## _wp11257

## Source details

**Canonical URL:** [_wp11257](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11257.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11257.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11257.pdf.json)

---

### I. INTRODUCTION — Inflation dynamics in Asian economies
- Managing inflation pressures is traditionally one of the macroeconomic challenges Asian economies have been facing, particularly in recent years.
- Inflation, as measured by consumer price indices, gathered momentum throughout 2007 and accelerated sharply in the first half of 2008.
- Inflationary pressures started receding towards the end of 2008 as a result of the global financial and economic crisis, but they have started to build up in some Asian countries as early as late 2009 (Figure 1).
- Inflation generally accelerated even further towards the end of October, owing mainly to still accommodative monetary conditions and higher oil and food prices.
- Higher global commodity prices often tend to translate directly into higher headline inflation due to high weights these items have on the CPI baskets of a number of Asian economies (Figure 3).
- Core inflation has mostly been on the move as well, reflecting overheating and potentially second-round effects (Figure 2).

Key empirical points and statistics cited
- By the end of 2007, headline inflation in emerging East Asia reached 5.3 percent, double the rate at the start of the year, and 6.9 percent by May 2008.
- The shares of food and energy in the average emerging Asian CPI basket are nearly 40 percent and 10 percent, respectively.
- In India and Indonesia, the CPI shares of food and energy are higher than the Asian average.
- Over the last decade, simple contemporaneous correlations between headline inflation and core inflation, and between core inflation and food and energy prices, have been quite high.

### Data and Methodology
- Two-stage quantitative analysis:
  - Stage 1: Global VAR (GVAR) framework (Pesaran, Schuermann, and Weiner (2004)) estimated for 33 countries using quarterly data from 1986:Q1 to 2010:Q1. Endogenous variables per country include output, consumer and producer price inflation, money supply, the nominal exchange rate, and the short term interest rate. Global oil and food prices treated as exogenous for all countries except China, India, and the United States.
  - Stage 2: Country-level structural VAR (SVAR) estimated for two sub-samples: 1986:Q1–1999:Q4 and 2000:Q1–2010:Q1. Seven variables: GDP, consumer price and producer price inflation, bilateral U.S. dollar exchange rate, real narrow money (or short-term interest rate), a food and oil commodity price index, and foreign (trade-weighted) GDP. Variables taken in first differences for stationarity. Identification via Choleski decomposition with specified Wold orderings for large economies (China, India) and for smaller economies (global demand and commodity prices exogenous).
- GVAR advantages: explicitly accounts for trade and financial linkages, constructs weakly exogenous country-specific foreign variables, uses generalized impulse response functions and forecast error variance decompositions (Koop, Pesaran, and Potter, 1996; Pesaran and Shin, 1998), does not require identification restrictions or variable ordering.

### Main Conclusions (threefold summary)
- Conclusion 1:
  - Over the last two decades, the main driving forces of inflation in Asia have been supply shocks and monetary shocks, while demand pressures have played a relatively smaller role.
  - The region’s role in determining global commodity prices has been rising; among ASEAN economies other than Indonesia, commodity prices play a particularly important role in driving inflation.
  - In higher-income economies (Australia, Hong Kong SAR, Japan, and New Zealand), output gaps tend to be more important.
  - Across the region, while foreign factors sometimes play an important role, most shocks are domestically driven.
- Conclusion 2:
  - The relative roles of key inflation drivers have changed over time:
    - The role of supply shocks in driving inflation appears to have fallen slightly in recent years.
    - The role of output gaps has increased.
    - The impact of monetary shocks on inflation in Asia has diminished, particularly in economies with clear monetary objectives and flexible exchange rate regimes (examples given: Indonesia, Korea, the Philippines, and Thailand).
- Conclusion 3:
  - Demand-driven inflation spillovers from China to the region are both significant and large, directly via higher imported goods prices and indirectly through higher commodity prices.

### Empirical Results — What Drives Inflation in Asia?
- Overview:
  - Supply shocks and monetary shocks account for most of the variation in Asia’s inflation over the last two decades.
- Supply shocks:
  - Supply shocks explain about 45 percent of the inflation fluctuations in Asia.
  - About three-quarters of that supply-shock contribution reflects commodity price shocks.
  - Commodity prices contribute more to inflation in economies with higher oil intensity (barrels of oil consumption divided by GDP in constant U.S. dollars).
  - Contribution of commodity prices to inflation is smaller for high-income commodity exporters (Australia and New Zealand), where they contribute less than 10 percent to fluctuations in inflation.
- Demand and monetary shocks:
  - Demand shocks explain 55 percent of fluctuations of inflation in Asia, of which:
    - Nearly three-quarters reflects the impact of monetary shocks.
    - One-quarter reflects the effect of output gaps.
  - Changes in money supply and interest rates explain about 25 percent of inflation fluctuations.
  - Changes in exchange rates explain about 15 percent of inflation fluctuations (with larger roles in economies experiencing large currency swings, e.g., Indonesia and Korea).
  - Changes in the output gap account for about 15 percent of Asia’s inflation fluctuations.
- Geographic origins:
  - More than 60 percent of inflation fluctuations in Asia have a domestic origin.
  - Global factors account for about 30 percent of inflation in Asia.
  - Regional factors account for slightly less than 10 percent of inflation in Asia; however, accounting for indirect impacts via commodity prices raises the regional contribution to about 20 percent.
  - Regional demand explains about 45 percent of demand-driven changes in world fuel prices and 30 percent of demand-driven fluctuations in food prices.

### Changes Over Time (SVAR results, sub-sample comparison)
- SVAR results broadly consistent with GVAR; inflation variance decomposition estimates between the two methodologies differ by no more than 10 percent.
- Key time-evolution findings:
  - The role of output gaps in driving inflation has grown over time:
    - In emerging Asia, the correlation between core inflation and the output gap rose to 0.7 over the past decade, from 0.2 in the previous two decades.
    - On average in Asia over the last decade, output gaps explained about 20 percent of inflation fluctuations, up from about 5 percent in the previous decade.
  - The contribution of monetary shocks to inflation has diminished over time, especially in economies with relatively clear monetary objectives and flexible exchange rate regimes (Indonesia, Korea, the Philippines, Thailand).

### Structural Stability and Break Tests
- Structural stability tests applied to country-specific error correction models using tests: PK sup, PK msq (Ploberger and Kramer (1992)), NY (Nyblom (1989)), QLR (Quandt likelihood ratio), MW (mean Wald), APW (Andrews and Ploberger (1994)), and heteroskedasticity-robust variants.
- Findings:
  - With PK tests, the null hypothesis of parameter stability rejected at most in 13 out of 157 cases.
  - Robust tests yield rejection rates slightly higher than PK tests; non-robust NY, QLR and APW tests yield a much larger number of rejections.
  - Evidence of some structural instability, largely explained by error variances rather than coefficient breaks.
  - Impulse response and forecast error variance analyses use bootstrapped means and confidence bounds to address changing error variances.

### China Inflation Spillovers — Direct and Indirect Channels
- Channels:
  - Direct: trade linkages and imported goods prices.
  - Indirect: China’s demand influences commodity prices, which then affect other Asian economies.
- Quantitative contributions to commodity price demand-driven shocks:
  - The region’s demand explains about 45 percent of demand-driven shocks to oil and about 30 percent of demand-driven changes in food prices; about one-third of these effects come from China.
- Generalized impulse response function (GIRF) results for a positive 1 percent shock to China’s GDP (shock ~ 3 standard errors; causes 1 percent increase in GDP on impact):
  - China’s inflation response:
    - Increase by about 0.5 percent on impact.
    - Increase by 0.7 percent on average per quarter in the first year.
    - Increase by 0.15 percent on average per quarter in the second year.
  - Regional inflation response:
    - On impact, region’s inflation increases by 0.1 percent.
    - Effects become more pronounced over the first two years, averaging about 0.2 percent increase in every quarter.
    - Overall cumulative impact on Asia’s inflation after two years amounts to 1.3 percent (slightly lower than China’s cumulative impact of 1.5 percent).
  - Transmission is relatively slow, reflecting high competitiveness in product markets that limits immediate pass-through of input price changes to consumer prices.
- Additional summary statistic:
  - 2.5 percent on average within two years.

### Generalized Impulse Response Findings (commodity-price shocks)
- A positive 1 percent shock to China’s GDP: bootstrap mean estimates with 90 percent bootstrap error bounds (Figures 13).
- A positive 1 percent shock to commodity prices: bootstrap mean estimates with 90 percent bootstrap error bounds (Figure 14).
- Commodity prices increase by about 0.7 percent on impact after the shock.
- In the first and the second year, the shock brings 4 and 3.5 percent increase in commodity prices in cumulative terms.
- A positive 1 percent shock to commodity prices, on average, brings about 0.05 percent increase in inflation in the region on impact.
- The same commodity-price shock yields a cumulative of 0.1 percent increase in inflation after a year, and 0.15 percent increase after two years.

### Regional Heterogeneity in Inflation Spillovers from China
- ASEAN countries and India:
  - Inflation increases 0.2 and 0.3 percent every quarter in the first and the second year after the shock (ASEAN: 0.2 percent; India: 0.3 percent per quarter).
  - Inflation rises by about 2 percent in ASEAN and 1.5 percent in India after two years in cumulative terms.
- Developed economies in the region (Japan, Australia, New Zealand):
  - Spillovers are more limited with only 0.05 percent average increase in inflation per quarter within two years.
- NIEs:
  - Somewhat shielded with an average of 0.1 percent increase in inflation in each quarter.
  - About a 1 percent cumulative increase in inflation after two years.
- ASEAN (commodity-importing economies except Malaysia): impact of commodity-price shock on inflation is about 0.25 percent on average.
- Advanced economies in Asia (Japan, Australia, New Zealand): impact of commodity-price shock on inflation is about 0.01 percent increase per quarter.

### Commodity Price Channel and Timing
- The demand shock in China has a statistically significant impact on commodity prices (0.7 percent on impact).
- Indirect effects via commodity prices amplify inflation spillovers to other Asian economies.
- Indirect impacts transmit faster than direct impacts:
  - In the first year after the shock, about 35 percent of the spillovers are caused by movements in commodity prices.
- Within two years, about 20 percent of the spillovers from an inflationary demand shock in China are due to higher commodity prices.
- Given China’s increasing impact on world commodity prices, indirect spillovers are likely to amplify going forward.

### Direct and Indirect Spillovers (aggregate findings)
- The paper calculates direct and indirect spillovers by subtracting the indirect effect brought by the rise in commodity prices from the total impact.
- Average change per quarter within two years (in percent) is shown for economies including Hong Kong, New Zealand, Australia, Japan, Korea, Singapore, Thailand, Philippines, India, Malaysia, Indonesia.
- Direct and indirect components are presented separately; overall conclusion: notable exposure of regional inflation to both direct import-price effects and indirect commodity-price-induced effects.

### Policy Implications and Concluding Remarks
- Inflation dynamics across Asia, including China and India, are mainly driven by domestic supply shocks; however, the contribution of demand factors has risen in recent years.
- If the influence of demand factors on inflation continues to grow, policymakers will need to give increasing priority to managing inflation relative to promoting growth.
- The contribution of monetary shocks to inflation has diminished over time, perhaps reflecting improvements in monetary frameworks in many countries (greater clarity and transparency with respect to monetary objectives and instruments; greater exchange rate flexibility).
- Additional moves toward clearer, more transparent monetary frameworks and greater exchange rate flexibility may help reduce the level and volatility of inflation across the region.
- Policymakers need to pay increasing attention to:
  - The influence of global commodity prices on domestic prices and domestic economic conditions.
  - The implications of domestic conditions for global commodity prices—especially important for China as the largest commodity importer in the region.
- An inflationary shock in China has a significant and large impact on commodity prices, which quickly feeds into inflation of other Asian economies; combined with direct effects through changes in import prices, economies in the region are exposed to notable inflation spillovers from China.

*Source: IMF working paper content (GVAR and SVAR analysis of inflation dynamics in Asia, 1986:Q1–2010:Q1).*

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

### _wp11257 - References .............................................................................................................

### I. INTRODUCTION — Inflation dynamics in Asian economies
- Managing inflation pressures is traditionally one of the macroeconomic challenges Asian economies have been facing, particularly in recent years.
- Inflation, as measured by consumer price indices, gathered momentum throughout 2007 and accelerated sharply in the first half of 2008.
- Inflationary pressures started receding towards the end of 2008 as a result of the global financial and economic crisis, but they have started to build up in some Asian countries as early as late 2009 (Figure 1).
- Inflation generally accelerated even further towards the end of October, owing mainly to still accommodative monetary conditions and higher oil and food prices.
- Higher global commodity prices often tend to translate directly into higher headline inflation due to high weights these items have on the CPI baskets of a number of Asian economies (Figure 3).
- Core inflation has mostly been on the move as well, reflecting overheating and potentially second-round effects (Figure 2).

### Key empirical points and statistics cited
- By the end of 2007, headline inflation in emerging East Asia reached 5.3 percent, double the rate at the start of the year, and 6.9 percent by May 2008.
- The shares of food and energy in the average emerging Asian CPI basket are nearly 40 percent and 10 percent, respectively, both of which are higher than the average for emerging economies worldwide.
- In India and Indonesia, the CPI shares of food and energy are higher than the Asian average.
- Over the last decade, simple contemporaneous correlations between headline inflation and core inflation, on the one hand, and between core inflation and food and energy prices on the other hand, have been quite high.

### References to figures and analysis in the source
- Figure 1: timing of inflationary pressures (late 2009 build-up in some countries).
- Figure 2: movements in core inflation and evidence of overheating and second-round effects.
- Figure 3: shares of food and energy in CPI baskets across Asian economies and implication for headline inflation sensitivity to commodity prices.

*Source: _wp11257 - References .............................................................................................................*

### 0.8 and 0.4, respectively). This suggests that changes in food and energy prices feed through quickly to core

### Inflation Dynamics in Asia: Supply, Demand, and Spillovers (GVAR and SVAR Analysis)

### Data and Methodology
- Two-stage quantitative analysis:
  - Stage 1: Global VAR (GVAR) framework (Pesaran, Schuermann, and Weiner (2004)) estimated for 33 countries using quarterly data from 1986:Q1 to 2010:Q1. Endogenous variables per country include output, consumer and producer price inflation, money supply, the nominal exchange rate, and the short term interest rate. Global oil and food prices treated as exogenous for all countries except China, India, and the United States.
  - Stage 2: Country-level structural VAR (SVAR) estimated for two sub-samples: 1986:Q1–1999:Q4 and 2000:Q1–2010:Q1. Seven variables: GDP, consumer price and producer price inflation, bilateral U.S. dollar exchange rate, real narrow money (or short-term interest rate), a food and oil commodity price index, and foreign (trade-weighted) GDP. Variables taken in first differences for stationarity. Identification via Choleski decomposition with specified Wold orderings for large economies (China, India) and for smaller economies (global demand and commodity prices exogenous).
- GVAR advantages highlighted: explicitly accounts for trade and financial linkages, constructs weakly exogenous country-specific foreign variables, uses generalized impulse response functions and forecast error variance decompositions (Koop, Pesaran, and Potter, 1996; Pesaran and Shin, 1998), does not require identification restrictions or variable ordering.

### Main Conclusions (threefold summary from the paper)
- Conclusion 1:
  - Over the last two decades, the main driving forces of inflation in Asia have been supply shocks and monetary shocks, while demand pressures have played a relatively smaller role.
  - The region’s role in determining global commodity prices has been rising; among ASEAN economies other than Indonesia, commodity prices play a particularly important role in driving inflation.
  - In higher-income economies (Australia, Hong Kong SAR, Japan, and New Zealand), output gaps tend to be more important.
  - Across the region, while foreign factors sometimes play an important role, most shocks are domestically driven.
- Conclusion 2:
  - The relative roles of key inflation drivers have changed over time:
    - The role of supply shocks in driving inflation appears to have fallen slightly in recent years.
    - The role of output gaps has increased.
    - The impact of monetary shocks on inflation in Asia has diminished, particularly in economies with clear monetary objectives and flexible exchange rate regimes (examples given: Indonesia, Korea, the Philippines, and Thailand).
- Conclusion 3:
  - Demand-driven inflation spillovers from China to the region are both significant and large, directly via higher imported goods prices and indirectly through higher commodity prices.

### Empirical Results — What Drives Inflation in Asia?
- Overview:
  - Supply shocks and monetary shocks account for most of the variation in Asia’s inflation over the last two decades.
- Supply shocks:
  - Supply shocks explain about 45 percent of the inflation fluctuations in Asia.
  - About three-quarters of that supply-shock contribution reflects commodity price shocks.
  - Commodity prices contribute more to inflation in economies with higher oil intensity (barrels of oil consumption divided by GDP in constant U.S. dollars).
  - Contribution of commodity prices to inflation is smaller for high-income commodity exporters (Australia and New Zealand), where they contribute less than 10 percent to fluctuations in inflation.
- Demand and monetary shocks:
  - Demand shocks explain 55 percent of fluctuations of inflation in Asia, of which:
    - Nearly three-quarters reflects the impact of monetary shocks.
    - One-quarter reflects the effect of output gaps.
  - Changes in money supply and interest rates explain about 25 percent of inflation fluctuations.
  - Changes in exchange rates explain about 15 percent of inflation fluctuations (with larger roles in economies experiencing large currency swings, e.g., Indonesia and Korea).
  - Changes in the output gap account for about 15 percent of Asia’s inflation fluctuations.
- Geographic origins:
  - More than 60 percent of inflation fluctuations in Asia have a domestic origin.
  - Global factors account for about 30 percent of inflation in Asia.
  - Regional factors account for slightly less than 10 percent of inflation in Asia; however, accounting for indirect impacts via commodity prices raises the regional contribution to about 20 percent.
  - Regional demand explains about 45 percent of demand-driven changes in world fuel prices and 30 percent of demand-driven fluctuations in food prices.

### Changes Over Time (SVAR results, sub-sample comparison)
- The SVAR results broadly consistent with GVAR; inflation variance decomposition estimates between the two methodologies differ by no more than 10 percent.
- Key time-evolution findings:
  - The role of output gaps in driving inflation has grown over time:
    - In emerging Asia, the correlation between core inflation and the output gap rose to 0.7 over the past decade, from 0.2 in the previous two decades.
    - On average in Asia over the last decade, output gaps explained about 20 percent of inflation fluctuations, up from about 5 percent in the previous decade.
  - The contribution of monetary shocks to inflation has diminished over time, especially in economies with relatively clear monetary objectives and flexible exchange rate regimes (Indonesia, Korea, the Philippines, Thailand).

### Structural Stability and Break Tests
- Structural stability tests applied to country-specific error correction models using tests: PK sup, PK msq (Ploberger and Kramer (1992)), NY (Nyblom (1989)), QLR (Quandt likelihood ratio), MW (mean Wald), APW (Andrews and Ploberger (1994)), and heteroskedasticity-robust variants.
- Findings:
  - With PK tests, the null hypothesis of parameter stability rejected at most in 13 out of 157 cases.
  - Robust tests yield rejection rates slightly higher than PK tests; non-robust NY, QLR and APW tests yield a much larger number of rejections.
  - Evidence of some structural instability, largely explained by error variances rather than coefficient breaks.
  - Impulse response and forecast error variance analyses use bootstrapped means and confidence bounds to address changing error variances.

### China Inflation Spillovers — Direct and Indirect Channels
- Channels:
  - Direct: trade linkages and imported goods prices.
  - Indirect: China’s demand influences commodity prices, which then affect other Asian economies.
- Quantitative contributions to commodity price demand-driven shocks:
  - The region’s demand explains about 45 percent of demand-driven shocks to oil and about 30 percent of demand-driven changes in food prices; about one-third of these effects come from China.
- Generalized impulse response function (GIRF) results for a positive 1 percent shock to China’s GDP (shock ~ 3 standard errors; causes 1 percent increase in GDP on impact):
  - China’s inflation response:
    - Increase by about 0.5 percent on impact.
    - Increase by 0.7 percent on average per quarter in the first year.
    - Increase by 0.15 percent on average per quarter in the second year.
  - Regional inflation response:
    - On impact, region’s inflation increases by 0.1 percent.
    - Effects become more pronounced over the first two years, averaging about 0.2 percent increase in every quarter.
    - Overall cumulative impact on Asia’s inflation after two years amounts to 1.3 percent (slightly lower than China’s cumulative impact of 1.5 percent).
  - Transmission is relatively slow, reflecting high competitiveness in product markets that limits immediate pass-through of input price changes to consumer prices.

*Source: IMF working paper content (GVAR and SVAR analysis of inflation dynamics in Asia, 1986:Q1–2010:Q1).*

### 2.5 percent on average within two years.

### _wp11257 - 2.5 percent on average within two years.

### Generalized Impulse Response Findings
- A positive 1 percent shock to China’s GDP: bootstrap mean estimates with 90 percent bootstrap error bounds (Figures 13).
- A positive 1 percent shock to commodity prices: bootstrap mean estimates with 90 percent bootstrap error bounds (Figure 14).
- Commodity prices increase by about 0.7 percent on impact after the shock.
- In the first and the second year, the shock brings 4 and 3.5 percent increase in commodity prices in cumulative terms.
- A positive 1 percent shock to commodity prices, on average, brings about 0.05 percent increase in inflation in the region on impact.
- The same commodity-price shock yields a cumulative of 0.1 percent increase in inflation after a year, and 0.15 percent increase after two years.

### Regional Heterogeneity in Inflation Spillovers from China
- ASEAN countries and India:
  - Inflation increases 0.2 and 0.3 percent every quarter in the first and the second year after the shock (ASEAN: 0.2 percent; India: 0.3 percent per quarter).
  - Inflation rises by about 2 percent in ASEAN and 1.5 percent in India after two years in cumulative terms.
- Developed economies in the region (Japan, Australia, New Zealand):
  - Spillovers are more limited with only 0.05 percent average increase in inflation per quarter within two years.
- NIEs:
  - Somewhat shielded with an average of 0.1 percent increase in inflation in each quarter.
  - About a 1 percent cumulative increase in inflation after two years.
- ASEAN (commodity-importing economies except Malaysia): impact of commodity-price shock on inflation is about 0.25 percent on average.
- Advanced economies in Asia (Japan, Australia, New Zealand): impact of commodity-price shock on inflation is about 0.01 percent increase per quarter.

### Commodity Price Channel and Timing
- The demand shock in China has a statistically significant impact on commodity prices (0.7 percent on impact).
- Indirect effects via commodity prices amplify inflation spillovers to other Asian economies.
- Indirect impacts transmit faster than direct impacts:
  - In the first year after the shock, about 35 percent of the spillovers are caused by movements in commodity prices.
- Within two years, about 20 percent of the spillovers from an inflationary demand shock in China are due to higher commodity prices.
- Given China’s increasing impact on world commodity prices, indirect spillovers are likely to amplify going forward.

### Direct and Indirect Spillovers (Figure 15)
- The paper calculates direct and indirect spillovers by subtracting the indirect effect brought by the rise in commodity prices from the total impact.
- Average change per quarter within two years (in percent) is shown for economies including Hong Kong, New Zealand, Australia, Japan, Korea, Singapore, Thailand, Philippines, India, Malaysia, Indonesia.
- Direct and indirect components are presented separately; overall conclusion: notable exposure of regional inflation to both direct import-price effects and indirect commodity-price-induced effects.

### Policy Implications and Concluding Remarks
- Inflation dynamics across Asia, including China and India, are mainly driven by domestic supply shocks; however, the contribution of demand factors has risen in recent years.
- If the influence of demand factors on inflation continues to grow, policymakers will need to give increasing priority to managing inflation relative to promoting growth.
- The contribution of monetary shocks to inflation has diminished over time, perhaps reflecting improvements in monetary frameworks in many countries (greater clarity and transparency with respect to monetary objectives and instruments; greater exchange rate flexibility).
- Additional moves toward clearer, more transparent monetary frameworks and greater exchange rate flexibility may help reduce the level and volatility of inflation across the region.
- Policymakers need to pay increasing attention to:
  - The influence of global commodity prices on domestic prices and domestic economic conditions.
  - The implications of domestic conditions for global commodity prices—especially important for China as the largest commodity importer in the region.
- An inflationary shock in China has a significant and large impact on commodity prices, which quickly feeds into inflation of other Asian economies; combined with direct effects through changes in import prices, economies in the region are exposed to notable inflation spillovers from China.

*Source: Authors’ calculations; content as presented in the source PDF.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11257.pdf_