## _wp1506

## Source details

**Canonical URL:** [_wp1506](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1506.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1506.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1506.pdf.json)

---

### Introduction and research questions
- After bouncing back from the historic low of the Great Recession in 2010, world trade grew by less than 3.0 percent in 2012 and 2013, compared to the pre-crisis average of 7.1 percent (1987-2007).
- Research questions highlighted:
  - Whether the slower growth in trade simply reflects the sluggishness of GDP or whether there is a deeper structural shift in the relationship between trade and GDP.
  - If there is a structural shift:
    - Whether it is a post-crisis phenomenon or reflects longer-term factors.
    - What those longer-term factors might be.

### Empirical strategy and model
- Empirical approach:
  - Error Correction Model (ECM): Δ ln m_t = α + β Δ ln y_t + γ ln m_{t-1} + δ ln y_{t-1} + ε_t.
    - Short-run trade elasticity = β.
    - Speed of adjustment = –γ.
    - Long-run trade elasticity = –δ/γ.
- Data and estimation:
  - Annual analysis: 1970-2013.
  - Quarterly analysis: 1991-2013, with focus on 2001q1-2007q4 and 2008q1-2013q4.
  - Non-linear Wald tests indicate significant structural breaks; recursive estimation tests confirm results.
  - Residuals are stationary; Breusch-Godfrey LM tests generally accept no serial correlation.

### Key empirical findings — elasticities and adjustment
- Long-run trade elasticities (selected periods):
  - 1970-1985: 1.3
  - 1986-2000: 2.2
  - 2000s: 1.3
  - Entire sample (1970-2013): 1.7 (long-run elasticity = −δ/γ = 1.70*** from pooled yearly estimation)
- Quarterly estimates for the 2000s:
  - 2001q1-2007q4: 1.5
  - 2008q1-2013q4: 0.7
- Short-run elasticities (β) (annual):
  - 1970-1985: 2.1
  - 1986-2000: 2.8
  - 2000-2013: 3.4
- Quarterly short-run and impact elasticities:
  - 1990s quarterly short-run: 1.5
  - 2000s quarterly short-run: 2.6
  - Impact elasticities: 4.0 (2001q1-2007q4) and 2.2 (2008q1-2013q4)
- Speed of adjustment (–γ):
  - Long 1990s: 0.58
  - 2000s: 0.31
  - 1970-1985: lower and not statistically significant
- Selected table estimates (preserved exactly):
  - Pooled yearly (1970-2013): α = -0.43** (0.17); β = 2.82*** (0.36); −γ = 0.12** (0.05); δ = 0.20** (0.09); −δ/γ = 1.70***; Breusch-Godfrey LM test = 9.67**; Rsquared = 0.740; N = 43.
  - Quarterly pooled (1991q2-2000q4): −δ/γ = 2.40***; (2001q1-2007q4) −δ/γ = 1.49***; (2008q1-2013q4) −δ/γ = 0.68***; (2001q1-2013q4) −δ/γ = 1.21***; Rsquared = 0.838 (cols (1)-(3)) and 0.812 (col (4)); N = 90 per column.

### Stylized facts on trade growth and model performance
- Global trade volume growth:
  - 13 percent in 2010.
  - 6.2 percent in 2011.
  - 2.8 percent in 2012.
  - 3.0 percent in 2013.
  - Pre-crisis average (1987-2007): 7.1 percent.
- World GDP growth in recent years hovered around 3 percent.
- Regional performance:
  - Euro Area imports: declined by 1.1 percent in 2012; increased by 0.3 percent in 2013.
  - Imports of emerging economies increased by 5.6 percent in 2013.
- ECM predictive performance:
  - The model "performs reasonably well in predicting trade growth over the past 43 years."
  - Long-run portion explains most of total growth for much of the sample; cyclical effects dominate during crises and recoveries.

### Decomposition of the slowdown (structural vs cyclical) and projections
- Decomposition method:
  - Predicted import growth decomposed into long-run component (from cointegration residual and long-run predicted log level of imports) and short-term (cyclical) component (model-predicted minus long-run).
- 2012–2013 decomposition (selected values):
  - Short-term component in 2012 and 2013: -0.9 percent for both years.
  - Long-run import growth:
    - 2012: 4.2 percent.
    - 2013: 3.9 percent.
    - Pre-Great Recession average long-term import growth: 5.8 percent.
    - Long 1990s long-term import growth average: 7.3 percent.
- 2013 vs benchmarks:
  - ECM import growth in 2013 was 1.9 percentage points lower than average ECM growth in the 2000s.
    - Approximately 48 percent of the 1.9 percentage-point decline is imputed to lower long-run growth (3.9 percent in 2013 vs 4.8 percent in the 2000s).
    - Remaining 52 percent attributed to cyclical factors (cyclical -0.9 percent in 2013 vs 0.1 percent in the 2000s).
  - Comparing 2013 vs the long 1990s:
    - Decline in long-run growth explains about 81 percent of the lower import growth; cyclical explains about 19 percent.
- Projections and scenarios:
  - WEO April 2014 projects world trade volumes to grow at an average annual rate of 5.2 percent in 2014-2019.
  - Model-based predictions:
    - WEO projections are lower than predictions using 2000-2013 estimates.
    - WEO projections are higher than predictions using 2008-2013 estimates.
  - Policy-relevant implication: if 2008-2013 estimates reflect a permanent structural change, current trade projections (e.g., WEO 5.2 percent) could be overly optimistic.
  - Formal tests support a structural break between pre- and post-Great Recession; however, 2008-2013 may be too short to rule out cyclical explanations.

### Regional decomposition and country contributions
- World trade elasticity is a weighted average of regional/country elasticities; changes can arise via three channels:
  - increase in import share of a region with lower elasticity;
  - increase in relative growth of a region with lower elasticity;
  - decrease in a region’s own import elasticity to its GDP.
- Group and country highlights:
  - Both advanced economies and emerging/low income economies show lower trade responsiveness in the 2000s relative to the 1990s; formal tests show structural breaks for both groups.
  - Key country contributions to change in world elasticity in the long 1990s:
    - China: 13 percent
    - United States: 20 percent
  - Contributions in the 2000s:
    - China: 32 percent
    - United States: 8 percent
- Selected country elasticity changes (yearly data):
  - United States: declined from 3.7 to 1.8 between the long 1990s and the 2000s (table reports 3.68 for 1986-2000 and 1.77*** for 2001-2013).
  - China: declined from 1.5 to 1.1 between the long 1990s and the 2000s (table reports 1.54*** for 1986-2000 and 1.10*** for 2001-2013).
- Shares in world imports (selected):
  - Advanced Economies: 1970-1985 = 75 ; 1986-2000 = 78 ; 2001-2013 = 71
  - United States: 1970-1985 = 18 ; 1986-2000 = 20 ; 2001-2013 = 19
  - Emerging Markets and Developing Economies: 1970-1985 = 25 ; 1986-2000 = 22 ; 2001-2013 = 28
  - China: 1970-1985 = 1 ; 1986-2000 = 3 ; 2001-2013 = 8

### Sectoral decomposition — goods, manufacturing, services, commodities
- Main driver of decline:
  - Fall in goods trade elasticity in the 2000s, driven by manufacturing trade—not by changing goods vs services composition.
- Selected numeric elasticities:
  - World manufacturing-trade long-run elasticity:
    - long 1990s: 2.6
    - 2000s: 0.8
  - United States manufacturing-trade elasticity:
    - 1990s: 2.8
    - 2000s: 1.1
  - China manufacturing-trade elasticity:
    - 1990s: 1.2
    - 2000s: 0.7
  - World services trade long-run elasticity:
    - long 1990s: 1.8
    - 2000s: 2.2
  - World commodity trade long-run elasticity:
    - long 1990s: 1.7
    - 2000s: 2.3
- Composition shares:
  - Share of services trade in total trade stable at about one-fifth in the past two decades.
  - Implication: changing composition of trade goods vs services does not explain the lower overall elasticity; decline in manufacturing-trade elasticity is principal.

### Vertical specialization and global supply chains — circumstantial evidence
- Hypothesis:
  - Expansion of global supply chains in the 1990s increased vertical specialization and boosted measured trade responsiveness to income; slowdown of that expansion in the 2000s reduced trade elasticity.
- China evidence:
  - Share of imports of parts and components in China’s merchandise exports:
    - Peak mid-1990s: 60 percent.
    - Current share reported in the source: approximately 35 percent.
  - Fall is more pronounced as a share of manufacturing exports.
  - WTO and authors’ reconciliation:
    - WTO (1995-2009) reports foreign inputs in China’s exports increased by 13.8 percent between 2000 and 2009 and China became more downstream between 1995-2008.
    - Authors note WTO figures include fuel and raw materials; commodity prices rose in 2000s; OECD-WTO TiVA data and further analysis show foreign inputs in exports declined by 3 percent after 2005.
  - Domestic input availability linked in part to growing FDI (Kee and Tang, 2014); geographical re-sourcing within China increased internal sourcing.
- World foreign value added trends:
  - Ratio of foreign value added to domestic value added increased by 8.4 percentage points between 1995 and 2005.
  - Increased by only 2.5 percentage points between 2005 and 2012.
  - Interpretation: global supply chains expanded at a slower pace after 2005.
- Gross vs value-added elasticities:
  - Long-run elasticities of gross trade to GDP decreased over time, approaching the lower and more stable estimates of trade elasticities in value-added terms (based on seven-year rolling estimates).

### Changes in composition of aggregate demand
- Differential import intensity:
  - Investment has a larger import content than private consumption and government spending.
- Evolution of world GDP composition:
  - Share of investment grew substantially faster than consumption before the Great Recession and then declined.
  - Government spending mirrored investment; private consumption was flat up to 2009 and then moderately increased.
- Conclusions on composition channel:
  - Weak investment helps explain the low trade elasticity for the post-Great Recession period.
  - Changing composition of demand cannot, on its own, explain the decline in long-run world trade elasticities throughout the 2000s.

### Changes in protectionism — measured effects and caveats
- Measured restrictiveness:
  - WTO indicators: share of world trade covered by new import-restricting measures since Great Recession modest; 2013 value = 1.3 percent for all WTO members and observers.
  - World Bank TTBs provide a similar picture.
- ECM augmented with protectionism (World Bank TTBs):
  - World-level protection variable not significant; long-run world elasticity essentially unaltered.
  - China: protection effects not significant.
  - United States: coefficient for long-run elasticity in the 2000s marginally larger when including TTBs (suggesting contingent protection may have had a minor role in declining US import responsiveness).
- Caveats:
  - Opaque non-tariff measures are not captured; slower pace of liberalization, rather than a surge in protection, may contribute to declining responsiveness.

### Conclusion and policy implications
- Main conclusions:
  - Rise of long-run world trade elasticity in the long 1990s and its decline in the 2000s; decline set in before the Great Recession.
  - Both long-term (structural) and short-term (cyclical) components contribute to the global trade slowdown; structural factors play a large role.
  - Slower pace of expansion of global supply chains is an important determinant of the trade slowdown.
  - No evidence that change in goods-vs-services composition or rise in protectionism are primary determinants of the lower trade elasticity in the 2000s.
  - Changing composition of GDP (smaller investment share after the crisis) can explain lower trade elasticity post-Great Recession but not the historical decline since early 2000s.
- Broader implications:
  - The changing long-term trade-income relationship partly reflects changing international production relations; a slower pace of finer international division of labor could mean world trade contributes less to global growth than in the long 1990s.
  - Considerable scope exists to enhance the international division of labor by integrating regions at the margin of global supply chains (South Asia, Africa, South America); timing and mechanisms are uncertain.

### Appendix 1 — selected decomposition formulas and table figures
- Definition: long-run world income elasticity of trade, σW, equals percentage change in total world imports (Wm) over percentage change in real world income (Wy).
- Decompositions:
  - By trade category: σW is weighted average of category elasticities εj weighted by mj/mW.
  - By region: σW equals sum over regions of (region import elasticity to own income) × (region share in world imports) × (elasticity of region income to world income).
- Selected Table 4 model predictions and components (preserved exactly):
  - 1970-1985: Trade growth (model prediction) 4.7% — Long term component 4.7% — Short term component -0.1%
  - 1986-2000: Trade growth (model prediction) 7.2% — Long term component 7.3% — Short term component -0.2%
  - 2001-2013: Trade growth (model prediction) 4.9% — Long term component 4.8% — Short term component 0.1%
  - 2001-2010: Trade growth (model prediction) 5.2% — Long term component 4.9% — Short term component 0.2%
  - 2001-2007: Trade growth (model prediction) 7.5% — Long term component 5.8% — Short term component 1.8%
  - 2011-2013: Trade growth (model prediction) 4.1% — Long term component 4.4% — Short term component -0.3%
  - 2011: Trade growth (model prediction) 5.9% — Long term component 5.1% — Short term component 0.9%
  - 2012: Trade growth (model prediction) 3.3% — Long term component 4.2% — Short term component -0.9%
  - 2013: Trade growth (model prediction) 3.0% — Long term component 3.9% — Short term component -0.9%
- Model vs WEO 2014-2015 comparisons (preserved table line):
  - World: WEO projections 2014-2015 = 4.6 ; Model predictions 2014-2015 using 2000-2013 estimates = 5.8 ; Model predictions 2014-2015 using 2008-2013 estimates = 2.3
  - Advanced Economies: WEO = 3.8 ; Model (2000-2013) = 6.0 ; Model (2008-2013) = 2.1
  - Emerging Markets and Developing Economies: WEO = 5.8 ; Model (2000-2013) = 6.5 ; Model (2008-2013) = 4.0

*Source: Excerpt from IMF working paper _wp1506 (provided PDF content).*

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

### _wp1506 - References

### Introduction
- After bouncing back from the historic low of the Great Recession in 2010, world trade grew by less than 3.0 percent in 2012 and 2013, compared to the pre-crisis average of 7.1 percent (1987-2007).
- The observed slowdown in trade has received attention (from, among others, Davies, 2013, and Krugman, 2013), but had not been subject to rigorous analysis in the source excerpt.

### Research questions highlighted
- Whether the slower growth in trade simply reflects the sluggishness of GDP or whether there is a deeper structural shift in the relationship between trade and GDP.
- If there is a structural shift:
  - Whether it is a post-crisis phenomenon or reflects longer-term factors.
  - What those longer-term factors might be.

### Empirical strategy (as stated in the source)
- The authors use a simple empirical strategy based on an error correction model.
- They first estimate the relationship between trade and income for the period 1970- (text truncated in source).

*Source: Excerpt from the provided PDF content unit.*

### 2013. While the global trade slowdown is a recent phenomenon, the analysis of a long time

### _wp1506 - 2013. While the global trade slowdown is a recent phenomenon, the analysis of a long time

### Key findings
- Long-term trade elasticity:
  - 1.3 between 1970 and 1985.
  - 2.2 in the period 1986-2000.
  - 1.3 in the 2000s.
  - For the entire sample (1970-2013), the long-run elasticity (- δ/γ) is 1.7.
- Quarterly estimates for the 2000s:
  - 1.5 in 2001q1-2007q4.
  - 0.7 in 2008q1-2013q4.
- Short-run elasticities (β):
  - 2.1 for 1970-1985 (annual data).
  - 2.8 for 1986-2000 (annual data).
  - 3.4 for 2000-2013 (annual data).
  - Quarterly: 1.5 for 1990s and 2.6 for the 2000s; impact elasticities 4.0 (2001q1-2007q4) and 2.2 (2008q1-2013q4).
- Speed of adjustment (–γ):
  - 0.58 in the long 1990s (if previous-period trade was 1 percent higher than predicted, current-period trade would be reduced by 0.58 percent to restore equilibrium).
  - 0.31 for the 2000s.
  - Speed of adjustment was lower, but not statistically significant, for 1970-1985.

### Stylized facts on trade growth (recent decades)
- Global trade volume growth:
  - 13 percent in 2010 (post-recession rebound).
  - 6.2 percent in 2011.
  - 2.8 percent in 2012.
  - 3.0 percent in 2013.
- Pre-crisis average trade growth: 7.1 percent (1987-2007).
- World GDP growth in recent years hovered around 3 percent.
- Regional performance:
  - Euro Area imports: declined by 1.1 percent in 2012; increased by 0.3 percent in 2013.
  - Imports of emerging economies increased by 5.6 percent in 2013.
- Structural observation:
  - Import volumes as a share of real GDP have been roughly constant for nearly a decade in China and the US.
  - In China, the share of imports of parts and components in total exports decreased from 60 percent in the mid-1990s to 35 percent today.

### Empirical approach
- Model used: Error Correction Model (ECM) of the form Δ ln m_t = α + β Δ ln y_t + γ ln m_{t-1} + δ ln y_{t-1} + ε_t.
  - Short-run trade elasticity = β.
  - Speed of adjustment = –γ.
  - Long-run trade elasticity = –δ/γ.
- Data:
  - Annual analysis: 1970-2013.
  - Quarterly analysis: 1991-2013, with focus on 2001q1-2007q4 (pre-Great Recession) and 2008q1-2013q4 (post-Great Recession).
- Tests:
  - Non-linear Wald tests indicate significant structural breaks: the long 1990s (1986-2000) differ from preceding and subsequent periods.
  - Recursive estimation tests confirm results (details not presented).
  - Residuals are stationary; Breusch-Godfrey LM tests generally accept no serial correlation in residuals.

### Decomposition of the slowdown and projections
- Decomposition result (2000s focus):
  - Close to half of the slow trade performance in 2013 results from the lower long-run component of trade growth; the rest is explained by cyclical factors.
- Projection insight:
  - Predicted trade growth for 2014-2019 depends on whether the 2008-2013 estimates are “transitional” or reflect a genuine change in the long-term trade-income association.
  - If the lower post-2008 elasticities are structural, GDP pick-up may not produce the dramatic trade growth of the 1990s.

### Possible explanations for the decline in trade elasticity
- Primary candidate: changes in international vertical specialization (global supply chains):
  - Rise in long-run elasticity during the 1990s coincided with fragmentation of production into global supply chains.
  - Decrease in the 2000s associated with deceleration of that process.
  - Evidence highlighted: Chinese exporters now use more domestically produced inputs; the share of Chinese imports of parts and components in total exports fell from 60 percent in the mid-1990s to 35 percent today.
- Other potential contributors (not found to be as prominent in preliminary evidence): protectionism and changing aggregate demand structure.

### Relation to existing literature and main conclusion
- Relation to literature:
  - Builds on work documenting historical variation in world trade elasticity and analyses of the 2008-2009 trade collapse.
  - Complements studies that attribute the 2008-2009 collapse to sharp contraction in aggregate demand and amplifying factors (inventory effects, trade credit constraints).
- Main conclusion:
  - The trade-income relationship experienced a structural change: the short-run impact of income shocks on trade has increased over time, while the long-run responsiveness of trade to income has attenuated.
  - The global trade slowdown is driven in significant part by this lower long-run component, implying that even with stronger GDP growth, trade may not return to the rapid expansion seen in the 1990s.

*Source: IMF working paper content (excerpt).*

### Section V will investigate the potential causes of the changing long-run relationship between

### _wp1506 - Section V will investigate the potential causes of the changing long-run relationship between

### A. Cyclical versus Structural Factors — decomposition and key empirical findings
- Purpose: decompose import growth into structural (long-run) and cyclical (short-run) components using the estimated ECM framework.
- Key methodology points (as presented):
  - Predicted import growth uses estimated coefficients from equation (1) to produce equation (2).
  - The cointegration residual gives the long-run relationship; long-run predicted log level of imports is used to compute long-run import growth (equations (5)–(7) in the source).
  - Short-term component of import growth is obtained by subtracting predicted long-run growth from the model-predicted import growth (equation (8)).
- Empirical performance of the ECM:
  - The model "performs reasonably well in predicting trade growth over the past 43 years."
  - Long-run portion of import growth tends to explain most of total growth for most of the sample.
  - Cyclical effects dominate during crisis and post-crisis recovery episodes (e.g., 2008–2009 collapse and 2010 rebound).
- 2012–2013 slowdown decomposition:
  - Short-term component of import growth (cyclical) in 2012 and 2013: -0.9 percent for both years (Table 3 referenced).
  - Long-run import growth components:
    - 2012 long-term import growth: 4.2 percent.
    - 2013 long-term import growth: 3.9 percent.
    - Pre-Great Recession average long-term import growth: 5.8 percent.
    - Long 1990s long-term import growth average: 7.3 percent.
- Focus on 2013 relative to benchmarks:
  - ECM import growth in 2013 was 1.9 percentage points lower than average ECM growth in the 2000s.
  - Comparing 2013 vs the 2000s:
    - Approximately 48 percent of the 1.9 percentage-point decline is imputed to lower long-run growth (long-run 3.9 percent in 2013 vs 4.8 percent in the 2000s).
    - Remaining 52 percent is attributed to cyclical factors (cyclical was -0.9 percent in 2013 vs 0.1 percent in the 2000s).
  - Comparing 2013 vs the long 1990s:
    - Decline in long-run growth explains about 81 percent of the lower import growth.
    - Cyclical component explains about 19 percent.
  - Interpretation: using the 2000s as the benchmark (described as the “new normal”), cyclical factors explain a larger share of the 2013 decline than when comparing to the long 1990s.

### B. Trade projections and scenario comparison
- Central observation: lower long-run elasticity of trade with respect to GDP has contributed to lower-than-expected world trade performance; this structural change implies moderate trade growth even if cyclical conditions improve.
- WEO baseline:
  - World Economic Outlook (WEO) April 2014 projects world trade volumes to grow at an average annual rate of 5.2 percent in the period 2014-2019.
- Model-based comparisons:
  - Using regression estimates for 2000-2013 and 2008-2013 yields different implied trade growth predictions:
    - WEO projections are lower than predictions using the longer-period (2000-2013) implied trade elasticity.
    - WEO projections are higher than predictions using the 2008-2013 estimates.
  - Policy-relevant implication:
    - If the 2008-2013 estimates reflect a further structural and permanent change in the trade-income relationship (rather than a transitional phenomenon), then current trade projections (e.g., WEO 5.2 percent) could be overly optimistic.
- Uncertainty:
  - Formal tests indicate a structural break between pre- and post-Great Recession trade-income relationship (supporting the more pessimistic scenario).
  - But 2008-2013 sample may be too short to rule out cyclical explanations; hence, final assessment remains uncertain.

### V. Factors explaining the decline in trade elasticity — overview
- Objective: investigate possible causes of the lower long-run trade elasticity:
  - (i) Changes in the composition of world trade (goods vs services).
  - (ii) Changes in the structure of trade (international fragmentation of production / vertical specialization).
  - (iii) Changes in the composition of GDP (e.g., share of investment).
  - (iv) Changes in the trade regime (e.g., rise in protectionism).
- Caveats:
  - Explanations are neither exclusive nor exhaustive.
  - Other possible influences (not directly tested here) include rising wages in some developing countries and the shale gas revolution in the United States.

### V.A Regional decomposition of the long-run trade elasticity — empirical highlights
- Decomposition logic: world trade elasticity is a weighted average of regional (or country) elasticities; change in world elasticity can result from:
  - an increase in import share of a region with lower elasticity;
  - an increase in relative growth of a region with lower elasticity;
  - a decrease in a region’s own import elasticity to its GDP.
- Group and country findings:
  - Both advanced economies and emerging/low income economies show lower trade responsiveness in the 2000s relative to the 1990s; formal tests show structural breaks for both groups.
  - A few countries drive much of the change in world long-run trade elasticity due to their large trade shares and/or faster growth:
    - China and the United States are particularly important:
      - Contribution to change in world trade elasticity in the long 1990s: China 13 percent, United States 20 percent.
      - Contribution in the 2000s: China 32 percent, United States 8 percent.
  - Country-level elasticity changes (yearly data reported):
    - United States trade elasticity declined from 3.7 to 1.8 between the long 1990s and the 2000s.
    - China trade elasticity declined from 1.5 to 1.1 between the long 1990s and the 2000s.
    - (Quarterly data provide slightly different figures but confirm substantial declines.)
  - Euro Area: displayed an increasing elasticity of imports to its own income over the 2000s.
  - Latin America and the Caribbean: decreasing trade elasticity but limited weight in explaining the change in world trade responsiveness.

### V.B Changes in the composition of world trade — goods vs services and sector findings
- Decomposition of world trade elasticity into components shows main driver of decline:
  - Fall in goods trade elasticity in the 2000s, driven by manufacturing trade—not by changing composition of trade between goods and services.
- Key numeric results:
  - World manufacturing-trade long-run elasticity:
    - long 1990s: 2.6
    - 2000s: 0.8
  - United States manufacturing-trade elasticity:
    - 1990s: 2.8
    - 2000s: 1.1
  - China manufacturing-trade elasticity:
    - 1990s: 1.2
    - 2000s: 0.7
  - World services trade long-run elasticity:
    - long 1990s: 1.8
    - 2000s: 2.2
  - World commodity trade long-run elasticity:
    - long 1990s: 1.7
    - 2000s: 2.3
- Composition share:
  - Share of services trade in total trade has been stable at about one-fifth in the past two decades.
  - Implication: changing composition of trade (goods vs services) does not explain the lower overall elasticity; instead, the decline in manufacturing-trade elasticity is the principal factor.

### V.C Changes in vertical specialization — circumstantial evidence on global supply chains
- Hypothesis: the expansion of global supply chains in the 1990s (ICT-driven) increased vertical specialization and boosted measured trade responsiveness to income; the slowing of that process in the 2000s could reduce trade elasticity.
- Circumstantial evidence from China–United States manufacturing supply chains:
  - Falling share of imports of parts and components in China’s merchandise exports:
    - Peak in mid-1990s: 60 percent.
    - Current share reported in the source: approximately 35 percent.
  - Fall is even more pronounced as a share of manufacturing exports.
  - Interpretation: substitution of domestic inputs for foreign inputs by Chinese firms; corroborated by evidence of increasing domestic value added in Chinese firms (Kee and Tang, 2014).
- Implication for elasticities:
  - The diminishing importance of cross-border parts-and-components trade reduces measured responsiveness of manufacturing trade to income.
  - The long 1990s’ higher trade elasticity is consistent with rapid expansion of vertical specialization; the 2000s’ lower elasticity is consistent with a slowdown in that expansion.
- Additional note:
  - The paper finds that the elasticity of durables trade decreased from 2.7 in the long 1990s to a lower value in the 2000s (table reference in source).

*https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1506.pdf*

### 0.8 in the 2000s. This is consistent with the changing structure of global supply chains, which are more

### _wp1506 - 0.8 in the 2000s. This is consistent with the changing structure of global supply chains, which are more

### Vertical specialization and global supply chains
- WTO (2014 a and b) findings cited:
  - Foreign inputs contained in China’s exports increased by 13.8 percent between 2000 and 2009 (“Evolution of GVC participation and its components, selected Asian economies,” in WTO 2014b).
  - Between 1995 and 2008 China’s position became more downstream (Figure C.7 in WTO 2014a).
- Reconciling differences with the paper’s conclusions:
  - WTO figures include imports of parts and components as well as fuel and raw materials.
  - Commodity input prices increased significantly in the 2000s; lumping components and raw materials can create the impression of higher import dependence.
  - The WTO data (to which the authors had access) show that China became more upstream since 2005 and that foreign inputs contained in exports actually declined by 3 percent after 2005.
- Domestic input availability and FDI:
  - Enhanced availability of inputs domestically is linked in part to growing foreign direct investment in these industries (Kee and Tang, 2014).
  - Geographical re-sourcing within China: coastal regions began to source relatively more from the Chinese interior as transport and communication costs fell faster domestically than with the rest of the world.
  - Trade integration may be shifting toward greater internal trade rather than international trade captured by official statistics.

### Trends in trade composition and import shares
- United States:
  - US manufacturing imports as a share of GDP have been stable at around 8 percent since the turn of the century, after nearly doubling in the preceding decade-and-a-half.
- China:
  - Chinese manufacturing imports as a share of GDP rose from 10 percent at the beginning of the long 1990s to almost 30 percent in the early 2000s and have sharply declined since then.

### Global supply chains and trade elasticities
- Change in foreign value added in world gross exports:
  - Ratio of foreign value added to domestic value added increased by 8.4 percentage points between 1995 and 2005.
  - Increased by only 2.5 percentage points between 2005 and 2012 (Figure 10).
  - Interpretation: global supply chains are expanding at a slower pace after 2005.
- Methodology:
  - Long-run trade elasticities in value-added terms estimated on a seven-year rolling basis and compared with gross trade elasticities calculated similarly.
  - Intuition: if slower expansion of global supply chains contributes to the trade slowdown, the gap between gross and value-added trade elasticities should close over time.
- Empirical finding:
  - World long-run elasticities of gross trade to GDP decreased over time, approaching the lower and more stable estimates of trade elasticities in value-added terms (Figure 11).
- Data note:
  - Data on world domestic value added and foreign value added in gross exports from the OECD-WTO dataset are only available starting from 1995 and for selected years.
  - The authors use a time series created by Duval et al. (2014) by interpolating the OECD-WTO data; further details are in the technical appendix of Duval et al. (2014).

### Changes in the composition of aggregate demand
- Differential import intensity:
  - Investment has a larger import content than private consumption and especially government spending.
- Evolution of world GDP composition (Figure 12):
  - Share of investment in world aggregate demand grew substantially faster than private and public consumption before the Great Recession and then declined.
  - Government spending as a share of GDP was virtually the mirror image of investment.
  - Share of private consumption was flat up to 2009 and then moderately increased.
- Conclusions on composition channel:
  - Weak investment helps explain the low trade elasticity for the post-Great Recession period documented in Section III.
  - However, the changing composition of demand cannot, on its own, explain the decline in long-run world trade elasticities throughout the 2000s.
  - If demand composition were the driving channel, the world trade elasticity in the pre-Great Recession period should have been increasing (due to rising investment share), which is not found in the data.
- Supporting evidence:
  - Boz et al. (2014) provide evidence consistent with the investment channel using “import-intensity adjusted” demand models based on Bussiere et al. (2013).

### Changes in protectionism
- Measured changes in trade restrictiveness:
  - WTO trade restrictiveness indicators show a modest increase in the share of world trade covered by new import-restricting measures since the Great Recession.
  - For 2013, this share was at 1.3 percent for all WTO member and observer countries (Figure 13).
  - World Bank data on Temporary Trade Barriers (TTBs) provide a similar picture.
- Empirical test:
  - The authors augment the error correction model (ECM) to include a protectionism variable using the World Bank’s TTBs dataset.
  - Results:
    - For the world, the estimate of the long-run elasticity of world trade with respect to income is essentially unaltered and remains significant; the world-level protection variable is not significant.
    - For China, the augmented framework yields similar non-significant protection effects.
    - For the United States, the estimated coefficient for the long-run trade elasticity in the 2000s is marginally larger when including TTBs, suggesting contingent protection may have had a minor role in the decline of import responsiveness in the United States.
- Caveats:
  - Changes in opaque non-tariff measures are not captured by standard measures of protection and may play a role the analysis cannot account for.
  - A slower pace of trade liberalization in the 2000s relative to the long 1990s, rather than a surge in protection, may contribute to explaining the declining responsiveness of world trade to GDP.

### Conclusion and implications
- Main findings:
  - Rise of the long-run world trade elasticity in the long 1990s and its decline in the 2000s, with the decline setting in before the Great Recession.
  - Both long-term and short-term components contribute to the global trade slowdown; structural factors play a large role in recent low world trade growth rates.
  - Slower pace of expansion of global supply chains is an important determinant of the trade slowdown.
  - No evidence that change in composition of trade or rise in protectionism are primary determinants of the lower trade elasticity in the 2000s.
  - Changing composition of GDP, especially smaller investment share after the crisis, can explain the lower trade elasticity in the post-Great Recession period but not the historical decline since the early 2000s.
- Broader implications:
  - The changing long-term relationship between trade and income partly reflects changing international production relations.
  - If a finer international division of labor is isomorphic to factor-augmenting technical change, a slower pace of its expansion could indicate that world trade is contributing less to global growth today than in the long 1990s.
  - There remains considerable scope to enhance the international division of labor by integrating regions at the margin of global supply chains, such as South Asia, Africa and South America; timing and mechanisms for seizing these opportunities are uncertain.

*Source: Excerpt from the provided IMF working paper content.*

### Appendix 1 - Decomposition of the World Trade Elasticity

### Appendix 1 - Decomposition of the World Trade Elasticity

### Definition and basic decomposition by trade category
- Long-run world income elasticity of trade, σW, is defined as the percentage change in total world imports (or exports) in volume terms (Wm) over the percentage change in real world income (Wy): Wm/Wy = σW (expressed in the source as W/ / WW WW mm yy σ Δ = Δ).
- Decomposition across two trade categories (1 and 2, e.g., goods and services):
  - Wm = m1 + m2; Wy = y1 + y2 (subscripts used to differentiate aggregate variables from components).
  - σW can be written as a weighted average of the elasticities of different trade categories: the elasticity of goods/services imports to global income (εj) weighted by the share mj/mW (with j = 1, 2).
- Key implications from this decomposition:
  - For given trade weights, σW declines if the responsiveness to GDP of goods trade and/or services trade declines.
  - For given trade elasticities, σW declines if the share of world imports of the trade category with lower elasticity increases over time.

### Decomposition by region (or country)
- σW can also be expressed as a weighted average of regions’ import elasticities to their own income.
  - For two regions (region 1 and 2), the world elasticity equals the sum across regions of: (elasticity of region i's imports to its own income) × (region i’s share in world imports) × (elasticity of region i’s income to world income).
- Three channels that can reduce σW:
  - A decline in a region’s elasticity of imports to its own income.
  - An increasing import share of a region with lower trade elasticity.
  - An increasing relative growth of a region with lower trade elasticity.

### Empirical estimates and key numeric results from estimations (selected values preserved exactly)
- Table 1 (Results of Estimations Using Yearly Data), pooled without dummy variables (1970-2013):
  - α = -0.43** (0.17)
  - Short-run elasticity (β) = 2.82*** (0.36)
  - Speed of adjustment (-γ) = 0.12** (0.05)
  - δ = 0.20** (0.09)
  - Long-run elasticity (-δ/γ) = 1.70***
  - Breusch-Godfrey LM test for serial correlation = 9.67**
  - Rsquared = 0.740
  - N = 43
- Table 1, pooled with dummy variables for separate periods (1991q2-2000q4; 2001q1-2007q4; 2008q1-2013q4; 2001q1-2013q4), selected long-run elasticities (−δ/γ):
  - Period (1) 1991q2-2000q4: long-run elasticity = 2.40***; α = -3.09*** (0.80)
  - Period (2) 2001q1-2007q4: long-run elasticity = 1.49***; α = -0.54 (0.46)
  - Period (3) 2008q1-2013q4: long-run elasticity = 0.68***; α = 1.01** (0.43)
  - Period (4) 2001q1-2013q4: long-run elasticity = 1.21***; α = -0.07 (0.11)
  - Rsquared for these quarterly specifications = 0.838 (columns (1)-(3)) and 0.812 (column (4)); N = 90 per column.
- Table 4 (Decomposition of Trade Growth — model predictions and components), periods and values:
  - 1970-1985: Trade growth (model prediction) 4.7% — Long term component 4.7% — Short term component -0.1%
  - 1986-2000: Trade growth (model prediction) 7.2% — Long term component 7.3% — Short term component -0.2%
  - 2001-2013: Trade growth (model prediction) 4.9% — Long term component 4.8% — Short term component 0.1%
  - 2001-2010: Trade growth (model prediction) 5.2% — Long term component 4.9% — Short term component 0.2%
  - 2001-2007: Trade growth (model prediction) 7.5% — Long term component 5.8% — Short term component 1.8%
  - 2011-2013: Trade growth (model prediction) 4.1% — Long term component 4.4% — Short term component -0.3%
  - 2011: Trade growth (model prediction) 5.9% — Long term component 5.1% — Short term component 0.9%
  - 2012: Trade growth (model prediction) 3.3% — Long term component 4.2% — Short term component -0.9%
  - 2013: Trade growth (model prediction) 3.0% — Long term component 3.9% — Short term component -0.9%
  - World / Advanced Economies / Emerging Markets and Developing Economies (tabular line in source):
    - World 4.6               5.8                     2.3
    - Advanced Economies 3.8               6.0                     2.1
    - Emerging Markets and Developing Economies 5.8               6.5                     4.0
    (columns correspond to: WEO projections 2014-2015 (percent) / Model predictions 2014-2015 using 2000-2013 estimates (percent) / Model predictions 2014-2015 using 2008-2013 estimates (percent) as laid out in the source.)
- Table 5 / Table 6 (Long Run Elasticity of Imports to GDP, by Country and Region — selected entries preserved exactly):
  - World: 1986-2000 = 2.18*** ; 2001-2013 = 1.31***
  - Advanced Economies (share in world imports, percent): 1970-1985 = 75 ; 1986-2000 = 78 ; 2001-2013 = 71
  - United States (share in world imports, percent): 1970-1985 = 18 ; 1986-2000 = 20 ; 2001-2013 = 19
  - Emerging Markets and Developing Economies (share in world imports, percent): 1970-1985 = 25 ; 1986-2000 = 22 ; 2001-2013 = 28
  - China (share in world imports, percent): 1970-1985 = 1 ; 1986-2000 = 3 ; 2001-2013 = 8
  - Selected country-region long-run elasticities from yearly estimations:
    - World: 1986-2000 2.18*** ; 2001-2013 1.31***
    - United States: 1986-2000 3.68 ; 2001-2013 1.77***
    - China: 1986-2000 1.54*** ; 2001-2013 1.10***
  - From quarterly estimations (selected):
    - World: 1991q2-2000q4 long-run elasticity = 2.40***
    - World: 2001q1-2013q4 long-run elasticity = 1.21***

### Assessing the impact of protectionism (selected entries)
- Table 7 (Assessing the Impact of Protectionism on Long Run Elasticity of Imports to GDP):
  - World:
    - 1986-2000: total imports long-run elasticity 2.18*** ; accounting for protectionism 2.04***
    - 2001-2013: total imports long-run elasticity 1.31*** ; accounting for protectionism 1.3***
  - United States:
    - 1986-2000: total imports long-run elasticity 3.68 ; accounting for protectionism 2.62***
    - 2001-2013: total imports long-run elasticity 1.77*** ; accounting for protectionism 1.86***
  - China:
    - 1986-2000: total imports long-run elasticity 1.54*** ; accounting for protectionism -
    - 2001-2013: total imports long-run elasticity 1.10*** ; accounting for protectionism 1.07***

### Figures and conceptual charts (sources and notes preserved)
- Figures referenced and data sources:
  - Figures 1–13 include Total World Trade Volumes and Growth Rates, Import/Export Volumes (2005=100), Share of Import Volumes in Real GDP, Average Growth Rates, Decomposition of Trade Growth, Trade Predictions, China’s Share of Imports of Parts and Components in Exports of Merchandise and Manufacturing (percent), Manufacturing Imports, Ratio of Foreign Value Added to Domestic Value Added in World Gross Exports (percent), Long Run Elasticities (7-year periods), Share of Real Investment and Consumption in World Real GDP (index, 2000=100), Trade Covered by Import-Restrictive Measures of All WTO Members and Observers (percent of world merchandise imports).
  - Data sources explicitly cited in figures: IMF World Economic Outlook (WEO), IMF IFS, UN Comtrade, Duval et al. (2014) and OECD-WTO TiVA Database, WTO.
  - Note on classification for parts and components (UN Comtrade’s BEC): sum of BEC categories 42, 53, 22 as listed in the source.

*Source: Appendix 1 - Decomposition of the World Trade Elasticity, _wp1506 - Appendix 1 - Decomposition of the World Trade Elasticity*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1506.pdf_