## _wp12176

## Source details

**Canonical URL:** [_wp12176](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12176.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12176.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12176.pdf.json)

---

### I. INTRODUCTION
- Pacific Island countries (PICs) are vulnerable to commodity price shocks, posing challenges to monetary policy.
- The spike in global commodity prices in 2007−08 led to rise in headline inflation.
- The global financial crisis that intensified in September 2008 exerted downward pressure on both imported and domestic inflation through lower commodity prices and weak demand conditions.
- Authorities loosened monetary policy, but:
  - there have been limited reductions in bank lending rates and private credit growth has been anemic.
  - these sluggish responses raised a key question about the effectiveness of monetary policy transmission mechanisms in PICs and the appropriate nominal anchor given the renewed surge in global commodity prices.
- Scope of the paper:
  - an assessment of the impact of international commodity price shocks, and monetary and exchange rate transmission in PICs to explore whether the exchange rate could help absorb exogenous shocks.
  - next section reviews the key objectives of monetary policy and how it is conducted in PICs.
  - Section III contains an empirical assessment of monetary and exchange transmission mechanisms in PICs, focusing on the responses of real GDP and headline inflation to external shocks, exchange rate changes, and monetary aggregates or interest rates.
  - a model-based approach is applied to Tonga to consider the appropriate nominal anchor to maintain macroeconomic stability in a Small Island Country (SIC).

### Exchange rate and policy mix
- Conceptual points:
  - The exchange rate could be used as an additional tool to cushion against exogenous shocks.
  - With a conventional exchange rate peg:
    - the burden of addressing the impact of exogenous shocks falls on monetary and fiscal policies.
    - given the limited degree of capital mobility, monetary policy enjoys a certain degree of freedom despite the exchange rate peg regime.
  - If monetary policy transmission is weak and prospects for improving its effectiveness are limited in the short run:
    - counter-cyclical policies will require an appropriate mix of monetary, fiscal and exchange rate policies.
    - monetary policy could be closely coordinated with fiscal policy to achieve the desired impact but that coordination is often constrained by the lack of fiscal buffers.
- Policy implication:
  - pegged exchange rate regimes should not preclude considerations to introduce some flexibility to increase the role of the exchange rate in absorbing external shocks and to provide additional freedom for monetary policy.

### Exchange rate regimes and inflation drivers (PICs)
- De jure exchange rate regimes:
  - adjustable pegs or bands to a basket of currencies except PNG that is a managed floating regime.
- Inflation drivers and pass-through:
  - There is a large weight of fuel and food items in the CPI basket in PICs, implying a high pass-through of global commodity prices to headline inflation.
  - The share of imported items in the CPI basket is more than 40 percent in Solomon Islands and Tonga (countries that disaggregate the CPI basket between domestic and imported goods).
  - Central banks of PICs have appreciated their currencies in nominal and real effective terms over the last 2 years to cushion the impact of higher global food and fuel prices.
  - Prior work (Dunn and others (2011)) shows that the monetary transmission mechanism in PICs is weak.

### Panel VAR evidence on transmission mechanisms (1995–2011, annual data)
- Model setup and identification:
  - Panel VAR covers Fiji, Papua New Guinea, Samoa, Solomon Islands, Tonga, and Vanuatu.
  - Ordering (recursive VAR): global fuel and food prices (exogenous) → exchange rate and monetary indicators → real GDP → headline inflation.
  - All variables in log-differences, except interest rates. Lag length selected by standard information criteria = 2 years.
- Key empirical results:
  - Higher global food and fuel prices raise headline inflation; global food prices have a larger and more persistent impact than global fuel prices.
  - The pass-through of the exchange rate to headline inflation is 60 percent within 4 one year, with a complete pass-through within the second year.
  - The impact on headline inflation explained by monetary impulses (broad money, domestic credit, reserve money, or interest rates) is relatively small in the baseline model.
  - Higher interest rates are associated with greater inflation, possibly indicating reverse causation where exchange rate changes and inflation determine the level of interest rates.
  - Real GDP is not well explained by the shocks considered except by its own innovations; global commodity prices appear to affect economic activity more than global GDP, although these effects are not statistically significant.
  - A panel VAR with global GDP instead of global commodity prices was statistically insignificant and not reported.

### FPAS (small New Keynesian / DSGE-style) model for Tonga: structure and estimation
- Model features:
  - Baseline extends Berg, Karam, and Laxton (2006 a, b) to include global oil and food prices, a Phillips-curve specification for headline and domestic price inflation, external demand (partner country GDP growth), and two closure rules: (i) Taylor-rule (interest rate targeting) and (ii) modified exchange rate targeting following Parrado (2004).
  - Four behavioral equations:
    1. aggregate demand (IS) with lending conditions proxied by real credit growth;
    2. Phillips curve linking inflation to past/expected inflation, output gap, fuel prices, and exchange rate;
    3. uncovered interest parity (UIP) in real terms with a historical average risk premium ρ* and smoothness parameter zδ;
    4. modified Taylor-rule for policy interest rate or exchange rate.
  - Exchange rate targeting is modeled as the change of NEER (or equivalent real exchange rate gap) being a function of inflation gap and output gap.
- Estimation:
  - Bayesian techniques with diffused priors; variables seasonally adjusted with X12 except interest rate and exchange rate; expressed in gap terms (Hodrick-Prescott deviations).
- Policy evaluation:
  - Loss function: L = Σ_{t=0}^∞ [λπ(πt − πT)^2 + λy(ygapt)^2], with λπ and λy as relative weights.

### FPAS model results and policy comparison for Tonga
- Shock identification and dynamics:
  - Impulse responses and variance decompositions indicate demand and supply shocks are nearly equally important for inflation dynamics.
  - Global commodity price and external demand shocks have significant impacts on output and inflation, requiring policy responses.
  - Exchange rate changes have a greater influence on inflation; interest rates affect aggregate demand more than exchange rate changes, but the impact is small.
- Relative performance of policy rules (variability under alternative rules):
  - Macroeconomic Volatility Under Alternative Policy Rules
    - GDP Gap — Exchange rate targeting: 0.34
    - Headline Inflation — Exchange rate targeting: 1.01
    - GDP Gap — Interest rate targeting: 1.09
    - Headline Inflation — Interest rate targeting: 4.76
  - Conclusion: An exchange rate based targeting approach does much better at stabilizing both output and inflation in Tonga.

### Policy implications and recommendations
- Use of exchange rate flexibility:
  - Exchange rate flexibility could be a useful tool for macroeconomic stabilization in PICs.
  - External shocks, particularly global commodity shocks, have a significant impact on headline inflation both directly and indirectly through intermediate imports.
  - The high degree of exchange rate pass-through and the weak monetary transmission mechanism suggest exchange rate changes are more effective than monetary policy at affecting inflation.
  - Monetary impulses have a weak impact on headline inflation whether measured by broad money, reserve money, interest rates, or domestic credit—likely reflecting excess liquidity in the banking system and structural impediments to lending (weak contract enforcement and bankruptcy procedures).
  - Use of exchange rate flexibility for short-term domestic stability within the de jure exchange rate basket regime requires a comfortable level of international reserves and should avoid large deviations from medium-term fundamentals.
- Strengthening monetary transmission through structural reforms:
  - Enhance interest rate pass-through to lending rates by addressing excess liquidity and structural problems, including introducing greater competition in the financial system.
  - Develop financial markets: collateralized interbank lending is not widely available, and secondary markets for government and central bank paper are virtually non-existent.
  - Greater exchange rate flexibility may provide greater independence to monetary policy, given limited interest-sensitive cross-border capital flows but some possible responsiveness of remittances and nonresident deposits to domestic interest rates.

### Appendix I: Identification of VAR
- Structural VAR modeling:
  - Structural form equation: G(L)Y t = C(L) X t + ε t
  - G(L) is a n × n matrix polynomial in the lag operator.
  - C(L) is a n × k matrix polynomial in the lag operator.
  - Y t is a n ×1 vector of endogenous variables.
  - X t is a k ×1 vector of exogenous foreign variables.
  - ε t is a n ×1 vector of structural disturbances, with var(ε t) = Λ, where Λ is a diagonal matrix and the diagonal elements are the variances of structural disturbances.
  - Structural disturbances are assumed to be mutually uncorrelated.
- Corresponding reduced form VAR:
  - Y t = A(L)Y t + B(L)X t + μ t
  - A(L) and B(L) are matrices polynomial.
  - μ t is a vector of reduced form disturbances, with var(μ t ) = Σ.
- Variables and controls:
  - Exogenous vector X t contains: X t  ̳ = [Fuel Food]
  - Fuel and Food are WEO global fuel prices and food prices.
  - Endogenous variables Y t include: Y t  ̳ = [Y E M HCPI]
    - Components: real GDP (Y), exchange rate (E), a monetary aggregate or interest rate (M), and headline consumer price index (HCPI).
  - Choice of monetary policy variable:
    - Baseline model uses interest rates as the money policy variable.
    - Other monetary aggregates are tested.
    - Domestic credit to the private sector has the largest explanatory power among tested monetary aggregates, albeit still relatively small compared to the exchange rate impact on inflation.
- Identification scheme:
  - Recursive VAR using Cholesky decomposition (as in Sims, 1980) to orthogonalize reduced form disturbances.
- Key methodological references cited include Kim and Roubini (2000); Sims and Zha (2006); Sims (1980); Berg, Karam, and Laxton (2006a,b); Laxton, Rose, and Scott (2009); Carabenciov et al. (2008).

*Source: _wp12176 (IMF content provided)*

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

### _wp12176 - References

### I. INTRODUCTION
- Pacific Island countries (PICs) are vulnerable to commodity price shocks, posing challenges to monetary policy.
- The spike in global commodity prices in 2007−08 led to rise in headline inflation.
- The global financial crisis that intensified in September 2008 exerted downward pressure on both imported and domestic inflation through lower commodity prices and weak demand conditions.
- Authorities loosened monetary policy, but:
  - there have been limited reductions in bank lending rates and private credit growth has been anemic.
  - these sluggish responses raised a key question about the effectiveness of monetary policy transmission mechanisms in PICs and the appropriate nominal anchor given the renewed surge in global commodity prices.

### Exchange rate and policy mix
- The exchange rate could be used as an additional tool to cushion against exogenous shocks.2
- With a conventional exchange rate peg:
  - the burden of addressing the impact of exogenous shocks falls on monetary and fiscal policies.
  - given the limited degree of capital mobility, monetary policy enjoys a certain degree of freedom despite the exchange rate peg regime.
- If monetary policy transmission is weak and prospects for improving its effectiveness are limited in the short run:
  - counter-cyclical policies will require an appropriate mix of monetary, fiscal and exchange rate policies.3
  - monetary policy could be closely coordinated with fiscal policy to achieve the desired impact but that coordination is often constrained by the lack of fiscal buffers.
- Policy implication:
  - pegged exchange rate regimes should not preclude considerations to introduce some flexibility to increase the role of the exchange rate in absorbing external shocks and to provide additional freedom for monetary policy.

### Scope and structure of the paper
- The paper provides:
  - an assessment of the impact of international commodity price shocks, and monetary and exchange rate transmission in PICs to explore whether the exchange rate could help absorb exogenous shocks.
- Structure:
  - next section reviews the key objectives of monetary policy and how it is conducted in PICs.
  - Section III contains an empirical assessment of monetary and exchange transmission mechanisms in PICs, focusing on the responses of real GDP and headline inflation to external shocks, exchange rate changes, and monetary aggregates or interest rates.
  - the following section more formally considers the appropriate nominal anchor to maintain macroeconomic stability in a Small Island Country (SIC) using a model-based approach applied to Tonga.

*Source: _wp12176 - References*

### Section IV concludes with some policy implications.

### _wp12176 - Section IV concludes with some policy implications.

### Exchange rate regimes and inflation drivers
- The de jure exchange rate regimes in PICs are adjustable pegs or bands to a basket of currencies except PNG that is a managed floating regime.
- There is a large weight of fuel and food items in the CPI basket in PICs, implying a high pass-through of global commodity prices to headline inflation.
- The share of imported items in the CPI basket is more than 40 percent in Solomon Islands and Tonga (countries that disaggregate the CPI basket between domestic and imported goods).
- Central banks of PICs have appreciated their currencies in nominal and real effective terms over the last 2 years to cushion the impact of higher global food and fuel prices.
- Dunn and others (2011) show that the monetary transmission mechanism in PICs is weak.

### Panel VAR evidence on transmission mechanisms (1995–2011, annual data)
- Model setup and identification:
  - Panel VAR covers Fiji, Papua New Guinea, Samoa, Solomon Islands, Tonga, and Vanuatu.
  - Ordering (recursive VAR): global fuel and food prices (exogenous) → exchange rate and monetary indicators → real GDP → headline inflation.
  - All variables in log-differences, except interest rates. Lag length selected by standard information criteria = 2 years.
- Key empirical results:
  - Higher global food and fuel prices raise headline inflation; global food prices have a larger and more persistent impact than global fuel prices.
  - The pass-through of the exchange rate to headline inflation is 60 percent within 4 one year, with a complete pass-through within the second year.
  - The impact on headline inflation explained by monetary impulses (broad money, domestic credit, reserve money, or interest rates) is relatively small in the baseline model.
  - Higher interest rates are associated with greater inflation, possibly indicating reverse causation where exchange rate changes and inflation determine the level of interest rates.
  - Real GDP is not well explained by the shocks considered except by its own innovations; global commodity prices appear to affect economic activity more than global GDP, although these effects are not statistically significant.
  - A panel VAR with global GDP instead of global commodity prices was statistically insignificant and not reported.

### FPAS (small New Keynesian / DSGE-style) model for Tonga: structure and estimation
- Model features:
  - Baseline extends Berg, Karam, and Laxton (2006 a, b) to include global oil and food prices, a Phillips-curve specification for headline and domestic price inflation, external demand (partner country GDP growth), and two closure rules: (i) Taylor-rule (interest rate targeting) and (ii) modified exchange rate targeting following Parrado (2004).
  - Four behavioral equations: (1) aggregate demand (IS) with lending conditions proxied by real credit growth; (2) Phillips curve linking inflation to past/expected inflation, output gap, fuel prices, and exchange rate; (3) uncovered interest parity (UIP) in real terms with a historical average risk premium ρ* and smoothness parameter zδ; (4) modified Taylor-rule for policy interest rate or exchange rate.
  - Exchange rate targeting is modeled as the change of NEER (or equivalent real exchange rate gap) being a function of inflation gap and output gap.
- Estimation:
  - Bayesian techniques with diffused priors; variables seasonally adjusted with X12 except interest rate and exchange rate; expressed in gap terms (Hodrick-Prescott deviations).
- Policy evaluation:
  - Loss function: L = Σ_{t=0}^∞ [λπ(πt − πT)^2 + λy(ygapt)^2], with λπ and λy as relative weights.

### FPAS model results and policy comparison for Tonga
- Shock identification and dynamics:
  - Impulse responses and variance decompositions indicate demand and supply shocks are nearly equally important for inflation dynamics.
  - Global commodity price and external demand shocks have significant impacts on output and inflation, requiring policy responses.
  - Exchange rate changes have a greater influence on inflation; interest rates affect aggregate demand more than exchange rate changes, but the impact is small.
- Relative performance of policy rules (variability under alternative rules):
  - Macroeconomic Volatility Under Alternative Policy Rules
    - GDP Gap — Exchange rate targeting: 0.34
    - Headline Inflation — Exchange rate targeting: 1.01
    - GDP Gap — Interest rate targeting: 1.09
    - Headline Inflation — Interest rate targeting: 4.76
  - Conclusion: An exchange rate based targeting approach does much better at stabilizing both output and inflation in Tonga.

### Policy implications and recommendations
- Exchange rate flexibility could be a useful tool for macroeconomic stabilization in PICs:
  - External shocks, particularly global commodity shocks, have a significant impact on headline inflation both directly and indirectly through intermediate imports.
  - The high degree of exchange rate pass-through and the weak monetary transmission mechanism suggest exchange rate changes are more effective than monetary policy at affecting inflation.
  - Monetary impulses have a weak impact on headline inflation whether measured by broad money, reserve money, interest rates, or domestic credit—likely reflecting excess liquidity in the banking system and structural impediments to lending (weak contract enforcement and bankruptcy procedures).
  - Use of exchange rate flexibility for short-term domestic stability within the de jure exchange rate basket regime requires a comfortable level of international reserves and should avoid large deviations from medium-term fundamentals.
- Strengthening monetary transmission through structural reforms:
  - Enhance interest rate pass-through to lending rates by addressing excess liquidity and structural problems, including introducing greater competition in the financial system.
  - Develop financial markets: collateralized interbank lending is not widely available, and secondary markets for government and central bank paper are virtually non-existent.
  - Greater exchange rate flexibility may provide greater independence to monetary policy, given limited interest-sensitive cross-border capital flows but some possible responsiveness of remittances and nonresident deposits to domestic interest rates.

*Source: _wp12176 - Section IV concludes with some policy implications.*

### Appendix I: Identification of VAR

### Appendix I: Identification of VAR

### Structural VAR Modeling
- Structural form equation:
  - G(L)Y t = C(L) X t + ε t
- Definitions and assumptions:
  - G(L) is a n × n matrix polynomial in the lag operator.
  - C(L) is a n × k matrix polynomial in the lag operator.
  - Y t is a n ×1 vector of endogenous variables.
  - X t is a k ×1 vector of exogenous foreign variables.
  - ε t is a n ×1 vector of structural disturbances, with var(ε t) = Λ, where Λ is a diagonal matrix and the diagonal elements are the variances of structural disturbances.
  - Structural disturbances are assumed to be mutually uncorrelated.

- Corresponding reduced form VAR:
  - Y t = A(L)Y t + B(L)X t + μ t
  - A(L) and B(L) are matrices polynomial.
  - μ t is a vector of reduced form disturbances, with var(μ t ) = Σ.

### Variables and controls
- Exogenous vector X t contains:
  - X t  ̳ = [Fuel Food]
  - Fuel and Food are WEO global fuel prices and food prices.
  - These variables are included to control for changes in overall global economic conditions and fluctuations in the prices of PIC‘s main imported commodities.
- Endogenous variables Y t include:
  - Y t  ̳ = [Y E M HCPI]
  - Where the components are: real GDP (Y), exchange rate (E), a monetary aggregate or interest rate (M), and headline consumer price index (HCPI).
- Choice of monetary policy variable:
  - Baseline model uses interest rates as the money policy variable.
  - Other monetary aggregates are tested.
  - Domestic credit to the private sector has the largest explanatory power among tested monetary aggregates, albeit still relatively small compared to the exchange rate impact on inflation.

### Identification scheme
- Identification approach:
  - Recursive VAR using Cholesky decomposition (as in Sims, 1980) to orthogonalize reduced form disturbances.
  - This recursive ordering is noted as a popular and convenient method to recover structural form parameters from estimated reduced form parameters.

### Key methodological references (selected from source)
- Approaches and prior work referenced include:
  - Kim and Roubini (2000); Sims and Zha (2006) for structural VAR setup.
  - Sims (1980) for recursive VAR and Cholesky orthogonalization.
  - Additional modeling and policy-analysis guides and applications cited (e.g., Berg, Karam, and Laxton 2006a,b; Laxton, Rose, and Scott 2009; Carabenciov et al. 2008).

*Source: Appendix I: Identification of VAR (from provided IMF content)*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12176.pdf_