## _wp11176

## Source details

**Canonical URL:** [_wp11176](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11176.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11176.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11176.pdf.json)

---

### Overview and objectives
- Monetary transmission transmits central bank policy actions to the real sector; the interest rate channel is central, especially under inflation-targeting frameworks.
- Interest rate pass-through describes how changes in central banks’ policy rates transmit to longer-term retail interest rates on loans and deposits.
- Main objective: explain cross-country differences in pass-through with country-specific structural and macroeconomic characteristics.
- Method: estimate country-specific pass-through coefficients using an autoregressive distributed lags (ARDL) approach for seventy economies; relate estimated pass-through coefficients to macroeconomic variables and indicators of financial market structure.

### Theoretical and econometric framework
- Long-run markup pricing relationship:
  - i_R = i_M · β + α, where i_R and i_M are retail and market rates; β is long-run pass-through; α is markup.
  - Under perfect markets and risk-neutral banks β = 1; empirical evidence usually finds β < 1.
- Short-run and out-of-equilibrium dynamics captured via an error-correction and ARDL representation:
  - Short-run pass-through measured by contemporaneous β_0 or the first statistically significant lagged β_k (k=0,1,2).
  - Long-run β = (Σ_{k=0}^n β_k) / (1 − Σ_{k=1}^n α_k).
- Required conditions for estimates:
  - Σ_{k=0}^n β_k ≥ 0.
  - Σ_{k=1}^n α_k ≤ 1.
- Sample and estimation details:
  - Initial dataset: 81 countries, monthly data December 2005 to March 2010.
  - Money market rates: three-month treasury bill rates (or three-month interbank rates if t-bill unavailable).
  - Retail lending rates: one- to two-year corporate loan rates in domestic currency; average lending rates used where unavailable.
  - Variables log-differenced; two-month lags in ARDL; final sample after filtering: 70 countries.

### Cross-country patterns in pass-through
- General patterns:
  - Large cross-country heterogeneity in short-run and long-run pass-through magnitudes.
  - Pass-through is stronger in the long run than in the short run.
  - More advanced economies tend to have stronger long-run pass-through (greater persistence of market interest rates).
  - Flexible exchange regimes have stronger pass-through than fixed regimes.
- Regional and group findings:
  - Middle East and Central Asia (MEA) have the weakest short-run and long-run pass-through; Islamic banking practices may contribute.
  - North America (United States and Canada) exhibit significantly higher pass-through.
  - Micro states, including small island economies, tend to have weaker pass-through due to shallower financial markets.

### Determinants: empirical specification and variables
- Main hypothesis: long-run pass-through differences are explained by macroeconomic conditions and financial market structure.
- Explanatory variables included (selected):
  - Macroeconomic: per capita GDP, GDP growth, Inflation (annual CPI), Average money market rates (MMR), Market volatility (standard deviation of market rates divided by average rates).
  - Financial sector: liquidity (liquid assets to short-term liabilities / total assets), Overhead costs, NPL (non-performing loans to total loans), ROE, banking system size, deposits, private sector credit.
  - Dummies: Island (small–island economies), MEast (Middle East and Central Asia), ER (exchange regime dummy, ER = 1 for flexible regimes).
- Regression approach: log of long-run pass-through coefficient regressed on macroeconomic and financial-structure variables; macroeconomic and financial regressions run separately.

### Key regression results (dependent variable: log of long-run pass-through coefficient)
- Equation (7)
  - Const: -1.01 (1.0)
  - Per capita GDP: 0.37 *** (0.08)
  - Inflation: 0.48 *** (0.15)
  - Market volatility: -0.28 *** (0.08)
  - Middle East dummy: -2.0 *** (0.37)
  - Island: -0.77 *** (0.26)
  - R squared: 0.41
  - # observations: 69

- Equation (7')
  - Const: -1.40 (0.9)
  - Per capita GDP: 0.37 *** (0.08)
  - Average money market rate: 0.69 *** (0.17)
  - Market volatility: -0.24 *** (0.07)
  - Middle East dummy: -1.71 *** (-0.33)
  - Island: -0.64 *** (0.23)
  - R squared: 0.46
  - # observations: 69

- Equation (8)
  - Const: 7.09 (1.03)
  - Liquidity: -0.32 ** (0.14)
  - Overhead costs: 0.69 *** (0.17)
  - NPL: -0.32 *** (0.12)
  - ROE: -0.21 ** (0.09)
  - Market volatility: -0.19 ** (0.07)
  - Middle East dummy: -1.28 *** (0.47)
  - Exchange regime dummy: 0.63 ** (0.2)
  - Island: -0.64 *** (0.23)
  - R squared: 0.69
  - # observations: 43

- Significance notation:
  - ** denotes significance at 5 percent confidence level.
  - *** denotes significance at 1 percent confidence level.

### Main empirical findings (summarized)
- Variables associated with stronger pass-through:
  - Per capita GDP (positive coefficient).
  - Inflation and average money market rates (positive; MMR specification has better fit and larger coefficient).
  - Overhead costs (positive and significant).
  - Flexible exchange regimes (positive ER dummy).
- Variables associated with weaker pass-through:
  - Market volatility (negative and statistically significant).
  - Excess liquidity (negative relationship).
  - Poor credit quality (higher NPLs; negative coefficient).
  - Higher ROE (negative coefficient, consistent with market power reducing pass-through).
  - Small island states and Middle East and Central Asian countries show weaker pass-through.

### Interpretations and mechanisms
- Market volatility reduces reliability of market signals, prompting banks to delay passing rate changes to customers.
- Excess banking liquidity reduces interbank trading and informativeness of market rates, weakening pass-through.
- High overhead costs at the country level may reflect structural frictions that induce higher net interest margins; banks may pass through market rate increases to preserve margins.
- Poor credit quality raises risk premia; banks absorb cost changes into risk premiums rather than passing them through.
- Higher ROE indicates market power and weaker competition, reducing responsiveness of lending rates to fund costs.
- Flexible exchange regimes allow domestic rates to attract capital and equilibrate yields, strengthening pass-through; fixed regimes lack this buffer.

### Policy implications and recommendations
- Monetary framework choice:
  - If pass-through is weak and not easily improved (e.g., small island states, low-income economies), frameworks relying on strong interest rate pass-through (such as inflation targeting) may not be optimal.
  - Alternative intermediate targets may include monetary aggregates, exchange rate, or other nominal anchors.
- Central bank priorities:
  - Strengthening monetary transmission helps achieve price stability objectives; stronger and faster pass-through eases central bank control.
  - Successful disinflation may weaken pass-through (positive relationship between inflation and pass-through), posing a trade-off for central banks.
- Structural and regulatory policy levers:
  - Improve financial market development and reduce market volatility to strengthen interest rate transmission.
  - Address structural frictions that drive high overhead costs (improvements in regulatory and legal environment, information sharing, removal of obstacles to intermediation).
  - Reduce excess banking liquidity where appropriate.
  - Improve credit quality and reduce NPLs.
  - Strengthen competition in banking.
- Compensation for weakening pass-through:
  - If pass-through weakens due to disinflation or reduced overhead costs, central banks can compensate by:
    - Strengthening competition,
    - Mopping up excess liquidity,
    - Improving credit quality,
    - Developing financial markets to reduce interest rate volatility.

*Source: _wp11176 - IMF Working Paper (excerpts from Introduction, Determinants, and Conclusions)*

### 1. INTRODUCTION ........................................................................................................

### 1. INTRODUCTION

### Overview
- Monetary transmission is the mechanism that transmits central bank policy actions to the real sector and determines the effectiveness of monetary policy.
- The interest rate channel is a traditional and central component of monetary transmission, especially important under inflation-targeting frameworks.
- Interest rate pass-through describes how changes in central banks’ policy rates transmit to longer-term retail interest rates on loans and deposits.
- Numerical estimates of pass-through coefficients vary widely across countries, markets, and time periods.

### Objectives and contribution
- Main objective: explain cross-country differences in pass-through with country-specific structural and macroeconomic characteristics.
- Method: estimate country-specific pass-through coefficients using an autoregressive distributed lags approach, expanding coverage to seventy economies from all parts of the world and of all levels of development.
- Second step: relate estimated pass-through coefficients to macroeconomic variables and indicators of financial market structure.

### Key findings (from the Introduction summary)
- Per capita GDP and inflation have positive effects on pass-through.
- Market volatility has a negative effect on pass-through.
- Among financial market variables:
  - Credit quality, overhead costs, and banking competition strengthen pass-through.
  - Excess banking liquidity impedes pass-through.
- Small island states, countries in Middle East and Central Asia, and economies with fixed exchange regimes tend to have weaker pass-through.
- Policy implications follow from these findings.

---

### Monetary Transmission Mechanism and Interest Rate Pass-Through

### 2.1 Background (channels of transmission)
- Four main channels of monetary transmission (Mishkin, 1996):
  - Exchange rate channel: important at early stages of financial development; exchange rate reacts to monetary policy and feeds into demand and supply.
  - Interest rate channel: transmits policy rates to retail lending and deposit rates via money markets, affecting investment and savings decisions.
  - Asset price channel: works through prices of bonds, stocks, and real estate; policy-induced interest rate changes alter cost of capital and net worth.
  - Credit channel: transmits policy by influencing credit supply rather than interest rates; important where loans have limited substitutability (Egert and MacDonald, 2009).
- Channels can operate simultaneously and complement each other (example: quantitative easing leveraged asset price and credit channels when the federal funds rate was close to zero).

### 2.2 Model of interest rate pass-through (markup pricing / econometric form)
- Cost of funds approach (De Bondt, 2002): money market rates reflect marginal or opportunity costs of funds; banks price retail products with a markup for maturity and risk transformation.
- Long-run markup pricing relationship (notation preserved):
  - i_R = i_M · β + α  (equation (1)), where i_R and i_M are retail and market rates; β is long-run pass-through; α is markup.
  - Under perfect markets and risk-neutral banks β = 1; empirical evidence usually finds β < 1.
- Error-correction (out-of-equilibrium) and autoregressive distributed lag representations:
  - Short-run dynamics represented by equation (2) with ρ measuring speed of adjustment and γ short-run pass-through.
  - Expanded ARDL form shown as equation (3) with β_k short-run interaction elasticities and α_k persistence terms.
  - Long-run pass-through computed as in equation (4): long-run β = (Σ_{k=0}^n β_k) / (1 − Σ_{k=1}^n α_k).
- Required conditions:
  - Σ_{k=0}^n β_k ≥ 0  (equation (5)): cumulative impact of market rates on retail rates nonnegative.
  - Σ_{k=1}^n α_k ≤ 1  (equation (6)): stability condition implying convergence of retail rates.

### 2.3 Sample selection and estimation details
- Initial dataset: 81 countries, monthly data from December 2005 to March 2010.
- Interest rate series:
  - Money market rates: three-month treasury bill rates (or three-month interbank lending rates if t-bill unavailable).
  - Retail lending rates: interest rates on one- to two-year corporate loans in domestic currency; where unavailable, average lending rates were used.
- Countries with controlled or nearly fixed interest rates (either money market or retail rates) were excluded.
- Time series preprocessing:
  - All variables log-differenced to smooth series and remove unit roots (unit roots present in all interest rate series).
- Econometric specifics:
  - Equation (3) estimated individually for each country with two-month lags.
  - Omitted were seven cases failing conditions (5) and (6), and four cases with very poor goodness of fit (R2 < 0.1) and all short-run coefficients β_k statistically insignificant.
  - Final sample size after filtering: 70 countries.
  - In the out-of-equilibrium representation retail rates respond to market rates contemporaneously (k=0) or with lag (k=1, 2).
  - Contemporaneous interaction parameter β_0 generally significant and used for short-run pass-through measure; otherwise the first statistically significant lagged β_k (β_1 or β_2) was used.
  - Long-run β coefficients obtained using equation (4).

- Definitions and classifications used in sample description:
  - Micro states defined as countries with population less than 2 million (Imam, 2009).
  - Exchange rate regime classification follows IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions (2009): free-floating, floating, and other managed arrangements classified as flexible; the rest as fixed regimes.
  - Classification by level of development based on IMF’s World Economic Outlook Database (April 2010); low-income countries are those eligible for financing from the IMF’s Poverty Reduction and Growth Trust.

### 2.4 Pass-through estimates (summary from text and Figure 1)
- Estimation confirms large cross-country heterogeneity in pass-through magnitudes.
- Regional and group patterns reported:
  - Middle East and Central Asia (MEA) appears to have the weakest pass-through, short-run and long-run; possible influence of Islamic banking rules restricting payment of interest.
  - North America (United States and Canada) exhibit significantly higher pass-through than other groups, reflecting highly competitive and developed financial markets.
  - Micro states, including small island economies, tend to have weaker pass-through due to shallower and less developed financial markets.
  - More advanced economies tend to have stronger pass-through in the long run, but not necessarily in the short run—implying stronger persistence of market interest rates (higher Σ α_k in equation (4)).
  - Pass-through is stronger in countries with flexible exchange regimes than in those with fixed regimes.
  - Pass-through is stronger in the long run than in the short run.

---

*Source: _wp11176 - 1. INTRODUCTION (IMF Working Paper content extract)*

### 3. D

### 3. D DETERMINANTS OF INTEREST RATE PASS-THROUGH

### 3.1 Brief Summary of Literature
- Cross-country heterogeneity of interest rate pass-through has been widely studied; structural determinants have received less attention due to data limitations.
- Cottarelli and Kourelis (1994) (31 developed and emerging economies) found that higher inflation, capital mobility, and money market development (proxied by volatility of money market rates or by the size of the market for short-term securities) result in stronger pass-through.
- Mojon (2000) and Sander and Kleimeier (2004b) obtained similar results for inflation and money market volatility.
- Literature is unanimous on the positive role of competition in the banking system for pass-through; measures used include banking concentration, private ownership of banks, indices of banking deregulation, return on equity (ROE).
- Sorensen and Werner (2006) report negative relationships between pass-through elasticities and banks’ excess liquidity, excess capital, rigidity of funding costs (share of deposits in total liabilities), and interest rate risk (maturity mismatch). They also find portfolio diversification (share of noninterest income) and credit risk (loan provisioning) improve pass-through.
- Sander and Kleimeier find that worse banking sector health (higher nonperforming loans as a percent of total loans) reduces pass-through. Mojon reports high operating costs (staff costs to gross income) impede pass-through.
- Summary table (qualitative): literature reports positive (+) and negative (−) impacts of variables such as Inflation, Capital Mobility, Market Volatility, Competition, Excess Liquidity, Excess Capital, Rigidity of Funding, Interest Rate Risk, Portfolio Diversification, Credit Risk, NPL, Operating Costs across different studies (Cottarelli and Kourelis; Mojon; Sander and Kleimeier; Sorensen and Werner).

### 3.2 Variable Selection and Methodology
- Main hypothesis: cross-country differences in long-run pass-through are explained by structural country characteristics (macroeconomic conditions and financial market structure).
- Coverage: expanded to 70 countries from all regions, including low income countries and small island economies; averages mostly for 2006–2009 (monthly data used for average money market interest rates and volatility).
- Explanatory variables considered:
  - Macroeconomic: GDP per capita (income), GDP growth, Inflation (annual CPI inflation), Average money market rates (MMR), Market volatility (standard deviation of market interest rates divided by average rates).
  - Financial sector indicators: size of banking system (total assets and liabilities, private sector credit, deposits as share of GDP), bank capitalization (capital to total assets, capital to risk weighted assets), asset quality (non-performing loans to total loans, NPL), profitability and cost structure (ROE, returns on assets, net interest margin, cost-to-income ratio, overhead costs to total assets), liquidity (liquid assets to short-term liabilities, liquid assets to total assets).
  - Dummies: geographic region, country size, exchange regime (ER = 1 for flexible regimes, 0 for fixed), Island (small–island economies), MEast (Middle East and Central Asian countries).
- Regressions: pass-through coefficients (log of long-run pass-through coefficient) regressed on sets of macroeconomic and financial sector variables. Macroeconomic and financial regressions run separately to preserve degrees of freedom and limit multicollinearity.
- Equations estimated (notation described in text):
  - Equation (7): includes income (per capita GDP), inf (inflation), volat (market volatility), Island, MEast, and additional controls.
  - Equation (7’) is a variation of (7) substituting inflation with average money market rates (MMR).
  - Equation (8): includes financial-structure variables such as liq (liquidity), Overhead (overhead costs), NPL, ROE, ER (exchange regime dummy), and market volatility.
- Data sources: IMF International Financial Statistics; IMF World Economic Outlook, Spring 2010; World Bank Financial Development and Structure Database (2010); IMF Country Reports.
- Money market volatility defined as: standard deviation of money-market rates divided by average interest rate.

### 3.3 Results
- Per capita GDP:
  - Positive coefficient; more advanced countries have stronger pass-through (financial markets better developed; interest rate channel functions better).
- GDP growth:
  - Insignificant in all specifications.
- Inflation and average money market rates:
  - Inflation and MMR are positively associated with stronger pass-through.
  - Inflation becomes insignificant when included jointly with interest rates due to high correlation.
  - Fit is better for specification with interest rates (MMR) which also has a larger coefficient than inflation.
- Market volatility:
  - Negative and statistically significant in all three equations; higher volatility weakens pass-through.
  - Interpretation: volatility reduces reliability of market signals, prompting banks to delay passing changes to customers.

- Liquidity:
  - Negative relationship with pass-through (Equation (8)); excessively liquid markets weaken pass-through by reducing interbank trading and the informativeness of rates.

- Overhead costs:
  - Positive and statistically significant (1 percent level) where included; higher overheads associated with stronger pass-through.
  - Discussion: contrary to Mojon’s finding for a small set of European countries. Possible interpretation here is that high country-level overheads reflect structural frictions (regulatory/legal obstacles, weak financial infrastructure, information asymmetry) that induce higher net interest margins; banks pass through market rate increases more strongly to preserve margins.

- Credit quality (NPL):
  - Negative coefficient: higher NPLs weaken pass-through. Lower NPLs improve pass-through.
  - Mechanism: poor credit quality increases risk premia, so more of a cost increase is absorbed in risk premiums rather than passed through; profit-maximizing banks may be less responsive to market rate changes when credit quality is poor.

- ROE (bank profitability / market power):
  - Negative coefficient: higher ROE weakens pass-through, consistent with market power reducing elasticity of lending rates to fund costs.

- Exchange regime:
  - Flexible exchange regimes associated with stronger pass-through (positive coefficient on ER dummy). Mechanism: higher domestic rates attract capital and lead to appreciation until yields equilibrate; fixed regimes lack this buffer and may see liquidity effects that dampen pass-through.
  - Note: binary ER variable may not capture full variation in exchange regime flexibility or capital account openness.

- Regional and small-island effects:
  - Middle East and Central Asian countries: more sluggish pass-through.
  - Small island economies: more sluggish pass-through (consistent with prevalence of fixed exchange regimes).

- Regression table (dependent variable: log of long-run pass-through coefficient). Standard errors in parentheses.
  - Equation (7)
    - Const: -1.01 (1.0)
    - Per capita GDP: 0.37 *** (0.08)
    - Inflation: 0.48 *** (0.15)
    - Market volatility: -0.28 *** (0.08)
    - Middle East dummy: -2.0 *** (0.37)
    - Island: -0.77 *** (0.26)
    - R squared: 0.41
    - # observations: 69
  - Equation (7')
    - Const: -1.40 (0.9)
    - Per capita GDP: 0.37 *** (0.08)
    - Average money market rate: 0.69 *** (0.17)
    - Market volatility: -0.24 *** (0.07)
    - Middle East dummy: -1.71 *** (-0.33)
    - Island: -0.64 *** (0.23)
    - R squared: 0.46
    - # observations: 69
  - Equation (8)
    - Const: 7.09 (1.03)
    - Liquidity: -0.32 ** (0.14)
    - Overhead costs: 0.69 *** (0.17)
    - NPL: -0.32 *** (0.12)
    - ROE: -0.21 ** (0.09)
    - Market volatility: -0.19 ** (0.07)
    - Middle East dummy: -1.28 *** (0.47)
    - Exchange regime dummy: 0.63 ** (0.2)
    - Island: -0.64 *** (0.23)
    - R squared: 0.69
    - # observations: 43
  - Significance: ** denotes significance at 5 percent confidence level. *** denotes significance at 1 percent confidence level.

Key interpretations and implications:
- Stronger pass-through is associated with higher per capita GDP, higher inflation or market interest rates, higher overhead costs, and flexible exchange regimes.
- Weaker pass-through is associated with higher market volatility, excess liquidity, higher NPLs, and higher ROE (indicative of weaker competition).
- Policy-relevant inferences:
  - Improving financial market development and reducing market volatility may strengthen the transmission of monetary policy via interest rates.
  - Addressing structural frictions that drive high overhead costs and improving credit quality (reducing NPLs) could influence pass-through dynamics.
  - Exchange regime design and degree of capital account openness affect the sensitivity of lending rates to market rates; finer classification of exchange regimes and inclusion of capital account openness variables could yield further insight.

*Source: _wp11176 - 3. D*

### 4. C

### 4. C
ONCLUSIONS AND POLICY IMPLICATIONS

### Key empirical findings on interest rate pass-through
- Per capita GDP, inflation, interest rates, credit quality, overhead costs and competition among banks facilitate the pass-through process.
- Market volatility and excess liquidity in the banking system impede the pass-through process.
- Countries with fixed exchange rate arrangements tend to have weak pass-through.
- Middle Eastern and Central Asian countries and small island states tend to have weak pass-through.

### Implications for choice of monetary framework
- If pass-through is weak and cannot be improved (for example because a country is a small island state or is a low income economy that would require considerable time to develop its financial markets and strengthen transmission), a monetary framework that relies on strong interest rate pass-through, such as inflation targeting, may not be an optimal option.
- Alternative intermediate targets that might be more appropriate in such cases include monetary aggregates, exchange rate or other nominal anchor.

### Policy challenges for central banks
- The primary objective of monetary policy is to achieve and maintain low inflation, defined as a specific inflation target (a point or a range) or more generally as price stability.
- Strengthening transmission is in a central bank’s interest because it provides connection between monetary policy instruments and banks’ objectives; the stronger and faster the pass-through, the easier it will be for a central bank to achieve its objective.
- There is a positive relationship between inflation and the strength of pass-through, implying that by being successful in reducing inflation a central bank contributes to the weakening of monetary transmission.
- Improvements to the regulatory and legal environment, promotion of information sharing and removal of obstacles to banking intermediation—which reduce overhead costs—would also weaken pass-through, making it harder for a central bank to control inflation.

### Policy options to compensate for weakening pass-through
- Weakening of pass-through owing to successful disinflation or reduced overhead costs could be compensated through other means, such as:
  - strengthening competition,
  - mopping up excess liquidity (if any),
  - improving credit quality,
  - developing financial markets to reduce interest rate volatility.

*IMF Working Paper — 4. C ONCLUSIONS AND POLICY IMPLICATIONS*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11176.pdf_