## wp1870

## Source details

**Canonical URL:** [wp1870](https://www.imf.org/-/media/files/publications/wp/2018/wp1870.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp1870.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp1870.pdf.json)

---

### Glossary
- ECCB    Eastern Caribbean Central Bank
- ECCU    Eastern Caribbean Currency Union
- GFC    Global Financial Crisis
- IFS     IMF’s International Finance Statistics
- MSR    Minimum Saving Deposit Rate
- WEO    IMF’s World Economic Outlook
- XCD    Eastern Caribbean Dollar

### I. Introduction — scope and contribution
- Central premise: the “impossible trinity” — open capital account, fixed exchange rate, and monetary autonomy cannot be fully achieved simultaneously.
- Paper investigates both international and domestic interest-rate transmission channels in the ECCU.
- Key contributions:
  - Empirical panel regression evidence pointing to a low long-run international pass-through coefficient of the U.S. interest rate for the ECCU, below the average for fixed-exchange-rate countries.
  - Retail interest rates in the ECCU respond to changes in the MSR.
  - Survey-based evidence and examination of macro and bank-level impediments to transmission and capital market development.
- Structure: background and stylized facts (section II); international transmission analysis (section III); domestic transmission analysis and policy implications (remaining sections).

### Stylized facts and ECCU institutional features
- ECCU membership: Antigua and Barbuda, Dominica, Grenada, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Anguilla, Montserrat.
- Currency peg: XCD$2.70=US$1 since 1976.
- ECCB framework:
  - Quasi-currency board: 60 percent of monetary demand liabilities required to be backed with foreign exchange reserves; operational target 80 percent; actual reserve ratio above 90 percent since early 2000s.
  - Reserve requirement ratio: 6 percent since ECCB inception.
  - Discount rate: adjusted only four times historically; remained at 6.5 percent over the last 15 years.
  - MSR history: instituted mid-1980s at 4 percent; reduced to 3 percent in 2002 and to 2 percent in 2015.
  - MSR applies to EC dollar-denominated savings deposits in commercial banks.
  - Savings deposits carrying up to 3 percent interest as a share of total saving deposits increased from 46 percent at end-2003 to 96 percent by end-2015.
  - Share of saving deposits ≈ about half of commercial banks’ liabilities to the private sector over the last decade.

### International transmission — methodology and key estimates
- Channels identified: interest rate channel, interest rate gap channel, exports-to-base channel, capital flows channel, exchange rate channel.
- Focus: interest rate channel (given peg, low non-FDI capital flows, and data limitations for other channels).
- Data and model:
  - Unbalanced panel of 178 economies, annual data 1990–2014.
  - Base country short-term money market rates; U.S. rates proxy global rates for flexible regimes.
  - De facto exchange rate regimes classified into fixed, intermediate, flexible.
  - Panel regression with fixed effects (Frankel et al. (2004) framework), log-first-differences for interest and exchange rate variables, controls for capital controls (Ito-Chinn), inflation differential, crisis dummies (Laeven and Valencia (2012)), etc.
- Full-sample estimated elasticities (long-run) of domestic retail lending rates to base short-term rates:
  - Fixed exchange rate regimes: 0.48
  - Intermediate regimes: 0.43
  - Flexible regimes: 0.32
- Subsample findings:
  - Caribbean and small states: estimated coefficient small but positive.
- ECCU subsample result:
  - Elasticity of domestic lending rates with respect to the U.S. interest rate estimated small and with a negative sign.
  - Crisis dummy coefficient statistically significant; transmission lower during crisis episodes.
  - Inflation differential and capital account openness statistically insignificant in ECCU subsample.
  - Interpretation: limited U.S. pass-through → weak interest rate channel and potential prominence of “exports-to-base” channel bypassing domestic interest rates.
- Caution: weak early-period domestic financial development may bias measured transmission (Mishra and Montiel, 2012).

### Domestic transmission — channels, impediments, and methods
- Domestic channels summarized:
  - Interest rate channel: policy rate → market rates → retail deposit & lending rates → savings, investment, consumption → aggregate demand/prices.
  - Credit availability channel: quantity effects via bank-lending and balance-sheet channels.
  - Exchange rate and asset price channels: limited effectiveness in ECCU due to peg and shallow capital markets.
- ECCU-specific impediments weakening domestic transmission:
  - Elevated commercial bank concentration (top three banks’ assets share high) → interest-rate stickiness.
  - Existence of credit unions in some countries → competition for established clients.
  - Excess liquidity in banking system → neutralizes transmission of ECCB discount rate.
  - MSR changes directly affect retail deposit rates by setting a lower bound.
  - High bank operating costs → banks absorb rate changes; reduce pass-through.
  - High NPLs, low profitability, episodes of undercapitalization, absence of credit bureau, high contract enforcement costs, debtor-friendly foreclosure/insolvency rules → weaken credit channel.
- Methodology for domestic pass-through:
  - Marginal cost pricing / markup model (De Bondt, 2002); ADL/ARDL and event-study methods.
  - Sample: unbalanced panel of 76 countries (1980–2017Q1) for cross-country comparison; ECCU series from IFS used for longer horizon.
  - ECCU models: ADL(2,6) with 6 lags for main ECCU specification; uniform ADL(1,1) for global sample comparison.
  - Variables: IFS lending interest rate (maximum rate charged on prime loans) and IFS deposit interest rate (maximum on three-month time deposits); all variables differenced.

### Empirical pass-through findings — deposit and lending rates
- Deposit-rate findings:
  - ECCU deposit interest rates are more sensitive to the MSR than to the U.S. policy rate and align broadly with the Advanced Economies average.
  - Models with ECCB discount rate failed convergence in several ECCU economies (discount rate rarely changed historically).
  - Transmission of Canada’s policy rate to the ECCU: virtually zero.
  - Average long-run elasticity of deposit rates with respect to the MSR: about 0.85.
    - Interpretation: a one-percentage point downward change in the MSR would lower deposit rates by about 0.85 percentage points over the long run.
  - Deposit-rate sensitivity from panel estimates: a decline (increase) in the MSR by 1 ppt on average leads to about 0.5 ppt decline (increase) in deposit interest rates (panel regression result).
  - Coefficients on the US policy rate and the ECCB discount rate not statistically significant in deposit regressions.
- Lending-rate findings:
  - Changes in the MSR historically had a larger effect on lending rates than adjustments in the U.S. policy rate, Canada’s policy rate, and the ECCB discount rate.
  - Average ECCU lending-rate pass-through in uniform 1-lag model: ECCU average falls below the full sample average and many subsamples; ADL(6) adjusted model: long-term transmission close to Emerging and Developing Economies.
  - Country heterogeneity (long-term MSR → lending rate coefficients, selected):
    - St. Lucia: 1.6
    - Dominica: 1.2
    - St. Vincent and the Grenadines: 0.25
    - Dominica (ADL country estimate): 0.7 (reported elsewhere)
    - St. Kitts and Nevis: 0.6
    - Grenada: 0.1
  - ECCU average implication: a one percentage point decline in the MSR would lead to 0.4 percentage point decline in lending rates across the ECCU on average.
  - Panel regression lending-rate sensitivity (1989–2016): a 1 ppt decline (increase) in deposit rates → about 0.46 ppt decline (increase) in lending rates, ceteris paribus.
  - Combined panel implication: a one percentage point decline in the MSR → about ½ ppt decline in deposit rates and about ¼ ppt decline in lending rates, holding other factors constant.
  - Time variation: 2001–2016 subsample shows stronger sensitivity — a 1 ppt decline in deposit rates on average associated with 1.2 ppt decline in lending rates; rolling-window regressions indicate increasing sensitivity and better model fit in recent subperiods.
- Survey evidence (May 2015 MSR reduction):
  - Survey of 11 banks (48 percent of ECCU banking sector assets as of end-2016) on MSR reduction from 3 to 2 percent (May, 2015).
  - Responding banks reported passing through between 10 to 80 bps of the 100 bps decline in the MSR into prime mortgage rates.
  - Survey supports hypothesis that limited new loan issuance restrains reflection of reduced interest rates in official lending rate statistics.

### Panel and country regression findings — quantitative details
- Panel regression (annual, independent ECCU members, 1985–2016) — key quantitative results:
  - Deposit rate reaction to MSR: a 1 ppt change in MSR → about 0.5 ppt change in deposit rates (panel).
  - Deposit → lending: a 1 ppt decline in deposit rates → about 0.46 ppt decline in lending rates (panel).
  - Asset quality deterioration associated with lower deposit interest rates.
  - Increased capital buffers and profitability associated with higher deposit interest rates offered in the following year.
  - Higher credit growth to private sector and larger loan-to-deposit ratios associated with lower lending interest rates.
  - Increase in excess liquidity appears to put downward pressure on lending rates.
  - Higher NPLs and operating costs increase lending rates.
- Appendix Table A1 (ADL(2,6) country regressions, D. Deposit rate and D. Lending rate):
  - Country-level MSR t-6 coefficients (deposit regressions):
    - Angtigua and Barbuda: 0.380 (0.196)
    - Dominica: 0.137 (0.485)
    - Grenada: -0.00993 (0.975)
    - St. Kitts and Nevis: -0.01391 (0.960)
    - St. Lucia: 1.596*** (7.40e-05)
    - St. Vincent and the Grenadines: 0.313* (0.0530)
  - Country-level MSR t-6 coefficients (lending regressions):
    - Angtigua and Barbuda: 0.747** (0.0363)
    - Dominica: 1.719*** (1.50e-09)
    - Grenada: 1.048*** (0.00465)
    - St. Kitts and Nevis: 0.834*** (0.000344)
    - St. Lucia: 2.308*** (2.69e-05)
    - St. Vincent and the Grenadines: 0.568** (0.0438)
  - Time dummy (second half of 2003) often significant and positive across deposit and lending regressions (examples: St. Lucia deposit time dummy 1.238*** (0.00456); Angtigua lending time dummy 1.387*** (0.00112)).
  - R-squared examples (deposit regressions): St. Lucia 0.349; Dominica 0.121; St. Vincent and the Grenadines 0.162. (lending regressions): Dominica 0.345; St. Kitts and Nevis 0.338; Angtigua and Barbuda 0.235.
- Appendix Table A2 (panel regressions — selected coefficients and model diagnostics):
  - Change in ECCU MSR (own regressor) coefficients reported across models: 0.444* (0.0687), 0.415* (0.0878), 0.524* (0.0815), 0.482* (0.0865), 0.543*** (4.24e-05), 0.411*** (4.65e-05), 0.382*** (0.000141).
  - Change in US policy rate coefficients vary by specification, examples:
    - 0.0880** (0.0314), 0.0940** (0.0280), -0.0617 (0.271), -0.109** (0.0239), -0.125** (0.0136), -0.000248 (0.996).
  - Change in deposit rate coefficients large and significant in several specifications: 0.446*** (2.19e-07), 0.461*** (1.11e-07), 0.452*** (2.91e-07), 1.222*** (0), 1.343*** (0).
  - Change in excess reserves: -0.0725** (0.0190) in one specification.
  - Change in NPL ratio t-1: -0.0503** (0.0466) in one specification.
  - Change in loan-to-deposit ratio: -0.0362*** (0.00851) in one specification.
  - Model coverage: number of countries = 6; start dates vary by model (examples: 1989, 2004, 2006, 2011) with end date 2016 for all models; observations per model range (examples: 170, 848, 474, 617); R-squared examples: 0.047, 0.140, 0.300, 0.681.

### Main conclusions and policy implications
- Empirical conclusions:
  - Low estimated long-run pass-through of the U.S. interest rate to ECCU retail rates; below average for fixed-exchange-rate peers.
  - No robust evidence of significant transmission from the U.S. and Canada’s policy rates, nor the ECCB discount rate, to retail rates.
  - Significant pass-through from the MSR to deposit rates and measurable transmission to lending rates; heterogeneity across ECCU members.
  - Survey evidence corroborates empirical findings for pass-through into prime mortgage and prime consumer lending rates for established clients.
- Policy implications:
  - Findings indicate a limited degree of monetary policy independence in the ECCU via the MSR.
  - MSR can be used more actively to adjust monetary policy consistent with objectives.
  - Given current conditions—high NPLs, low banking sector profitability, declining credit growth to private sector, and growing operating costs—a looser monetary policy stance could be warranted.
  - A temporary reduction in the MSR could:
    - Ease monetary conditions,
    - Support credit recovery,
    - Potentially improve economic growth.

*Source — wp1870 - Section III provides analysis of the international transmission channel, of the U.S. interest (IMF working paper content as provided).*

### References _______________________________________________________________________________________ 28

### wp1870 - References _______________________________________________________________________________________ 28

### Glossary
- ECCB    Eastern Caribbean Central Bank
- ECCU    Eastern Caribbean Currency Union
- GFC    Global Financial Crisis
- IFS      IMF’s International Finance Statistics
- MSR    Minimum Saving Deposit Rate
- WEO    IMF’s World Economic Outlook
- XCD    Eastern Caribbean Dollar

### I.   INTRODUCTION
- The inability to simultaneously opt for an open capital account, a fixed exchange rate regime, and monetary autonomy is a central postulate in international macroeconomics known as the “impossible trinity”.
- Accordingly, countries may choose only two of the tree options of the impossible trinity (see Obstfeld, Shambaugh and Taylor, 2004, for an historical perspective).
- The choice of the policy mix, or the policy constraint, is fundamental to the conduct of macroeconomic policy and economic stability, especially for small open economies (Ghosh and Ostry, 2009).
- The literature has mostly focused on estimating international2 and the domestic interest-rate channels of monetary policy transmission to gauge the existence of the trilemma empirically.
- Contribution of this paper:
  - Investigates empirically both the international and domestic interest-rate channels in the ECCU.
  - Findings based on panel regression analyses point to a low long-run international pass-through coefficient of the U.S. interest rate, falling short of the average for countries with fixed exchange rates.
  - ECCU retail interest rates are found to respond to changes in the minimum saving deposit rate (MSR).
  - Relies on survey-based evidence and examines potential causes of limited monetary policy transmission in the ECCU, concluding that both macro-economic and bank-level factors play a role in interest rates transmission and capital market development.
- Structure:
  - The remainder of the paper is structured as follows. section II presents some background and stylized facts about the ECCU, focusing on the conduct of monetary policy in the union.

*Source: wp1870 - References _______________________________________________________________________________________ 28*

### Section III provides analysis of the international transmission channel, of the U.S. interest

### wp1870 - Section III provides analysis of the international transmission channel, of the U.S. interest

### Literature review and stylized facts
- Recent cross-country studies of global interest-rate transmission cited: Frankel et al. (2004), Shambaugh (2004), Bleaney et al. (2013).
- Empirical findings summarized:
  - Frankel et al. (2004): Using data over 1970-1999, interest rates in levels adjust more slowly under floating rates, but in the long run there is full transmission whatever the regime.
  - Shambaugh (2004): Annual data for 155 countries over 1973-2000 — transmission of interest rate changes weakest for non-pegs with capital controls; strongest for pegs without capital controls; followed by non-pegs without capital controls; transmission also quite strong for pegs with capital controls.
  - Bleaney et al. (2013): Using data for 126 countries from 1990 — countries with credible pegs without capital controls follow foreign interest rates closely.
- Differences by country income/development:
  - Advanced economies: developed capital and asset markets, strong links to international capital markets, more flexible exchange rate regimes; monetary policy transmits through interest rate, credit availability, asset price, and exchange rate channels; studies often find close to hundred percent pass-through of policy rates to retail lending rates (Saborowski and Weber, 2013; De Bondt, 2002).
  - Emerging and developing economies (e.g., ECCU): less developed fixed-income, equity, and real estate markets; imperfect links to international capital markets; less flexible exchange rate regimes; often inactive exchange rate and asset price channels; bank-lending channel likely more effective (Mishra and Montiel, 2012).
- Small-state evidence:
  - Pacific Islands (Dunn et al., 2011): low degree of pass-through of policy rates to commercial bank rates, with varied magnitudes across countries.
  - Caribbean (Kendall, 2001; Ramlogan, 2007; Haughton and Iglesias, 2012): heterogeneous responses of lending rates to policy changes; credit channel important in some cases; complete pass-through of 90-day T-bill rate to lending rates observed only for St. Lucia and Trinidad and Tobago in Haughton and Iglesias (2012); deposit-rate transmission often incomplete; evidence of downward adjustment rigidity in some lending rates indicative of collusive pricing.
- ECCU institutional and historical facts:
  - ECCU membership: eight members (Antigua and Barbuda, Dominica, Grenada, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Anguilla, Montserrat).
  - Common central bank since 1983: Eastern Caribbean Central Bank (ECCB).
  - Currency peg: XCD$2.70=US$1 since 1976.
  - ECCB operates as a quasi-currency board: lending to members limited by ECCB Agreement Act; 60 percent of monetary demand liabilities required to be backed with foreign exchange reserves; operational target 80 percent reserve coverage; actual reserve ratio above 90 percent since early 2000s.
  - Capital account liberalization progressive since mid-1990s; some legal/regulatory restrictions remain (e.g., alien land holding licenses).
- ECCB monetary instruments and practice:
  - Instruments: reserve requirements, open market operations, discount rate, foreign exchange reserves; credit allocation to priority sectors and differentiated reserve requirements by deposit type are part of the operational framework.
  - Discount rate: adjusted only four times historically; remained at 6.5 percent over the last 15 years.
  - Reserve requirement ratio: remained at 6 percent since ECCB inception.
  - Minimum savings deposit rate (MSR): instituted in mid-1980s at 4 percent; reduced to 3 percent in 2002 and to 2 percent in 2015.
  - Operational facts: MSR applies to EC dollar-denominated savings deposits in commercial banks; savings deposits carrying up to 3 percent interest as a share of total saving deposits increased from 46 percent at end-2003 to 96 percent by end-2015 (end-2015 includes deposits carrying less than 3 percent since MSR reduced in May 2015).
  - Share of saving deposits has constituted about half of commercial banks’ liabilities to the private sector over the last decade.
- Effects and concerns about MSR:
  - MSR provided stable returns for savers but may cause misallocation of resources by setting a lower bound on saving deposit interest rates.
  - Possible circumvention via other financial intermediaries or offshore operations.
  - Polius (2002): level of savings not influenced by MSR, but MSR may increase cost of investment and reduce long-run investment.
  - Mounsey and Polius (2015) and IMF (2017): MSR may be distortionary in a protracted downturn, harmful to banking sector and wider economy, impede optimal bank response.
  - Distributional effect: equivalent to transfer of income from lenders to borrowers if deposit rates are held above equilibrium.
  - Banks narrowed MSR eligibility to individuals/households, redirecting corporate funds to other deposit instruments.

### International monetary policy transmission pass-through estimates
- Identified channels through which foreign interest rates affect domestic economy:
  - Interest rate channel: direct effect of base interest rates on domestic retail interest rates (uncovered and real interest rate parity).
  - Interest rate gap channel: foreign rate moves domestic rate and changes expectations of risk premium, altering spreads.
  - Exports-to-base channel: economic dependence on base country implies exports channel as primary transmission.
  - Capital flows channel: base-country rate increases shrink capital available outside base country.
  - Exchange rate channel: base rate moves exchange rate, affecting the economy.
- Paper focus: the interest rate channel — reasoned by:
  - Relatively low non-FDI capital flows in the ECCU and the exchange rate peg make capital flows and exchange rate channels ineffective.
  - Interest rate gap channel estimation limited by data.
  - Exports-to-base channel already studied elsewhere; authors focus on interest-rate pass-through.
- Data and model for international pass-through:
  - Unbalanced panel of 178 economies, annual data covering 1990-2014.
  - Base country short-term money market rates used as base interest rates; for flexible regimes global interest rates proxied by U.S. interest rates.
  - De facto exchange rate regimes from IMF AREAER classified into fixed, intermediate, and flexible categories.
  - Panel regression with fixed effects following Frankel et al. (2004): dependent variable nominal domestic lending interest rates; base interest rate (lagged) and controls (capital account barriers, inflation differential, hyperinflation periods, exchange rate fluctuations, banking/sovereign/debt crises). Interest rate and exchange rate variables defined as first log-differences (log-difference defined as first difference of ln(1+i/100), where i is interest rate).
  - Controls include Ito-Chinn capital controls index, Laeven and Valencia (2012) crisis episodes, data from WEO, IFS, WDI.
- Key international pass-through results:
  - Estimated full-sample pass-through coefficients (elasticities of domestic retail lending rates to base country short-term interest rates, lagged):
    - Fixed exchange rate regimes: 0.48
    - Intermediate regimes: 0.43
    - Flexible regimes: 0.32
  - Interpretation: countries with less flexible exchange rate regimes display higher levels of transmission; results consistent with the impossible trinity.
  - Subsample findings:
    - Difference between fixed and intermediate coefficients narrows when sample restricted to emerging market economies.
    - For Caribbean and small states the estimated coefficient is small but positive.
  - ECCU-specific international-pass-through findings:
    - For the ECCU subsample, elasticity of domestic lending rates with respect to U.S. interest rate estimated small and with a negative sign.
    - Possible explanation: diverging trends — gradual decline of U.S. policy rate vs relatively stable ECCU lending rates over a large part of sample period.
    - Coefficient on crisis dummy (banking, sovereign default, debt restructuring episodes) statistically significant in ECCU subsample; interest rate transmission lower during crisis episodes.
    - Inflation differential and capital account openness statistically insignificant in ECCU subsample.
    - Overall interpretation: limited U.S. interest rate pass-through suggests a weak interest rate channel and potentially important “export to base” channel that bypasses domestic interest rates.
  - Cautionary note: limited development of domestic financial system in early transition period may affect estimated strength of transmission (Mishra and Montiel, 2012).

### Domestic monetary policy transmission pass-through estimates
- Motivation: Limited international pass-through implies potential high degree of domestic monetary policy independence; empirical testing of domestic channels on deposit and lending rates using ADL and event-study methods.
- Domestic transmission channels summarized:
  - Interest rate channel: policy rate → market rates → retail deposit and lending rates → savings, investment, consumption → aggregate demand and prices. Strength depends on competition, alternative funding sources, capital/money market depth, liquidity, operating costs.
    - ECCU-specific factors that weaken transmission via interest rate channel:
      - Elevated commercial bank concentration (share of top three banks’ assets high) may promote interest-rate stickiness.
      - Availability of credit unions in some countries (e.g., Dominica, St. Vincent and the Grenadines) increases competition for established clients.
      - Excess liquidity in ECCU banking system neutralizes transmission of ECCB discount rate to retail rates.
      - MSR changes would directly affect retail rates by changing the lower bound on savings deposit rates, bypassing money market channel.
      - Higher bank operating costs reduce transmission as banks absorb rate changes to cover costs.
  - Credit availability channel: monetary stance affects quantity of credit irrespective of rates via bank-lending and balance-sheet channels; ECCU impediments include high NPLs, low profitability, instances of undercapitalization, absence of credit bureau, high contract enforcement costs, debtor-friendly foreclosure/insolvency rules.
  - Exchange rate channel: quasi-currency board arrangement of ECCU diminishes effectiveness of exchange rate channel.
  - Asset price channel: shallow capital markets and bank-dominated financial intermediation inhibit asset-price-channel effectiveness; MSR history biases portfolios toward banks, reducing asset-price channel strength.
- Data and methodology for domestic pass-through:
  - Marginal cost pricing / markup model (De Bondt, 2002): long-run pass-through coefficient ߚ; error-correction and ADL/ARDL specifications used.
  - Sample: unbalanced panel of 76 countries (11 advanced, 61 emerging/developing) for quarterly frequency 1980–2017Q1; ECCU series use IFS for longer horizon (IFS series preferred over ECCB series because ECCB series begin in 2006).
  - Variables: IFS lending interest rate = maximum rate charged by commercial banks on prime loans, weighted by loan amounts; IFS deposit interest rate = maximum rate offered on three-month time deposits, weighted by deposit amounts.
  - All variables differenced to remove unit roots; most countries estimated with one-quarter lag structure; ECCU models used ADL(2,6) with 6 lags for main ECCU specification; ECCU model includes dummy flagging second half of 2003 when deposit/lending rates spiked due to exogenous factors.
  - Four policy interest rates analyzed for transmission into retail rates: MSR, U.S. policy rate, ECCU discount rate, Canadian policy rate.
- Deposit rate pass-through results (high-level summary from provided text):
  - ECCU deposit interest rates are more sensitive to the MSR than to the U.S. policy rate and align broadly with the Advanced Economies average.
  - Models using the ECCB discount rate failed to satisfy convergence conditions for several ECCU economies (explained by very few historical changes in the discount rate).
  - Transmission of Canada’s policy interest rate to the ECCU estimated at virtually zero.
  - Average long-run elasticity of deposit rates with respect to the MSR: about 0.85 (i.e., a one-percentage point downward change in the MSR would lower deposit rates by about 0.85 percentage points over the long run).
  - Average long-run pass-through coefficient from the U.S. policy rate: (text truncated in source prior to numeric summary) — the source states it is smaller than the MSR pass-through and limited.
- Additional identified institutional/structural impediments to domestic transmission in ECCU (from analysis):
  - Bank concentration, excess liquidity, high operating costs, high NPLs, low bank profitability, undercapitalization episodes, absence of credit bureau, high contract enforcement costs, and debtor-friendly insolvency/foreclosure rules all constrain interest-rate and credit channels.
  - MSR provides a direct transmission path to deposit rates but also creates distortions and possible misallocation effects.
- Methodological note on interpretation:
  - Some model specifications failed convergence conditions and were excluded from averaging.
  - Short-run and long-run pass-through estimates derived from ADL/ARDL specifications and computed via prescribed formulae for long-run pass-through.

*Italic: Source — wp1870 - Section III provides analysis of the international transmission channel, of the U.S. interest (IMF working paper content as provided).*

### 0.03 over the long run.

### wp1870 - 0.03 over the long run.

### Estimated long-run pass-through: overview
- Estimated long-run pass-through coefficients vary across ECCU member countries.
- Selected point estimates:
  - St. Lucia: 1.6
  - Dominica: 1.2
  - St. Vincent and the Grenadines: 0.25
- Model selection and lag choices:
  - Alternative model for ECCU with 6-quarter lags guided by Akaike information criteria (AIC).
  - ADL(1,1) for the global sample aligns with AIC for many countries and the literature.
  - Similar lag length selection results reported by BIC criteria.
  - Uniform lag applied to all regressions for comparability; dashed ECCU estimates indicate ADL model with 6 lags where noted.
- Numeric reference from text: 0.03 over the long run.

### Lending rate pass-through assessment
- Historical comparisons:
  - Changes in the MSR have had a larger effect on lending rates than adjustments in the U.S. policy rate, Canada’s policy rate, and the ECCB discount rate.
  - Models with ECCB discount rate and Canada’s policy rate failed to converge and satisfy condition (3); results presented use models with the U.S. policy rate and the MSR.
- Average ECCU pass-through:
  - Average estimated MSR pass-through into lending rates in the ECCU is broadly on par with comparator country averages using the uniform 1-lag model.
  - Uniform model with one lag: average ECCU lending rate pass-through falls below the full sample average and those for Caribbean, small states, Emerging and Developing economies, Latin America, and Advanced Economies subsamples.
  - Adjusted model with 6 lags: long-term transmission of the MSR to lending rates coefficient close to Emerging and Developing Economies.
- Country heterogeneity (long-term coefficients from MSR to retail lending rates):
  - Dominica: 0.7
  - St. Kitts and Nevis: 0.6
  - Grenada: 0.1
  - ECCU average implication: a one percentage point decline in MSR would lead to 0.4 percentage point decline in lending rates across the ECCU on average.

### Pass-through to deposit rates (comparisons and patterns)
- MSR versus U.S. policy rate:
  - Pass-through from changes in the MSR has been greater for deposit rates than for lending interest rates.
  - Transmission of the U.S. policy rate is inconsistent across ECCU countries.
  - Main model results indicate MSR transmission into deposit rates is largely on par with the advanced economies average.
- Cross-subsample ratios (model-based estimate with 1 lag) referenced for:
  - ECCU, Caribbean, Small States, EMDE, LAC, Sample average, AE.
  - ECCU alternative model specification (ADL with 6 lags) shown as dashed line in comparative charts.

### Survey-based assessment of May 2015 MSR reduction
- Context and data:
  - Survey of ECCU banks conducted to assess transmission of the May 2015 change in the MSR.
  - Responses included 11 banks, constituting 48 percent of the ECCU banking sector assets as of end-2016.
  - The minimum saving rate was reduced from 3 to 2 percent effective May, 2015.
- Survey findings:
  - A large portion of the 100 bps decline in the MSR was transmitted into lower prime mortgage rates and, to a lesser degree, into prime consumer lending rates.
  - Responding banks reported passing through to prime consumer mortgage rates between 10 to 80 bps of the 100 bps decline in the MSR.
  - Survey results support the hypothesis that limited new loan issuance restrains the reflection of reduced interest rates in official lending rate statistics.

### Panel model estimation: determinants and quantitative effects
- Model specifications:
  - Deposit rates regression (equation 5): first differences of deposit interest rates regressed on change in MSR, change in US policy rate, sum of ECCU-specific factors, country fixed effects, and error term.
  - Lending rates regression (equation 6): first differences of lending interest rates regressed on changes in deposit interest rates, change in MSR, change in US policy rate, ECCU-specific factors, country fixed effects, and error term.
  - Models estimated with country fixed effects using an unbalanced panel for the 6 independent ECCU economies covering period from 1985 to 2016.
  - Interest rate data from IMF’s IFS; macro indicators from IMF’s WEO; other ECCU variables largely from the ECCB.
  - Annual frequency data; contemporaneous interest rate variables used; first differences to remove unit roots.
- ECCU-specific control variables included in ܥ set:
  - Banking sector concentration: Herfindahl-Hirschman Index (available for 2011-2016).
  - Excess liquidity: banks’ excess reserves held at the ECCB in percent of total deposits and ratio of liquid-to-total assets.
  - Banking sector capitalization: capital adequacy ratio and tier 1 ratio.
  - Bank asset quality: NPLs and provisioning for NPLs.
  - Banking sector efficiency: operating costs-to-non-interest expenses, operating expenses-to-total expenses.
- Quantitative panel results:
  - Deposit rate sensitivity: a decline (increase) in the MSR by 1 ppt on average leads to about 0.5 ppt decline (increase) in deposit interest rates.
  - Coefficients on the US policy rate and the ECCB discount rate are not statistically significant.
  - Asset quality deterioration associated with lower deposit interest rates.
  - Increased capital buffers and profitability associated with higher deposit interest rates offered in the following year.
  - Lending rate sensitivity: estimates based on 1989 – 2016 suggest a 1 ppt decline (increase) in deposit rates leads to about 0.46 ppt decline (increase) in lending rates, ceteris paribus.
  - Combined implication: a one percentage point decline in the MSR would lead to about ½ ppt decline in deposit rates and about ¼ ppt decline in lending rates, holding other factors constant.
  - Note: a conservative interpretation; coefficients estimated over shorter periods suggest fuller transmission (see footnote indicating possible ~½ ppt decline in lending rates in some estimates).
- Other associations:
  - Higher credit growth to private sector and larger loan-to-deposit ratios associated with lower lending interest rates.
  - Increase in excess liquidity appears to put downward pressure on lending rates.
  - Higher NPLs and operating costs increase lending rates.
- Time variation:
  - Sensitivity of lending rates to changes in deposit rates has increased over time based on 15-year rolling window regressions.
  - Results based on the 2001-2016 subsample: a 1 ppt decline in deposit rates on average is associated with 1.2 ppt decline in lending rates.
  - Rolling-window regressions show better model fit and larger R-squared for more recent subsamples.

### Conclusions and policy implications
- Main empirical findings:
  - Low estimated long-run pass-through coefficient of the U.S. interest rate to ECCU retail rates; falls short of the average for countries with fixed exchange rates.
  - No evidence of significant transmission from the U.S. and Canada’s policy rates, nor the ECCB discount rate, to retail rates.
  - Significant pass-through from the MSR to deposit rates, and some transmission to lending rates.
  - Survey-based assessment supports empirical results, especially for prime mortgage and prime consumer lending rates to well-established clients.
- Policy implications for monetary policy in the ECCU:
  - Findings support a limited degree of monetary policy independence in the ECCU via the MSR.
  - MSR can be used more actively to adjust monetary policy in accordance with policy objectives.
  - Given current conditions—high NPLs, low banking sector profitability, declining credit growth to private sector, and growing operating costs—a looser monetary policy stance could be warranted.
  - A temporary reduction in the MSR could:
    - Ease monetary conditions,
    - Support credit recovery,
    - Potentially improve economic growth.

*Source: wp1870 - 0.03 over the long run.*

### Appendix Table A1. Regression Results

### Appendix Table A1. Regression Results

### Transmission of Minimum Saving Deposit Rate to Retail Interest Rates — Deposit Rates (Dependent variable: D. Deposit rate)
- Countries: Angtigua and Barbuda, Dominica, Grenada, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines
- Model: ADL(2,6). Interest rate variables taken in first differences. T-n refers to lagged variables. Time dummy flags interest rate volatility in the second half of 2003 due to exogenous factors.
- Key estimated coefficients (coefficient followed by p-value in parentheses):
  - Deposit rate t-1:
    - Angtigua and Barbuda: -0.0460 (0.626)
    - Dominica: -0.110 (0.235)
    - Grenada: -0.103 (0.268)
    - St. Kitts and Nevis: -0.230** (0.0173)
    - St. Lucia: -0.137 (0.106)
    - St. Vincent and the Grenadines: 0.0702 (0.435)
  - Deposit rate t-2:
    - Angtigua and Barbuda: -0.0248 (0.792)
    - Dominica: -0.282*** (0.00282)
    - Grenada: -0.240** (0.0104)
    - St. Kitts and Nevis: 0.0433 (0.650)
    - St. Lucia: -0.149* (0.0715)
    - St. Vincent and the Grenadines: -0.283*** (0.00220)
  - MSR (contemporaneous):
    - Angtigua and Barbuda: 0.0923 (0.752)
    - Dominica: 0.138 (0.486)
    - Grenada: 0.244 (0.443)
    - St. Kitts and Nevis: 0.127 (0.649)
    - St. Lucia: 0.279 (0.445)
    - St. Vincent and the Grenadines: 0.0586 (0.715)
  - MSR t-1:
    - Angtigua and Barbuda: 0.0965 (0.741)
    - Dominica: 0.259 (0.189)
    - Grenada: 0.374 (0.241)
    - St. Kitts and Nevis: 0.220 (0.431)
    - St. Lucia: 0.316 (0.388)
    - St. Vincent and the Grenadines: 0.324** (0.0458)
  - MSR t-2:
    - Angtigua and Barbuda: -0.0469 (0.873)
    - Dominica: 0.213 (0.284)
    - Grenada: 0.144 (0.652)
    - St. Kitts and Nevis: 0.121 (0.665)
    - St. Lucia: 0.00834 (0.982)
    - St. Vincent and the Grenadines: -0.0630 (0.699)
  - MSR t-3:
    - Angtigua and Barbuda: 0.0320 (0.925)
    - Dominica: 0.105 (0.649)
    - Grenada: -0.0391 (0.915)
    - St. Kitts and Nevis: -0.386 (0.237)
    - St. Lucia: 0.627 (0.140)
    - St. Vincent and the Grenadines: -0.0119 (0.950)
  - MSR t-4:
    - Angtigua and Barbuda: 0.281 (0.405)
    - Dominica: 0.236 (0.299)
    - Grenada: 0.169 (0.643)
    - St. Kitts and Nevis: 0.111 (0.732)
    - St. Lucia: -1.010** (0.0181)
    - St. Vincent and the Grenadines: 0.0693 (0.709)
  - MSR t-5:
    - Angtigua and Barbuda: 0.120 (0.722)
    - Dominica: 0.183 (0.418)
    - Grenada: 0.0671 (0.854)
    - St. Kitts and Nevis: 0.223 (0.493)
    - St. Lucia: 0.561 (0.193)
    - St. Vincent and the Grenadines: 0.0510 (0.783)
  - MSR t-6:
    - Angtigua and Barbuda: 0.380 (0.196)
    - Dominica: 0.137 (0.485)
    - Grenada: -0.00993 (0.975)
    - St. Kitts and Nevis: -0.01391 (0.960)
    - St. Lucia: 1.596*** (7.40e-05)
    - St. Vincent and the Grenadines: 0.313* (0.0530)
  - Time dummy:
    - Angtigua and Barbuda: 0.585* (0.0862)
    - Dominica: 0.257 (0.256)
    - Grenada: 0.261 (0.473)
    - St. Kitts and Nevis: 0.474 (0.146)
    - St. Lucia: 1.238*** (0.00456)
    - St. Vincent and the Grenadines: 0.217 (0.247)
  - Constant:
    - Angtigua and Barbuda: -0.0534 (0.187)
    - Dominica: -0.0280 (0.295)
    - Grenada: -0.0363 (0.401)
    - St. Kitts and Nevis: -0.0444 (0.247)
    - St. Lucia: -0.0546 (0.278)
    - St. Vincent and the Grenadines: -0.0323 (0.146)
- Sample and fit:
  - Observations: 121 for each country
  - R-squared:
    - Angtigua and Barbuda: 0.054
    - Dominica: 0.121
    - Grenada: 0.084
    - St. Kitts and Nevis: 0.118
    - St. Lucia: 0.349
    - St. Vincent and the Grenadines: 0.162

### Transmission of Minimum Saving Deposit Rate to Retail Interest Rates — Lending Rates (Dependent variable: D. Lending rate)
- Key estimated coefficients (coefficient followed by p-value in parentheses):
  - Lending rate t-1:
    - Angtigua and Barbuda: -0.200** (0.0246)
    - Dominica: -0.0743 (0.342)
    - Grenada: -0.256*** (0.00672)
    - St. Kitts and Nevis: -0.0129 (0.887)
    - St. Lucia: -0.210** (0.0175)
    - St. Vincent and the Grenadines: -0.206** (0.0292)
  - Lending rate t-2:
    - Angtigua and Barbuda: -0.0435 (0.599)
    - Dominica: -0.234*** (0.00331)
    - Grenada: -0.0458 (0.627)
    - St. Kitts and Nevis: -0.104 (0.254)
    - St. Lucia: -0.128 (0.140)
    - St. Vincent and the Grenadines: -0.0118 (0.900)
  - MSR (contemporaneous):
    - Angtigua and Barbuda: 0.0236 (0.947)
    - Dominica: 0.161 (0.538)
    - Grenada: 0.003540 (0.992)
    - St. Kitts and Nevis: 0.238 (0.294)
    - St. Lucia: 0.0246 (0.963)
    - St. Vincent and the Grenadines: -0.00226 (0.994)
  - MSR t-1:
    - Angtigua and Barbuda: 0.00269 (0.994)
    - Dominica: -0.0521 (0.842)
    - Grenada: 0.630* (0.0852)
    - St. Kitts and Nevis: 0.148 (0.515)
    - St. Lucia: 0.00505 (0.992)
    - St. Vincent and the Grenadines: 0.0787 (0.778)
  - MSR t-2:
    - Angtigua and Barbuda: 0.282 (0.424)
    - Dominica: 0.138 (0.596)
    - Grenada: -0.207 (0.573)
    - St. Kitts and Nevis: -0.0729 (0.748)
    - St. Lucia: 0.03620 (0.945)
    - St. Vincent and the Grenadines: 0.126 (0.652)
  - MSR t-3:
    - Angtigua and Barbuda: -0.514 (0.214)
    - Dominica: 0.0277 (0.927)
    - Grenada: -0.313 (0.462)
    - St. Kitts and Nevis: -0.874*** (0.00130)
    - St. Lucia: -0.619 (0.313)
    - St. Vincent and the Grenadines: -0.364 (0.260)
  - MSR t-4:
    - Angtigua and Barbuda: 0.326 (0.432)
    - Dominica: 0.179 (0.551)
    - Grenada: 0.259 (0.538)
    - St. Kitts and Nevis: 0.524* (0.0574)
    - St. Lucia: 0.204 (0.740)
    - St. Vincent and the Grenadines: 0.279 (0.390)
  - MSR t-5:
    - Angtigua and Barbuda: 0.391 (0.344)
    - Dominica: 0.294 (0.330)
    - Grenada: 0.355 (0.401)
    - St. Kitts and Nevis: 0.366 (0.187)
    - St. Lucia: 0.739 (0.229)
    - St. Vincent and the Grenadines: 0.109 (0.737)
  - MSR t-6:
    - Angtigua and Barbuda: 0.747** (0.0363)
    - Dominica: 1.719*** (1.50e-09)
    - Grenada: 1.048*** (0.00465)
    - St. Kitts and Nevis: 0.834*** (0.000344)
    - St. Lucia: 2.308*** (2.69e-05)
    - St. Vincent and the Grenadines: 0.568** (0.0438)
  - Time dummy:
    - Angtigua and Barbuda: 1.387*** (0.00112)
    - Dominica: 0.509* (0.0923)
    - Grenada: 0.601 (0.157)
    - St. Kitts and Nevis: 0.749*** (0.00765)
    - St. Lucia: 1.751*** (0.00582)
    - St. Vincent and the Grenadines: 0.393 (0.221)
  - Constant:
    - Angtigua and Barbuda: -0.0527 (0.277)
    - Dominica: 0.0085 (0.810)
    - Grenada: -0.0169 (0.733)
    - St. Kitts and Nevis: -0.0249 (0.422)
    - St. Lucia: -0.0605 (0.402)
    - St. Vincent and the Grenadines: -0.0323 (0.396)
- Sample and fit:
  - Observations: 121 for each country
  - R-squared:
    - Angtigua and Barbuda: 0.235
    - Dominica: 0.345
    - Grenada: 0.175
    - St. Kitts and Nevis: 0.338
    - St. Lucia: 0.270
    - St. Vincent and the Grenadines: 0.129

### Notes on Table A1
- Source: IMF staff estimates and calculations.
- MSR signifies minimum saving deposit rate.
- Results of ADL(2,6) model. P-values presented in parentheses. *** p<0.01, ** p<0.05, * p<0.1.

---

### Appendix Table A2. Panel Regressions Results

### Main panel estimates (selected coefficients across models)
- Dependent variable: Change in ECCU MSR (models labeled (1) through (15))
- Change in ECCU MSR (own regressor where included):
  - Models (1)–(7): 0.444* (0.0687), 0.415* (0.0878), 0.524* (0.0815), 0.482* (0.0865), 0.543*** (4.24e-05), 0.411*** (4.65e-05), 0.382*** (0.000141)
- Change in US policy rate (coefficients by model):
  - (1): 0.0880** (0.0314)
  - (2): 0.0940** (0.0280)
  - (3): -0.0617 (0.271)
  - (4): -0.0627 (0.260)
  - (5): -0.0275 (0.230)
  - (6): 0.485 (0.375)
  - (7): -0.340 (0.220)
  - (8): -0.109** (0.0239)
  - (9): -0.125** (0.0136)
  - (10): -0.109** (0.0242)
  - (11): -0.0534 (0.380)
  - (12): -0.0526 (0.355)
  - (13): -0.0412 (0.474)
  - (14): -0.0892 (0.172)
  - (15): -0.000248 (0.996)
  - Additional single coefficient shown: -0.0696* (0.0843)
- Change in ECCU discount rate:
  - Coefficients reported: -0.262 (0.206) and 0.253 (0.295)
- Change in deposit rate:
  - Models reporting large and highly significant coefficients:
    - 0.446*** (2.19e-07), 0.461*** (1.11e-07), 0.452*** (2.91e-07), 1.222*** (0), 1.343*** (0), 1.347*** (0), 1.299*** (0), 1.092*** (5.70e-05)
- Private credit growth:
  - Coefficients: -0.00891 (0.293), -0.00899 (0.304), -0.0196 (0.192)
  - Lagged private credit growth t-1: -0.0242* (0.0869)
- Change in HFI (industry concentration):
  - -3.80e-05 (0.812), -2.55e-05 (0.880)
- Change in excess reserves:
  - -0.0725** (0.0190), -0.0339 (0.303), -0.0329 (0.314), -0.0187 (0.194), -0.0119 (0.308), -0.00510 (0.892)
- Change in ROE:
  - 0.000313 (0.678), 0.000154 (0.670)
- Change in ROA t-1:
  - -0.0844 (0.388)
- Change in NPL ratio t-1:
  - -0.0503** (0.0466), -0.0490** (0.0493)
- Change in NPL ratio (contemporaneous):
  - 0.00678 (0.795)
- Change in Tier 1 ratio t-1:
  - 2.66e-06 (1.000)
- Change in loan-to-deposit ratio:
  - -0.0362*** (0.00851), -0.0306 (0.157), -0.0120 (0.608), 0.00494 (0.507)
- Constants by model (selected):
  - -0.123* (0.0623), -0.137** (0.0445), -0.0645 (0.408), -0.0687 (0.371), -0.0727** (0.0363), -0.159*** (1.67e-05), -0.144*** (0.000264), -0.0542 (0.460), 0.0426 (0.678), 0.0169 (0.868), 0.0241 (0.762), -0.0445 (0.583), 0.0739 (0.518), 0.120 (0.282), 0.0205 (0.801)

### Sample, fit, and coverage (panel)
- Number of observations by model: 170, 170, 848, 474, 363, 617, 317, 317, 390, 808, 808, 808, 474 (as reported across columns)
- R-squared by model: 0.047, 0.092, 0.140, 0.138, 0.300, 0.681, 0.655, 0.159, 0.170, 0.164, 0.584, 0.573, 0.589, 0.639, 0.231
- Number of countries: 6 (reported for each model)
- Approximate coverage (start date and end date by model):
  - Start dates shown: 1989, 1989, 2004, 2004, 2006, 2011, 2011, 1989, 1989, 1989, 2002, 2004, 2004, 2004, 2006
  - End dates shown: 2016 for all models listed

### Variable definitions and notes
- MSR = ECCU minimum saving deposit rate.
- Excess reserves defined as banks' excess reserves held at the ECCB, in percent of total deposits.
- HFI = Herfindahl-Hirschman Index, higher number signifies greater industry concentration.
- ROA = return on assets.
- NPL = non-performing-to-total assets ratio.
- CAR = capital adequacy ratio, defined as qualifying capital in percent of risk-weighted assets.
- Regression results on unbalanced annual panel data. Includes independent ECCU members. P-values in parentheses. *** p<0.01, ** p<0.05, * p<0.1.

*Source: IMF staff estimates and calculations; Authors' calculations.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp1870.pdf_