## _wp12126

## Source details

**Canonical URL:** [_wp12126](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12126.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12126.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12126.pdf.json)

---

### Introduction and paper objectives
- The study examines exchange rate pass-through (ERPT) into domestic prices in Maldives since 1994 using:
  - A nonparametric (Supply and Use Table) static estimation to calculate the share of imported inputs in total costs and the overall static sensitivity of the CPI to exchange rate changes.
  - A recursive VAR econometric estimation (quarterly data for 1994–2010) to assess the speed and dynamics of ERPT to consumer and producer price indexes.
- Research questions:
  - What is the extent of exchange rate pass through in the Maldives?
  - What is the speed of the pass through to different prices, and what are the key drivers?
  - What would be the likely impact of an exchange rate shock on prices?

### Key findings (nonparametric and VAR evidence)
- Nonparametric (Supply and Use Table) estimates:
  - Total ERPT into CPI (first and second round effects) is 0.79; interpretation: a 10 percent depreciation would lead to an increase in consumer prices of 7.9 percent (Table 1).
  - First and second round effects by major CPI groups (selected entries from Table 1):
    - Total Maldives: 1st round effect = 0.51; 2nd round effect = 0.28; Total impact = 0.79; CPI Weights = 1.000; Pass-through to CPI = 0.79
    - Food and non-alcoholic beverages: 1st round = 0.67; 2nd round = 0.22; Total impact = 0.89; CPI Weights = 0.330; Pass-through to CPI = 0.30
    - Tobacco and narcotics: 1st round = 1.00; 2nd round = 0.00; Total impact = 1.00; CPI Weights = 0.030; Pass-through to CPI = 0.03
    - Clothing and footwear: 1st round = 0.40; 2nd round = 0.48; Total impact = 0.89; CPI Weights = 0.060; Pass-through to CPI = 0.05
    - Housing, water, electricity, gas and other fuels: 1st round = 0.15; 2nd round = 0.67; Total impact = 0.81; CPI Weights = 0.190; Pass-through to CPI = 0.16
    - Furnishings, household equipment and routine maintenance: 1st round = 0.85; 2nd round = 0.04; Total impact = 0.89; CPI Weights = 0.050; Pass-through to CPI = 0.05
    - Health: 1st round = 0.41; 2nd round = 0.05; Total impact = 0.46; CPI Weights = 0.050; Pass-through to CPI = 0.02
    - Transport: 1st round = 0.24; 2nd round = 0.36; Total impact = 0.61; CPI Weights = 0.050; Pass-through to CPI = 0.03
    - Communications: 1st round = 0.16; 2nd round = 0.23; Total impact = 0.39; CPI Weights = 0.060; Pass-through to CPI = 0.02
    - Education: 1st round = 0.27; 2nd round = 0.16; Total impact = 0.43; CPI Weights = 0.030; Pass-through to CPI = 0.01
    - Hotels, cafes and restaurants: 1st round = 0.01; 2nd round = 0.32; Total impact = 0.33; CPI Weights = 0.010; Pass-through to CPI = 0.00
    - Miscellaneous goods and services: 1st round = 0.96; 2nd round = 0.01; Total impact = 0.97; CPI Weights = 0.080; Pass-through to CPI = 0.08
    - Religion: 1st round = 0.00; 2nd round = 0.12; Total impact = 0.12; CPI Weights = 0.000; Pass-through to CPI = 0.00
  - Out of 85 SUT products, about 29 are imports, 19 have import content exceeding 50 percent, and only 23 are fully domestic. Considering imported content of domestic inputs (2nd round), 62 products have total imported content over 50 percent.
- Accounting for substitution effects and distribution margins (Campa and Goldberg (2006) calibrated elasticity parameters):
  - Estimated ERPT coefficient for Maldives = 0.77 (Table 2).
  - Comparison context in Table 2 (selected figures):
    - Campa & Goldberg (2006) average ERPT into CPI = 0.26 for their sample of 21 (mainly OECD) countries.
    - For Maldives: Average Share of imported inputs = 45.9; Average distribution margins = 9.7; Imported input share of tradables in consumption = 0.85; Share of tradables in consumption = 0.65; Share of imported input cost in tradable production costs = 0.89; Share of imported input cost in nontradable production costs = 0.48; Exchange rate pass through into CPI = 0.77
- VAR estimation results (summary statements from paper):
  - ERPT has strong impacts on consumer and producer prices, though pass through is incomplete.
  - Shocks to producer and consumer price indexes arising from changes in the nominal effective exchange rate persist for most of the time horizon.
- Additional reported pass-through figures:
  - Table 3 reported pass-through into CPI computed as (D-C-B)/A*100 = 90.6 percent.

### Maldives macroeconomic and price context relevant to ERPT
- Inflation and commodity price movements:
  - Maldives inflation peaked at 17 percent on a y-o-y basis in July 2008 amid high global food and fuel prices.
  - By end 2010, inflation had decelerated and stabilized at 5 percent as international commodity prices fell.
  - Between January 2008 and March 2009, movement in the food group accounted for more than 50 percent of total inflation, while housing, water, electricity, and fuels accounted for a quarter.
  - Inflation excluding fish prices and core inflation (excluding food prices) remained lower than overall inflation between mid-2008 and end-2010.
- Regional variation:
  - Maldives topography (1,192 islands; total land area about 298 square kilometers spread over more than 90,000 square kilometers of ocean) produces strong regional variations in inflation.
  - Between 2007 and mid-2009, inflation in Malé was lower than in the atolls; since then trend reversed with Malé inflation driven by housing, water, electricity, fuel, and education costs.
- Exchange rate history:
  - Since 1994, Maldives exchange rate moved twice:
    - Devaluation of 8.8 percent on July 25, 2001 (fixed parity maintained).
    - April 2011 depreciation of almost 20 percent following introduction of an exchange rate band of +/- 20 percent around parity of Rufiyaa 12.8 per US dollar.
  - Devaluation episode of July 2001: despite negative underlying domestic inflation and a sharp fall in international commodity prices, average inflation rose to 3.7 percent in the year after the devaluation.

### Theoretical and empirical considerations affecting ERPT
- Theory:
  - First approximation: change in CPI from exchange rate shocks expected to be proportionate to share of imported goods in consumption.
  - Second-round channels can raise ERPT above import share if domestically produced goods use imported inputs or if tariffs/regulated prices (e.g., electricity indexed to fuel) adjust.
  - Pass-through can be lower than import share due to pricing-to-market, incomplete pass through to import prices, or negative income effects reducing demand.
- Empirical literature summarized:
  - Industrial/advanced economies generally show incomplete ERPT to consumer prices; smaller and more open economies exhibit higher ERPT.
  - Goldfajn and Werlang (2000): pass-through effects increase over time, reach maximum after 12 months, higher in emerging markets.
  - McCarthy (1999): ERPT to consumer prices modest in many countries; ERPT to producer and import prices significant though incomplete.
  - Gueorguiev (2003): Romania—ERPT ranging 60-70 percent for producer prices and 30-40 percent for consumer prices, varying by exchange rate benchmark.
  - Pricing-to-market and margin adjustments by firms commonly reduce observed ERPT.
  - Large devaluations/regime switches: Borensztein and De Gregorio (1999) find about 30 percent offset after three months and about 60 percent after two years.

### Interpretation and structural drivers in Maldives
- Structural features producing high ERPT estimates:
  - High import dependence in production and consumption.
  - Lack of domestic substitutes for many goods.
  - Relatively small distribution margins.
  - Evidence that certain distributors/retailers may pass through 100 percent of additional costs to consumers.
  - Non-tradable sectors (health, transport, construction, education) face substantial second-round impacts due to imported intermediate materials.

### V. Econometric estimation: objective and framework
- Purpose: undertake dynamic estimation of exchange rate pass through (ERPT) to complement static analysis and to inform effective policy response to inflation and exchange rate developments.
- Model focus: response of prices to exchange rate shocks using the nominal effective exchange rate (NEER) because Maldives operated a fixed exchange rate regime during most of the sample period.
- Analytical framework: recursive five-variable VAR system (drawn from McCarthy (1999)) with identified shocks from commodity prices, economic activity, and the exchange rate. Variables in order:
  - Change in all commodity prices (ComP_t)
  - Output gap (y_gap)
  - Exchange rate (Δe_t using NEER)
  - Producer prices (PPI)
  - Consumer prices (CPI)
- Two-stage inflation mechanism: shocks transmit first through commodity prices → output gap → exchange rate → producer prices → consumer prices.

### Data, estimation setup, and econometric issues
- Sample: quarterly observations from 1994Q3-2010Q4.
- Exchange rate regime context:
  - Fixed peg to the US dollar for most of the sample (modest parity change in 2001).
  - Study excludes post-April 2011 regime change due to insufficient data.
  - t = Jul y 2001 is the month when the rufiyaa/US$ exchange rate was devalued.
- Key data definitions:
  - All Commodity Price Index: Commodity Price Index, 2005 = 100, includes both Fuel and Non-Fuel Price Indices.
  - Output gap: deviation of log real GDP from trend (Hodrick Prescott filter); alternative specification uses real GDP growth.
  - NEER: trade-weighted index of bilateral exchange rates of major trading partners.
  - PPI and CPI: as reported by Maldives authorities (PPI reported from 1994; pre-1994 interpolated).
- Stationarity and preprocessing:
  - Unit root tests on log levels indicated nonstationarity; series transformed to achieve stationarity (Appendix II).
  - ADF test details: Augmented Dickey Fuller Test for unit root, with trend and intercept; selected lag length minimizes Schwartz Information Criterion; maximum lags truncated at 11.
  - Eight lags of the variables used in VAR estimation following Akaike Information Criterion selection.
  - All level variables transformed to natural logarithms and seasonally adjusted using X-12.
- Identification:
  - Structural shocks recovered via Cholesky decomposition of the VAR residual variance-covariance matrix (assumes zero contemporaneous correlations consistent with the recursive ordering).

### B. Estimation results — impulse responses and ERPT coefficients
- Impulse response estimation:
  - IRFs computed over a two-year (eight-quarter) horizon, standardized to correspond to the response to a 1% shock in the exchange rate.
  - ERPT coefficient computed as cumulative change in a price index divided by cumulative change in the exchange rate over the horizon.
- Key quantitative ERPT findings:
  - 74 percent of the change in the NEER is passed through to producer prices (PPI) by the end of the fourth quarter following the NEER change; speed of pass through is gradual.
  - 88 percent of the change in NEER is passed on to consumer prices (CPI) by the end of the fourth quarter; CPI pass through is slower initially than PPI but overall stronger and persistent through the fifth quarter before falling.
  - Nonparametric estimate: around 79 percent ERPT to CPI as a first round effect (Table 1).
  - Table 3 reported pass-through into CPI computed as (D-C-B)/A*100 = 90.6 percent.
- Sectoral context and interpretation:
  - PPI composition: resorts and hotels account for 57 percent of PPI weight; air transport accounts for 7.5 percent of PPI weight.
  - Dollar pricing and advance booking in tourism reduce immediate short-run pass through to PPI, allowing margin adjustments over time.
  - Persistence: most shocks to prices persist beyond the first year, implying a long horizon for monetary policy responses.

### Variance decomposition: drivers of price variation
- Method: variance decomposition of PPI and CPI from the VAR to determine proportion of forecast variance attributable to each variable’s shocks.
- Main findings (verbatim table-line strings preserved from source):
  - Heading: "Percentage change in forecast variance of PPI attributable to changes in:"
  - Table-like lines:
    - 15.40.02.692.00.0
    - 212.25.82.176.73.2
    - 324.54.92.165.82.7
    - 430.58.01.954.75.0
    - 537.08.82.147.05.1
    - 639.08.33.643.65.5
    - 740.38.93.741.65.4
    - 848.58.35.333.54.4
  - Heading: "Percentage change in forecast variance of CPI attributable to changes in:"
  - Table-like lines:
    - 13.80.71.459.834.3
    - 25.21.81.953.837.2
    - 35.34.01.853.435.5
    - 48.13.85.649.033.4
    - 515.13.84.945.131.1
    - 620.55.55.640.527.9
    - 727.38.75.733.125.2
    - 828.29.110.128.823.7
- Interpretation summary:
  - PPI: initially dominated by own shocks; beyond the first five quarters, international commodity price shocks are strong determinants of forecast variance.
  - CPI: a larger share of its variation is explained by other variables in the system (commodity price shocks and PPI shocks) rather than own shocks.
  - Output gap: explains a moderate proportion of variation in prices, reflecting a role for domestic demand but moderated by Maldives’s small open-economy link to the international cycle.

### VI. Summary, conclusions, and policy implications
- Summary of empirical results:
  - Nonparametric estimate: ERPT to CPI ≈ 79 percent (first round).
  - Recursive VAR estimate: ERPT to CPI = 88 percent by end of first year; PPI ERPT = 74 percent by end of fourth quarter.
  - Table 3 reports pass-through into CPI = 90.6 percent (calculation from devaluation event).
  - International commodity price shocks are an important source of variation in PPI and CPI in addition to exchange rate changes.
- Policy implications:
  - High ERPT to CPI implies that devaluation policies are likely to have substantial inflationary effects; monetary policy responses should be designed to mitigate such inflationary impacts.
  - Persistence of price shocks beyond the first year suggests monetary policy responses should take a long horizon into account.
  - Knowledge of the speed and dynamics of pass through informs timing and choice of policy intervention points.
- Suggested extension:
  - Include a monetary aggregate in the VAR system to assess how monetary policy changes impact price development and to inform appropriate monetary policy responses to future price shocks, including depreciation.

### Appendices (selected technical notes and figures)
- Appendix IV: Impulse Response Functions (Real GDP growth model)
  - Response variables and plotted horizons (horizon marks: 2, 4, 6, 8, 10) include DLOG(ALLCOM_SA), RGDPR_SA, DLOG(NEER_SA), DLOG(PPI_SA), DLOG(CPI_SA) responses to each other and to shocks.
  - Response scales and example axis ticks:
    - DLOG(ALLCOM_SA) vertical axis: -.10 -.05 .00 .05 .10
    - RGDPR_SA vertical axis: -4 -2 0 2 4 6 8 10
    - DLOG(NEER_SA) vertical axis: -.02 -.01 .00 .01 .02
    - DLOG(PPI_SA) and DLOG(CPI_SA) vertical axes: -.04 -.02 .00 .02 .04 .06 (for some panels)
  - Note on identification: "Response to Cholesky One S.D. Innovations ± 2 S.E."
- Appendix V — Estimated Cumulative Pass through coefficient
  - Graphic axis ranges:
    - Vertical axis: -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
    - Horizontal axis: 1 2 3 4 5 6 7 8
  - Series label shown: PPICPI
  - Caption: "Estimated Cumulative Pass Through Coefficient"
  - Source note in figure: "Sources: IMF Staff calculations."
- Appendix VI — Variance Decomposition of Producer and Consumer Price Indexes
  - Verbose table-line strings preserved as shown above.

*Source: Maldivian authorities, and Fund staff calculations (from the provided content of _wp12126 - 90.6 percent (Table 3)).*

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

### _wp12126 - References .............................................................................................................

### Introduction and paper objectives
- The study examines exchange rate pass-through (ERPT) into domestic prices in Maldives since 1994, using:
  - A nonparametric (Supply and Use Table) static estimation to calculate the share of imported inputs in total costs and the overall static sensitivity of the CPI to exchange rate changes.
  - A recursive VAR econometric estimation (quarterly data for 1994–2010) to assess the speed and dynamics of ERPT to consumer and producer price indexes.
- Research questions:
  - What is the extent of exchange rate pass through in the Maldives?
  - What is the speed of the pass through to different prices, and what are the key drivers?
  - What would be the likely impact of an exchange rate shock on prices?

### Key findings (nonparametric and VAR evidence)
- Nonparametric (Supply and Use Table) estimates:
  - Total ERPT into CPI (first and second round effects) is 0.79; interpretation: a 10 percent depreciation would lead to an increase in consumer prices of 7.9 percent (Table 1).
  - First and second round effects by major CPI groups (selected entries from Table 1):
    - Total Maldives: 1st round effect = 0.51; 2nd round effect = 0.28; Total impact = 0.79; CPI Weights = 1.000; Pass-through to CPI = 0.79
    - Food and non-alcoholic beverages: 1st round = 0.67; 2nd round = 0.22; Total impact = 0.89; CPI Weights = 0.330; Pass-through to CPI = 0.30
    - Tobacco and narcotics: 1st round = 1.00; 2nd round = 0.00; Total impact = 1.00; CPI Weights = 0.030; Pass-through to CPI = 0.03
    - Clothing and footwear: 1st round = 0.40; 2nd round = 0.48; Total impact = 0.89; CPI Weights = 0.060; Pass-through to CPI = 0.05
    - Housing, water, electricity, gas and other fuels: 1st round = 0.15; 2nd round = 0.67; Total impact = 0.81; CPI Weights = 0.190; Pass-through to CPI = 0.16
    - Furnishings, household equipment and routine maintenance: 1st round = 0.85; 2nd round = 0.04; Total impact = 0.89; CPI Weights = 0.050; Pass-through to CPI = 0.05
    - Health: 1st round = 0.41; 2nd round = 0.05; Total impact = 0.46; CPI Weights = 0.050; Pass-through to CPI = 0.02
    - Transport: 1st round = 0.24; 2nd round = 0.36; Total impact = 0.61; CPI Weights = 0.050; Pass-through to CPI = 0.03
    - Communications: 1st round = 0.16; 2nd round = 0.23; Total impact = 0.39; CPI Weights = 0.060; Pass-through to CPI = 0.02
    - Education: 1st round = 0.27; 2nd round = 0.16; Total impact = 0.43; CPI Weights = 0.030; Pass-through to CPI = 0.01
    - Hotels, cafes and restaurants: 1st round = 0.01; 2nd round = 0.32; Total impact = 0.33; CPI Weights = 0.010; Pass-through to CPI = 0.00
    - Miscellaneous goods and services: 1st round = 0.96; 2nd round = 0.01; Total impact = 0.97; CPI Weights = 0.080; Pass-through to CPI = 0.08
    - Religion: 1st round = 0.00; 2nd round = 0.12; Total impact = 0.12; CPI Weights = 0.000; Pass-through to CPI = 0.00
  - Out of 85 SUT products, about 29 are imports, 19 have import content exceeding 50 percent, and only 23 are fully domestic. Considering imported content of domestic inputs (2nd round), 62 products have total imported content over 50 percent.
- Accounting for substitution effects and distribution margins (Campa and Goldberg (2006) calibrated elasticity parameters):
  - Estimated ERPT coefficient for Maldives = 0.77 (Table 2).
  - Comparison context in Table 2 (selected figures):
    - Campa & Goldberg (2006) average ERPT into CPI = 0.26 for their sample of 21 (mainly OECD) countries.
    - For Maldives: Average Share of imported inputs = 45.9; Average distribution margins = 9.7; Imported input share of tradables in consumption = 0.85; Share of tradables in consumption = 0.65; Share of imported input cost in tradable production costs = 0.89; Share of imported input cost in nontradable production costs = 0.48; Exchange rate pass through into CPI = 0.77
- VAR estimation results (summary statements from paper):
  - ERPT has strong impacts on consumer and producer prices, though pass through is incomplete.
  - Shocks to producer and consumer price indexes arising from changes in the nominal effective exchange rate persist for most of the time horizon.

### Maldives macroeconomic and price context relevant to ERPT
- Inflation and commodity price movements:
  - Maldives inflation peaked at 17 percent on a y-o-y basis in July 2008 amid high global food and fuel prices.
  - By end 2010, inflation had decelerated and stabilized at 5 percent as international commodity prices fell.
  - Between January 2008 and March 2009, movement in the food group accounted for more than 50 percent of total inflation, while housing, water, electricity, and fuels accounted for a quarter.
  - Inflation excluding fish prices and core inflation (excluding food prices) remained lower than overall inflation between mid-2008 and end-2010.
- Regional variation:
  - Maldives topography (1,192 islands; total land area about 298 square kilometers spread over more than 90,000 square kilometers of ocean) produces strong regional variations in inflation.
  - Between 2007 and mid-2009, inflation in Malé was lower than in the atolls; since then trend reversed with Malé inflation driven by housing, water, electricity, fuel, and education costs.
- Exchange rate history:
  - Since 1994, Maldives exchange rate moved twice:
    - Devaluation of 8.8 percent on July 25, 2001 (fixed parity maintained).
    - April 2011 depreciation of almost 20 percent following introduction of an exchange rate band of +/- 20 percent around parity of Rufiyaa 12.8 per US dollar.
  - Devaluation episode of July 2001: despite negative underlying domestic inflation and a sharp fall in international commodity prices, average inflation rose to 3.7 percent in the year after the devaluation.

### Theoretical and empirical considerations affecting ERPT
- Theory:
  - First approximation: change in CPI from exchange rate shocks expected to be proportionate to share of imported goods in consumption.
  - Second-round channels can raise ERPT above import share if domestically produced goods use imported inputs or if tariffs/regulated prices (e.g., electricity indexed to fuel) adjust.
  - Pass-through can be lower than import share due to pricing-to-market, incomplete pass through to import prices, or negative income effects reducing demand.
- Empirical literature summarized:
  - Industrial/advanced economies generally show incomplete ERPT to consumer prices; smaller and more open economies exhibit higher ERPT.
  - Goldfajn and Werlang (2000): pass-through effects increase over time, reach maximum after 12 months, higher in emerging markets.
  - McCarthy (1999): ERPT to consumer prices modest in many countries; ERPT to producer and import prices significant though incomplete.
  - Gueorguiev (2003): Romania—ERPT ranging 60-70 percent for producer prices and 30-40 percent for consumer prices, varying by exchange rate benchmark.
  - Pricing-to-market and margin adjustments by firms commonly reduce observed ERPT.
  - Large devaluations/regime switches: Borensztein and De Gregorio (1999) find about 30 percent offset after three months and about 60 percent after two years.

### Interpretation and structural drivers in Maldives
- Structural features producing high ERPT estimates:
  - High import dependence in production and consumption.
  - Lack of domestic substitutes for many goods.
  - Relatively small distribution margins.
  - Evidence that certain distributors/retailers may pass through 100 percent of additional costs to consumers.
  - Non-tradable sectors (health, transport, construction, education) face substantial second-round impacts due to imported intermediate materials.

*Source: Maldivian authorities, and Fund staff calculations.*

### 90.6 percent (Table 3).

### _wp12126 - 90.6 percent (Table 3)

### V. Econometric estimation: objective and framework
- Purpose: undertake dynamic estimation of exchange rate pass through (ERPT) to complement static analysis and to inform effective policy response to inflation and exchange rate developments.
- Model focus: response of prices to exchange rate shocks using the nominal effective exchange rate (NEER) because Maldives operated a fixed exchange rate regime during most of the sample period.
- Analytical framework: recursive five-variable VAR system (drawn from McCarthy (1999)) with identified shocks from commodity prices, economic activity, and the exchange rate. Variables in order:
  - Change in all commodity prices (ComP_t)
  - Output gap (y_gap)
  - Exchange rate (Δe_t using NEER)
  - Producer prices (PPI)
  - Consumer prices (CPI)
- Two-stage inflation mechanism: shocks transmit first through commodity prices → output gap → exchange rate → producer prices → consumer prices.

### Data, estimation setup, and econometric issues
- Sample: quarterly observations from 1994Q3-2010Q4.
- Exchange rate regime context:
  - Fixed peg to the US dollar for most of the sample (modest parity change in 2001).
  - Study excludes post-April 2011 regime change due to insufficient data.
  - t = Jul y 2001 is the month when the rufiyaa/US$ exchange rate was devalued.
- Key data definitions:
  - All Commodity Price Index: Commodity Price Index, 2005 = 100, includes both Fuel and Non-Fuel Price Indices.
  - Output gap: deviation of log real GDP from trend (Hodrick Prescott filter); alternative specification uses real GDP growth.
  - NEER: trade-weighted index of bilateral exchange rates of major trading partners.
  - PPI and CPI: as reported by Maldives authorities (PPI reported from 1994; pre-1994 interpolated).
- Stationarity and preprocessing:
  - Unit root tests on log levels indicated nonstationarity; series transformed to achieve stationarity (Appendix II).
  - ADF test details: Augmented Dickey Fuller Test for unit root, with trend and intercept; selected lag length minimizes Schwartz Information Criterion; maximum lags truncated at 11.
  - Eight lags of the variables used in VAR estimation following Akaike Information Criterion selection.
  - All level variables transformed to natural logarithms and seasonally adjusted using X-12.
- Identification:
  - Structural shocks recovered via Cholesky decomposition of the VAR residual variance-covariance matrix (assumes zero contemporaneous correlations consistent with the recursive ordering).

### B. Estimation results — impulse responses and ERPT coefficients
- Impulse response estimation:
  - IRFs computed over a two-year (eight-quarter) horizon, standardized to correspond to the response to a 1% shock in the exchange rate.
  - ERPT coefficient computed as cumulative change in a price index divided by cumulative change in the exchange rate over the horizon.
- Key quantitative ERPT findings:
  - 74 percent of the change in the NEER is passed through to producer prices (PPI) by the end of the fourth quarter following the NEER change; speed of pass through is gradual.
  - 88 percent of the change in NEER is passed on to consumer prices (CPI) by the end of the fourth quarter; CPI pass through is slower initially than PPI but overall stronger and persistent through the fifth quarter before falling.
  - Nonparametric estimate: around 79 percent ERPT to CPI as a first round effect (Table 1).
  - Table 3 reported pass-through into CPI computed as (D-C-B)/A*100 = 90.6 percent.
- Sectoral context and interpretation:
  - PPI composition: resorts and hotels account for 57 percent of PPI weight; air transport accounts for 7.5 percent of PPI weight.
  - Dollar pricing and advance booking in tourism reduce immediate short-run pass through to PPI, allowing margin adjustments over time.
  - Persistence: most shocks to prices persist beyond the first year, implying a long horizon for monetary policy responses.

### Variance decomposition: drivers of price variation
- Method: variance decomposition of PPI and CPI from the VAR to determine proportion of forecast variance attributable to each variable’s shocks.
- Main findings:
  - PPI: initially dominated by own shocks; beyond the first five quarters, international commodity price shocks are strong determinants of forecast variance.
  - CPI: a larger share of its variation is explained by other variables in the system (commodity price shocks and PPI shocks) rather than own shocks.
  - Output gap: explains a moderate proportion of variation in prices, reflecting a role for domestic demand but moderated by Maldives’s small open-economy link to the international cycle.

### VI. Summary, conclusions, and policy implications
- Summary of empirical results:
  - Nonparametric estimate: ERPT to CPI ≈ 79 percent (first round).
  - Recursive VAR estimate: ERPT to CPI = 88 percent by end of first year; PPI ERPT = 74 percent by end of fourth quarter.
  - Table 3 reports pass-through into CPI = 90.6 percent (calculation from devaluation event).
  - International commodity price shocks are an important source of variation in PPI and CPI in addition to exchange rate changes.
- Policy implications:
  - High ERPT to CPI implies that devaluation policies are likely to have substantial inflationary effects; monetary policy responses should be designed to mitigate such inflationary impacts.
  - Persistence of price shocks beyond the first year suggests monetary policy responses should take a long horizon into account.
  - Knowledge of the speed and dynamics of pass through informs timing and choice of policy intervention points.
- Suggested extension:
  - Include a monetary aggregate in the VAR system to assess how monetary policy changes impact price development and to inform appropriate monetary policy responses to future price shocks, including depreciation.

*Source: Maldivian authorities and IMF Staff calculations (from the provided content of _wp12126 - 90.6 percent (Table 3)).*

### Appendix IV: Impulse Response Func

### Appendix IV: Impulse Response Func tions (Real GDP growth model)

### Impulse response plots and variables
- Response variables plotted (horizon marks: 2, 4, 6, 8, 10):
  - Response of DLOG(ALLCOM_SA) to DLOG(ALLCOM_SA)
  - Response of DLOG(ALLCOM_SA) to RGDPR_SA
  - Response of DLOG(ALLCOM_SA) to DLOG(NEER_SA)
  - Response of DLOG(ALLCOM_SA) to DLOG(PPI_SA)
  - Response of DLOG(ALLCOM_SA) to DLOG(CPI_SA)
  - Response of RGDPR_SA to DLOG(ALLCOM_SA)
  - Response of RGDPR_SA to RGDPR_SA
  - Response of RGDPR_SA to DLOG(NEER_SA)
  - Response of RGDPR_SA to DLOG(PPI_SA)
  - Response of RGDPR_SA to DLOG(CPI_SA)
  - Response of DLOG(NEER_SA) to DLOG(ALLCOM_SA)
  - Response of DLOG(NEER_SA) to RGDPR_SA
  - Response of DLOG(NEER_SA) to DLOG(NEER_SA)
  - Response of DLOG(NEER_SA) to DLOG(PPI_SA)
  - Response of DLOG(NEER_SA) to DLOG(CPI_SA)
  - Response of DLOG(PPI_SA) to DLOG(ALLCOM_SA)
  - Response of DLOG(PPI_SA) to RGDPR_SA
  - Response of DLOG(PPI_SA) to DLOG(NEER_SA)
  - Response of DLOG(PPI_SA) to DLOG(PPI_SA)
  - Response of DLOG(PPI_SA) to DLOG(CPI_SA)
  - Response of DLOG(CPI_SA) to DLOG(ALLCOM_SA)
  - Response of DLOG(CPI_SA) to RGDPR_SA
  - Response of DLOG(CPI_SA) to DLOG(NEER_SA)
  - Response of DLOG(CPI_SA) to DLOG(PPI_SA)
  - Response of DLOG(CPI_SA) to DLOG(CPI_SA)
- Response scales and example axis ticks shown in figures:
  - DLOG(ALLCOM_SA) response vertical axis: -.10 -.05 .00 .05 .10
  - RGDPR_SA response vertical axis: -4 -2 0 2 4 6 8 10 (horizontal appears as horizon)
  - DLOG(NEER_SA) response vertical axis: -.02 -.01 .00 .01 .02
  - DLOG(PPI_SA) and DLOG(CPI_SA) response vertical axes: -.04 -.02 .00 .02 .04 .06 (for some panels)
- Note on identification: "Response to Cholesky One S.D. Innovations ± 2 S.E."

### Appendix V — Estimated Cumulative Pass through coefficient
- Graphic axis range and labels:
  - Vertical axis: -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
  - Horizontal axis: 1 2 3 4 5 6 7 8
  - Series label shown: PPICPI
- Caption: "Estimated Cumulative Pass Through Coefficient"
- Source note in figure: "Sources: IMF Staff calculations."

### Appendix VI — Variance Decomposition of Producer and Consumer Price Indexes
- Heading: "Percentage change in forecast variance of PPI attributable to changes in:"
- Table-like lines as presented in source (verbatim numeric strings preserved):
  - All 
  - Commodity 
  - Price Index
  - Output gapNEERPPICPI
  - 15.40.02.692.00.0
  - 212.25.82.176.73.2
  - 324.54.92.165.82.7
  - 430.58.01.954.75.0
  - 537.08.82.147.05.1
  - 639.08.33.643.65.5
  - 740.38.93.741.65.4
  - 848.58.35.333.54.4
- Heading: "Percentage change in forecast variance of CPI attributable to changes in:"
- Table-like lines as presented in source (verbatim numeric strings preserved):
  - All 
  - Commodity 
  - Price Index
  - Output gapNEERPPICPI
  - 13.80.71.459.834.3
  - 25.21.81.953.837.2
  - 35.34.01.853.435.5
  - 48.13.85.649.033.4
  - 515.13.84.945.131.1
  - 620.55.55.640.527.9
  - 727.38.75.733.125.2
  - 828.29.110.128.823.7

*Source: IMF staff (appendices and figures as presented in the source content).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12126.pdf_