## _wp11275 - 2008. The cost of provisioning for bad loans could therefore partly explain the high lending interest rates and the high interest margins ## Source details **Canonical URL:** [_wp11275 - 2008. The cost of provisioning for bad loans could therefore partly explain the high lending interest rates and the high interest margins](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11275.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11275.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11275.pdf.json) --- ### Financial sector background and monetary framework - Bank of Ghana Act (2002) mandates price stability as primary objective and created the monetary policy committee (MPC). - Formal inflation targeting started in May 2007; preceding transition involved shifting from reserve money targeting to broader indicator analysis. - Key financial liberalizations and developments: - Liberalization of exchange and interest rate controls. - Partial opening of external capital account allowing foreigners participation in longer-end domestic bond market. - Introduction and spread of payment instruments (credit/debit cards) and electronic banking. - Sample and data windows cited: - Wholesale interest rate series: 2004M12–2011M5 and 2004M12–2011M7 in figures. - Econometric samples: 2004M12—2010M4 (wholesale rates) and 2005Q1—2010Q1 (bank-specific retail rates). - Some wholesale estimations comprise only 65 observations. ### Illustrative bank optimization and pass-through model - Operational target: Bank of Ghana targets the interbank money market interest rate. - Risk-neutral bank chooses loan portfolio (L), deposit portfolio (D), and end-of-day reserves (R_t). - Intermediation cost function c(L) with c’>0 and c’’>0 (marginal cost increasing in loan portfolio). - Central bank policy rate described as prime rate with an asymmetric band: - deposit rate = prime rate − h’ - lending (central bank) rate = prime rate + h’’ - Key relationships: - Lending interest rate (i_L) depends on interbank rate (i_IB) and marginal intermediation cost ρ′(L). - Pass-through from interbank to lending rates is dampened when c’’>0 (steeply rising intermediation costs). - Interbank rate equation incorporates central bank liquidity stance, reserves (R_t), money market short u and settlement uncertainty v. ### Econometric representation for pass-through (generic) - Error-correction style specification (equation 3) with: - short-run dynamics via ∆ terms, - long-run elasticity denoted by ϕ, - speed of convergence denoted by μ. - Error term ε_t assumed normally distributed with zero mean and constant variance. ### Wholesale market empirical findings - Benchmarks: interbank money market interest rate and 91-day treasury bill interest rate. - Interbank rate findings: - Responds to changes in Bank of Ghana prime rate with a one-month lag (all Models). - Significant inertia: significant first lag of dependent variable. - Long-run parity: estimated constant α and slope δ not statistically different from zero and one, respectively (text statement). - Speed of adjustment estimate: μ = -0.07 (statistically significant) — described as slow. - Asymmetry: faster convergence when policy rate falling (Model 3) than when rising (Model 4). - Interbank and treasury bill rates were significantly higher than prime rate during 2008-2009 for prolonged periods. - 91-day treasury bill findings: - Strong pass-through from prime rate to 91-day TB in recent period (all Models). - Policy rate influences 91-day TB within a one-month lag; same-period effect often not significant. - Inertia in TB rate stronger than interbank rate. - Significant contemporaneous effect from interbank rate to treasury bill rate (all Models). - VAR between interbank and 91-day TB: - Changes in prime rate influence both wholesale rates with a one-month lag (about one-half of pass-through occurs in this period). - Convergence to long-run parity is complete but takes time; interbank equation has low adjusted R-squared indicating unexplained variation. ### Retail deposit and lending rate findings (bank-specific quarterly data) - Deposits: - Demand deposit rates changed infrequently and were excluded. - Time deposit interest rates: - 91-day TB significant in explaining both time and savings deposit rates (all Models). - Interbank rate significant for time deposits (Models 2 and 3 in Table 6). - Cross-rate influences: demand and savings deposit rate changes affect time deposit rates. - Long-term convergence: strong convergence between time deposit and wholesale rates in Models 5–7 (Table 6). - Speed of adjustment for time deposit to interbank gap in Model 7: -0.28 (highly significant); long-run convergence complete. - Fixed (bank-specific) effects important. - Savings deposit interest rates: - Long-run convergence less strong than for time deposits; convergence to TB and time-deposit rates significant but incomplete. - Bank-specific effects important (Models 6–7 in Table 7). - Retail lending interest rates: - Short-run effect from interbank rate is immediate (all Models). - Short-run pass-through from 91-day TB significant only when long-run effects excluded (Models 2–4). - Time deposit rate has significant short-run effect on lending rates (Models 3–7 in Table 8), indicating deposit cost channel to lending rates. - Long-run (level) effects (Models 5–7 in Table 8): - Retail lending rates converge to a weighted average of interbank and time deposit rates (Model 7). - Estimated speed of adjustment for lending rates: -0.30 (highly significant). - Pass-through to lending rates is incomplete in the long-run. - Bank-specific fixed effects significant — adjustment differs across banks. ### Simulation exercises and quantitative scenarios - Wholesale market VAR simulation (Table 5): - Shock: 500 basis points reduction in prime rate (reflecting Bank of Ghana monetary easing during October 2009–July 2010). - Result: interbank and 91-day TB adjust comparably within two months, then “overshoot” the policy rate and converge back within 24 months. - Treasury bill rate found more volatile than interbank rate in simulations. - Retail market simulation (using Model 7 in Table 6, Model 7 in Table 8, and wholesale simulation): - Assumption: demand and savings deposit rates unchanged for simplicity. - Retail time deposit rate: falls from 12 percent to 7 percent at trough — similar to 2009–2010 experience. - Retail lending rate: falls from 31 percent to 26 percent at trough — similar to actual experience during policy easing. - Timing: retail deposit and lending rates reach lowest points after 5–6 quarters following initial easing. - Spread between retail lending and time deposit rates remains relatively constant during adjustment. ### Conclusions, interpretation, and policy-relevant observations - Overall pass-through: - About one-half of a change in the policy interest rate is reflected in wholesale market rates with a one-month lag. - Wholesale interbank market long-run responses are protracted; deviations from policy rate are substantial and prolonged, weakening Bank of Ghana’s ability to control short-term money market rates. - Retail adjustment: - Banks’ retail rates adjust to changes in wholesale and policy rates but the process is slow and long-run adjustment is incomplete. - Simulations indicate retail time deposit and lending rates can and did respond materially to a 500 basis points policy easing, but only with multi-quarter lags. - Factors contributing to incomplete and slow pass-through: - High levels of problem loans and provisioning costs could explain elevated lending interest rates and margins. - Past high inflation and large fiscal deficits may raise concerns about sustainability of low inflation and fiscal consolidation, contributing to uncertainty and banks’ reluctance to lower retail lending rates rapidly. - Lack of competition / monopolistic competition in the banking system raises intermediation costs and mark-ups. - High financing needs of government and reliance on government bonds reduce competitive pressures for private lending business. - Market structure and institutional factors: settlement uncertainty, shallow money markets, and central bank liquidity management can add to interbank volatility and pass-through frictions. - Research and policy avenues highlighted: - Further research into short- and long-run pass-through nuances and how market development can speed and complete pass-through. - Address banking sector competition, provisioning and non-performing loan levels, payment/settlement efficiency, and central bank liquidity management to improve transmission. ### Key quantitative estimates and statistics - Wholesale interbank speed-of-adjustment estimate: μ = -0.07 (statistically significant). - Time deposit speed-of-adjustment (Model 7): -0.28 (highly significant). - Retail lending speed-of-adjustment (Model 7): -0.30 (highly significant). - Simulation shock used: 500 basis points reduction in prime rate. - Retail time deposit simulated fall: from 12 percent to 7 percent. - Retail lending simulated fall: from 31 percent to 26 percent. - Data windows and sample sizes: - Monthly samples noted: 2004M12–2011M5 and 2004M12–2011M7. - Quarterly sample for bank-specific retail rates: 2005Q1—2010Q1. - Observations in key statistics tables: 65 (some wholesale estimations), 65 (monthly Key Statistics observations), 384 (quarterly Key Statistics observations). *Italicized source: IMF working paper excerpt (content unit: _wp11275 - 2008. The cost of provisioning for bad loans could therefore partly explain the high lending interest rates and the high interest margins).* ### 2008. The cost of provisioning for bad loans could therefore partly explain the high lending ### _wp11275 - 2008. The cost of provisioning for bad loans could therefore partly explain the high lending interest rates and the high interest margins ### Financial sector background and monetary framework - Bank of Ghana Act (2002) mandates price stability as primary objective and created the monetary policy committee (MPC). - Formal inflation targeting started in May 2007; prior transition involved shifting from reserve money targeting to broader indicator analysis. - Key financial liberalizations and developments: - Liberalization of exchange and interest rate controls. - Partial opening of external capital account allowing foreigners participation in longer-end domestic bond market. - Introduction and spread of payment instruments (credit/debit cards) and electronic banking. - Sample and data windows cited in the analysis: - Wholesale interest rate series: 2004M12–2011M5 and 2004M12–2011M7 in figures. - Econometric samples often cover 2004M12—2010M4 (wholesale rates) and 2005Q1—2010Q1 (bank-specific retail rates). - Data sample for some wholesale estimations comprises only 65 observations. ### Illustrative bank optimization and pass-through model - Operational target: Bank of Ghana targets the interbank money market interest rate. - Risk-neutral bank chooses loan portfolio (L), deposit portfolio (D), and end-of-day reserves (R_t). - Intermediation cost function c(L) assumed with c’>0 and c’’>0 (marginal cost increasing in loan portfolio). - Central bank policy rate described as prime rate with an asymmetric band: - deposit rate = prime rate − h’ - lending (central bank) rate = prime rate + h’’ - Key model variables and relationships: - Lending interest rate (i_L) depends on interbank rate (i_IB) and marginal intermediation cost ρ′(L). - Pass-through from interbank to lending rates is dampened when c’’>0 (steeply rising intermediation costs). - Interbank rate equation incorporates central bank liquidity stance, reserves (R_t), and money market short u and settlement uncertainty v. ### Econometric representation for pass-through (generic) - Error-correction style specification used (equation 3) with short-run changes and lagged level term to estimate: - short-run dynamics via ∆ terms, - long-run elasticity denoted by ϕ (presented in text as ϕ or similar), - speed of convergence denoted by μ. - Error term ε_t assumed normally distributed with zero mean and constant variance. ### Wholesale market empirical findings - Wholesale benchmarks analyzed: interbank money market interest rate and 91-day treasury bill interest rate. - Interbank rate: - Responds to changes in Bank of Ghana prime rate with a one-month lag (all Models). - Significant inertia: significant first lag of dependent variable. - Long-run parity: estimated constant α and slope δ not statistically different from zero and one, respectively (text statement). - Speed of adjustment estimate: μ = -0.07 (statistically significant) — described as slow. - Evidence of asymmetry: faster convergence when policy rate falling (Model 3) than when rising (Model 4). - Volatility and deviations: interbank and treasury bill rates were significantly higher than prime rate during 2008-2009 for prolonged periods. - 91-day treasury bill: - Strong pass-through from prime rate to 91-day TB in recent period (all Models). - Policy rate influences 91-day TB within a one-month lag; same-period effect often not significant. - Inertia in TB rate stronger than interbank rate. - Significant contemporaneous effect from interbank rate to treasury bill rate (all Models). - Vector autoregressive (VAR) between interbank and 91-day TB: - Changes in prime rate influence both wholesale rates with a one-month lag (about one-half of pass-through occurs in this period). - Convergence to long-run parity is complete but takes time; interbank equation has low adjusted R-squared indicating unexplained variation. ### Retail deposit and lending rate findings (bank-specific quarterly data) - Retail deposits analyzed: time and savings deposit interest rates (demand deposit rates changed infrequently and were excluded). - Time deposit interest rates: - 91-day TB is significant in explaining both time and savings deposit rates (all Models). - Interbank rate significant for time deposits (Models 2 and 3 in Table 6). - Deposit-rate-setting exhibits cross-rate influences: demand and savings deposit rate changes affect time deposit rates. - Long-term convergence: strong convergence between time deposit and wholesale rates in Models 5–7 (Table 6). - In Model 7, speed of adjustment for time deposit to interbank gap: -0.28 (highly significant); long-run convergence complete. - Fixed (bank-specific) effects important. - Savings deposit interest rates: - Long-run convergence less strong than for time deposits; convergence to TB and time-deposit rates significant but incomplete. - Bank-specific effects important (Models 6–7 in Table 7). - Retail lending interest rates: - Short-run effect from interbank rate is immediate (all Models). - Short-run pass-through from 91-day TB significant only when long-run effects excluded (Models 2–4). - Time deposit rate has significant short-run effect on lending rates (Models 3–7 in Table 8), indicating deposit cost channel to lending rates. - Long-run (level) effects (Models 5–7 in Table 8): - Retail lending rates converge to a weighted average of interbank and time deposit rates (Model 7). - Estimated speed of adjustment for lending rates: -0.30 (highly significant). - Pass-through to lending rates is incomplete in the long-run. - Bank-specific fixed effects significant — adjustment differs across banks. ### Simulation exercises and quantitative scenarios - Wholesale market simulation using VAR estimates (Table 5): - Shock: 500 basis points reduction in prime rate (reflecting Bank of Ghana monetary easing during October 2009–July 2010). - Result: interbank and 91-day TB adjust comparably within two months, then “overshoot” the policy rate and converge back within 24 months. - Treasury bill rate found more volatile than interbank rate in simulations. - Retail market simulation using retail estimates (Model 7 in Table 6 and Model 7 in Table 8) and wholesale simulation: - Assumed demand and savings deposit rates unchanged for simplicity. - Retail time deposit rate: falls from 12 percent (pre-change) to 7 percent at trough — similar to what happened during 2009–2010. - Retail lending rate: falls from 31 percent (pre-change) to 26 percent at trough — similar to actual experience during policy easing. - Timing: retail deposit and lending rates reach lowest points after 5–6 quarters following initial easing. - Spread between retail lending and time deposit rates remains relatively constant during adjustment. ### Conclusions, interpretation, and policy-relevant observations - Overall pass-through: - About one-half of a change in the policy interest rate is reflected in wholesale market rates with a one-month lag. - Wholesale interbank market long-run responses are protracted; deviations from policy rate are substantial and prolonged, weakening the Bank of Ghana’s ability to control short-term money market rates. - Retail adjustment: - Banks’ retail rates adjust to changes in wholesale and policy rates but the process is slow and long-run adjustment is incomplete. - Simulations indicate retail time deposit and lending rates can and did respond materially to a 500 basis points policy easing, but only with multi-quarter lags. - Factors contributing to incomplete and slow pass-through: - High levels of problem loans and provisioning costs could explain elevated lending interest rates and margins. - Past high inflation and large fiscal deficits may raise concerns about sustainability of low inflation and fiscal consolidation, contributing to uncertainty and banks’ reluctance to lower retail lending rates rapidly. - Lack of competition / monopolistic competition in the banking system raises intermediation costs and mark-ups. - High financing needs of government and reliance on government bonds reduce competitive pressures for private lending business. - Market structure and institutional factors: settlement uncertainty, shallow money markets, and central bank liquidity management can add to interbank volatility and pass-through frictions. - Quantified speeds of adjustment and estimates cited in the analysis: - Wholesale interbank speed-of-adjustment estimate: μ = -0.07 (statistically significant). - Time deposit speed-of-adjustment (Model 7): -0.28 (highly significant). - Retail lending speed-of-adjustment (Model 7): -0.30 (highly significant). - Simulation shock used: 500 basis points reduction in prime rate. - Retail time deposit simulated fall: from 12 percent to 7 percent. - Retail lending simulated fall: from 31 percent to 26 percent. - Research and policy avenues highlighted: - Further research into short- and long-run pass-through nuances and how market development can speed and complete pass-through. - Addressing banking sector competition, provisioning and non-performing loan levels, payment/settlement efficiency, and central bank liquidity management to improve transmission. *Italicized source: IMF working paper excerpt (content unit: _wp11275 - 2008. The cost of provisioning for bad loans could therefore partly explain the high lending interest rates and the high interest margins).* ### REFERENCES ### _wp11275 - REFERENCES ### References - Addison, Ernest, 2008, “Lessons from Inflation Targeting in Developing Countries: The Case of Ghana,” International Symposium on Monetary Policy and Price Stability in the West African Economic and Monetary Union Zone: Constraints and Challenges, Dakar, Senegal. - Bondt, Gabe de, 2002, “Retail Bank Interest Rate Pass-Through: New Evidence at the Euro Area Level,” ECB Working Paper Series 02/136 (Frankfurt: European Central Bank). - Borio, Claudio E.V. and Wilhelm Fritz, 1995, “Response of Short-Term Bank Lending Rates to Policy Rates: A Cross-Country Perspective,” Bank for International Settlements, Monetary and Economic Department, Working Papers (International), No. 27, May. - Buchs, Thierry and Johan Mathisen, 2005, “Competition and Efficiency in Banking: Behavioral Evidence from Ghana,” IMF Working Paper 05/17 (Washington: International Monetary Fund). - Chen, Chuling, 2009, “Bank Efficiency in Sub-Saharan African Middle-Income Countries,” IMF Working Paper 09/14 (Washington: International Monetary Fund). - Cottarelli, Carlo and Angeliki Kourelis, 1994, ‘Financial Structure, Bank Lending Rates, and the Transmission Mechanism of Monetary Policy,” IMF Staff Papers, 41, pp. 587-623. - Ennis, Huberto M. and Todd Keister, 2008, “Understanding Monetary Policy Implementation,” Economic Quarterly, Federal Reserve Bank of Richmond, 94, 3, pp. 235-63. - Ghartey, Edward E., 2005, “Monetary Policy on Ghana’s Term Structure of Interest Rates: Effects and Implications,” No. 8, The Institute of Economic Affairs, Accra, Ghana. - Goodfriend, M., 1991, “Interest Rates and the Conduct of Monetary Policy,” Carnegie Rochester Series on Public Policy, 34, pp. 7-30. - Henckel, Timo, Alain Ize, and Arto Kovanen, 1999, “Central Banking Without Central Bank Money,” IMF Working Paper 99/92 (Washington: International Monetary Fund). - Hofmann, Boris and Paul Mizen, 2004, “Interest Rate Pass-Through and Monetary Transmission: Evidence from Individual Financial Institutions’ Retail Rates,’ Economica, 71, pp. 99-123. - International Monetary Fund, 2011, “Ghana: Financial System Stability Assessment Update,” IMF Country Report 11/131 (Washington: International Monetary Fund). - Karagiannis, Stelios, Yannis Panagopoulos, and Prodromos Vlamis, 2010, “Interest Rate Pass-Through in Europe and the US: Monetary Policy After the Financial Crisis,” Journal of Policy Modeling, 32, pp. 323-338. - Kwapil, Claudia and Johann Scharler, 2010, “Interest Rate Pass-Through, Monetary Policy Rules and Macroeconomic Stability,” Journal of International Money and Finance, 29, pp. 236-51. - Liu, Ming-Hua, Dimitri Margaritis, and Alireza Tourani-Rad, 2008, “Monetary Policy Transparency and Pass-Through of Retail Interest Rates,” Journal of Banking and Finance, 32, pp. 501-11. - Mishkin, Frederic S., 1999, “International Experiences With Different Monetary Policy Regimes,” Journal of Monetary Economics, 43, pp. 579-605. - Mishra, Prachi, Peter J. Montiel, and Antonio Spilimbergo, 2010, “Monetary Transmission in Low Income Countries,” Centre for Economic Policy Research, Discussion Paper Series, 7951, August. - Sander, Harald and Stefanie Kleimeier, 2006, “Interest Rate Pass-Through in the Common Monetary Area of the SACU Countries,” South African Journal of Economics, 74, 2, pp. 215-29. - Tieman, Alexander, 2004, “Interest Rate Pass-Through in Romania and Other Central European Countries,“ IMF Working Paper 04/211 (Washington: International Monetary Fund). ### Appendix Table 1: Variable Definitions - Variable name — Description - IB_RATE — Interbank interest rate, annual percent - BOG_RATE — Prime rate (Bank of Ghana's policy rate), annual percent - TB_91_D_RATE — Treasury bill interest rate (91 days maturity), annual percent - DD_RATE — Retail demand deposit interest rate, annual percent - SD_INT — Retail savings deposit interest rate, annual percent - TD_INT — Retail time deposit interest rate, annual percent - LEND_INT — Retail lending interest rate, annual percent - SPREAD # 1 — IB_RATE - BOG_RATE - SPREAD # 2 — TB_91_D_RATE - IB_RATE - SPREAD # 3 — TB_91_D_RATE - BOG_RATE ### Appendix Table 2. Key Statistics (Monthly Data), 2004M12–2010M4 - Variables: BOG_RATE IB_RATE TB_91_D - Me an15.6714.8415.15 - Me di an15.5013.9513.26 - Std. De v.2.144.025.68 - Skewness-0.060.520.54 - Kurtosi s1.692.301.71 - Jarque-Bera4.664.327.72 - Probability0.100.120.02 - Sum1,019965985 - Sum Sq. De v.2931,0332,067 - Obse rvati ons656565 - Sources: Bank of Ghana and author's estimates. ### Appendix Table 3. Key Statistics (Quarterly Data), 2005Q1–2010Q1 - Variables: BOG_RATE IB_RATE TB_91_D DD_INT SD_INT TD_INT LEND_INT - Me an15.5114.8315.190.612.529.5829.05 - Me di an15.4213.0112.680.252.009.3028.66 - Std. De v.2.094.085.880.822.405.333.83 - Ske wne ss0.010.560.512.201.390.570.40 - Kurtosi s1.642.081.549.865.753.332.57 - Jarque - Be ra29.7933.5350.631,063.09245.0822.2913.26 - Probability0.000.000.000.000.000.000.00 - Sum5,9575,6975,8342339683,67811,154 - Sum Sq. De v.1,6726,38313,2642582,20910,8665,619 - Obse rvati ons384384384384384384384 - Sources: Bank of Ghana and author's estimates. *Source: _wp11275 - REFERENCES* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11275.pdf_