## _wp07101 - References

## Source details

**Canonical URL:** [_wp07101 - References](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07101.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07101.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07101.pdf.json)

---

### Introduction and purpose
- Focus: examine deposit-lending interest rate spreads in a selected group of English-speaking African countries.
- Structure: background/theory; data and countries; single- and multivariate analysis of determinants of spreads; impact on growth and investment; summary, conclusions, and policy recommendations; appendix on behavior of spreads.

### Background and theoretical context
- Prior findings cited: higher spreads correlated with higher inflation, higher real interest rates, and less competition; institutional factors (creditor rights, corruption, reserves) matter more in developing countries.
- Potential drivers of spreads emphasized:
  - higher operating costs (personnel, reserves, governance, NPLs, inefficiency),
  - reduced competition/concentration,
  - uncertainty (inflation, interest-rate volatility, exchange-rate variability),
  - size and structure of banking sector,
  - public ownership.

### Data and measurement
- Two primary spread measures:
  - Raw spread = difference between a uniform lending rate (prime or equivalent) and a uniform deposit rate.
  - Net interest income to assets (net interest margin), data available for 1997-2004.
- Data sources and period:
  - WEO and IFS (1975-2004) for interest rates and macro aggregates.
  - BankScope for 1997-2004: net interest income to assets, concentration, public and foreign ownership shares, provisioning, NPLs, costs; Herfindahl indices constructed.
  - Sample countries (18): Botswana, Ethiopia, Gambia, Ghana, Kenya, Lesotho, Malawi, Mauritius, Mozambique, Namibia, Nigeria, Rwanda, Sierra Leone, Swaziland, Tanzania, Uganda, Zambia, Zimbabwe.
- Inflation adjustment formulas applied:
  - adjusted spread = spread / (1+inflation)
  - adjusted net interest margin = net interest margin / (1+inflation)
- Panel regressions used; regressions include AR(1) term due to serial correlation.

### Factors affecting spreads — single-variable findings (A–P)
- Time
  - Trend (1977-2003): annual increase in inflation-adjusted spreads of 0.29 percentage points.
  - Time variable significant for regressions starting before 1993; strongest positive trend during 1986-2000.
- Competition (concentration, number of banks)
  - No robust evidence that higher concentration raises spreads; simple regression of adjusted spreads on asset share of 3 largest banks produced a negative significant coefficient that lost significance when YEAR added.
  - Number of banks sometimes positively related to spreads (more banks associated with higher spreads, nearly significant at 95 percent in some tests).
  - Herfindahl indices generally insignificant.
- Size of the economy
  - Adjusted spreads: real GDP (US$) not significant.
  - Adjusted net interest income/assets: real GDP (US$) significant positive coefficient.
- Size of the banking sector
  - Bank assets/GDP not significant in regressions with spreads or average interest income (bank assets/GDP not robust).
- Operating costs
  - Measures examined: total costs, operating expenses/assets, salaries/assets.
  - Adjusted spreads: cost measures positive but not significant.
  - Adjusted net interest income/assets: salaries/assets and total operating expenses/assets significant (positive).
- Loan quality (NPLs, provisioning)
  - Adjusted spreads: NPLs not significant.
  - Adjusted net interest income/assets: NPLs significant positive; provisioning sometimes significant (negative) for net interest income.
- Capitalization
  - Higher capital-to-asset ratios associated with higher spreads and higher net interest margins (data limited to 26 observations 1998–2003; BankScope measure gave similar results).
- Inflation
  - Contrary to many other studies, higher inflation strongly associated with LOWER inflation-adjusted spreads.
  - Example: an additional 12 points of inflation associated with about a half point decline in inflation-adjusted spread (high significance reported).
  - End-of-period inflation more significant than year-average inflation.
  - Unadjusted spreads: inflation not significant; unadjusted net interest income to assets: inflation significant negative.
- Changes in inflation (volatility)
  - Annual changes and three-year variances of inflation insignificant for adjusted spreads and adjusted net interest income/assets.
- Broad money growth
  - Broad money growth associated with lower nominal spreads, but significance disappears when inflation included; when both included, inflation remains significant while broad money growth does not.
- Interest rates
  - Regressing adjusted spreads: lending rates positive and significant; deposit rates negative and significant.
  - Regressing adjusted net interest income/assets: deposit rates and other interest rates had positive coefficients, many significant — suggests compositional effects of lending and banking structure.
- Exchange rates and external sector
  - Exchange-rate variance (3- and 5-year) not significant for spreads, net interest margins, or adjusted net interest income/assets.
  - No relationship with foreign reserves, foreign liabilities, or FDI found.
- Public sector involvement
  - Public ownership share of banking sector assets: significant positive coefficient on adjusted spreads; positive but not always significant for adjusted net interest income/assets.
  - Fiscal expenditure/GDP positive and significant for net interest income/assets in some regressions; primary fiscal balance/GDP had a small negative coefficient in some adjusted spread regressions (not robust).
- Foreign involvement
  - Foreign ownership share of banks not significant in regressions of adjusted spreads or adjusted net interest income/assets.
  - In some regressions both foreign and public ownership coefficients positive (not significant).
- Governance and regulatory quality
  - Quality of regulatory framework had negative coefficient (better regulation → lower spreads) and was significant when regressing interest earnings to assets on it (limited data).
- Required reserve ratios
  - Limited data; simple regression yields positive coefficient (as expected) but not significant.

### Multivariate results
- Short-term (primarily 1998–2004)
  - Most robust determinants of adjusted interest rate spreads: inflation (end of period, negative coefficient), number of banks (positive), deposit rate (negative), public ownership share of banks (positive).
  - Example coefficients from multivariate regressions of Spread (1998 - 2004):
    - Inflation (end of period): −0.11 (t-statistics reported; robust across specifications).
    - Deposit rate: −0.13 to −0.15.
    - Number of banks: 0.08–0.09.
    - Public ownership share of banks: positive (coefficients 3.77 to 5.45 across specs).
  - Adjusted net interest income/assets multivariate results: deposit rate turned positive and significant; operating expenses/assets and real GDP (in US$) became significant in some specifications; NPLs near significance.
- Long-term (1977–2004 or similar)
  - Inflation and deposit rate remained robust (inflation end-of-period negative; deposit rate negative).
  - Fiscal primary balance/GDP sometimes significant (negative coefficient).
  - Year/time trend significant (positive trend), suggesting omitted structural change (candidate: improvements in supervision/regulation).
  - Bank assets/GDP and real GDP often insignificant for adjusted spreads in long-term regressions.

### Effects of spreads on macro outcomes
- Growth
  - Higher inflation-adjusted spreads negatively associated with subsequent GDP growth.
  - Simple regression: 5-year PPP growth on adjusted spread yields negative and significant coefficient.
  - This significance persists when controlling for initial PPP per capita and investment/GDP; persists when lending rate added.
  - Lag structure: regressions for subsequent n years show similar signs but less significance for n ≠ 5.
- Investment
  - Investment rate regressions on adjusted spreads with lags 0–5 years yield insignificant relationships.
  - Thus, spreads appear to affect growth via channels other than measured investment rate alone.

### Summary and conclusions (selected)
- Robust determinants of inflation-adjusted spreads in English-speaking Africa include: inflation (negative), deposit rates (negative), number of banks (positive), and public ownership share of banking assets (positive).
- Higher net interest income may reflect banking sector deepening into retail and SME lending or lending to riskier clients; this can raise net interest income without necessarily raising measured deposit-lending spreads.
- Time trend (especially 1986–2000) significant across regressions — suggests a structural change not captured by available variables (possible candidate: stronger regulation and supervision).
- Many variables commonly thought to raise spreads (exchange-rate variability, broad set of external variables, foreign ownership) were not robustly significant in this sample.
- Higher spreads associated with lower growth; effect not explained by a decline in measured investment rates.

### Policy recommendations (as reported)
- Do not automatically target lower spreads without diagnosing causes: higher spreads may reflect desirable developments (financial sector deepening, more retail/SME lending, improved supervision) and may be necessary to cover real costs and risks.
- When higher spreads stem from undesirable sources (public sector ownership inefficiencies, political interference), policy should address those root causes.
- Authorities should:
  - Investigate the composition of lending (retail/SME versus large enterprise/SOE) to interpret spreads and net interest incomes properly.
  - Monitor public ownership and market power dynamics; maintain competition and effective supervision.
  - Recognize that improved supervision/regulation may raise spreads by internalizing costs formerly borne by taxpayers; seek to minimize administrative costs while maintaining supervision effectiveness.
- Data gaps noted: comprehensive time series on retail/SME lending, public influence measures, effectiveness of regulation and supervision, and required reserve ratios — collecting these would improve diagnosis and policy design.

### Appendix — key descriptive statistics (selected)
- Average adjusted spread, 1975–2004:
  - All countries mean: 7.03
  - Country range (means): Botswana 3.53; Zambia 11.01; sample includes Botswana, Ethiopia, Gambia, Ghana, Kenya, Lesotho, Malawi, Mauritius, Mozambique, Namibia, Nigeria, Rwanda, South Africa, Seychelles, Sierra Leone, Swaziland, Tanzania, Uganda, Zambia, Zimbabwe.
- Spread variability:
  - Standard deviation of adjusted spreads (all countries): 4.17
  - Country examples: Seychelles Std. Dev. 0.42; Malawi Std. Dev. 6.02.
- Adjusted net interest income to assets, 1997–2004:
  - All countries mean: 5.87
  - Country range (means): Ethiopia 2.42; Uganda 8.42.
  - Std. Dev. (all): 2.35

*— _wp07101 - References*

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

### _wp07101 - References

### Introduction and purpose
- Focus: examine deposit-lending interest rate spreads in a selected group of English-speaking African countries.
- Structure: background/theory; data and countries; single- and multivariate analysis of determinants of spreads; impact on growth and investment; summary, conclusions, and policy recommendations; appendix on behavior of spreads.

### Background and theoretical context
- Prior findings cited: higher spreads correlated with higher inflation, higher real interest rates, and less competition; institutional factors (creditor rights, corruption, reserves) matter more in developing countries.
- Potential drivers of spreads emphasized: higher operating costs (personnel, reserves, governance, NPLs, inefficiency), reduced competition/concentration, uncertainty (inflation, interest-rate volatility, exchange-rate variability), size and structure of banking sector, public ownership.

### Data and measurement
- Two primary spread measures:
  - Raw spread = difference between a uniform lending rate (prime or equivalent) and a uniform deposit rate.
  - Net interest income to assets (net interest margin), data available for 1997-2004.
- Data sources and period:
  - WEO and IFS (1975-2004) for interest rates and macro aggregates.
  - BankScope for 1997-2004: net interest income to assets, concentration, public and foreign ownership shares, provisioning, NPLs, costs; Herfindahl indices constructed.
  - Sample countries (18): Botswana, Ethiopia, Gambia, Ghana, Kenya, Lesotho, Malawi, Mauritius, Mozambique, Namibia, Nigeria, Rwanda, Sierra Leone, Swaziland, Tanzania, Uganda, Zambia, Zimbabwe.
- Inflation adjustment formulas applied:
  - adjusted spread = spread / (1+inflation)
  - adjusted net interest margin = net interest margin / (1+inflation)
- Panel regressions used; regressions include AR(1) term due to serial correlation.

### Factors affecting spreads — single-variable findings (A–P)
- Time
  - Trend (1977-2003): annual increase in inflation-adjusted spreads of 0.29 percentage points.
  - Time variable significant for regressions starting before 1993; strongest positive trend during 1986-2000.
- Competition (concentration, number of banks)
  - No robust evidence that higher concentration raises spreads; simple regression of adjusted spreads on asset share of 3 largest banks produced a negative significant coefficient that lost significance when YEAR added.
  - Number of banks sometimes positively related to spreads (more banks associated with higher spreads, nearly significant at 95 percent in some tests).
  - Herfindahl indices generally insignificant.
- Size of the economy
  - Adjusted spreads: real GDP (US$) not significant.
  - Adjusted net interest income/assets: real GDP (US$) significant positive coefficient.
- Size of the banking sector
  - Bank assets/GDP not significant in regressions with spreads or average interest income (bank assets/GDP not robust).
- Operating costs
  - Measures examined: total costs, operating expenses/assets, salaries/assets.
  - Adjusted spreads: cost measures positive but not significant.
  - Adjusted net interest income/assets: salaries/assets and total operating expenses/assets significant (positive).
- Loan quality (NPLs, provisioning)
  - Adjusted spreads: NPLs not significant.
  - Adjusted net interest income/assets: NPLs significant positive; provisioning sometimes significant (negative) for net interest income.
- Capitalization
  - Higher capital-to-asset ratios associated with higher spreads and higher net interest margins (data limited to 26 observations 1998–2003; BankScope measure gave similar results).
- Inflation
  - Contrary to many other studies, higher inflation strongly associated with LOWER inflation-adjusted spreads.
  - Example: an additional 12 points of inflation associated with about a half point decline in inflation-adjusted spread (high significance reported).
  - End-of-period inflation more significant than year-average inflation.
  - Unadjusted spreads: inflation not significant; unadjusted net interest income to assets: inflation significant negative.
- Changes in inflation (volatility)
  - Annual changes and three-year variances of inflation insignificant for adjusted spreads and adjusted net interest income/assets.
- Broad money growth
  - Broad money growth associated with lower nominal spreads, but significance disappears when inflation included; when both included, inflation remains significant while broad money growth does not.
- Interest rates
  - Regressing adjusted spreads: lending rates positive and significant; deposit rates negative and significant.
  - Regressing adjusted net interest income/assets: deposit rates and other interest rates had positive coefficients, many significant — suggests compositional effects of lending and banking structure.
- Exchange rates and external sector
  - Exchange-rate variance (3- and 5-year) not significant for spreads, net interest margins, or adjusted net interest income/assets.
  - No relationship with foreign reserves, foreign liabilities, or FDI found.
- Public sector involvement
  - Public ownership share of banking sector assets: significant positive coefficient on adjusted spreads; positive but not always significant for adjusted net interest income/assets.
  - Fiscal expenditure/GDP positive and significant for net interest income/assets in some regressions; primary fiscal balance/GDP had a small negative coefficient in some adjusted spread regressions (not robust).
- Foreign involvement
  - Foreign ownership share of banks not significant in regressions of adjusted spreads or adjusted net interest income/assets.
  - In some regressions both foreign and public ownership coefficients positive (not significant).
- Governance and regulatory quality
  - Quality of regulatory framework had negative coefficient (better regulation → lower spreads) and was significant when regressing interest earnings to assets on it (limited data).
- Required reserve ratios
  - Limited data; simple regression yields positive coefficient (as expected) but not significant.

### Multivariate results
- Short-term (primarily 1998–2004)
  - Most robust determinants of adjusted interest rate spreads: inflation (end of period, negative coefficient), number of banks (positive), deposit rate (negative), public ownership share of banks (positive).
  - Example coefficients from multivariate regressions of Spread (1998 - 2004):
    - Inflation (end of period): −0.11 (t-statistics reported; robust across specifications).
    - Deposit rate: −0.13 to −0.15.
    - Number of banks: 0.08–0.09.
    - Public ownership share of banks: positive (coefficients 3.77 to 5.45 across specs).
  - Adjusted net interest income/assets multivariate results: deposit rate turned positive and significant; operating expenses/assets and real GDP (in US$) became significant in some specifications; NPLs near significance.
- Long-term (1977–2004 or similar)
  - Inflation and deposit rate remained robust (inflation end-of-period negative; deposit rate negative).
  - Fiscal primary balance/GDP sometimes significant (negative coefficient).
  - Year/time trend significant (positive trend), suggesting omitted structural change (candidate: improvements in supervision/regulation).
  - Bank assets/GDP and real GDP often insignificant for adjusted spreads in long-term regressions.

### Effects of spreads on macro outcomes
- Growth
  - Higher inflation-adjusted spreads negatively associated with subsequent GDP growth.
  - Simple regression: 5-year PPP growth on adjusted spread yields negative and significant coefficient.
  - This significance persists when controlling for initial PPP per capita and investment/GDP; persists when lending rate added.
  - Lag structure: regressions for subsequent n years show similar signs but less significance for n ≠ 5.
- Investment
  - Investment rate regressions on adjusted spreads with lags 0–5 years yield insignificant relationships.
  - Thus, spreads appear to affect growth via channels other than measured investment rate alone.

### Summary and conclusions (selected)
- Robust determinants of inflation-adjusted spreads in English-speaking Africa include: inflation (negative), deposit rates (negative), number of banks (positive), and public ownership share of banking assets (positive).
- Higher net interest income may reflect banking sector deepening into retail and SME lending or lending to riskier clients; this can raise net interest income without necessarily raising measured deposit-lending spreads.
- Time trend (especially 1986–2000) significant across regressions — suggests a structural change not captured by available variables (possible candidate: stronger regulation and supervision).
- Many variables commonly thought to raise spreads (exchange-rate variability, broad set of external variables, foreign ownership) were not robustly significant in this sample.
- Higher spreads associated with lower growth; effect not explained by a decline in measured investment rates.

### Policy recommendations (as reported)
- Do not automatically target lower spreads without diagnosing causes: higher spreads may reflect desirable developments (financial sector deepening, more retail/SME lending, improved supervision) and may be necessary to cover real costs and risks.
- When higher spreads stem from undesirable sources (public sector ownership inefficiencies, political interference), policy should address those root causes.
- Authorities should:
  - Investigate the composition of lending (retail/SME versus large enterprise/SOE) to interpret spreads and net interest incomes properly.
  - Monitor public ownership and market power dynamics; maintain competition and effective supervision.
  - Recognize that improved supervision/regulation may raise spreads by internalizing costs formerly borne by taxpayers; seek to minimize administrative costs while maintaining supervision effectiveness.
- Data gaps noted: comprehensive time series on retail/SME lending, public influence measures, effectiveness of regulation and supervision, and required reserve ratios — collecting these would improve diagnosis and policy design.

### Appendix — key descriptive statistics (selected)
- Average adjusted spread, 1975–2004:
  - All countries mean: 7.03
  - Country range (means): Botswana 3.53; Zambia 11.01; sample includes Botswana, Ethiopia, Gambia, Ghana, Kenya, Lesotho, Malawi, Mauritius, Mozambique, Namibia, Nigeria, Rwanda, South Africa, Seychelles, Sierra Leone, Swaziland, Tanzania, Uganda, Zambia, Zimbabwe.
- Spread variability:
  - Standard deviation of adjusted spreads (all countries): 4.17
  - Country examples: Seychelles Std. Dev. 0.42; Malawi Std. Dev. 6.02.
- Adjusted net interest income to assets, 1997–2004:
  - All countries mean: 5.87
  - Country range (means): Ethiopia 2.42; Uganda 8.42.
  - Std. Dev. (all): 2.35

*— _wp07101 - References*

### References

### _wp07101 - References

### Cited works
- Banco Central do Brasil (1999), “Juros e Spread Bancario no Brasil.”
- Barajas, Adolpho, Roberto Steiner, and Natalia Salazar (1999), “Interest Spreads in Banking in Colombia, 1974–96,” IMF Staff Papers, Vol. 46, No. 2 (June 1999).
- Beck, Thorsten (2000), “Impediments to the Development and Efficiency of Financial Intermediation in Brazil,” World Bank Policy Research Working Paper 2382 (June).
- Beck, Thorsten, Asli Demirgüç-Kunt, and Ross Levine (2005), “Bank Concentration and Fragility: Impact and Mechanics,” NBER Working Paper No. 11500.
- Beck, Thorsten and Ross Levine (2001), “Industry Growth and Capital Allocation: Does Having a Market- or Bank-Based System Matter?” mimeo.
- Berger, Allen N (1995), "The Relationship between Capital and Earnings in Banking," Journal of Money, Credit and Banking, Ohio State University Press, vol. 27(2), pp 432-56, May.
- Berger, Allen N. and Timothy H. Hannan (1989), “The Price-Concentration Relationship in Banking,” Review of Economics and Statistics, Vol. 71, No. 2, pp. 291-99.
- Bernanke, Ben (1990), “On the Predictive Power of Interest Rates and Interest Rate Spreads,” NBER Working paper 3486.
- Brock, Philip, and Liliana Rojas-Suárez (2000), “Understanding the Behavior of Bank Spreads in Latin America,” Journal of Development Economics, Vol. 63, No. 1, pp. 113–34.
- Chirwa, Ephraim and Montfort Mlachila (2004) “Financial Reforms and Interest Rate Spreads in the Commercial Banking System in Malawi,” IMF Staff Papers, Vol 51, No. 1.
- Claessens, Stijn, Asli Demirgüç-Kunt, and Harry Huizinga (2001), “How Does Foreign Entry Affect the Domestic Banking Market?” Journal of Banking and Finance, Vol 25, pp 891-911.
- Claessens, Stijn and Luc Laeven (2004), “What Drives Bank Competition? Some International Evidence,” Journal of Money, Credit, and Banking, Vo. 36, No. 3, June, pp. 133–152.
- Demirguc -Kunt, Asli and Huizinga, Harry (1999), “Determinants of Commercial Bank Interest Margins and Profitability: Some International Evidence,” World Bank Economic Review 13, 379-408.
- Demirguc -Kunt, Asli, Luc Laeven, and Ross Levine (2003), “The Impact of Bank Regulations, Concentration, and Institutions on Bank Margins,” mimeo.
- Detragiache, Enrica, Thierry Tressel, and Poonam Gupta (2006), “Foreign Banks in Poor Countries: Theory and Evidence,” Paper presented at the 7th Jacques Polak Annual Research Conference Hosted by the International Monetary Fund─November 9-10, 2006
- Gelos, R. Gaston (2006), “Banking Spreads in Latin America,” IMF Working Paper WP/06/44.
- Gischer, Horst and D. Johannes Juttner (2002), “Global Competition, Fee Income, and Interest Rate Margins of Banks,” Paper presented at the Financial Management Association Conference in Copenhagen in June 2002.
- Hanson, James A., and Roberto de Rezende Rocha (1986), “High Interest Rates, Spreads, and the Cost of Intermediation: Two Studies,” Industry and Finance Series 18. World Bank, Industry Department, Washington, D.C.
- Ho, Thomas S. Y. and Anthony Saunders (1981), “The Determinants of Bank Interest Margins: Theory and Empirical Evidence,” Journal of Financial and Quantitative Analysis, 16, 581-600.
- Kahn, M.S., and A. S. Senhadji (2000), “Financial Development and Economic Growth: An Overview,” IMF Working Paper WP/00/209.
- King, Robert G and Levine, R. (1993), "Finance and Growth: Schumpeter Might Be Right," Quarterly Journal of Economics, MIT Press, vol. 108(3), pp 717-37, August.
- La Porta, Rafael, Florencio Lopez de-Silanes, and Andrei Schleifer, (2002), “Government Ownership of Banks,” Journal of Finance, Vol LVII, Feb.
- Levine, Ross (1997). “Financial Development and Economic Growth,” Journal of Economic Literature 35(2):688–726.
- Randall, Ruby (1998), “Interest rate spreads in the Eastern Caribbean,” IMF Working Paper WP/98/59.
- Saxegaard, Magnus (2006), “Excess Liquidity and the Effectiveness of Monetary Policy: Evidence from Sub-Saharan Africa.” IMF Working Paper WP/06/115.

*Source: _wp07101 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07101.pdf_