk,
    - one-tenth by the general difficulty of collecting against security (including mortgages),
    - three quarters of the difference due to problems with the legal framework for secured transactions against movable property (Fleisig (1998)).
  - Macro consequences of defective secured transaction systems include: inefficient savings allocation; inefficient capital allocation (preference for assets usable as collateral); fragile banks exposed to credit risks; inability to use securitization for risk diversification; negative effects on equity for small/start-up businesses.
  - Cost estimates from country studies: defective systems of lending security impose costs of 5-10 percent of annual GNP; reform costs are a small fraction of that.
- Quantitative legal‑institutional indices and relationship with bank holdings
  - Constructed three components representing segmented credit markets:
    - "Getting credit" — quality and accessibility of credit information and quality of collateral and bankruptcy laws.
    - "Enforcing contracts" — cost and time delays in collection of overdue debt.
    - "Registering property" — cost and time delays of transferring title to real estate.
  - Overall raw index: equal weights attached to the three categories; final index ranges between 0 and 100 where 0/100 corresponds to the worst/best raw index.
  - Another index for importance of government debt on financial institutions’ balance sheets scaled 0/100 (lowest/highest government debt to total assets ratio).
  - Figure 13 result: a very strong negative relationship between the quality of law and institutions and the amount of government debt that financial institutions choose to hold. Most top-left countries (poor legal/institution quality, high holdings) are developing; most bottom-right (better legal/institution quality, low holdings) are industrialized. Chile noted as best among developing countries.
- Role of government debt in financial development and risks of devaluation/inflation
  - Government debt provides financial infrastructure and an informational benchmark for private bond markets.
  - Preconditions for government bond market to support private bond markets:
    - Low macroeconomic volatility, especially inflation volatility.
    - Sufficient volume of government debt issuance to generate a yield curve.
  - Repo markets and collateralized transactions:
    - Repo markets typically accept only government debt as acceptable collateral.
    - Repo market sizes (Bankscope bank balance sheet data, 2002/03): Brazil around 7 percent (repo loans as fraction of total assets); Mexico around 5 percent; other sizable markets in Philippines, Poland, and India.
  - Risks: volatile inflation or debt devaluation undermines the benchmark role of government debt and damages further financial market development (Brazil 1999, Argentina 2001/2 referenced as impediments).

### III. SUMMARY, CONCLUSIONS, AND POLICY/RESEARCH IMPLICATIONS
- Main conclusions
  - Theoretical arguments favoring debt devaluations as state-contingent fiscal responses contrast with practitioners’ views, especially in developing countries.
  - Empirical evidence: domestically held government debt and bank holdings of government debt are central in developing countries.
  - High bank exposure to government debt combined with weak legal and institutional frameworks for private lending implies that debt devaluation or default would likely produce severe damage to domestic financial intermediation.
  - Keeping government debt a safe investment for banks is critical to:
    - support existing limited financial intermediation;
    - support the development of private securities and bond markets; and
    - prevent banking collapses that cause severe output contractions.
- Policy implication / research recommendation
  - Further develop optimal fiscal policy theory incorporating financial-market and institutional frictions to improve its usefulness for applied policy analysis.
  - Guillermo Calvo’s work provides conceptual inspiration for modeling costly unanticipated inflation and the trade-offs between taxation and inflation.

### APPENDIX I — data classification, exposure definitions, and reserve requirement country notes
- BIS classification and biases
  - BIS classification (Bank for International Settlements (2004)) distinguishes securities by:
    - In domestic currency, targeted at resident investors: Issues by Residents = Domestic; Issues by non-residents = International
    - In domestic currency, targeted at non-resident investors: Issues by Residents = International; Issues by non-residents = International
    - In foreign currency: Issues by Residents = International; Issues by non-residents = International
  - Classification covers the borrower side of securities issues, not actual ownership.
  - BIS is conservative in classifying domestic securities:
    - Issues by non-residents targeted at domestic investors, whether in domestic or foreign currency, are classified as international—even though many holders may be domestic.
    - Issues in domestic currency targeted at foreign investors are classified as international, despite evidence that ultimate holders of developing countries’ domestic currency debt are often domestic residents.
  - Data omissions and biases:
    - Notes and money market instruments issued by non-residents in domestic currency and in domestic markets are not included — biasing results against domestically held debt.
    - Local issues in foreign currency are generally classified as international; BIS makes exceptions for Argentina, Peru, and the Philippines.
    - Brady bonds are missing from the BIS dataset; authors merged BIS with a JP Morgan dataset on outstanding Brady bonds using the Merrill Lynch (1994) guide for classification.
- Bank holdings of government debt — definition and computation
  - Exposure measure defined as: (Financial institutions’ net credit to government) / (financial institutions’ net total assets)
  - Percentage figures computed from the IMF’s International Financial Statistics.
  - Averages are for 1998-2002, except for countries in the euro area where 1999-2003 was used.
  - Numerator: sum of all entries representing net credit to the public sector by deposit money banks, other banking institutions and nonbank financial institutions (the latter series do not exist for all countries).
  - Denominator: sum of the net total assets of these three groups, after canceling out credit items between them.
  - To obtain net figures, deduct from both numerator and denominator the sum of all entries representing credit by the public sector to these institutions.
  - Crisis-year exclusions from averages: Argentina 2002; Indonesia 1998.
- Reserve requirements — country notes and exact values reported
  - Brazil:
    - 1996: RR remains over 80 percent on demand deposits.
    - 1999: April: RR on demand deposits up to 100 percent, on other acc. claims up to 60 percent. October: RR on bank deposits reduced from 65 to 45 percent
    - 2001: RR on demand deposits 45 percent, RR on time deposits 10 percent.
    - 2004: No de jure RR (source: IMF Staff Economists).
  - China:
    - 2004: No de jure RR (source: IMF Staff Economists).
  - Czech Republic:
    - 2004: No de jure RR (source: IMF Staff Economists).
  - India:
    - 1990: Cash reserve ratio (CRR) 10 percent.
    - 1998: CRR raised to 10.5%. lowered to 10 percent raised to 11%.
    - 1999: CRR lowered to 10.5%, then to 10 percent and to 9 percent in November.
    - 2000: CRR lowered to 8 percent.
    - 2001: CRR lowered to 7.5 percent.
    - 2004: Statutory Liquidity Ratio (SLR): RR of 25 percent in the form of government bond holdings.
  - Indonesia:
    - 1999: No RR.
    - 2004: No de jure RR (source: IMF Staff Economists).
  - Jamaica:
    - 2004: Actual holdings of government bonds far exceed RR (source: IMF Staff Economists).
  - Korea:
    - 2001: RR=2.9 percent.
  - Lebanon:
    - 2004: No de jure RR (source: IMF Staff Economists).
  - Mexico:
    - 1991: Liquidity coefficients on foreign currency deposits up to 50 percent dep. on maturities.
    - 1995: Change from a daily zero RR system to cumulative zero RR over 28 day period.
    - 1999: RR=0 percent.
  - Philippines:
    - 1991: RR rationalized.
    - 1993: RR for bank deposits lowered from 25 to 22 percent.
    - 1994: RR lowered to 17 percent.
    - 1995: RR lowered to 15 percent
    - 1997: RR lowered to 13 percent
    - 1998: RR lowered to 10 percent.
    - 1999: RR=11 percent for commercial banks, RR=0% for rural and cooperative banks.
    - 2002: RR=9 percent
  - Poland:
    - 2004: No RR on banks. (source: IMF Staff Economists).
  - Russia:
    - 2004: RR=3.5 percent.
  - Turkey:
    - 1994: Comprehensive reform of RR. RR=10.8 percent.
    - 1995: RR=9.3 percent.
    - 1996: RR=8 percent for domestic currency deposits, RR=11 percent for foreign currency.
    - 2004: Actual holdings of government bonds far exceed RR (source: IMF Staff Economists).

*Source: _wp0557 - Appendix I (PDF) — canonical URL provided in the content unit*

### Appendix I  ............................................................................................................

### _wp0557 - Appendix I

### Figures (listed with page numbers)
- 1a. Latin America–Domestic Share of Marketable Government Debt .................................................................. 19
- 1b. Latin America–Marketable Government Debt to GDP Ratios ................................................................. 19
- 2a. Asia – Domestic Share of Marketable Government Debt ........................................................................... 20
- 2b. Asia – Marketable Government Debt to GDP Ratios ............................................................................... 20
- 3a. Other Developing Countries – Domestic Share of Marketable Government Debt ................................... 21
- 3b. Other Developing Countries – Marketable Government Debt to GDP Ratios ....................................... 21
- 4a. Small Industrialized Countries – Domestic Share of Marketable Government Debt ............................... 22
- 4b. Small Industrialized Countries – Marketable Government Debt to GDP Ratios .................................... 22
- 5a. Scandinavia – Domestic Share of Marketable Government Debt .......................................................... 23
- 5b. Scandinavia – Marketable Government Debt to GDP Ratios ................................................................... 23
- 6a. Large Industrialized Countries – Domestic Share of Marketable Government Debt ............................. 24
- 6b. Large Industrialized Countries – Marketable Government Debt to GDP Ratios .................................. 24
- 7. Latin American Financial Institutions – Credit to Public Sector/Total Assets ........................................ 25
- 8. Asian Financial Institutions – Credit to Public Sector/Total Assets ...................................................... 25
- 9. Eastern European Financial Institutions – Credit to Public Sector / Total Assets ................................ 26
- 10. Middle Eastern/African Financial Institutions – Credit to Public Sector/Total Assets ........................ 26
- 11. Industrialized Countries Financial Institutions – Credit to Public Sector/Total Assets ........................ 27
- 12. Small European Countries Financial Institutions – Credit to Public Sector/Total Assets .................... 27
- 13. Government Debt on Balance Sheets and Legal/Institutional Quality .................................................. 28

### Sections and Pagination
- Appendix I .................................................................................................................................. 13
- References ................................................................................................................................... 15

*Source: _wp0557 - Appendix I (PDF) — canonical URL provided in the content unit*

### 1.   Reserve Requirements ..............................................................................................

### _wp0557 - 1.   Reserve Requirements ..............................................................................................

### I. INTRODUCTION
- Policy question: why policymakers treat debt devaluation (default or inflation-induced debt devaluation) as a last resort despite academic literature suggesting state-contingent capital levies can reduce distortionary taxation costs.
- Practical constraints emphasized:
  - Bonds with explicit state-contingent returns are rarely available.
  - Default or debt devaluation is typically seen as a policy of last resort and avoided in fiscal practice.
- Hypothesis: features of financial markets absent from many theoretical models—particularly asymmetric information and legal/institutional imperfections—explain policymakers’ reluctance to use debt devaluation.
- Key mechanism highlighted:
  - Government nominal debt can serve a collateral-like function for financial intermediation.
  - Producers borrow capital from intermediaries that require a minimum fraction of loans to be covered by government debt as collateral.
  - Fiscal policy then creates two distortions:
    - Positive labor tax → suboptimal work effort.
    - Collateral requirement → suboptimal level of capital.
  - Trade-off: volatile inflation (to create state-contingent real returns) improves the first distortion but undermines the second by eroding government debt used as collateral.
- Empirical motivation: concerns about damage to domestic banks have driven delayed restructurings in practice (International Monetary Fund (2002) evidence).

### II. GOVERNMENT DEBT IN DEVELOPING COUNTRIES
- Central empirical findings and arguments:
  - Domestically held government debt is of greater and growing importance than external debt for almost every emerging market country in the sample.
  - Banks and pension funds are the major holders of government debt; banks are particularly important.
  - High bank holdings of government debt are associated with weak legal and institutional frameworks.
  - Large bank holdings of government debt are increasingly voluntary rather than the result of financial repression or binding reserve requirements.

- Specific evidence and observations:
  - Data source: BIS (2004) dataset on individual securities issues (merged with a JP Morgan Brady bond dataset because BIS excludes Brady bonds).
  - Time coverage for figures referenced: quarterly data from 1994Q1 through 2003Q4.
  - Trend: developing countries’ marketable debt-to-GDP ratios were around 20 percent in the mid-1990s and exhibited a positive growth trend; the domestic share of marketable debt has risen and now represents in excess of 70 percent of overall debt in almost all major developing countries (conservative domestic classification).
  - Exceptions cited where domestic debt share stayed below 50 percent due to severe crises and devaluations: Russia and Argentina; Mexico (1995) and Thailand (1997) experienced such episodes but returned to domestic debt shares of over 70 percent afterwards.
  - Bank exposure metric: average ratio (1998-2002) of financial institutions’ net credit to the government to their total assets.
    - Industrialized countries: ratio typically around 10 percent.
    - Developing countries: ratios of 20-40 percent are common; ratios above 50 percent observed in some large developing countries.
    - Typical bank capitalization ratio cited as less than 10 percent of assets, implying government debt losses can render banks insolvent (Argentina example referenced).
  - Reserve requirements and financial repression:
    - Reserve requirements "no longer exist in several key developing countries, and have been falling in others, to the point that they are actually no longer binding."
    - Binding reserve requirements tend to affect pension funds in some countries.
    - Conclusion: large holdings of government debt by banks must therefore be largely voluntary in many countries.
  - Legal and institutional constraints on lending:
    - Problems: difficulty, cost, and uncertainty in creating security interests over movable property; imperfect perfection and registration; slow and expensive enforcement.
    - Empirical contrasts:
      - United States: 70 percent of all loans secured by collateral; almost 50 percent of all credit secured by movable collateral; movable capital represents 67 percent of U.S. industrial capital stock and 75 percent of gross business investment.
      - Argentina: only around 10 percent of all bank credit is secured at all; practically none secured by movable collateral.
    - Lending consequences: when acceptable collateral is available, private creditors offer six to eight times more credit, two to ten times longer maturity, and 30 to 50 percent lower interest rates (Asian Development Bank (2000)).
    - Example: loans for agricultural machinery in Argentina face lending rates around 60 percent; decomposition suggests:
      - about one sixth of the rate difference explained by macroeconomic risk,
      - one-tenth by the general difficulty of collecting against security (including mortgages),
      - three quarters of the difference due to problems with the legal framework for secured transactions against movable property (Fleisig (1998)).
    - Macro consequences of defective secured transaction systems include: inefficient savings allocation, inefficient capital allocation (preference for assets usable as collateral), fragile banks exposed to credit risks, inability to use securitization for risk diversification, and negative effects on equity for small/start-up businesses.
    - Cost estimates from country studies: defective systems of lending security impose costs of 5-10 percent of annual GNP; reform costs are a small fraction of that.
  - Quantitative legal-institutional indices:
    - Constructed three components representing segmented credit markets:
      - "Getting credit" — quality and accessibility of credit information and quality of collateral and bankruptcy laws.
      - "Enforcing contracts" — cost and time delays in collection of overdue debt.
      - "Registering property" — cost and time delays of transferring title to real estate.
    - Overall raw index: equal weights attached to the three categories; final index ranges between 0 and 100 where 0/100 corresponds to the worst/best raw index.
    - Another index for importance of government debt on financial institutions’ balance sheets scaled 0/100 (lowest/highest government debt to total assets ratio).
    - Figure 13 result: a very strong negative relationship between the quality of law and institutions and the amount of government debt that financial institutions choose to hold. Most top-left countries (poor legal/institution quality, high holdings) are developing; most bottom-right (better legal/institution quality, low holdings) are industrialized. Chile noted as best among developing countries.
  - Role of government debt in financial development:
    - Government debt provides financial infrastructure and an informational benchmark for private bond markets.
    - Preconditions for government bond market to support private bond markets:
      - Low macroeconomic volatility, especially inflation volatility (state-contingent inflation undermines this).
      - Sufficient volume of government debt issuance to generate a yield curve.
    - Repo markets and collateralized transactions:
      - Repo markets typically accept only government debt as acceptable collateral.
      - Found sizable repo markets in Brazil and Mexico (using Bankscope bank balance sheet data): Brazil around 7 percent (repo loans as fraction of total assets) and Mexico around 5 percent in 2002/03; other sizable markets in Philippines, Poland, and India.
    - Risks of state-contingent inflation or default:
      - Volatile inflation or debt devaluation undermines the benchmark role of government debt and damages further financial market development (Brazil 1999, Argentina 2001/2 referenced as impediments).

### III. SUMMARY AND CONCLUSIONS
- Main conclusions:
  - Theoretical literature advocating debt devaluations as state-contingent fiscal responses contrasts sharply with practitioners’ views, especially in developing countries.
  - Empirical evidence indicates domestically held government debt and bank holdings of government debt are central in developing countries.
  - High bank exposure to government debt combined with weak legal and institutional frameworks for private lending implies that debt devaluation or default would likely produce severe damage to domestic financial intermediation.
  - Keeping government debt a safe investment for banks is critical to:
    - support existing limited financial intermediation;
    - support the development of private securities and bond markets; and
    - prevent banking collapses that cause severe output contractions.
- Policy implication / research recommendation:
  - It would be beneficial to further develop optimal fiscal policy theory incorporating these financial-market and institutional frictions to improve its usefulness for applied policy analysis.
  - Guillermo Calvo’s work provides conceptual inspiration for modeling costly unanticipated inflation and the trade-offs between taxation and inflation.

*Source: _wp0557 - 1.   Reserve Requirements ..............................................................................................*

### APPENDIX I

### APPENDIX I

### Domestic Versus External Debt — BIS classification and biases
- BIS classification system (Bank for International Settlements (2004)) distinguishes securities by:
  - In domestic currency, targeted at resident investors: Issues by Residents = Domestic; Issues by non-residents = International
  - In domestic currency, targeted at non-resident investors: Issues by Residents = International; Issues by non-residents = International
  - In foreign currency: Issues by Residents = International; Issues by non-residents = International
- Classification covers the borrower side of securities issues, not actual ownership.
- The BIS system is conservative in classifying domestic securities:
  - Issues by non-residents targeted at domestic investors, whether in domestic or foreign currency, are classified as international—even though many holders may be domestic.
  - Issues in domestic currency targeted at foreign investors are classified as international, despite evidence that ultimate holders of developing countries’ domestic currency debt are often domestic residents.
- Data omissions and resulting biases:
  - Notes and money market instruments issued by non-residents in domestic currency and in domestic markets are not included due to lack of data — this biases results against domestically held debt.
  - Local issues in foreign currency are generally classified as international; BIS makes exceptions for Argentina, Peru, and the Philippines.
  - Brady bonds are missing from the BIS dataset; the authors remedy this by merging the BIS data set with a JP Morgan data set on outstanding Brady bonds, using the Merrill Lynch (1994) guide for classification of securities.

### Bank Holdings of Government Debt — definition and computation
- Exposure measure defined as:
  - (Financial institutions’ net credit to government) / (financial institutions’ net total assets)
- Data source and averaging:
  - Percentage figures computed from the IMF’s International Financial Statistics.
  - Averages are for 1998-2002, except for countries in the euro area where 1999-2003 was used.
- Numerator and denominator construction:
  - Numerator: sum of all entries representing net credit to the public sector by deposit money banks, other banking institutions and nonbank financial institutions (the latter series do not exist for all countries).
  - Denominator: sum of the net total assets of these three groups, after canceling out credit items between them.
  - To obtain net figures, deduct from both numerator and denominator the sum of all entries representing credit by the public sector to these institutions.
- Crisis-year exclusions from averages:
  - Argentina 2002
  - Indonesia 1998

### Reserve requirements — country notes and exact values reported
- Brazil:
  - 1996: RR remains over 80 percent on demand deposits.
  - 1999: April: RR on demand deposits up to 100 percent, on other acc. claims up to 60 percent. October: RR on bank deposits reduced from 65 to 45 percent
  - 2001: RR on demand deposits 45 percent, RR on time deposits 10 percent.
  - 2004: No de jure RR (source: IMF Staff Economists).
- China:
  - 2004: No de jure RR (source: IMF Staff Economists).
- Czech Republic:
  - 2004: No de jure RR (source: IMF Staff Economists).
- India:
  - 1990: Cash reserve ratio (CRR) 10 percent.
  - 1998: CRR raised to 10.5%. lowered to 10 percent raised to 11%.
  - 1999: CRR lowered to 10.5%, then to 10 percent and to 9 percent in November.
  - 2000: CRR lowered to 8 percent.
  - 2001: CRR lowered to 7.5 percent.
  - 2004: Statutory Liquidity Ratio (SLR): RR of 25 percent in the form of government bond holdings.
- Indonesia:
  - 1999: No RR.
  - 2004: No de jure RR (source: IMF Staff Economists).
- Jamaica:
  - 2004: Actual holdings of government bonds far exceed RR (source: IMF Staff Economists).
- Korea:
  - 2001: RR=2.9 percent.
- Lebanon:
  - 2004: No de jure RR (source: IMF Staff Economists).
- Mexico:
  - 1991: Liquidity coefficients on foreign currency deposits up to 50 percent dep. on maturities.
  - 1995: Change from a daily zero RR system to cumulative zero RR over 28 day period.
  - 1999: RR=0 percent.
- Philippines:
  - 1991: RR rationalized.
  - 1993: RR for bank deposits lowered from 25 to 22 percent.
  - 1994: RR lowered to 17 percent.
  - 1995: RR lowered to 15 percent
  - 1997: RR lowered to 13 percent
  - 1998: RR lowered to 10 percent.
  - 1999: RR=11 percent for commercial banks, RR=0% for rural and cooperative banks.
  - 2002: RR=9 percent
- Poland:
  - 2004: No RR on banks. (source: IMF Staff Economists).
- Russia:
  - 2004: RR=3.5 percent.
- Turkey:
  - 1994: Comprehensive reform of RR. RR=10.8 percent.
  - 1995: RR=9.3 percent.
  - 1996: RR=8 percent for domestic currency deposits, RR=11 percent for foreign currency.
  - 2004: Actual holdings of government bonds far exceed RR (source: IMF Staff Economists).

*Source: APPENDIX I, _wp0557 - APPENDIX I*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2005/_wp0557.pdf_