## 1.   Testing Delinking of Imports and Exports: Full Sample, Alternative Specifications

## Source details

**Canonical URL:** [1.   Testing Delinking of Imports and Exports: Full Sample, Alternative Specifications](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07214.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07214.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07214.pdf.json)

---

### I. Introduction and research focus
- Focus: analyze shifts in China’s trade structure and implications for the link between exports and imports, rather than estimating equilibrium exchange rate effects on the trade balance.
- Motivating observation: China’s rapid expansion of its trade surplus and debate over its causes (undervalued renminbi vs. structural shifts and changing role in production chains).
- Two competing views:
  - Conventional processing-trade/world-assembler view: export changes largely offset by import changes; limited impact of external demand or exchange-rate shocks on the trade balance.
  - Alternative view advanced here: the export–import link has weakened as domestic production capacity increased, implying slower exports may not be offset by similar import declines.

### Key empirical observations
- Shift in export composition and domestic content:
  - Exports of capital equipment and intermediate products increased to account for more than 40 percent of China’s total exports (up from 10–15 percent a decade ago).
  - Increased domestic production capacity raised domestic sourcing of exports, weakening the export–import linkage in many sectors (notably home electrical appliances, ordinary machinery; to a lesser extent precision apparatus).
- Product sophistication and elasticities:
  - Exports’ sophistication index rose significantly over the last decade; import sophistication increased more slowly.
  - High-tech products account for slightly less than 40 percent of China’s total exports and more than 50 percent of its total imports.
  - Exports of more sophisticated products tend to be more sensitive to aggregate demand and price changes (higher demand- and price-elasticities).
- Processing trade and margins:
  - Share of assembly operations in processing trade balance declined to about 10 percent in 2006 from more than 30 percent in the late 1990s.
  - Margins (domestic value added per dollar exported, i.e., trade balance divided by exports) rose from the teens in the mid-1990s to more than 40 percent last year.
- Trade surplus and recent magnitudes:
  - Over the past five years, the trade surplus has grown over 5 folds in dollar terms and more than tripled as a percent of GDP.
  - The most significant increase occurred over the last three years, with China’s trade surplus rising by over 5 percent of GDP.
  - At end-2006, the trade surplus is estimated at around $215 billion (BOP basis) or more than 8 percent of GDP.
  - Expansion driven by manufacturing trade surplus: electronics, machinery, iron and steel, textiles and clothing, and organic chemicals are main contributors; steel products moved into surplus for the first time in 2006.

### Recent developments in China’s external trade (composition and dynamics)
- Import slowdown and composition:
  - Import growth started to lag export growth by large margins in early 2005.
  - Imports of intermediate goods slowed the most: parts and components and semi-finished goods account for almost half of the slowdown in import growth between 2003 and 2006.
  - China’s trade deficit with the rest of Asia has shrunk over the last two years even as trade surpluses with the U.S. and the EU continue to grow.
- Structural drivers:
  - Continued strength of investment and technological upgrade raised domestic production capacity of key intermediate products several folds since 2000.
  - Foreign direct investment contributed strongly (large FDI increases in chemicals, electronics, precision instruments between 1999 and 2005 noted).
- Sophistication percentiles and shares:
  - Products divided into three percentiles: low-tech ≤ 30th percentile; medium-tech between 30th and 70th percentiles; high-tech > 70th percentile.
  - High-tech products’ share in total exports increased by 20 percent between 1995 and 2005 while low-tech declined by almost 20 percentage points.

### Empirical approach — goals, data, and estimation
- Goals:
  - Test whether imports of parts and components have “delinked” from exports of associated final products.
  - Analyze how product characteristics (especially sophistication level) affect trade responses to aggregate shocks and implications for trade elasticities.
- Data and sample:
  - Disaggregated trade data at five-digit SITC Revision 2 from UN Comtrade, nominal values annual 1992–2005.
  - Directly relate 105 pairs of inputs and associated outputs within SITC codes 7, 87, and 88.
    - These 105 pairs account for, on average, around half of total imports of parts and components or roughly 15 percent of total imports over this time period.
    - Roughly a quarter of pairs are in electronics; remaining in machinery and transport equipment.
  - Balanced panel of 105 imported parts over 1993–2005 = 1,365 observations; first-differenced panel for 1994–2005 = 1,260 observations.
- Deflators and proxies:
  - Import and export values deflated using SITC-specific U.S. import price indices (proxy for world prices).
  - Domestic demand for the final product estimated as: total manufacturing imports + total domestic industrial production − total manufacturing exports.
  - Relative prices measured as ratio of Chinese sectoral production price indices to U.S. sector-specific import price indices; adjusted for RMB/USD exchange rate changes and tariff reductions.
- Econometric specification (log levels, notation from source):
  - M_it = f(X_it, DD_it, (P_it^F / P_it)) + η_i + ε_it
    - M_it: imports of intermediate product i at time t.
    - X_it: exports of associated final product.
    - DD_it: domestic demand.
    - (P_it^F / P_it): relative price term.
    - η_i: product-specific fixed effect; ε_it: idiosyncratic error.
  - Expected signs:
    - β (on X_it): positive if foreign demand/processing trade drives imports.
    - φ (on DD_it): positive if imports are associated with domestic demand.
    - γ (on relative prices): positive if higher relative prices in China increase imports.
- Estimators and diagnostics:
  - Pooled OLS: correct signs; only exports significant.
  - Random effects: all three regressors correctly signed and highly significant; random effects rejected by Hausman test at 1 percent.
  - Fixed effects: regressors highly significant; evidence of first-order serial correlation in residuals.
  - First-differenced OLS: eliminates time-invariant product effects; regressors correctly signed and highly significant; addresses unit-root concerns and serial correlation.

### Empirical results — delinking and domestic-demand linkage
- Delinking evidence:
  - Sub-sample analysis (1994–1999 vs. 2000–2005; each with 630 observations) indicates the relationship between imported inputs and associated exports has weakened in the more recent period.
  - Imports of parts and components have delinked from exports of final products in the sample industries examined.
- Domestic demand role:
  - Imported inputs have become more strongly associated with domestic demand over time.
  - In recent years domestic demand growth has become strongly associated with, and export growth completely disassociated from, import growth (first-differenced OLS results).
- Robustness:
  - Results robust across pooled OLS, random effects, fixed effects, and first-differenced OLS estimators.
  - IV robustness checks (instrumenting aggregate derived domestic demand with real domestic credit) produce coefficient estimates closely resembling main results; no evidence of weak instruments and Hausman test does not reject validity of IV model in log levels.

### Product sophistication and heterogeneous elasticities — empirical findings
- Sophistication measure:
  - Sophistication index coded 0, 1, 2 for low tech, medium tech, high tech; used to interact with demand and RER variables.
- Export-side (first specification, volumes):
  - More sophisticated products: exports increase more in response to a given increase in foreign demand.
  - More sophisticated products: exports drop more for a given appreciation of the real effective exchange rate.
  - Interaction coefficients a5 and a6 are statistically significant at the 5 percent level.
- Import-side:
  - More sophisticated products: imports increase more in response to a rise in domestic demand.
  - More sophisticated products: imports tend to increase less (or decrease more) in response to a given appreciation of the real effective exchange rate.
  - Interaction coefficient b6 significant for import-to-GDP ratio in full sample and for import volumes in 1994–2002 subsample; b6 becomes insignificant when both interaction terms included in some specifications.
- Second specification (value shares to GDP):
  - Export-to-GDP: more sophisticated products see larger drops for a given RER appreciation; the increase in export-to-GDP with foreign demand for more sophisticated products is not statistically significant in some specifications.
  - Import-to-GDP: more sophisticated products see larger increases in import-to-GDP in response to rising domestic demand; coefficient b6 significant at the 5 percent level in the full sample.

### Quantitative highlights from reported tables (values preserved exactly)
- Table 4 (Export Elasticities, sample 1994–2005; dependent = log first differenced exports):
  - dln_FD x SL (a5): 1.7** 0.75; 1.6** 0.75; 2.1** 0.86; 2.1** 0.86
  - dln_RER x SL (a6): -0.3* 0.18; -0.3* 0.18; -0.5* 0.24; -0.5* 0.24
  - Number of observations: 12,903; Number of Sectors: 1,523; R2: 0.01
- Table 5 (Import Elasticities, sample 1994–2005; dependent = log first differenced imports):
  - dln_DD x SL (b6): 0.3*** 0.10; 0.1 0.2; 0.7*** 0.17; 0.4** 0.2
  - dln_REER x SL (b7): -0.6*** 0.16; -0.5* 0.3; -0.9*** 0.22; -0.5** 0.3
  - Number of observations: 12,903; Number of Sectors: 1,532; R2: 0.005
- Table 6 (Export-GDP Ratio; coefficients scaled up by 1000):
  - RER_CH: -0.1*** 0.0; -0.1*** 0.0; 0.1*** 0.03; 0.1*** 0.0
  - FD_CH: 0.7*** 0.1; 0.5** 0.1; 0.7*** 0.1; 0.5** 0.1
  - FD_CH x SL: 0.1 0.1 0.1 0.1
  - RER_CH x SL: -0.2*** 0.03; -0.2*** 0.0
  - Number of observations: 19,579; Number of sectors: 1,534; R2: 0.002–0.003
- Table 7 (Import-GDP Ratio; coefficients scaled up by 1000):
  - DD_CH: 0.3*** 0.14; 0.1 0.04; 0.3** 0.04; 0.0 0.04
  - FD_CH: 0.4*** 0.67; 0.4** 0.14; 0.4** 0.14; 0.4*** 0.14
  - RER_CH: 0.1*** 0.24; 0.2** 0.05; 0.3*** 0.06; 0.1*** 0.05
  - DD_CH x SL: 0.2*** 0.05; 0.2*** 0.07
  - RER_CH x SL: -0.2*** 0.05; 0.1 0.08
  - Number of observations: up to 22,143; Number of Sectors: up to 1,733; R2: 0.003–0.006

### Interpretations and mechanisms
- Delinking mechanism:
  - Increased domestic production capabilities and movement away from simple processing trade make imports increasingly determined by domestic production and consumption needs rather than final exports.
  - Where delinking occurs, macroeconomic shocks (global demand shifts or exchange-rate changes) can have larger effects on China’s trade balance and domestic economy because imported intermediates are less tied to foreign demand.
- Heterogeneous product responses:
  - Exports of more sophisticated products: stronger response to foreign demand; larger negative response to RER appreciation.
  - Imports of more sophisticated products: stronger response to domestic demand; smaller or negative response to RER appreciation (reflecting limited domestic capacity at the high end and import substitution at the lower end).
  - Electronics sector: domestic capability still lags imports, showing weaker evidence of parameter heterogeneity across time relative to other sectors.

### Policy implications and recommendations
- Macroeconomic vulnerability and rebalancing:
  - China’s trade balance may be more sensitive to external shocks (global demand or exchange rates) than commonly assumed, increasing vulnerability to volatile net exports.
  - Policy priority: hasten rebalancing of China’s growth away from potentially volatile net exports toward a more sustainable path driven by domestic demand.
- Empirical practice:
  - Empirical analyses of trade dynamics and equilibrium exchange rates should account for shifts in production and trade structure, allowing for structural breaks and heterogeneity by product sophistication to avoid biased estimates (e.g., overestimating renminbi undervaluation if changing trade elasticities are ignored).
- Regional and global implications:
  - China’s reduced imports of intermediate goods from Asia could weaken intra-regional trade linkages; expansion of China’s domestic market might create opportunities for regional economies, particularly for higher-tech goods China cannot produce domestically.
  - Rising Chinese labor costs and evolving comparative advantage could open opportunities for lower-income Southeast Asian countries to occupy lower-end production networks.
  - China’s transition toward greater domestic production could shift it from price taker to price setter for intermediate goods, with potential dampening influence on global inflation through effects on intermediate goods supply and prices.
- Research recommendation:
  - Further research on implications for regional production networks and global prices is suggested.

### IV. Conclusion — Main empirical takeaways
- Within the sample, imports have delinked from exports and have become more responsive to domestic demand in recent years.
- Coefficient behavior across sub-samples:
  - Coefficient on exports is highly significant and of similar magnitude in both sub-samples.
  - Estimated coefficient on domestic demand: insignificantly different from zero in the earlier period; highly significant and much larger in the latter period.
  - Point estimates on exports and domestic demand are more than three standard deviations away across sub-samples; formal tests strongly reject the hypothesis that these coefficients are the same across the two time periods.
- Estimation robustness:
  - First-differenced OLS (addressing first-order serial correlation) yields robust and starker evidence of delinking.
  - IV checks (instrumenting derived domestic demand with real domestic credit) corroborate main findings.

*Source: _wp07214 (IMF Working Paper) — Testing Delinking of Imports and Exports: Full Sample, Alternative Specifications*

### 1.   Testing Delinking of Imports and Exports: Full Sample, Alternative Specifications ...... 21

### 1.   Testing Delinking of Imports and Exports: Full Sample, Alternative Specifications

### I. Introduction and research focus
- Focus: analyze shifts in China’s trade structure and implications for the link between exports and imports, rather than estimating equilibrium exchange rate effects on the trade balance.
- Motivating observation: China’s rapid expansion of its trade surplus and debate over its causes (undervalued renminbi vs. structural shifts and changing role in production chains).

### Key conceptual points
- Conventional "world assembler" (processing trade) view:
  - China uses abundant and cheap labor to assemble and process imported inputs into final goods for export.
  - Under this view, changes in external demand or the exchange rate have limited impact on the trade balance because export changes are offset by similar import changes.
  - Prior econometric analyses typically find relatively small effects of relative price changes on China’s exports and trade balance, attributed to China’s role as a processing center (Shu and Yip (2006)).
- Alternative view from this paper:
  - The tight link between exports and imports has weakened over time as domestic production capacity has increased.
  - A slowdown in exports may not be offset by a commensurate decline in imports, implying macroeconomic shocks could have more potent effects on China’s trade balance and domestic economy.

### Major empirical observations from the paper
- Shift in export composition:
  - Exports of capital equipment and intermediate products have increased significantly and now account for more than 40 percent of China’s total exports.
- Product sophistication and elasticities:
  - Exports of more sophisticated products tend to be more sensitive to aggregate demand and price changes, implying higher demand- and price-elasticities for aggregate trade as the export bundle becomes more sophisticated.
- Domestic content:
  - Increased domestic production capacity has enabled increased domestic sourcing of exports, weakening the traditional export-import linkage, especially for lower technological sophistication products.

### Relation to prior literature and methodology
- Builds on literature emphasizing disaggregated data to understand trade patterns (Feenstra and Rose (2000), Schott (2004), Hummels and Klenow (2005), Schott (2006)).
- Uses the product-level sophistication measurement methodology proposed by Rodrik (2006) to explore implications of shifting product composition for China’s aggregate trade.
- Cites empirical findings in the literature:
  - Lemoine and Ünal-Kesenci (2002): China’s specialization pattern enhanced technological transfer, including increased local content in processing trade.
  - Hiratsuka (2005): by 2001, China advanced in nondurable goods and light machineries.
  - Gao and Ha (2006): large structural changes with machinery and transport equipment increasing share of exports; more natural resource-based and high-tech products being imported.
  - Schott (2004, 2006): product-level evidence on specialization and relative sophistication; prices received for exports have been declining over time.
  - Rodrik (2006): sophistication indices based on per capita GDP of exporting countries; China’s export bundle more sophisticated than expected and has become more sophisticated over time.

### Implications highlighted
- The weakening of the export-import nexus implies:
  - Greater sensitivity of China’s trade balance to changes in global demand and relative prices.
  - Macroeconomic shocks may have larger domestic effects because decreases in exports may not produce offsetting declines in imports.
- The rising sophistication of exports implies:
  - A shift in the types of imports required for manufacturing toward more sophisticated inputs.
  - Heterogeneous responses of different product categories to price and demand shocks.

### Paper organization (as stated)
- Section II: discusses recent developments in China’s trade and production structure, including evidence that a slowdown in intermediate imports has not been matched by a commensurate slowdown in final exports, increased domestic content in exports, and rising sophistication of the export bundle.

*Source: _wp07214 - 1.   Testing Delinking of Imports and Exports: Full Sample, Alternative Specifications ...... 21*

### Section III uses detailed trade data to conduct two separate sets of empirical analyses that

### _wp07214 - Section III uses detailed trade data to conduct two separate sets of empirical analyses that

### Recent developments in China’s external trade
- Trade surplus developments
  - Over the past five years, the trade surplus has grown over 5 folds in dollar terms and more than tripled as a percent of GDP.
  - The most significant increase has occurred over the last three years, with China’s trade surplus rising by over 5 percent of GDP.
  - At end-2006, the trade surplus is estimated to have stood at around $215 billion (BOP basis) or more than 8 percent of GDP.
  - Expansion driven by manufacturing trade surplus: electronics, machinery, iron and steel, textiles and clothing, and organic chemicals are the main contributors.
  - Steel products: the trade balance in steel products, traditionally in deficit, moved into surplus for the first time in 2006.

- Import slowdown and composition
  - Import growth started to lag export growth by large margins in early 2005.
  - Imports of intermediate goods slowed the most: parts and components and semi-finished goods account for almost half of the slowdown in import growth between 2003 and 2006.
  - Resulting regional effects: China’s trade deficit with the rest of Asia has shrunk over the last two years even as trade surpluses with the U.S. and the EU continue to grow.

- Structural drivers
  - Continued strength of investment and technological upgrade have raised domestic production capacity of key intermediate products several folds since 2000.
  - Foreign direct investment has played an important role (examples in the source text note large FDI increases in chemicals, electronics, precision instruments between 1999 and 2005).
  - Supported by rising domestic production capacity, import content of exports has declined in many sectors (notably home electrical appliances, ordinary machinery; to a lesser extent precision apparatus).
  - Final-product exports have continued to grow strongly despite slowdown in imports of intermediate inputs.

- Processing trade and margins
  - Processing trade is shifting away from simple assembly toward operations using more domestic inputs.
  - Share of assembly operations in processing trade balance declined to about 10 percent in 2006 from more than 30 percent in the late 1990s.
  - Margins (domestic value added per dollar exported, i.e., trade balance divided by exports) rose from the teens in the mid-1990s to more than 40 percent last year.

- Changes in trade composition and sophistication
  - Over the past decade, share of labor-intensive consumption goods in exports declined more than 20 percentage points.
  - Exports of capital goods and parts and components rose sharply to account for more than 40 percent of total exports (compared to 10–15 percent a decade ago).
  - Product sophistication:
    - Exports’ sophistication index rose significantly over the last decade; import sophistication increased more slowly.
    - China tends to import more sophisticated goods than it exports.
    - Sectoral differences: domestic production capability near or exceeding imported goods in intermediate products (manufacturing materials, chemicals, nonfuel crude materials); for electronics, domestic capability continues to lag imports.
  - Sophistication percentiles and shares:
    - Products divided into three percentiles: low-tech ≤ 30th percentile; medium-tech between 30th and 70th percentiles; high-tech > 70th percentile.
    - On the export side, high-tech products’ share in total exports increased by 20 percent between 1995 and 2005 while low-tech declined by almost 20 percentage points.
    - Today, high-tech products account for slightly less than 40 percent of China’s total exports and more than 50 percent of its total imports.

---

### Evaluating trade equations using sectoral data — goals and approach
- Purpose
  - Explore how changes in China’s production structure may have affected trade flows and their responses to macroeconomic shocks.
  - Two empirical tests:
    1. Examine whether imports of parts and components have “delinked” from exports of associated final products (consistent with rising domestic production capacity).
    2. Analyze how product characteristics (especially sophistication level) affect the response of trade flows to aggregate shocks, and implications for trade elasticities.

- Key empirical findings (high level)
  - Delinking: imports of parts and components have delinked from exports of final products in recent years for the sample examined.
  - Domestic demand linkage: imported inputs have become more strongly associated with domestic demand.
  - Sophistication and sensitivity:
    - For more sophisticated products, both exports and imports are more sensitive to aggregate demand shocks.
    - Exports of more sophisticated products are also more sensitive to real exchange rate changes.
    - Given rising sophistication, aggregate trade may have become increasingly sensitive to demand and price fluctuations.

---

### A. The delinking of imports from exports — data, estimation, and results
- Data and sample
  - Disaggregated trade data at five-digit SITC Revision 2 from UN Comtrade, nominal values annual 1992–2005.
  - Able to directly relate 105 pairs of inputs and associated outputs within SITC codes 7, 87, and 88.
    - These 105 pairs account for, on average, around half of total imports of parts and components or roughly 15 percent of total imports over this time period.
    - Roughly a quarter of pairs are in electronics; remaining in machinery and transport equipment.
  - Deflators and proxies:
    - Import and export values deflated using SITC-specific U.S. import price indices (proxy for world prices).
    - Domestic demand for the final product estimated as: total manufacturing imports + total domestic industrial production − total manufacturing exports.
    - Alternative domestic demand measures used in robustness checks (sectoral-level and aggregate national accounts); deflated using manufacturing PPI in China and GDP deflator.
    - Relative prices measured as ratio of Chinese sectoral production price indices to U.S. sector-specific import price indices; adjusted for RMB/USD exchange rate changes and tariff reductions.
  - Panel: balanced panel of 105 imported parts over 1993–2005 = 1,365 observations.

- Econometric specification
  - Regression (log levels) tested:
    - M_it = f(X_it, DD_it, (P_it^F / P_it)) with product-specific fixed effect eta_i and idiosyncratic error epsilon_it.
    - Notation in the source: equation (1) where M_it denotes imports of intermediate product i at time t; X_it denotes exports of associated final product; DD_it denotes domestic demand; (P_it^F / P_it) is the relative price term; product-specific effect iη and residual itε.
  - Interpretation:
    - Coefficient β on exports (X_it): expected positive if foreign demand (processing trade/assembly) drives imports.
    - Coefficient φ on domestic demand (DD_it): expected positive if imports are associated with domestic demand.
    - Coefficient γ on relative prices: expected positive if higher relative prices in China increase imports.
  - Estimators and diagnostics:
    - Pooled OLS (column 1): correct signs; only exports significant.
    - Random effects (column 2): all three regressors correctly signed and highly significant; random effects rejected by Hausman test at 1 percent.
    - Fixed effects (column 3): allows product-specific effects correlated with regressors; regressors highly significant; evidence of first-order serial correlation in residuals.
    - First-differenced OLS (column 4): eliminates time-invariant product effects; all regressors correctly signed and highly significant; first-differencing addresses potential unit root concerns and serial correlation.
    - First-differencing reduces sample by one year: resulting sample 1,260 observations between 1994 and 2005.

- Empirical results and interpretation
  - Robustness: basic results robust across pooled OLS, random effects, fixed effects, and first-differenced OLS estimators.
  - Delinking evidence:
    - Sub-sample analysis splitting the sample into 1994–1999 and 2000–2005 (each with 630 observations) indicates the relationship between imported inputs and associated exports has weakened in the more recent period.
    - Imports of parts and components have delinked from exports of final products in the sample industries examined.
  - Domestic demand role:
    - Imported inputs have become more strongly associated with domestic demand over time.
  - Implications:
    - Where delinking occurs, macroeconomic shocks may have more potent effects on China’s trade balance and domestic economy because imported intermediates are less directly tied to foreign demand.
    - Given rising product sophistication and its greater sensitivity to demand and price shocks, aggregate trade elasticities could increase even absent further price or trade regime changes.

*Italic: Content derived from _wp07214 - Section III uses detailed trade data to conduct two separate sets of empirical analyses that (IMF PDF).*

### conclusions. However, the two sub-samples display markedly different properties. In

### IV. CONCLUSION

### Main empirical findings
- Within the sample, imports have delinked from exports and have become more responsive to domestic demand in recent years.
- The coefficient on exports is highly significant and of similar magnitude in both sub-samples, while the estimated coefficient on domestic demand displays significant heterogeneity:
  - domestic demand coefficient is insignificantly different from zero in the earlier period and is both highly significant and much larger in the latter period.
  - Point estimates on exports and domestic demand are more than three standard deviations away across sub-samples; formal tests of parameter heterogeneity strongly reject the hypothesis that these coefficients are the same across the two time periods.
- Using a first-differenced OLS estimator (appropriate due to first-order serial correlation in fixed effects residuals) yields robust results and even starker evidence: in recent years domestic demand growth has become strongly associated with, and export growth completely disassociated from, import growth.
- IV robustness checks (instrumenting aggregate derived domestic demand with real domestic credit) produce coefficient estimates closely resembling the main results, with no evidence of weak instruments and a Hausman test not rejecting the validity of the IV model in log levels.

### Product sophistication and heterogeneous elasticities
- The “sophistication index” (values 0, 1, 2 for low tech, medium tech, high tech) is used to test whether products of different technical sophistication respond differently to foreign and domestic demand and to trade-weighted real exchange rate (RER) fluctuations.
- Export-side results (first specification, volumes):
  - The more sophisticated a product is, the more its exports tend to increase in response to a given increase in foreign demand.
  - The more sophisticated a product is, the more its exports tend to drop for a given appreciation of the real effective exchange rate.
  - Both interaction coefficients a5 and a6 are statistically significant at the 5 percent level.
- Import-side results:
  - The more sophisticated a product is, the more its imports tend to increase in response to a rise in domestic demand.
  - Imports of more sophisticated products tend to increase less (or decrease more) in response to a given appreciation of the real effective exchange rate.
  - Interaction coefficient b6 is significant for import-to-GDP ratio in the full sample and for import volumes in the 1994–2002 subsample; however, b6 becomes insignificant when both interaction terms are included in some specifications.
- Second specification (value shares to GDP):
  - Export-to-GDP: more sophisticated products see larger drops for a given RER appreciation; the increase in export-to-GDP with foreign demand for more sophisticated products is not statistically significant in some specifications.
  - Import-to-GDP: more sophisticated products see larger increases in import-to-GDP in response to rising domestic demand; coefficient b6 is significant at the 5 percent level in the full sample.

### Quantitative highlights from reported tables (selected coefficients and statistics; values preserved exactly)
- Table 4 (Export Elasticities, sample 1994–2005; dependent = log first differenced exports):
  - dln_FD x SL (a5): 1.7** 0.75; 1.6** 0.75; 2.1** 0.86; 2.1** 0.86 (shown across specifications)
  - dln_RER x SL (a6): -0.3* 0.18; -0.3* 0.18; -0.5* 0.24; -0.5* 0.24
  - Number of observations: 12,903; Number of Sectors: 1,523; R2: 0.01 (varies by column)
- Table 5 (Import Elasticities, sample 1994–2005; dependent = log first differenced imports):
  - dln_DD x SL (b6): 0.3*** 0.10; 0.1 0.2; 0.7*** 0.17; 0.4** 0.2 (across specifications)
  - dln_REER x SL (b7): -0.6*** 0.16; -0.5* 0.3; -0.9*** 0.22; -0.5** 0.3
  - Number of observations: 12,903; Number of Sectors: 1,532; R2: 0.005 (varies)
- Table 6 (Export-GDP Ratio; coefficients scaled up by 1000):
  - RER_CH: -0.1*** 0.0; -0.1*** 0.0; 0.1*** 0.03; 0.1*** 0.0
  - FD_CH: 0.7*** 0.1; 0.5** 0.1; 0.7*** 0.1; 0.5** 0.1
  - FD_CH x SL: 0.1 0.1 0.1 0.1
  - RER_CH x SL: -0.2*** 0.03; -0.2*** 0.0
  - Number of observations: 19,579; Number of sectors: 1,534; R2: 0.002–0.003
- Table 7 (Import-GDP Ratio; coefficients scaled up by 1000):
  - DD_CH: 0.3*** 0.14; 0.1 0.04; 0.3** 0.04; 0.0 0.04
  - FD_CH: 0.4*** 0.67; 0.4** 0.14; 0.4** 0.14; 0.4*** 0.14
  - RER_CH: 0.1*** 0.24; 0.2** 0.05; 0.3*** 0.06; 0.1*** 0.05
  - DD_CH x SL: 0.2*** 0.05; 0.2*** 0.07 (in interaction specifications)
  - RER_CH x SL: -0.2*** 0.05; 0.1 0.08
  - Number of observations: up to 22,143; Number of Sectors: up to 1,733; R2: 0.003–0.006

### Interpretations and mechanisms discussed
- The delinking is consistent with increased domestic production capabilities and a move away from reliance on processing trade; imports are increasingly determined by domestic production and consumption needs rather than final exports.
- Product composition has shifted toward more sophisticated products for both exports and imports:
  - Exports of more sophisticated products are more sensitive to foreign demand and more negatively affected by RER appreciations.
  - Imports of more sophisticated products are more responsive to domestic demand increases and less responsive to RER appreciations, reflecting limited domestic production capacity for high-end goods and greater import substitution at the lower end.
- The relative sophistication of sectors (e.g., electronics) affects the pace at which exports delink from imports; electronics shows weaker evidence of parameter heterogeneity across time.

### Policy implications and recommendations
- China’s trade balance may have become more sensitive to external shocks (shifts in global demand or exchange rates) than commonly assumed; this increases vulnerability to volatile net exports.
- There is a policy need to hasten the rebalancing of China’s growth away from potentially volatile net exports toward a more sustainable path driven by domestic demand.
- Empirical analyses of trade dynamics and equilibrium exchange rates should account for shifts in production and trade structure, allowing for structural breaks and heterogeneity based on product sophistication; failing to do so may bias estimates (e.g., potentially overestimating the extent of renminbi undervaluation in analyses that ignore changing trade elasticities).
- Regional implications:
  - As China specializes in more parts of production chains, its imports of intermediate goods from Asia could fall, potentially decreasing intra-regional trade linkages.
  - Expansion of China’s domestic market could create new opportunities for regional economies, particularly for higher-tech products that China cannot produce domestically in the near term.
  - Rising Chinese labor costs and evolving comparative advantage could present opportunities for lower-income Southeast Asian countries to occupy the lower end of production networks.
- Global markets:
  - China’s transition could shift it from price taker to price setter for intermediate goods, potentially exerting a dampening influence on global inflation via effects on intermediate goods supply and prices.
  - Further research is suggested on implications for regional production networks and global prices.

*Source: IMF Working Paper (conclusions and section on product sophistication and trade flows, extracted from the supplied content).*

### REFERENCES

### _wp07214 - REFERENCES

### Works cited
- Bayoumi, Tamim, 1999, “Estimating Trade Equations from Aggregate Bilateral Data,” IMF Working Paper No. 99/74 (Washington: International Monetary Fund).  
- Cerra, Valerie, and Sweta Chaman Saxena, 2003, “How Responsive is Chinese Export Supply to Market Signals,” China Economic Review, Vol. 14, Issue 3, pages 350–370.  
- Cerra, Valeria, and Anuradha Dayal-Gulati, 1999, “China’s Trade Flows: Changing Price Sensitivities and the Reform Process,” IMF Working Paper No. 99/1 (Washington: International Monetary Fund).  
- Coudert, Virginie, and Cécile Couharde, 2005, “Real Equilibrium Exchange Rate in China,” CEPII Working Paper No. 2005-01 (Paris: Centre d'Etudes Prospectives et d'Informations Internationales).  
- Dark, Deborah, and John Hawkins, 2005, “Why Have Imports of Australia’s Goods Increased So Much?” Australia Treasury.  
- Dunaway, Steven, Lamin Leigh, and Xiangming Li, 2006, “How Robust are Estimates of Equilibrium Real Exchange Rates: the Case of China,” IMF Working Paper No. 06/220 (Washington: International Monetary Fund).  
- Feenstra, Robert C., and Andrew K. Rose, 2000 “Putting Things in Order: Trade Dynamics and Product Cycles,” Review of Economics and Statistics, Vol. 82, pages 369–382.  
- Gao, Ting, and Jiming Ha, 2006, “Reaching New Heights: Analysis of China’s Foreign Trade Structure,” China International Capital Corporation Limited.  
- Goldstein, Morris, 2004, paper presented at the International Monetary Fund’s Seminar on China’s Foreign Exchange System, “Adjusting China’s Exchange Rate Policies,” Dalian, China, May.  
- Hiratsuka, Daisuke, 2005, “The ‘Catching Up’ Process of Manufacturing in East Asia,” IDE Discussion Paper No. 22 (Chiba: Institute of Developing Economies).  
- Hummels, David, and Peter Klenow, 2005, “The Variety and Quality of a Nation’s Exports,” American Economic Review.  
- Lee, Jaewoo, and Menzie D. Chinn, 2002, “Current Account and Real Exchange Rate Dynamics in the G-7 Countries,” IMF Working Paper No. 02/130 (Washington: International Monetary Fund).  
- Lemoine, Francoise, and Deniz Ünal-Kesenci, 2002, “China in the International Segmentation of Production Process,” CEPII Working Paper No 2002-02 (Paris: Centre d'Etudes Prospectives et d'Informations Internationales).  
- Rodrik, Dani, 2006, “What’s So Special About China’s Exports,” NBER Working Paper No. 11947 (Cambridge, Massachusetts: National Bureau of Economic Research).  
- Schott, Peter K., 2004, “Across-Product Versus Within-Product Specialization in International Trade,” Quarterly Journal of Economics, Vol. 119, No. 2, Pages 647–678.  
- Schott, Peter K., 2006, “The Relative Sophistication of Chinese Exports,” NBER Working Paper No. 12173 (Cambridge, Massachusetts: National Bureau of Economic Research).  
- Schott, Peter K., 2004, “Across-Product Versus Within-Product Specialization in International Trade,” Quarterly Journal of Economics, Vol. 119, No. 2, Pages 647–678.  
- Shu, Chang, and Raymond Yip, 2006, “Impact of Exchange Rate Movement on the Mainland Economy,” Hong Kong Monetary Authority, China Economic Issues.

*Source: _wp07214 - REFERENCES*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp07214.pdf_