## 1.    Bank Credit to the Private Sector, 2003

## Source details

**Canonical URL:** [1.    Bank Credit to the Private Sector, 2003](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0644.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0644.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0644.pdf.json)

---

### Introduction: scope and motivation
- Financial intermediation in Latin America is low by international standards; systems remain largely bank-based.
- After strong credit growth in the early 1990s, bank lending had not recovered from mid-1990s banking crises (Singh et al., 2005).
- This paper analyzes determinants of bank interest margins using bank- and country-level data from 85 countries, including 14 Latin American economies, and micro data from over 2,200 banks.
- Time coverage for bank-specific data: 1999–2002 (with reserve requirement data partly for 2002 and partly for 2003).

### Literature and candidate determinants of spreads
- Country-level factors likely to influence spreads:
  - Creditor rights and legal framework quality: better recovery rates and shorter collateral repossession expected to reduce spreads.
  - Degree of banking competition: more intense competition should lower spreads; empirical correlation between concentration/competition measures and spreads is weak.
  - Macroeconomic environment: volatility and inflation can raise spreads; theoretical prediction that margins rise with variance of interest rates.
  - Taxation and reserve requirements: taxation of financial intermediation (including unremunerated reserve requirements) drives a wedge and can increase spreads.
  - Availability of borrower information: better credit information and accounting standards lower default risk and spreads.
  - Banking regulations and mandated lending: barriers to entry and mandated subsidized lending tend to increase spreads.
- Bank-specific factors:
  - Bank strategies, size (economies of scale), overhead costs, capitalization, and ownership (foreign vs. domestic) affect margins.
  - Empirical evidence on foreign bank impact is mixed.

### Descriptive evidence and stylized facts
- Measurement focus: ex-post net interest margins (total interest income minus total interest expense, divided by interest-bearing assets); complements ex-ante spreads.
- Key cross-country observations:
  - Legal protection: Latin America’s average score on legal indices is below developing-country average; negative correlation between legal protection index (legprop) and mean net interest margins.
  - Country risk and macro volatility: Latin America’s macro risk on average similar to other emerging markets; positive correlation between country risk (ICRG score; higher = lower risk) and margins, with Latin American margins above fitted values.
  - Competition: Behavioral H statistic (Panzar and Rosse) — Latin American average ≈ 0.67 (roughly equal to total-sample average); no robust cross-country correlation between H statistic and interest margins.
  - Overhead and personnel costs: strong positive correlation with net interest margins; Latin American banks exhibit relatively high overhead and personnel costs.
  - Reserve requirements: strongly correlated with intermediation spreads; reserve requirements on demand deposits relatively high in Latin America (data mostly for 2003 and partly for 2002).
  - Deposit rates: net interest margins are positively correlated with deposit rates; Latin America has relatively high deposit rates.
  - Taxation: bank profit and revenue taxation in Latin America is relatively low by international standards; financial transaction taxes in Argentina, Bolivia, Brazil, Colombia, Peru, and Venezuela likely contribute to high spreads (data limitations).
  - Information availability: corporate transparency and availability of borrower information are more limited in Latin America; mild negative correlation between information availability index and net interest margins.

- Table 1 — Selected country-level means (1999–2002 averages unless noted; reserve requirements partly 2002/2003):
  - Latin American Median: 8.0; 6.1; 2.6; 13.80; 4.0; 8.0; 4.0; 66.0; 13.0; 2.2; 7.2; 11.6
  - Other EM Median: 5.0; 3.6; 1.7; 6.00; 5.4; 3.3; 4.4; 67.6; 18.8; 3.3; 4.7; 6.4
  - (Columns correspond to: Interest Margin; Overhead as % of total assets; Personnel expenses as % of total assets; Reserve requirement (2002/03) (percent of demand deposits); Legal Protection; Annual Inflation; Avail. of Inform.; Country risk score (ICRG); Average Bank Profit Tax Rate; Annual GDP Growth; Std. Dev of real ex-change rate; Deposit Rate — presented verbatim.)

### Econometric specification and variables
- Regression framework:
  - Net Interest Margin_i,j = a + b·Bank_j + c·Comp_i + d·ResReq + e·Deposit Rate + f·Taxes + g·Legal_i + h·Info_i + i·Macro_i + ε_i,j
- Variable definitions:
  - Net interest margin = (interest income − interest expense) / interest-bearing assets.
  - Bank: bank-specific variables (bank size, bank equity, overhead costs, foreign ownership).
  - Comp: competition measures (five-bank-concentration ratio, H-Statistic).
  - ResReq: required reserves on demand deposits (unremunerated reserves where available).
  - Deposit Rate: average rate paid on deposits.
  - Taxes: taxes paid by the bank.
  - Legal: legprop and prop indices measuring legal environment and creditor rights.
  - Info: indices measuring availability of information about potential borrowers.
  - Macro: GDP growth, inflation, inflation volatility, country risk ratings.

### Main empirical findings — Bank-level determinants
- Bank size:
  - Larger banks associated with lower net interest margins.
  - Ln(Size) coefficient (Table 2, column 1): -0.604 (t = (4.43)**).
  - Ln(Size) coefficient (Table 3, column (1) RE): -0.412 (z = (7.27)***).
- Overhead costs:
  - Strong positive association with higher net interest margins.
  - Overhead coefficient (Table 2, column 1): 68.814 (t = (4.41)**).
  - Overhead coefficient (Table 3, column (1) RE): 62.663 (z = (30.28)***).
- Capitalization:
  - Ratio of equity to total assets not consistently associated cross-sectionally but positive in several specs.
  - Equt/Tot.Assets (Table 2, column 1): 5.583 (t = (3.06)**).
  - Equity/Assets (Table 3, column (1) RE): 3.479 (z = (6.64)***).
- Ownership:
  - Foreign ownership dummy not significant in full cross-sectional sample; in panel regressions foreign dummy enters significantly with a negative sign (foreign banks charge lower spreads ceteris paribus).

### Main empirical findings — Country-level determinants
- Reserve requirements and deposit rates:
  - Strong positive associations with higher spreads.
  - A fall in the deposit rate of one percentage point associated with a drop in net interest margins by 0.1–0.2 percentage points.
  - A reduction in reserve requirements on demand deposits by 10 percentage points would reduce net interest margins by an average of 0.4–0.7 percentage points.
  - Cross-country coefficient on Res. Req. (Table 2, column 1): 0.063 (t = (3.66)**).
  - Cross-country coefficient on Deposit rate (Table 2, column 1): 0.152 (t = (2.37)**).
  - Panel coefficient on Res. Req. (Table 3, column (1) RE): 0.053 (z = (2.25)**).
  - Panel coefficient on Deposit Rate (Table 3, column (1) RE): 0.105 (z = (5.94)***).
- Economic activity:
  - Higher GDP growth correlated with lower spreads.
  - A one-percent increase in GDP growth associated with a 0.2–0.5 percentage point lower net interest margin (examples of cross-country coefficients include -0.296, -0.324, -0.357, -0.310, -0.175, -0.370, -0.532).
  - Panel GDP growth (Table 3, column (1) RE): -0.095 (z = (3.16)***).
- Market concentration and competition:
  - Five-bank concentration ratio positively associated with higher interest margins; H statistic not significant.
  - Overhead appears to mediate some concentration effects; procedure used to regress overhead on concentration and competition and include residuals in main estimations.
- Legal, informational, and tax indicators:
  - Enter with expected signs but not consistently significant after controls.
  - Panel results: profit taxes matter — a ten percentage point increase in profit taxes raises net interest margins by around 0.15 percentage points.
  - Panel results: foreign ownership dummy negative and significant.

### Robustness and alternative specifications
- Instrumenting deposit rates with lagged levels makes deposit-rate coefficient insignificant; estimated effect of reserve requirements becomes larger and more significant.
- Using money market rate instead of deposit rates yields similar results but with somewhat higher estimated reserve-requirement coefficient; smaller sample.
- Panel regressions for 1999–2002 broadly confirm cross-sectional results and strengthen evidence for legal environment, profit taxation, and foreign ownership effects.
- Results robust to exclusion of outliers (Malawi, Venezuela, Brazil) though some coefficients fall (equity/total assets, overhead, deposit rates); GDP growth and profit tax rate often lose significance in restricted samples.
- Alternative spread definition (Spreads2 = [(total interest income/total loans) − (total interest expense/total deposits)]*100) yields qualitatively similar findings:
  - Bank size loses statistical significance.
  - Effect of reserve requirements substantially larger: a 10 percentage point increase in reserve requirements associated with an increase in spreads between 2 and 7 percentage points (Table AII, Appendix).

### Decomposition: Latin America vs other developing countries
- Total difference in average net interest margins: 3.3 percentage points.
- Contributions to the 3.3 percentage points gap (based on column 4 of Table 4):
  - Overhead costs: about 1.7 percentage points.
  - Reserve requirements: about 0.5 percentage points.
  - Higher interest rates: about 0.4 percentage points.
  - Less developed property rights: about 0.2 percentage points.
  - Lower GDP growth: about 0.15 percentage points.
  - Other factors: insignificant fraction.

### Selected model fit and statistics (reported values)
- Table 2 (cross-country regressions, 2002), column (1):
  - Observations: 953
  - # of countries: 58
  - R-squared: 0.48
  - Ln(Size): -0.604
  - Equt/Tot.Assets: 5.583
  - Overhead: 68.814
  - Res. Req.: 0.063
  - Deposit rate: 0.152
- Table 3 (panel regressions, 1999–2002), column (1) RE:
  - Observations: n4018
  - # of countries: 59
  - Ln(Size): -0.412
  - Equity/Assets: 3.479
  - Overhead: 62.663
  - Res. Req.: 0.053
  - Deposit Rate: 0.105
  - GDP growth: -0.095

### Policy implications and conclusions
- Factors likely explaining higher spreads in Latin America (relative to other developing countries):
  - Less efficient banks (reflecting weaker competition).
  - Relatively higher levels of interest rates.
  - Higher reserve requirements.
  - A less supportive legal environment increasing intermediation costs.
- Policies likely to reduce spreads in Latin America:
  - Promote bank competition and efficiency.
  - Provide a macroeconomic environment conducive to lower equilibrium interest rates.
  - Reduce reserve requirements.
- Data gap and suggested research:
  - The role of financial transaction taxes in influencing spread levels could not be investigated due to limited cross-country data and is proposed as a profitable area for future research.

*Source: _wp0644 (IMF Working Paper excerpt).*

### 1.    Bank Credit to the Private Sector, 2003 ..........................................................................

### 1.    Bank Credit to the Private Sector, 2003

### Major sections
- 1.    Bank Credit to the Private Sector, 2003 .............................................................................3
- 2.    Ex-Ante Banking Spreads Worldwide................................................................................4
- 3.    Net    Interest    Margins ...........................................................................................................4
- 4.    Ex-Ante Banking Spreads in Latin America ......................................................................5
- 5.    Legal Protection and Interest Margins..............................................................................11
- 6.    Country Risk and Net Interest Margins ............................................................................12
- 7.    Average Bank Overhead and Net Interest Margins ..........................................................13
- 8.    Reserve Requirements on Demand Deposits and Net Interest Margins...........................14
- 9.    Deposit Rates and Net Interest Margins ...........................................................................15
- 10.  Availability of Information about Companies and Net Interest Margins .........................16
- 11.  Net Interest Margins: Contribution of Different Factors in Explaining Difference 
    Between Latin America’s Average and the Average for Developing Countries...........21

### Tables and empirical material
- Tables
  - 1.    Factors Influencing Level of Spreads ...............................................................................10
  - 2.    Determinants of Net Interest Margins: Cross-Country Regressions, 2002 ......................19
  - 3.    Determinants of Net Interest Margins: Panel Regressions for 1999–2002 ......................22

- Appendix tables
  - Table AI. Determinants of Net Interest Margins: Cross-Country Regressions, 2002 
    (excluding         Brazil,         Malawi, and Venezuela) ....................................................28
  - Table AII.  Determinants of Spreads: Cross-Country Regressions, 2002 
    (alternative         spread         definition) ..........................................................................29

### Thematic emphasis for indexing and crawling
- Focus on banking spreads, net interest margins, and their determinants across countries and specifically for Latin America.
- Empirical approaches include cross-country regressions for 2002 and panel regressions for 1999–2002.
- Coverage of institutional and market factors: legal protection, country risk, bank overhead, reserve requirements, deposit rates, and availability of firm-level information.
- Comparative analysis: decomposition of factors explaining differences between Latin America’s average and the average for developing countries.

*Source: _wp0644 - 1.    Bank Credit to the Private Sector, 2003 (IMF PDF).*

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

### _wp0644 - References  ............................................................................................................

### Introduction: scope and motivation
- Financial intermediation in Latin America is low by international standards; systems remain largely bank-based.
- After strong credit growth in the early 1990s, bank lending had not recovered from mid-1990s banking crises (Singh et al., 2005).
- Bank lending as a percentage of GDP is low relative to industrialized countries and other emerging markets (Figure 1).
- Financial intermediation is viewed as an important obstacle to growth; literature indicates financial intermediation is correlated with and a causal factor for economic performance (see Levine, 2004).
- This paper analyzes determinants of bank interest margins using bank- and country-level data from 85 countries, including 14 Latin American economies, and micro data from over 2,200 banks.
- Time coverage for bank-specific data: 1999–2002 (with reserve requirement data partly for 2002 and partly for 2003).

### Literature review: candidate determinants of spreads
- Country-level factors likely to influence spreads:
  - Creditor rights and legal framework quality: better recovery rates and shorter collateral repossession expected to reduce spreads (Laeven and Majnoni, 2003; DLL).
  - Degree of banking competition: more intense competition should lower spreads, but empirical correlation between concentration/competition measures and spreads is weak.
  - Macroeconomic environment: volatility and inflation can raise spreads; dealership model (Ho and Saunders, 1981) predicts margins rise with variance of interest rates; several empirical studies support positive correlation between spreads and inflation/interest-rate level.
  - Taxation and reserve requirements: taxation of financial intermediation (including unremunerated reserve requirements) drives a wedge between borrowing and lending rates and can increase spreads (Honohan, 2003).
  - Availability of borrower information: better credit information and accounting standards lower default risk and spreads (Chu and Schechtman, 2003).
  - Banking regulations and mandated lending: barriers to entry and mandated subsidized lending tend to increase spreads.
- Bank-specific factors:
  - Bank strategies (fee income vs. interest income), size (economies of scale), overhead costs, capitalization, and ownership (foreign vs. domestic) affect margins.
  - Empirical evidence on foreign bank impact is mixed (Martínez Peria and Mody, 2004; Claessens et al., 2001; Detragiache et al., 2005).
- Prior Latin America studies emphasize taxation, operating costs, imperfect competition, macroeconomic volatility, and reserve requirements as key spread determinants (Brock and Rojas Suárez, 2000a,b; Martínez Peria and Mody, 2000; Dick, 1999; Barajas, Steiner, and Salazar, 1999; Catão, 1988; Brock and Franken, 2003; Banco Central do Brasil studies).

### Descriptive evidence: cross-country comparisons and stylized facts
- Data sources and coverage:
  - Bank-specific variables from Bankscope; analyses use data for 1999–2002.
  - Sample sizes in many estimations: 30–60 countries and 1,000–1,300 banks.
- Measurement focus:
  - Primary focus on ex-post net interest margins (total interest income minus total interest expense, divided by interest-bearing assets); complements ex-ante spreads.
- Key descriptive correlations and observations:
  - Legal protection: Latin America’s average score on legal indices is below developing-country average; negative correlation between legal protection index (legprop) and mean net interest margins (Figure 5).
  - Macroeconomic volatility/country risk: Latin America’s macro risk on average similar to other emerging markets; positive correlation between country risk (ICRG score; higher = lower risk) and margins, but Latin American margins lie above fitted values (Figure 6).
  - Competition measures:
    - Behavioral H statistic (Panzar and Rosse approach): Latin American average ≈ 0.67 (roughly equal to total-sample average).
    - No robust cross-country correlation between H statistic and interest margins.
    - Strong positive correlation between bank-level administrative/personnel costs and net interest margins; Latin American banks exhibit relatively high overhead and personnel costs, indicating inefficiencies (Figure 7).
  - Reserve requirements:
    - Reserve requirements on demand deposits are strongly correlated with intermediation spreads.
    - Reserve requirements are relatively high in Latin America; data mostly for 2003 and partly for 2002. Reserve requirements shown to be an important factor in high spreads (Figure 8).
  - Level of interest rates:
    - Net interest margins are positively correlated with deposit rates; relatively high deposit rates in Latin America are a likely contributor to high margins (Figure 9).
  - Taxation:
    - Bank profit and revenue taxation in Latin America is relatively low by international standards.
    - Financial transaction taxes present in Argentina, Bolivia, Brazil, Colombia, Peru, and Venezuela are likely to contribute to high spreads though comparable cross-country data are lacking.
  - Information availability:
    - Corporate transparency and availability of borrower information are more limited in Latin America than in most emerging markets; mild negative correlation between information availability index (Global Competitiveness Report) and net interest margins (Figure 10).

- Table 1 — Selected country-level means (1999–2002 averages unless noted; reserve requirements partly 2002/2003):
  - Columns (in order) and Latin American Median and Other EM Median rows reproduced exactly as in source:
    - Latin American Median: 8.0; 6.1; 2.6; 13.80; 4.0; 8.0; 4.0; 66.0; 13.0; 2.2; 7.2; 11.6
    - Other EM Median: 5.0; 3.6; 1.7; 6.00; 5.4; 3.3; 4.4; 67.6; 18.8; 3.3; 4.7; 6.4
  - (These entries correspond to the table headings: Interest Margin; Overhead as % of total assets; Personnel expenses as % of total assets; Reserve requirement (2002/03) (percent of demand deposits); Legal Protection; Annual Inflation; Avail. of Inform.; Country risk score (ICRG); Average Bank Profit Tax Rate; Annual GDP Growth; Std. Dev of real ex-change rate; Deposit Rate — presented verbatim.)

### Econometric estimations: specification and main explanatory variables
- Regression framework (bank j in country i):
  - Net Interest Margin_i,j = a + b·Bank_j + c·Comp_i + d·ResReq + e·Deposit Rate + f·Taxes + g·Legal_i + h·Info_i + i·Macro_i + ε_i,j
- Definitions used in the specification:
  - Net interest margin: interest income minus interest expense divided by interest-bearing assets.
  - Bank: bank-specific variables (bank size, bank equity, overhead costs, foreign ownership).
  - Comp: country-level competition measures (e.g., five-bank-concentration ratio, H-Statistic).
  - ResReq: rate of required reserves on demand deposits (unremunerated reserves used where available).
  - Deposit Rate: average rate paid on deposits.
  - Taxes: taxes paid by the bank.
  - Legal: indices measuring legal environment and enforceability of contracts and creditor rights (legprop and prop).
  - Info: indices measuring availability of information about potential borrowers.
  - Macro: macroeconomic variables (GDP growth, inflation, inflation volatility, country risk ratings).
- Estimation approach:
  - Initial simple cross-sectional regressions focus on 2002 for coverage; further estimations incorporate the full panel and bank fixed effects (details to follow beyond excerpt).

### Main empirical messages (from descriptive and initial econometric discussion)
- Latin America exhibits:
  - Higher interest rate levels.
  - Less efficient banks (higher overhead and personnel costs).
  - Larger reserve requirements.
- These factors have a significant impact on spreads in the region.
- Latin American countries do not differ markedly from peers in several other determinants: inflation, bank profit taxation, average bank size, and average equity as percentage of total bank assets.
- Bank-level variables play an important role in determining spreads; subsequent econometric sections explore their relative importance controlling for country-level factors.

*Source: _wp0644 - References  ............................................................................................................*

### section is partly motivated by the fact that we do not have time-varying information about all

### Determinants of Net Interest Margins: Cross-Country and Panel Results (excerpt)

### Bank-level determinants
- Larger banks are associated with lower net interest margins:
  - Ln(Size) coefficient in cross-country regression (Table 2, column 1): -0.604 (t = (4.43)**).
  - Ln(Size) coefficient in panel regression (Table 3, column (1) RE): -0.412 (z = (7.27)***).
- Overhead costs are strongly positively associated with higher net interest margins:
  - Overhead coefficient in cross-country regression (Table 2, column 1): 68.814 (t = (4.41)**).
  - Overhead coefficient in panel regression (Table 3, column (1) RE): 62.663 (z = (30.28)***).
- The ratio of equity to total assets is not consistently associated with interest margins in cross-sectional estimations but shows positive coefficients in several specifications:
  - Equt/Tot.Assets (Table 2, column 1): 5.583 (t = (3.06)**).
  - Equity/Assets (Table 3, column (1) RE): 3.479 (z = (6.64)***).
- The dummy for foreign ownership is not significant in the full cross-sectional sample; data on ownership is only available for a subset of banks (note supplied by Laura Kodres).
- Measures of nonperforming loans and similar bank-quality proxies did not enter estimations systematically.

### Country-level determinants
- Interest rate level and reserve requirements are strongly positively associated with higher spreads:
  - Estimated effect: a fall in the deposit rate of one percentage point would be associated with a drop in net interest margins by 0.1–0.2 percentage points.
  - Estimated effect: a reduction in reserve requirements on demand deposits by 10 percentage points would reduce net interest margins by an average of 0.4–0.7 percentage points.
  - Cross-country coefficient on Res. Req. (Table 2, column 1): 0.063 (t = (3.66)**).
  - Cross-country coefficient on Deposit rate (Table 2, column 1): 0.152 (t = (2.37)**).
  - Panel coefficient on Res. Req. (Table 3, column (1) RE): 0.053 (z = (2.25)**).
  - Panel coefficient on Deposit Rate (Table 3, column (1) RE): 0.105 (z = (5.94)***).
- Economic activity is negatively correlated with spreads:
  - A one-percent increase in GDP growth is associated with a 0.2–0.5 percentage point lower net interest margin (regression results).
  - Cross-country estimates show GDP growth coefficients in the range reported (examples across specifications include -0.296, -0.324, -0.357, -0.310, -0.175, -0.370, -0.532 in Table 2 where present).
  - Panel GDP growth (Table 3, column (1) RE): -0.095 (z = (3.16)***).
- Market concentration and competition:
  - Higher concentration (five-bank concentration ratio) is associated with higher interest margins; the H statistic does not enter estimations significantly.
  - Procedure used: regress overheads on concentration and competition measures and include residuals in main estimations; the five-bank concentration ratio enters significantly with expected positive sign.
- Legal, informational, and tax indicators:
  - Taxation (profit tax rate), indices of information availability, and legal framework enter with expected signs but are not consistently significant after controlling for other factors—measurement problems are a possible reason.
  - In panel regressions (1999–2002), profit taxes matter: a ten percentage point increase in profit taxes raises net interest margins by around 0.15 percentage points (panel results).
  - In panel regressions, the foreign ownership dummy enters significantly with a negative sign, indicating foreign banks charge lower spreads ceteris paribus (Martínez Peria and Mody (2004) consistent).

### Robustness and alternative specifications
- Instrumenting deposit rates with lagged levels renders the deposit-rate coefficient insignificant; the estimated effect of reserve requirements becomes larger and more significant (results not shown).
- Using a prevailing money market rate instead of deposit rates yields broadly similar results but with a somewhat higher estimated coefficient on reserve requirements; sample is substantially smaller due to data availability.
- Panel regressions for 1999–2002 broadly confirm cross-sectional results but strengthen the case for legal environment, bank profit taxation, and foreign ownership effects.
- Results robust to exclusion of outliers (Malawi, Venezuela, Brazil): coefficients on some variables fall (e.g., equity/total assets, overhead, deposit rates); GDP growth and profit tax rate often lose statistical significance in these restricted samples.
- Alternative spread definition:
  - Spreads2 = [(total interest income/total loans)-(total interest expense/total deposits)]*100.
  - Using Spreads2 yields qualitatively similar results; bank size loses statistical significance; effect of reserve requirements is substantially larger: a 10 percentage point increase in reserve requirements is associated with an increase in spreads between 2 and 7 percentage points (Table AII, Appendix).

### Difference between Latin America and other developing countries
- Decomposition (based on column 4 of Table 4) of the average net interest margin gap (Latin America vs other developing countries):
  - Total difference in average net interest margins: 3.3 percentage points.
  - Overhead costs explain about 1.7 percentage points of the 3.3 percentage points total difference.
  - Reserve requirements contribute about 0.5 percentage points.
  - Higher interest rates explain about 0.4 percentage points.
  - Less developed property rights contribute about 0.2 percentage points.
  - Lower GDP growth contributes about 0.15 percentage points.
  - Other factors account for an insignificant fraction of the difference.

### Key statistics and model fit (selected from tables)
- Table 2 (cross-country regressions, 2002), column (1):
  - Observations: 953
  - # of countries: 58
  - R-squared: 0.48
  - Ln(Size): -0.604
  - Equt/Tot.Assets: 5.583
  - Overhead: 68.814
  - Res. Req.: 0.063
  - Deposit rate: 0.152
- Table 3 (panel regressions, 1999–2002), column (1) RE:
  - Observations: n4018
  - # of countries: 59
  - Ln(Size): -0.412
  - Equity/Assets: 3.479
  - Overhead: 62.663
  - Res. Req.: 0.053
  - Deposit Rate: 0.105
  - GDP growth: -0.095

### Policy implications and conclusions (from study)
- Factors likely explaining higher spreads in Latin America (relative to other developing countries):
  - Less efficient banks (reflecting weaker competition),
  - Relatively higher levels of interest rates,
  - Higher reserve requirements,
  - A less supportive legal environment contributes to larger intermediation costs.
- Policies likely to reduce spreads in Latin America:
  - Promote bank competition and efficiency,
  - Provide a macroeconomic environment conducive to lower equilibrium interest rates,
  - Reduce reserve requirements.
- Data gap noted: the role of financial transaction taxes in influencing spread levels could not be investigated due to limited data availability and is proposed as a profitable area for future research.

*Source: _wp0644 - section is partly motivated by the fact that we do not have time-varying information about all*

### REFERENCES

### REFERENCES

### Referenced works
- Afanasieff, Tarsila, Priscilla Lhacer, and Márcio Nakane, 2002, “The Determinants of Bank Interest Spreads in Brazil,” BCB Working Paper No. 46
- Albuquerque, Pedro H. 2001, “Os Impactos Economicos da CPMF: Teoria e Evidencia,” unpublished manuscript, University of Wisconsin
- Banco Central do Brasil, 1999, “Juros e Spread Bancário no Brasil”
- Banco Central do Brasil, 2000, “Economia Bancaria e Crédito—Avaliaçao de um ano do Projeto Juros e Spread Bancario”
- Banco Central do Brasil, 2002, “Economia Bancaria e Crédito—Avaliaçao de 2 anos do Projeto Juros e Spread Bancario”
- Banco Central do Brasil, 2002, “Economia Bancaria e Crédito—Avaliaçao de 3 anos do Projeto Juros e Spread Bancario”
- Banco Central do Brasil, 2003, “Economia Bancaria e Crédito—Avaliaçao de 4 anos do Projeto Juros e Spread Bancario”
- Barajas, Adolfo, Roberto Steiner, and Natalia Salazar, 1999, “Interest Spreads in Banking in Colombia, 1974–96,” IMF Staff Papers, Vol. 46, No. 2, pp. 196–224
- Beck, Thorsten, 2002, “Impediments to the Development and Efficiency of Financial Intermediation in Brazil” (unpublished; Washington: World Bank)
- Belaisch, Agnès, 2003, “Do Brazilian Banks Compete?” IMF Working Paper 03/113 (Washington: International Monetary Fund)
- Brock, Philip, and Helmut Franken, 2003, “Bank Interest Margins Meet Interest Rate Spreads: How Good is Balance Sheet Data for Analyzing the Cost of Financial Intermediation” (Seattle: University of Washington)
- Brock, Philip, and Liliana Rojas-Suárez, 2000a, “Understanding the Behavior of Bank Spreads in Latin America,” Journal of Development Economics, Vol. 63, No. 1, pp. 113–34
- Brock, Philip, and Liliana Rojas-Suárez (eds.), 2000b, “Why So High? – Understanding Interest Rate Spreads in Latin America,” Inter-American Development Bank, Washington D.C.
- Cardoso, Eliana, 2003, “Implicit and Explicit Taxation of Financial Intermediaries in Brazil: The Effects of Reserve Requirements on Bank Spreads,” in Taxation of Financial Intermediation, ed. by Patrick Honohan (New York and Oxford: Oxford University Press)
- Catão, Luis, 1998, “Intermediation Spreads in a Dual Currency Economy: Argentina in the 1990s,” IMF Working Paper 98/90 (Washington: International Monetary Fund)
- Chamley, Christophe and Patrick Honohan, 1993, “Financial Repression and Bank Intermediation,” Savings and Development, Vol. 17, No. 3, pp. 113–34
- Chu, Victorio and Ricardo Schechtman, 2003, “O Use de Informaçoes no Credito Bancario,” in: “Economia Bancária e Crédito – Avaliaçao de 4 Anos do Projeto Juros e Spread Bancario,” Banco Central do Brasil (Brasilia)
- Claessens, Stijn, Asli Demirgüç-Kunt, and Harry Huizinga, 2001, “How Does Foreign Entry Affect Domestic Banking Markets?,” Journal of Banking and Finance Vol. 25, No. 5, pp. 891–911
- Claessens, Stijn and Luc Laeven, 2004, “What Drives Bank Competition? Some International Evidence,” Journal of Money, Credit, and Banking, Vol. 36, No. 3, June, pp. 133–152
- Coelho, I., Liam Ebrill, and Victoria Summers, 2001, “Bank Debit Taxes in Latin America: An Analysis of Recent Trends,” IMF Working Paper 01/67 (Washington: International Monetary Fund)
- Costa, Ana C.A. and Eduardo Lundberg, 2004, “Direcionamentos de crédito no Brasil: uma avaliaçao das aplicaçoes obrigatórias em crédito rural e habitacional,” unpublished manuscript, Banco Central do Brasil (Brasilia)
- Costa, Ana C.A. and Marcio I. Nakane, 2004, “Revisitando a Metodologia de Decomposição do Spread Bancário do Brasil,” working paper, Universidade de São Paulo (São Paulo)
- Dell’Ariccia, Giovanni, and Robert Márquez, 2004, ”Information and Bank Credit Allocation,” Journal of Financial Economics Vol. 72, No.1, pp. 185–214
- Demirgüç-Kunt, Asli, and Harry Huizinga, 1999, “Determinants of Commercial Bank Interest Margins and Profitability: Some International Evidence,” World Bank Economic Review Vol. 13, No. 2, pp. 379–408
- Demirgüç-Kunt, Asli, Luc Laeven, and Ross Levine, 2004, “Regulations, Market Structure, Institutions, and the Cost of Financial Intermediation,” Journal of Money, Credit, and Banking Vol. 36, No. 3, pp. 593–622
- Detragiache, Enrica, Poonam Gupta, and Thierry Tressel, 2005, “Foreign Banks in Poor Countries: Theory and Evidence,” unpublished manuscript, IMF (Washington: International Monetary Fund)
- Dick, Astrid, 1999, “Banking Spreads in Central America: Evolution, Structure, and Behavior,” HIID Development Discusssion Papers (Cambridge: Harvard Institute for International Development)
- Gelos, Gaston, and Jorge Roldós, 2004, “Consolidation and Market Structure in Emerging Market Banking Systems,” Emerging Markets Review Vol. 5, No.1, pp. 39–59
- Gelos, Gaston and Shang-Jin Wei, 2005, “Transparency and International Portfolio Holdings,” Journal of Finance, Vol. 60, No. 6, pp. 2987–3020
- Goldfajn, Ilan, Katherine Hennings, and Helio Mori, 2003, “Brazil’s Financial System: Resilience to Shocks, No Currency Substitution, but Struggling to Promote Growth,” unpublished manuscript, Brazilian Central Bank (Brasilia)
- Ho, Thomas, and Anthony Saunders, 1981, “The Determinants of Bank Interest Margins: Theory and Empirical Evidence,” Journal of Financial and Quantitative Analysis Vol. 16, No.4, pp. 581–600
- Honohan, Patrick, 1999, “Designing the Taxation of Financial Intermediation,” unpublished manuscript (Washington: World Bank)
- Honohan, Patrick, 2003, “The Accidental Tax: Inflation and the Financial Sector,” in: “Taxation of Financial Intermediation,” ed. by Patrick Honohan, (New York and Oxford: Oxford University Press)
- Inter-American Development Bank, 2004, “Unlocking Credit – The Quest for Deep and Stable Bank Lending” (Washington, D.C.)
- Kirilenko, Andrei, and Victoria Perry, 2004, ”On the Financial Disintermediation of Bank Transaction Taxes,” unpublished manuscript, IMF (Washington: International Monetary Fund)
- Kodres, Laura, Matt Pritsker, and Marco Souto, 2005, “Foreign Bank Entry in Developing Countries: Helpful or Harmful?,” unpublished manuscript, IMF (Washington: International Monetary Fund)
- Laeven, Luc and Giovanni Majnoni, 2003, “Does Judicial Effiicency Lower the Cost of Credit?,” World Bank Policy Research Paper 3159 (Washington: World Bank)
- Levine, Ross, 2004, “Finance and Growth: Theory and Evidence,” NBER Working Paper 10766 (Cambridge: National Bureau of Economic Research)
- Martínez Peria, Soledad and Ashoka Mody, 2002, “How Foreign Participation and Market Concentration Impact Bank Spreads: Evidence from Latin America,” Journal of Money, Credit, and Banking, Vol. 36, No. 2, pp. 511–537
- Nakane, Marcio, 2003, “Concorrencia e Spread Bancario: Uma Revisao da Evidencia para o Brasil,” in: Banco Central do Brasil, “Economia Bancaria e Crédito – Avaliaçao de 4 anos do Projeto Juros e Spread Bancario” (Brasilia)
- Panzar, John C. and James N. Rosse, 1987, “Testing for ‘Monopoly’ Equilibrium,” Journal of Industrial Economics Vol. 35, No. 4, pp. 443–456
- Pinheiro, Armando C. and Célia Cabral, 2001, “Credit Markets in Brazil: The Role of the Judiciary and Other Institutions,” in: “Defusing Default: Incentives and Institutions”, ed. by Marco Pagano, (Washington: Inter-American Development Bank)
- Saunders, Anthony and Liliana Schumacher, 2000, “The Determinants of Bank Interest Rate Margins: An International Study,” Journal of International Money and Finance Vol. 19, No. 6, pp. 813–832
- Singh, Anoop, Agnès Belaisch, Charles Collyns, Paula De Masi, Reva Krieger, Guy Meredith, and Robert Rennhack, 2005, “Stabilization and Reform in Latin America: A Macroeconomic Perspective of the Experience Since the 1990s,” (Washington: International Monetary Fund)
- Wong, Kit Pong, 1997, “On the Determinants of Bank Interest Margins Under Credit and Interest Rate Risks,” Journal of Banking and Finance Vol. 21, No. 2, pp. 251–271
- World Bank, 2004a, “Brazil: Access to Financial Services,” Report No. 27773-BR (Washington, D.C.)
- World Bank, 2004b, ”Bank Regulation and Supervision Database,” http://www.worldbank.org/research/projects/bank_regulation.htm

### Appendix — Table AI. Determinants of Net Interest Margins: Cross-Country Regressions, 2002 (excluding Brazil, Malawi, and Venezuela)
- Dependent variables and columns: (1) (2) (3) (4) (5) (6) (7) (8) (9)
- Ln(Size): -0.525; -0.476; -0.221; -0.531; -0.484; -0.581; -0.518; -0.461; -0.389
  - t-statistics (in parentheses): (4.42)**; (3.82)***; (-1.77)**; (4.37)***; (2.99)***; (3.25)***; (3.86)***; (3.30)***; (3.55)***
- Equ./Tot. Assets: 4.248; 4.685; 6.605; 3.780; 3.974; 3.272; 3.907; 3.374; 4.220
  - t-statistics: (1.94)*; (1.81)*; (3.46)**; (1.72)*; (1.48); (1.15); (1.70)*; (1.39); (1.32)
- Overhead: 53.058; 40.114; 85.155; 53.104; 56.177; 47.880; 51.389; 51.520
  - t-statistics: (3.71)**; (4.27)***; (14.28)** *; (3.64)***; (3.34)***; (2.69)**; (3.53)***; (3.35)***
- Mkt. share: 1.392; (blank); 2.089; (blank); 2.342
  - t-statistics: (1.14); (1.33); (0.73)
- Profit tax rate: 1.130; 1.262; 1.473; 1.086; 1.188; 2.025; 1.096; 0.751; 1.456
  - t-statistics: (1.40); (1.70)*; (1.77); (1.34); (1.31); (2.25)**; (1.42); (0.89); (1.90)*
- Res. Req.: 0.075; 0.058; 0.040; 0.090; 0.101; 0.055; 0.090; 0.089; 0.082
  - t-statistics: (1.75)*; (1.19); (1.01); (2.14)**; (2.44)**; (0.63); (2.25)**; (2.10)**; (1.74)*
- Deposit rate: 0.096; (blank); 0.024; 0.077; 0.133; 0.028; 0.086; 0.119; 0.066
  - t-statistics: (6.18)**; (1.20); (2.83)***; (1.89)*; (0.54); (2.32)**; (2.11)**; (1.91)*
- Money mkt rate: (blank); 0.075
  - t-statistics: (3.55)***
- Inflation: (blank); (blank); 0.031; -0.033; 0.047; 0.009; 0.044; 0.017
  - t-statistics: (0.89); (0.40); (0.59); (0.18); (0.91); (0.33)
- GDP growth: (blank); (blank); 0.098; 0.073; 0.241; 0.063; 0.126; 0.021
  - t-statistics: (0.83); (0.61); (0.85); (0.52); (0.91); (0.18)
- Pred. Overhead: (blank); (blank); (blank); (blank); (blank); (blank); (blank); (blank); 68.463
  - t-statistic: (2.74)**
- 5-Bank-Concentration Ratio: (blank) ... - 0.030
  - t-statistic: (1.79)*
- PROP: 0.169
  - t-statistic: (0.47)
- Legprop: -0.022
  - t-statistic: (0.10)
- Avail. Of Inf.: 0.317
  - t-statistic: (0.66)
- ICRG score: 0.030
  - t-statistic: (0.55)
- St. dev ( infl.): -0.001
  - t-statistic: (3.01)***
- Foreign: -0.790**
  - t-statistic: (1.98)
- Observations: 808; 586; 293; 781; 672; 571; 770; 714; 546
- R-squared: 0.36; 0.38; 0.53; 0.36; 0.37; 0.30; 0.35; 0.34; 0.41
- Note: Robust t statistics in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%

### Appendix — Table AII. Determinants of Spreads: Cross-Country Regressions, 2002 (alternative spread definition)
- Columns: (1) (2) (3) (4) (5) (6) (7) (8) (9)
- Ln(Size): -0.524; 0.178; -0.318; -0.010; -0.035; 0.962; 0.058; 0.473; 0.682
  - t-statistics: (0.65); (0.16); (0.41); (0.01); (0.04); (1.13); (0.07); (0.61); (0.74)
- Equity/Tot. Assets: 29.361; 34.279; 6.844; 30.023; 28.720; 37.830; 31.015; 33.009; 33.392
  - t-statistics: (2.29)**; (2.58)**; (0.68); (2.60)**; (2.34)**; (3.20)***; (2.71)***; (2.82)***; (2.00)*
- Overhead: 71.534; 79.362; 28.447; 68.243; 63.275; 78.594; 70.706; 58.501
  - t-statistics: (2.70)***; (2.91)***; (0.94); (2.69)***; (2.57)**; (3.06)***; (2.76)***; (2.39)**
- Mkt Share: (blank); (blank); 6.452; (blank); (blank); (blank); -0.576; (blank); -14.887
  - t-statistics: (1.10); (0.08); (1.46)
- Profit tax rate: -0.460; -8.137; -6.115; 0.511; -2.494; -0.311; -0.397; 1.054; 2.584
  - t-statistics: (0.21); (1.17); (1.73)*; (0.22); (0.97); (0.13); (0.19); (0.41); (0.98)
- Res. Req.: 0.342; 0.328; 0.690; 0.315; 0.218; 0.337; 0.281; 0.275; 0.332
  - t-statistics: (4.83)***; (4.74)***; (9.31)***; (4.38)***; (2.67)**; (5.11)***; (3.98)***; (3.98)***; (3.64)***
- Deposit rate: 0.850; (blank); 0.588; 0.634; 1.172; 0.619; 0.850; 0.909; 0.363
  - t-statistics: (3.25)***; (5.91)***; (2.29)**; (3.11)***; (3.25)***; (3.25)***; (3.12)***; (1.24)
- Money mkt Interest rate: (blank); (blank); (blank); (blank); (blank); (blank); (blank); 0.893
  - t-statistic: (5.53)***
- Inflation: (blank); (blank); (blank); (blank); 0.168; -0.580; 0.162; -0.105; 0.524; 0.430
  - t-statistics: (0.50); (1.25); (0.80); (0.36); (1.86)*; (1.30)
- GDP growth: (blank); (blank); (blank); (blank); (blank); -1.196; -1.296; -1.426; -1.197; -0.577; -1.501
  - t-statistics: (2.69)***; (3.74)***; (5.05)***; (2.79)***; (1.20); (4.75)***
- Pred. Overh.: (blank for most); 102.694
  - t-statistic: (4.11)***
- 5-bank-conc.: (blank for most); -0.013
  - t-statistic: (0.25)
- PROP: 1.202
  - t-statistic: (0.97)
- Legprop: -0.197
  - t-statistic: (0.25)
- Avail: -1.210
  - t-statistic: (1.37)
- ICRG: -0.217
  - t-statistic: (0.89)
- St. dev (Infl): -0.000
  - t-statistic: (0.04)
- Foreign: 1.679
  - t-statistic: (0.99)
- Observations: 806; 592; 292; 780; 681; 584; 770; 723; 576
- R-squared: 0.26; 0.33; 0.38; 0.28; 0.31; 0.32; 0.29; 0.29; 0.31
- Dependent variable defined as Spreads2 = ((Interest income/Total Loans)-(Interest Expenses/Total Deposits))*100. Robust t statistics are given in parentheses. * denotes significant at the 10 percent level; ** significant at the five percent level; *** significant at the one percent level.

*Source: _wp0644 - REFERENCES (PDF).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0644.pdf_