## _wp0964

## Source details

**Canonical URL:** [_wp0964](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0964.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0964.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0964.pdf.json)

---

### I. Introduction — context and research focus
- During the 1990s, large trade imbalances developed: the United States ran persistent deficits, while Japan, the euro area first, and later emerging Asia and oil-exporting countries ran surpluses.
- By the 2000s the United States absorbed an overwhelming share of the world’s current account surpluses; net U.S. liabilities were about a fifth of U.S. GDP.
- Debate on sources of global imbalances is polarized:
  - Some view imbalances as largely “equilibrium” phenomena generated by growth and financial development differentials.
  - Others view imbalances as originating in economic distortions requiring policy adjustments.
- Paper focus:
  - Investigates emerging Asia’s role in emergence and evolution of global trade imbalances.
  - Considers two exogenous factors in emerging Asia: (1) a regional productivity slowdown in the non-tradable sector in the second half of the 1990s; (2) a desired net foreign asset (NFA) increase in the first half of the 2000s (reflecting official reserve accumulation).
  - Assesses ability of these factors to explain U.S. and emerging Asia trade balance evolution from 1995 to 2007 and broader macroeconomic dynamics.
- Key interpretation points:
  - Financial linkages between emerging Asia and the United States are more important than real linkages in the model.
  - Emerging Asia’s role in the emergence of global imbalances in the mid-late 1990s is more limited than often assumed.
  - Both an investment “drought” and a saving “glut” characterize emerging Asia’s contribution to global imbalances depending on the time period considered.

### II. Methodology and model overview
- Modeling approach:
  - Feed exogenous paths for productivity and desired NFA in emerging Asia into a multi-region DGE model and compare simulated and actual paths for trade balance, real effective exchange rate, terms of trade, real interest rate, and national accounts.
- Model features:
  - Flexible-price, two-sector (tradable/non-tradable), five-region DGE model with incomplete international asset markets, home bias in consumption and investment, distribution sector price discrimination, capital accumulation, and non-zero desired NFA in steady state.
  - Five regions: emerging Asia (AS: China, India, Hong Kong, Malaysia, the Philippines, Singapore, South Korea, Taiwan, Thailand), United States (US), Japan (JA), euro area (EA, defined as EU-15), and rest of the world (ROW).
  - Two household types: forward-looking (Ricardian) and liquidity-constrained (share = 0.10 in all regions).
  - Single internationally traded asset: U.S.-denominated nominal bond (net and gross foreign assets identical).
  - Perfect foresight solution (DYNARE “simul”).
- Data sources and aggregation:
  - Most series from the World Economic Outlook database of the IMF.
  - Real effective exchange rate series from the BIS.
  - NFA data from an updated Lane and Milesi-Ferretti (2007) dataset; official reserves from the IFS database of the IMF.
  - Emerging Asia aggregates: weighted averages (current dollar GDP weights used for exchange rates and interest rates when available; PPP-valued GDP weights for other variables).
- Time coverage and normalizations:
  - Productivity quarterly path covers 1995 to 2004; world economy assumed in steady state in 1994:Q4.
  - Normalizations: trade balance and NFA as share of GDP (deviation from steady state); terms of trade and real effective exchange rate deviations with steady state = 100 in 1994:Q1; real interest rate deviation with steady state = actual short rate in 1994:Q4; real GDP growth reported as deviation from sample mean over 1994-2005.
- Persistence parameter used in some processes: 0.001 per quarter (i.e., with an autoregressive coefficient of 0.999).

### III. Scenarios simulated
- Scenario 1 — Non-tradable productivity slowdown (1995–99 specification):
  - Negative, temporary but persistent productivity shock to non-tradable TFP in the emerging Asia block.
  - Rationale: evidence of a marked productivity slowdown in the non-tradable (service) sector during 1995–99 with rebound in 2000–04 (Table 1 evidence).
  - Transmission highlights:
    - Decline in non-tradable productivity → persistent increase in relative price of non-tradables → substitution and complementarity effects reduce demand for non-tradables and for tradables, and foreign bias implies imports fall more → terms of trade improve and trade balance moves into surplus.
    - Three forces determine net trade-balance effect: substitution effects (push toward surplus), consumption smoothing (opposes surplus), and general equilibrium interactions.
- Scenario 2 — Desired NFA increase (official reserve accumulation, 2001–06 specification):
  - Represents large increase in official reserves in emerging Asia after 2000.
  - Modeling choice: regional steady-state (desired) NFA position grows as a percent of GDP between 2001:Q1 and 2006:Q4 at a constant quarterly rate so that cumulative desired NFA change equals cumulative change in official reserves over same period.
  - For simplicity the change is specified to be persistent but ultimately temporary, reverting to zero with autoregressive coefficient of 0.99 (constructed NFA path reported in Figure 2, Panel B).
  - Correlation between regional NFA series and official reserve series over 2000–2006 is 0.96.

### IV. Key simulation findings (effects, fit, and limitations)
- Productivity slowdown scenario (non-tradable TFP decline):
  - Fits regional macroeconomic developments around the 1997-98 financial crisis: matches actual GDP growth in emerging Asia from 1996 to 2002 and captures composition of growth (sharp increase in foreign demand, investment collapse, consumption decline).
  - Predicts trade balance surplus via dominance of substitution effects and investment postponement.
  - Shortcomings relative to data:
    - Predicts investment and consumption patterns too volatile and too smooth, respectively; investment recovers faster than in the data and reduces the trade surplus counterfactually.
    - Misses a large and persistent terms of trade deterioration in the data (model predicts improvement).
    - Misses the sharp devaluation during the crisis (predicts real exchange rate behavior correctly for 1994–1997 and 1998–2004, but not the crisis spike).
    - Produces very limited spillovers to the United States (reproduces the 1998 deterioration of the U.S. trade balance but little broader macro impact).
    - Predicts a small and counterfactual real interest rate fluctuation (initial fall then rise), opposite to observed falling rates.
- Desired NFA increase scenario (reserve accumulation):
  - By construction matches the NFA path and fits the trade balance path relatively well between 2000 and 2004 (though generates smaller surpluses than data).
  - Captures a smooth and persistent terms of trade decline and trend depreciation in the real exchange rate from 1999 to 2004.
  - Can generate a persistent decline in investment (“investment drought”).
  - Shortcomings relative to data:
    - Misses falling real interest rates observed in the data; model produces an excessively smooth real interest rate response.
    - Misses real consumption and GDP growth dynamics (these remain essentially constant in the model).
    - Produces more sizable spillovers to the United States than the productivity scenario: simulated U.S. trade balance, NFA position, terms of trade, and investment qualitatively consistent with data, with cumulative U.S. NFA position reaching almost 20 percent of GDP by 2004 in simulations, but no effects on U.S. real interest rate, consumption, or GDP growth in the model.
- Combined insight:
  - Neither factor alone generates the strong downward pressure on U.S. interest rates observed in the first half of the 2000s within the frictionless model.
  - Both an “investment drought” (productivity-driven) and a “saving glut” (desired NFA increase/reserve accumulation) are necessary to match different aspects of the data.
  - Financial linkages (and realistic financial frictions and policy distortions) likely amplify spillovers; the frictionless model understates these effects.

### V. Model structure, calibration specifics, and exact parameters
- Households and firms parameters:
  - Rate of time preference: .997
  - Depreciation rate: 0.02
  - Intertemporal elasticity of substitution: 1.00
  - Inverse of Frisch elasticity of labor: 1.50
  - Share of liquidity-constrained consumers: 0.10
- Production and final goods parameters:
  - Tradable intermediate goods substitution between factors of production: 0.75; bias towards capital: 0.60
  - Nontradable intermediate goods substitution between factors of production: 0.75; bias towards capital: 0.50
  - Final consumption goods substitution between domestic and imported goods: 4
    - Bias towards domestic goods: US 0.87; JA 0.49; AS 0.15; EA 0.04; ROW 0.20
  - Substitution between tradables and nontradables (final consumption): 0.50
    - Bias towards tradable goods: US 0.35; JA 0.36; AS 0.37; EA 0.36; ROW 0.30
  - Final investment goods substitution between domestic and imported goods: 4
    - Bias towards domestic goods (investment): US 0.87; JA 0.63; AS 0.06; EA 0.92; ROW 0.14
    - Substitution between tradables and nontradables (investment): 0.50
    - Bias towards tradable goods (investment): US 0.75; JA 0.77; AS 0.82; EA 0.75; ROW 0.73
- Real rigidities:
  - Capital accumulation adjustment cost parameter: 0.01 (all regions)
  - Distribution costs: 0.30 (all regions)
- Price and wage markups:
  - Tradables price markups: 1.18 (all regions)
  - Nontradables price markups: 1.35 (all regions)
  - Wage markup: 1.20 (all regions)
- International linkages (selected entries preserved exactly where available):
  - Substitution between consumption imports: 5 (all regions)
  - Substitution between investment imports: 5 (all regions)
  - Net foreign liabilities intermediation cost parameters:
    - Maximum of financial intermediation cost function: 0.05 (all regions)
    - Steepness of financial intermediation cost function: 0.1 (all regions)
  - Selected bias parameters (consumption/imports and investment/imports) reported in calibration tables (full matrices in Appendix Tables A4–A5).
- Steady-state national accounts (percents of GDP; columns US JA AS EA ROW):
  - Private Consumption: 68.9, 59.64, 69.33, 58.99, 67.83
  - Forward-looking consumers: 62.64, 52.56, 54.71, 54.73, 62.00
  - Liquidity-constrained consumers: 6.31, 7.08, 14.60, 4.25, 5.83
  - Private Investment: 15.94, 21.08, 19.34, 15.32, 15.73
  - Public Expenditure: 14.51, 19.78, 12.1, 25.5, 16.8
  - Trade balance: 0.60, -0.41, -0.64, 0.18, -0.37
  - Imports: 11.46, 11.49, 26.73, 17.41, 22.98
    - Consumption Goods (imports): 7.33, 8.06, 11.32, 15.01, 12.13
    - Investment Goods (imports): 4.13, 3.43, 15.40, 2.40, 10.85
  - Government Debt: 61.5, 80.0, 55.0, 60.0, 60.0
  - Net Foreign Assets: -51.08, 55.03, 49.02, -11.3, 26.4
  - Share of World GDP (percent): 30.05, 11.48, 9.83, 22.80, 25.84

### VI. Policy implications and recommended research directions
- In a frictionless DGE framework, relative-price and demand-composition mechanisms alone are insufficient to replicate several observed data features, notably falling real interest rates and consumption/growth dynamics in emerging Asia and the United States.
- To better understand the observed dynamics—reserve accumulation, investment drought, saving glut, and U.S. external position—models should incorporate realistic financial frictions and policy distortions, including:
  - Reserve currency role of the U.S. dollar.
  - Sterilized foreign exchange intervention.
  - Other financial linkages that can amplify spillovers between emerging Asia and the United States.
- Introducing such features is a promising avenue for future research to understand the macroeconomic consequences of the increase in official reserves in emerging Asia and their global spillovers.

*Italic: Content summarized from _wp0964 - References.................................................................................................39*

### References.................................................................................................39

### _wp0964 - References.................................................................................................39

### I. INTRODUCTION
- During the 1990s, large trade imbalances developed in different regions of the world, with the United States running persistent deficits, and Japan and the euro areafirst, and later emerging Asia and oil-exporting countries, running surpluses (Figure 1).
- Today, the United States absorbs an overwhelming share of the world’s current account surpluses, and net U.S. liabilities remain close to record-highs, representing about afifth of U.S. GDP.
- Debate on sources and resolution of external imbalances is polarized:
  - Some view global imbalances as largely “equilibrium” phenomena generated by growth and financial development differentials (examples cited: Engel and Rogers (2006), Blanchard (2006), Caballero, Fahri, and Gourinchas (2008), Mendoza, Rull, and Quadrini (2007), Fogli and Perri (2006), McGrattan and Prescott (2007)).
  - Others view imbalances as originating in economic distortions requiring policy adjustments (examples cited: IMF (2005 and 2006), Blanchard, Giavazzi and Sa (2005), Mussa (2004), Obstfeld and Rogoff (2007), Roubini and Setser (2004), Yoshitomi (2007)).
- Relative productivity differences across world regions likely played a non-negligible role in trade imbalances:
  - Glick and Rogoff (1995) estimate that a one percent increase in country-specific productivity decreases the current account balance by 0.15 percent of GDP.
  - Other studies (Bems, Dedola and Smets (2007); Edwards (2007); Corsetti, Dedola, and Leduc (2006)) find even larger elasticities between productivity shocks and imbalances.
  - Productivty differences can explain U.S. deficit only if concentrated either in the U.S. non-tradable sector or the tradable sector of the rest of the world.
- The sharp shift into surplus of the emerging Asia trade balance during the 1997-98 financial crisis is not fully accounted for by TFP differences among advanced economies.
  - Hypotheses: Bernanke (2005) “saving glut” vs. Rajan (2006) “investment draught”.
  - Strong global monetary liquidity creation in first half of 2000s weakened link between short and long-term interest rates, particularly in the United States.
- Paper focus:
  - Investigates emerging Asia’s role in emergence and evolution of global trade imbalances.
  - Two potential explanatory factors considered as given (not explained): a regional productivity slowdown in non-tradable sector in second half of 1990s, and a desired net foreign asset (NFA) increase in first half of 2000s.
  - Assess ability of these factors to explain U.S. and emerging Asia trade balance evolution from 1995 to 2007 and broader macroeconomic dynamics.
- Rationale for factors:
  - New productivity data show a marked slowdown in second half of 1990s in the service sector of economies at center-stage of 1997-98 crisis.
  - Spectacular increase in official reserves in emerging Asia in first half of 2000s may have contributed directly (trade surpluses to accumulate reserves) and indirectly (pressure on global capital markets via Asian demand for foreign assets and associated monetary liquidity creation).
- Model and approach:
  - Feed a non-tradable productivity growth decline and a desired NFA increase to the model used by Cova et al (2008) and compare model simulations with data for the United States and an aggregate of emerging Asian countries.
  - Model is a five-region dynamic general equilibrium (DGE) model: United States, Japan, euro area (EU-15), rest of the world, and emerging Asia (China, India, Hong Kong, Malaysia, the Philippines, Singapore, South Korea, Taiwan, and Thailand).
  - Model limitations: does not incorporate realistic financial features like reserve-currency status of the U.S. dollar, or policy distortions such as sustained sterilized foreign exchange intervention; does not treat oil-exporting economies separately.
- Main simulation results (summary):
  - A non-tradable productivity slowdown in the second half of the 1990s and an increase in desired NFA in the first half of the 2000s can partially account for regional macroeconomic dynamics in emerging Asia and some of the U.S. trade balance deficits during this period.
  - The productivity slowdown fits Asian regional macroeconomic developments around the 1997-98 financial crisis but has limited spillover to the United States.
  - The desired NFA increase after 2001 can partially fit regional macroeconomic developments and has significant spillover effects onto the United States.
  - Neither factor alone is capable of generating strong downward pressure on U.S. interest rates in the first half of the 2000s in the model.
- Interpretation and implications:
  - Financial linkages between emerging Asia and the United States are more important than real linkages.
  - Emerging Asia’s role in the emergence of global imbalances in the mid-late 1990s is more limited than often assumed.
  - Both an investment “drought” and a saving “glut” characterize emerging Asia’s contribution to global imbalances depending on the time period considered.
  - Results cast doubt on the notion that “home grown” growth in emerging Asia can have significant effects on the rest of the world, particularly the United States.

### II. METHODOLOGY
- Approach:
  - Feed to the model exogenous paths for productivity and desired NFA in emerging Asia.
  - Compare actual and simulated paths for the trade balance, the real effective exchange rate, the terms of trade, the real interest rate, and the national accounts.
- Model description:
  - Uses a flexible price version of the Bank of Italy model by Cova and Pisani used by Cova et al (2008).
  - Multi-country, two-sector (tradable and non-tradable) DGE model with incomplete international asset markets, home bias in consumption and investment, international price discrimination (distribution sector), capital accumulation, and non-zero net foreign asset positions in steady state (desired NFA).
  - Five regions: emerging Asia (AS), the United States (US), Japan (JA), the euro area (EA, defined here as EU-15), and rest of the world (ROW). Emerging Asia comprises China, India, Hong Kong, Malaysia, the Philippines, Singapore, South Korea, Taiwan, and Thailand.
- Scenarios:
  - Scenario 1: non-tradable productivity slowdown in emerging Asia.
    - Table 1 reports 5-year averages of annual TFP and labor productivity growth for the aggregate of emerging Asian countries, and for China, India, and group excluding China and India.
    - Labor productivity data are by sector: “Industry” = tradable sector, “Service” = non-tradable sector. TFP data are economy-wide.
    - Evidence (Table 1) indicates a marked productivity slowdown in the non-tradable sector during 1995–99 compared to historical average annual growth, with a strong rebound during 2000-04.
    - Jaumotte and Spatafora (2006) interpret evidence in terms of convergence and catch up to income levels of advanced economies; note that productivity tends to decline over time after growth take-offs in many countries of the region, especially in services and particularly in China.
    - Non-tradable productivity started to decelerate in 1995-1996, before the 1997-98 financial crisis, consistent with the convergence interpretation; strong rebound during 2000-04 suggests slowdown was temporary.
    - Economy-wide TFP decelerated over same period, suggesting non-tradable labor productivity slowdown was not simply driven by investment boom-bust around the financial crisis.
    - Region-wide tradable productivity growth remained close to historical average throughout the 1990s, accelerating slightly in the first half of the 2000s.
  - Scenario calibration choices:
    - The non-tradable productivity slowdown is driven by countries directly affected by 1997-98 financial crises.
    - In China both tradable and non-tradable labor productivity are slightly above average during 1995-2004, with tradable productivity slowing somewhat in second half of 1990s, and aggregate TFP close to historical 1980-1994 average.
    - In India during 1995-2004, non-tradable (service) labor productivity accelerates strongly while tradable (industry) productivity falls sharply and aggregate TFP hovers around 1980-1994 average. Classification issues noted: many services in India (financial and IT) are tradable, making the service/tradable classification problematic.
    - For simulations of productivity slowdown, feed to model the simple average of labor productivity growth only in the non-tradable sector (the second column of Table 1).
    - Rationale: simple average of all countries excluding China and India would give too much weight to smaller “crisis” countries; a weighted average would give too little weight to slowdown in smaller crisis countries; excluding India from a weighted average would give China disproportionate weight.
    - Construction of non-tradable productivity level path follows Cova et al (2008): assume world economy in steady state in 1994:Q4 and that after 2004:Q4 non-tradable productivity reverts to trend at the rate of [text truncated in source].

### III. MODEL, DATA, AND CALIBRATION (AS NOTED)
- Model features reiterated:
  - Two-sector (tradable/non-tradable), five-region DGE with incomplete markets and desired NFA in steady state.
- Data notes:
  - Table 1 used to calibrate non-tradable productivity path is based on 5-year averages of annual TFP and labor productivity growth for the aggregate emerging Asia group and selected members.
  - Labor productivity by sector: Industry = tradable; Service = non-tradable.
- Calibration and solution:
  - The appendix (not fully reproduced here) contains details: The Model (I), Firms (A), Households (B), Government (C), Market Clearing and the Current Account (D), Measuring Output and Current Account (E), The Equilibrium (F), The Calibration (G), Solution (H), and Appendix Tables A1–A5 (steady-state national accounts; households and firms behavior; real rigidities; price and wage markups; international linkages).

### IV. KEY FINDINGS AND INTERPRETATIONS
- Findings from simulations:
  - A non-tradable productivity slowdown in emerging Asia (second half of the 1990s) fits regional macroeconomic developments around the 1997-98 financial crisis but produces limited spillovers to the United States.
  - A desired NFA increase in emerging Asia after 2001 (first half of the 2000s) can partially account for regional macroeconomic developments and has significant spillovers to the United States.
  - Neither shock alone generates strong downward pressure on U.S. interest rates in the first half of the 2000s within the model.
- Broader interpretations:
  - Financial linkages matter more than real linkages for transmission from emerging Asia to the United States in this framework.
  - Emerging Asia’s contribution to global imbalances was heterogeneous over time: characterized by both investment “drought” and saving “glut” depending on period.
  - Evidence casts doubt on the view that emerging Asia “home grown” growth has large effects on the rest of the world, particularly the United States.

*Italic: Content summarized from _wp0964 - References.................................................................................................39*

### 0.001 per quarter (i.e., with an autoregressive coefficient of 0.999). Further, for simplicity and

### _wp0964 - 0.001 per quarter (i.e., with an autoregressive coefficient of 0.999). Further, for simplicity and

### Data, calibration, and paths constructed
- Persistence parameter for a specified process: 0.001 per quarter (i.e., with an autoregressive coefficient of 0.999).
- Productivity in all other regions: assumed to remain at its steady state level throughout the period considered.
- Detrended productivity level path: computed by taking a sector-specific gross percent deviation from a linear trend.
  - Sector-specific linear trend: computed by cumulating the average annual rate of growth from 1980:Q1 to 1994:Q4.
  - Annual data converted to quarterly path: by taking a moving average of the annual observations.
  - Resulting quarterly productivity path covers 1995 to 2004 and is plotted in Figure 2, Panel A for non-tradable and tradable sectors.
- When both tradable and non-tradable paths are fed to the model as a robustness check, results are similar to using only non-tradable productivity (reasons: larger share of non-tradable sector and more volatile tradable productivity path).
- Spillover effects from TFP differences among advanced economies to emerging Asia found to be very small in the multi-region model used (Cova et al (2008)); reasons include large trade shares of the rest of the world block and offsetting sector TFP moves in euro area and Japan during this period.

### Scenario 1 — Productivity slowdown (non-tradable TFP)
- Scenario design:
  - Negative, temporary but persistent productivity shock to non-tradable TFP in the emerging Asia block.
  - Transmission tends to generate a trade balance surplus and an effective exchange rate appreciation.
- Mechanisms and channels:
  - Real effective exchange rate response driven by relative price of non-tradable goods.
  - Decline in non-tradable productivity → persistent increase in relative price of non-tradable goods.
  - With foreign bias in consumption and investment, the terms of trade also appreciate; as a result, the real exchange rate depreciates even more than the terms of trade.
- Trade balance surplus arises from net effect of three forces:
  1. Substitution effects:
     - Rising relative price of non-tradables → demand for non-tradables declines.
     - Complementarity between tradables and non-tradables → initial domestic demand for tradables also decreases.
     - Foreign bias in consumption/investment → demand for foreign tradables falls more than demand for domestic tradables → terms of trade improve.
     - Net effect of substitution: lower imports of foreign tradables → pushes trade balance into surplus.
  2. Consumption smoothing:
     - Households decrease consumption, but less than the decline in labor and capital income from falling productivity.
     - This force tends to reduce the trade balance surplus.
  3. (Third force implied by general equilibrium interactions; impulse responses are not reported here but available on request.)
- Normalizations and reporting conventions used for comparisons:
  - Trade balance and net foreign asset (NFA) position: share of GDP, deviation from steady state, normalized so steady state equals the data in 1994:Q4.
  - Terms of trade and real effective exchange rate: deviation from model steady state, normalized so steady state equals 100 in 1994:Q1.
  - Real interest rate: deviation from steady state, normalized so steady state equals actual short rate in 1994:Q4.
  - Real GDP: annual rate of growth, normalized so steady state equals data in 1994:Q4; actual real GDP growth reported as deviation from its sample mean over 1994-2005.
  - Consumption and investment: reported both as annual rate of growth and as a share of GDP; GDP shares are in deviation from steady state, normalized like the trade balance. Annual real consumption and investment growth normalized like GDP growth.
- Data sources:
  - World Economic Outlook database of the IMF for most series.
  - Real effective exchange rate series from the BIS.
  - Net foreign assets from an updated version of the Lane and Milesi-Ferretti (2007) dataset.
  - Emerging Asia-wide aggregates: weighted averages of individual country variables.
    - Current dollar GDP weights used for exchange rates and interest rates (available for only a small subset of countries).
    - Constant dollar (PPP-valued) GDP weights used for other variables.
  - All quarterly flow variables annualized in the model by multiplying the fourth-quarter value by four.

### Scenario 2 — Desired NFA increase (official reserve accumulation)
- Scenario design:
  - Represents the large increase in official reserves in emerging Asia after 2000 (Figure 2, Panel B).
  - Modeling notes:
    - In the model net and gross foreign assets are the same because only one internationally traded asset exists—the U.S. dollar-denominated nominal bond.
    - Regional NFA position tracks evolution of official reserves closely during the period considered.
    - Regional steady-state NFA position in the model (desired NFA position) is assumed to grow as a percent of GDP between 2001:Q1 and 2006:Q4 at a constant quarterly rate such that cumulative desired NFA change equals cumulative change in official reserves over the same period.
    - Although the change in the steady-state NFA position should be permanent, for simplicity the change is specified to be persistent but ultimately temporary, reverting to zero with autoregressive coefficient of 0.99.
  - The resulting exogenous path for desired NFA is reported in Figure 2, Panel B.
- Data and correlations:
  - Correlation between regional NFA series and official reserve series over 2000-2006 interval is 0.96.
  - NFA data from Lane and Milesi-Ferretti (2007); official reserve data from the IFS database of the IMF.
  - End-2006 is the last year for which these reserve data are available; NFA data extended to 2007 using the average annual growth rates of the preceding three years.

### Results overview and available outputs
- The results compare actual and simulated paths for:
  - Trade balance
  - Real effective exchange rate
  - Terms of trade
  - Real interest rate
  - National accounts (GDP, consumption, investment)
- Figures referenced:
  - Figure 3a and 3b: effects of the non-tradable productivity slowdown scenario.
  - Figure 4a and 4b: effects of the desired NFA increase scenario.
- Additional notes:
  - Impulse responses are not reported in the excerpt but are available on request from the authors.
  - Results of a combined productivity slowdown and increased desired NFA scenario are available from the authors on request.
  - All data used are from sources noted above; emerging Asia aggregates constructed as described.

*Source: _wp0964 - 0.001 per quarter (i.e., with an autoregressive coefficient of 0.999). Further, for simplicity and*

### 0.82 and 0.37, respectively.  The bias towards domestic goods in the traded good basket

### _wp0964 - 0.82 and 0.37, respectively.  The bias towards domestic goods in the traded good basket

### Major findings and conclusions
- A non-tradable productivity decline in emerging Asia (persistent from 1995 to 1999 in the scenario) combined with firms postponing investment pushes the trade balance into surplus; with the paper's calibration, the first and third effects dominate the consumption-smoothing effect, so when non-tradable productivity declines, the trade balance goes into surplus.
- The productivity-slowdown scenario:
  - Fits actual GDP growth in emerging Asia from 1996 to 2002 and broadly mirrors the productivity path.
  - Captures composition of growth from 1997 to 2002: sharp increase in foreign demand (trade balance adjustment during 1997-98 crises), investment collapse and consumption decline.
  - Predicts investment and consumption patterns that are too volatile and too smooth, respectively, relative to the data without significant adjustment costs or financial frictions; investment recovers faster than in the data and reduces the trade surplus counterfactually.
  - Misses a large and persistent terms of trade deterioration in the data, predicting an improvement rather than a decline in the second half of the 1990s.
  - Predicts real exchange rate behavior correctly from 1994–1997 and 1998–2004, but misses the sharp devaluation during the financial crisis.
  - Predicts a small and counterfactual fluctuation in the real interest rate (initially falls then increases), opposite to the data, consistent with technology shock transmission mainly through relative prices.
  - Produces very limited spillovers to the United States (reproduces the 1998 deterioration of the U.S. trade balance closely but little broader macroeconomic impact), attributed to: absence of financial features (reserve currency status, sterilized intervention), policy distortions, and the relatively small share of U.S. exports absorbed by Asia’s imports.

- A sustained increase in the desired level of Net Foreign Assets (NFA) in emerging Asia requires a sustained trade balance surplus to move actual NFA toward the new desired level; the adjustment is gradual because actual NFA need not equal desired NFA in all periods.
  - When actual NFA is smaller than desired NFA an international intermediation premium increases interest rates, inducing postponement of investment and consumption; lower domestic absorption leads to terms of trade deterioration and real exchange rate depreciation, switching demand toward domestic tradables to generate surpluses and accumulate NFA.
  - As actual NFA approaches desired NFA, the trade balance must move toward smaller surpluses or larger deficits because higher NFA generates higher capital income (requiring smaller trade surplus) and because unwinding the desired NFA shock requires larger deficits before reverting to initial steady state.
  - The constructed NFA scenario involves a constant increase in desired NFA from 2001 to 2006 which then unwinds gradually; under perfect foresight agents start accumulating NFA even before the change takes place, resulting in actual NFA at the beginning of simulation being above the still unchanged desired NFA.
- The desired-NFA-increase scenario:
  - Matches the NFA path in the data by construction and fits the trade balance path relatively well between 2000 and 2004, though it generates smaller surpluses than in the data.
  - Captures a smooth and persistent terms of trade decline and trend depreciation in the real exchange rate from 1999 to 2004.
  - Generates a counterfactual, excessively smooth real interest rate response (transmission mostly through relative price changes).
  - Can generate an investment growth decline that is even more persistent than in the data (“investment drought”).
  - Misses real consumption and GDP growth dynamics (these remain essentially constant in the model) and the falling real interest rates observed in the data; in the model domestic absorption falls driven by protracted fall in investment while net exports increase sharply.
  - Produces more sizable spillovers to the United States than the productivity slowdown scenario: simulated paths for the trade balance, NFA position, terms of trade, investment, and to a lesser extent the real exchange rate are qualitatively consistent and quantitatively comparable to the data, though simulated dynamics tend to lead the data (because of perfect foresight). Successive trade deficits cumulate to generate a U.S. NFA position reaching almost 20 percent of GDP by 2004; no effects on real interest rate, consumption, or GDP growth in the U.S. in the model.

- Overall conclusions:
  - Emerging Asia is unlikely to have been a key contributor to the emergence of global imbalances, though it played a role in their evolution since 1997–1998.
  - To understand emerging Asia’s role in the evolution of global imbalances, both an “investment drought” and a “saving glut” hypothesis are necessary to match the data.
  - Financial linkages between emerging Asia and the United States are likely more important than real linkages; “home grown” growth in emerging Asia may not have large effects on the rest of the world, particularly the United States, in a frictionless model.
  - Introducing realistic financial frictions and policy distortions (reserve currency role of the dollar, sterilized foreign exchange intervention) is a promising area for future research to understand consequences of the increase in official reserves in emerging Asia.

### Model structure and key mechanisms
- World economy model: frictionless DGE with five regional blocs (United States, Japan, Euro area, Emerging Asia, Rest of the World).
- Key features:
  - Two sectors per country: tradable and nontradable intermediate goods; final consumption and investment goods produced using these intermediates.
  - Two household types: forward-looking (Ricardian) and liquidity-constrained (non-Ricardian, share of liquidity-constrained consumers = 0.10 in all regions).
  - Forward-looking households hold domestic bonds and a US-denominated international bond (zero net supply worldwide); intermediation costs for international bond are increasing in net foreign asset position to ensure stationarity.
  - Real rigidities: capital adjustment costs, distribution costs, habit persistence; in baseline many adjustment costs set low or zero to reflect frictionless benchmark.
  - Firms face monopolistic competition and set mark-ups; distribution sector introduces wedge η between wholesale and consumer prices, causing market segmentation and price discrimination across export markets.
  - Perfect foresight solution (DYNARE “simul”) with agents anticipating shocks and adjusting from the start of simulations.

### Calibration and steady-state/statistical parameters (exact values preserved)
- Household and firm parameters (Table A2, per-region identical values unless specified):
  - Rate of time preference: .997
  - Depreciation rate: 0.02
  - Intertemporal elasticity of substitution: 1.00
  - Inverse of the Frisch elasticity of labor: 1.50
  - Share of liquidity-constrained consumers: 0.10
- Tradable intermediate goods:
  - Substitution between factors of production: 0.75
  - Bias towards capital: 0.60
- Nontradable intermediate goods:
  - Substitution between factors of production: 0.75
  - Bias towards capital: 0.50
- Final consumption goods:
  - Substitution between domestic and imported goods: 4
  - Bias towards domestic goods: US 0.87; JA 0.49; AS 0.15; EA 0.04; ROW 0.20
  - Substitution between tradables and nontradables: 0.50
  - Bias towards tradable goods: US 0.35; JA 0.36; AS 0.37; EA 0.36; ROW 0.30
- Final investment goods:
  - Substitution between domestic and imported goods: 4
  - Bias towards domestic goods: US 0.87; JA 0.63; AS 0.06; EA 0.92; ROW 0.14
  - Substitution between tradables and nontradables: 0.50
  - Bias towards tradable goods: US 0.75; JA 0.77; AS 0.82; EA 0.75; ROW 0.73
- Real rigidities (Table A3):
  - Capital accumulation adjustment cost parameter: 0.01 (all regions)
  - Distribution costs: 0.30 (all regions)
- Price and wage markups (Table A4):
  - Tradables price markups: 1.18 (all regions)
  - Nontradables price markups: 1.35 (all regions)
  - Wage markup: 1.20 (all regions)
- International linkages (Table A5):
  - Substitution between consumption imports: 5 (all regions)
  - Substitution between investment imports: 5 (all regions)
  - Bias towards imported consumption goods from:
    - US: 0.45 (to US), 0.23 (to JA), 0.04 (to AS), 0.39 (to EA)
    - JA: 0.06 (to US), ... , 0.12 (to AS), 0.03 (to EA), 0.01 (to ROW)
    - AS: 0.17 (to US), 0.38 (to JA), ... , 0.14 (to EA), 0.02 (to ROW)
    - EA: 0.16 (to US), 0.10 (to JA), 0.25 (to AS), ... , 0.58 (to ROW)
    - ROW: 0.61 (to US), 0.07 (to JA), 0.40 (to AS), 0.79 (to EA), ...
  - Bias towards imported investment goods from:
    - US: 0.53 (to US), 0.26 (to JA), 0.26 (to AS), 0.37 (to ROW)
    - JA: 0.06 (to US), ... , 0.13 (to AS), 0.05 (to EA), 0.04 (to ROW)
    - AS: 0.28 (to US), 0.25 (to JA), ... , 0.14 (to EA), 0.16 (to ROW)
    - EA: 0.16 (to US), 0.13 (to JA), 0.12 (to AS), ... , 0.43 (to ROW)
    - ROW: 0.50 (to US), 0.09 (to JA), 0.49 (to AS), 0.55 (to EA), ...
  - Net foreign liabilities intermediation cost parameters:
    - Maximum of financial intermediation cost function: 0.05 (all regions)
    - Steepness of financial intermediation cost function: 0.1 (all regions)
- Steady-state national accounts (Table A1, percents of GDP; columns US JA AS EA ROW):
  - Private Consumption: 68.9, 59.64, 69.33, 58.99, 67.83
  - Forward-looking consumers: 62.64, 52.56, 54.71, 54.73, 62.00
  - Liquidity-constrained consumers: 6.31, 7.08, 14.60, 4.25, 5.83
  - Private Investment: 15.94, 21.08, 19.34, 15.32, 15.73
  - Public Expenditure: 14.51, 19.78, 12.1, 25.5, 16.8
  - Trade balance: 0.60, -0.41, -0.64, 0.18, -0.37
  - Imports: 11.46, 11.49, 26.73, 17.41, 22.98
    - Consumption Goods (imports): 7.33, 8.06, 11.32, 15.01, 12.13
    - Investment Goods (imports): 4.13, 3.43, 15.40, 2.40, 10.85
  - Government Debt: 61.5, 80.0, 55.0, 60.0, 60.0
  - Net Foreign Assets: -51.08, 55.03, 49.02, -11.3, 26.4
  - Share of World GDP (percent): 30.05, 11.48, 9.83, 22.80, 25.84

### Policy implications and recommended research directions
- The model suggests that relative-price and demand-composition mechanisms alone (in a frictionless setting) are insufficient to explain consumption and growth dynamics and real interest rate movements observed in the data for emerging Asia and the United States.
- The scenario analysis implies that to better understand the observed dynamics (investment drought, saving glut, reserve accumulation, U.S. external position), models should incorporate realistic financial frictions and policy distortions, including:
  - Reserve currency role of the U.S. dollar.
  - Sterilized foreign exchange intervention.
  - Other financial linkages that can amplify spillovers between emerging Asia and the United States.

*Italic: Source: working paper content provided in the PDF chapter/section.*

### References

### _wp0964 - References

### Global imbalances, current account, and exchange rates
- Aizenman, Joshua, "Large hoarding of international reserves and the emerging global economic architecture," Keynote address July 2007, The Manchester School 76 (2008): 487-503.  
- Bernanke, Ben S., "The Global Saving Glut and the U.S. Current Account Deficit," Homer Jones Lecture, St. Louis, Missouri, April 14 (2005).  
- Blanchard, Olivier, "Current Account Deficits in Rich Countries," Paper presented at the IMF Seventh Jacques Polak Annual Research Conference on Capital Flows,Washington D.C., November 9--10 (2006).  
- Blanchard, Olivier, Francesco Giavazzi, and Filipa Sa, "The U.S. Current Account and the Dollar," in William Brainard and George Perry (eds.), Brookings Papers on Economic Activity 1 (2005):67-146.  
- Caballero, Ricardo J, Emmanuel Fahri, and Pierre-Olivier Gourinchas, "An Equilibrium Model of `Global Imbalances' and Low Interest Rates,"American Economic Review 98:1, (2008): 358-393.  
- Edwards, Sebastian, "On Current Account Surpluses and the Correction of Global Imbalances," working paper 12904, National Bureau of Economic Research (2007).  
- Engel, Charles and John H. Rogers, "The U.S. Current Account Deficit and the Expected Share of World Output," Journal of Monetary Economics 53 (2006): 1063-1093.  
- Fogli, Alessandra and Fabrizio Perri, "The `Great Moderation' and the US External Imbalance," Working Paper 12708, National Bureau of Economic Research (2006).  
- Mussa, Michael. "Exchange Rate Adjustments Needed to Reduce Global Payments Imbalances," in C. Fred Bergsten and John Williamson (eds.), Dollar Adjustment: How Far? Against What?, Institute for International Economics, , (2004): 113-138.  
- Obstfeld, Maurice and Kenneth Rogoff, "The Unsustainable US Current Account Position Revisited," in Richard Clarida (ed.), G7 Current Account Imbalances: Sustainability and Adjustment, University of Chicago Press (2007).  
- Rajan, Raghuram, "Perspectives on Global Imbalances, " Remarks At the Global Financial Imbalances Conference London, United Kingdom January 23, (2006).  
- Roubini, Nouriel and Brad Setser, "The U.S. as a Net Debtor: The Sustainability of the U.S. External Imbalances," unpublished manuscript, Stern School of Business, New York University (2004).  
- Yoshitomi, Masaru, "Global Imbalances and East Asian Monetary Cooperation," in Duck-Koo Chung and Barry Eichengreen (eds.), Toward an East Asian Exchange Rate Regime (2007)

### Productivity, trade balance, and external transmission mechanisms
- Bems, Rudolf, Luca Dedola, and Frank Smets, "US Imbalances: the role of technology and policy," Journal of International Money and Finance 26 (2007): 523-545.  
- Chakraborty, Suparna and Robert Dekle, "Can International Productivity Differences Explain U.S. Current Account Deficits?," this volume.  
- Corsetti, Giancarlo, Luca Dedola, and Sylvain Leduc, "Productivity, External Balance and Exchange Rates: Evidence on the Transmission Mechanism Among G7 Countries," in Lucrezia Reichlin and Kenneth West (eds.) NBER International Seminar on Macroeconomics (2006).  
- Cova, Pietro, Massimiliano Pisani, Nicoletta Batini, and Alessandro Rebucci, "Productivity and Global Imbalances: The Role of Nontradable Total Factor Productivity in Advanced Economies," IMF Staff Papers 55 (2008): 312-325.  
- Erceg, C., L. Guerrieri and C. Gust, 2002, "Productivity Growth and the Trade Balance in the 1990s: the Role of Evolving Perceptions," (unpublished; Washington, DC: Board of Governors of the Federal Reserve System).  
- Glick, Reuven and Kenneth Rogoff, "Global versus country-specific productivity shocks and the current account," Journal of Monetary Economics 35 (1995): 159-192.  
- Hunt, Ben and Alessandro Rebucci, "The US Dollar and the Trade Deficit: What Accounts for the Late 1990s?," International Finance 8 (2005): 399-434.  
- Mendoza, Enrique G, Rios-Rull, Jose V. and Vincenzo Quadrini, "Financial Integration, Financial Deepness and Global Imbalances," working paper 12909, National Bureau of Economic Reearch (2007)

### Oil shocks and external balances
- "Oil and Global Imbalances", World Economic Outlook, Chapter 2, Washington DC, April 2006.  
- Kilian, Lutz, Nikola Spatafora, and Alessandro Rebucci, "Oil Shocks and External Balances" (2008), CEPR working paper No. 6303.

### IMF World Economic Outlook and regional development patterns
- IMF, (2007) "Exchange Rates and External Imbalances", World Economic Outlook, Chapter3, Washington DC, April 2007.  
- Jaumotte, Florence and Nikola Spatafora, "Asia Rising: Patterns of Economic Development and Growth,". Chapter III, World Economic Outlook, September International Monetary Fund, (2006).

*Source: _wp0964 - References*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp0964.pdf_