## wpiea2025129-print-pdf

## Source details

**Canonical URL:** [wpiea2025129-print-pdf](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025129-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025129-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025129-print-pdf.pdf.json)

---

### Overview and research question
- Since the advent of the US-China trade tensions, debate has grown on the impact of rising barriers to trade on global and bilateral trade patterns.
- Two channels for “connector” countries are disentangled:
  - Trade reallocation: connector countries produce more goods domestically to export to the US—with rising domestic value added—as the US shifts away from China, likely using intermediate inputs imported from China.
  - Trade rerouting: connector countries serve as a one-stop place for Chinese exports to the US with minimal to no domestic value added (in the extreme case through transshipment) to circumvent trade barriers, especially for strategic products (e.g., electronics and chemicals).

### Scope and data
- Focus countries: India, Indonesia, Malaysia, Philippines, Thailand, and Vietnam.
- Primary MRIO database: Eora26 Global Supply Chain database covering 189 countries and 26 sectors spanning 1990 to 2022.
- Other data:
  - Orbis Cross-Border Investment database (greenfield FDI projects announced 2003–2022; focus on projects with latest status either completed or announced).
  - Matching characteristics in synthetic control: value-added measures 2010–2017, nominal GDP in US dollar based on purchasing power parity in 2010 and 2015, exports of goods as a share of GDP in 2010 and 2015 (from IMF World Economic Outlook Database).
- Data adjustment: to account for a visible seam between 2015 and 2016 in Eora, value added measures from 2016 onwards are adjusted so that the adjusted 2016 value equals the linearly-extrapolated value based on 2010–2015 while keeping the post-2016 trend unchanged.

### Analytical approach and inference
- Construct two value-added measures in a country’s exports to the US using MRIO:
  - Share of domestic value added in a country’s exports to the US.
  - Share of Chinese value added in a country’s exports to the US.
- Causal estimation: synthetic control approach (treatment period set as years from 2018 onwards).
- Hypothesis I (trade reallocation): a higher share of domestic value added in a country’s strategic sector exports to the US relative to the synthetic control (with statistical significance) implies trade reallocation.
- Hypothesis II (trade rerouting): a lower share of Chinese value added in a country’s strategic sector exports to the US relative to the synthetic control (with statistical significance) would reject significant trade rerouting.
- Statistical inference: placebo tests constructing a 10th−90th percentile range from the placebo distribution, excluding countries with pretreatment MSPE more than five times the MSPE of the country at hand. An estimated effect is statistically significant at the 90 percent confidence level if it lies outside the constructed 10th−90th percentile range.

### Main empirical findings (aggregate and country-level)
- Broad trade-flow observation (2013-17 vs 2018-22):
  - India, Malaysia, Thailand, and Vietnam: increases in shares in Chinese exports and increases in shares in US imports.
  - India, Indonesia, Malaysia, and Vietnam: increases in shares in Chinese imports.
  - All six countries: increases in shares in Chinese abroad-absorbed value added.
- General interpretation: rising imports from China together with rising exports to the US can reflect production relocation, import-back of processed products by Chinese firms, and growing domestic markets—so aggregate trade-flow correlations do not by themselves imply rerouting.

### Evidence of trade reallocation (strategic sectors)
- Strategic sectors mapped to MRIO sectors: (i) electrical and machinery; and (ii) petroleum, chemical and non-metallic mineral products.
- Electrical and machinery:
  - Vietnam: share of domestic value added increased since 2018, rising above 40 percent while the counterfactual shows a gradual decline.
  - Vietnam: increase of 6 percentage points in 2018, accumulating to more than 10 percentage points in 2022 (positive and statistically significant since 2018).
  - Malaysia: positive effect but not statistically significant at the 90 percent confidence level.
  - India, Indonesia, Philippines, Thailand: exhibited lower shares of domestic value added than their synthetic counterfactuals; Thailand’s negative effect is statistically significant.
- Petroleum, chemical and non-metallic mineral products:
  - Vietnam: increase of 7 percentage points in 2018, accumulating to almost 12 percentage points in 2022 (positive and statistically significant).
  - Malaysia: positive effect but not statistically significant.
  - Other countries: negative but not statistically significant effects.
- Overall: strongest and statistically significant trade reallocation evidence is for Vietnam in the strategic sectors examined.

### Evidence on trade rerouting (share of Chinese value added)
- Method: compare actual share of Chinese value added in a country’s exports to the US with a synthetic counterfactual (weighted average excluding the six countries of interest).
- Electrical and machinery:
  - Malaysia and Vietnam: observed lower share of Chinese value added since 2018 relative to synthetic counterfactuals.
  - Vietnam: negative effects are statistically significant at 90 percent confidence level from 2019 onwards.
  - India, Thailand, Philippines: positive and statistically significant effects during 2018-22.
  - Indonesia: positive and statistically significant effects during 2018-19 only.
- Petroleum, chemical and non-metallic mineral products:
  - Vietnam: negative and statistically significant effect on the share of Chinese value added in its exports to the US.
  - Other countries (except India): negative but not statistically significant effects.
- Interpretation:
  - For Vietnam, results reject significant trade rerouting: Vietnam shows statistically significant declines in Chinese value added in its strategic sector exports to the US, consistent with reallocation rather than transshipment.
  - Positive effects in other countries can reflect deeper supply chain integration with China (backward linkages) or relocation of assembly/packaging that sources inputs from China and exports to the US—not definitive evidence of tariff-circumventing rerouting.

### FDI dynamics and Vietnam
- Vietnam attracted more cumulative greenfield FDI projects committing Chinese capital in strategic sectors after 2018, likely supporting scaling up of domestic production.
- FDI inflows to Vietnam from other countries (including the US, Japan, and Korea) also significantly increased over the same period.
- Appendix G presents cumulative greenfield FDI project trends and synthetic-counterfactual estimates (Orbis Cross-border Investment Database used as source).

### Robustness checks, confounders, and identification threats
- Pandemic timing and design features:
  - Covid-19 pandemic spans only the later half of the post-treatment period; bulk of significant evidence is captured in 2018-2019.
  - If pandemic was a simultaneous global shock, it would affect countries in the synthetic control group similarly, mitigating confounding on relative value-added distribution.
- Non-strategic manufacturing replication:
  - Vietnam’s positive effects on non-manufacturing sectors are more muted and "not statistically significant until 2022", whereas strategic sectors show positive and statistically significant effects throughout 2018-22.
- Structural factors:
  - Value-added shares are slow-moving; pretreatment matching includes dynamics of the variable of interest over almost one decade plus economy and export sizes, which should account for gradual structural changes.
  - Failure to reject trade rerouting in some cases could reflect incentives for Chinese firms to relocate production to connector countries rather than pure transshipment.
- Placebo/synthetic-control construction:
  - Appendix E shows top countries used to construct synthetic counterfactuals (largest weights that add up to "70 percent"); frequently mentioned connector countries such as Mexico and Poland do not receive large weights in many synthetic controls.
  - Placebo distribution excludes countries with pretreatment MSPE > five times the treated country’s MSPE; effect on Vietnam remains statistically significant under these specifications.

### Interpretation, policy-relevant implications, and risks
- Distinguishing reallocation from rerouting matters for domestic gains:
  - Trade reallocation may boost domestic production and attract inward FDI, benefiting the connector country at least in the short term.
  - Trade rerouting with minimal domestic production raises vulnerability to countervailing trade restrictions and offers limited domestic gains.
- For Vietnam specifically:
  - Evidence indicates trade reallocation during 2018-2022: domestic production scaled up in strategic sectors, domestic content embedded in exports to the US increased, and Chinese value added in those exports declined.
  - Vietnam’s connector role manifested through FDI and Chinese production shifting to Vietnam—motives include tariffs and "lower labor costs".
  - Benefits could include higher employment and wages, technological spillovers, and deeper domestic firms’ global value chain integration, but Vietnam’s “dual economy” may limit technological diffusion and productivity spillovers.
- Broader risks of geoeconomic fragmentation:
  - In relative and short-term terms, non-aligned countries that maintain open trade to all partners may suffer the least from fragmentation.
  - Export-oriented, globally integrated countries remain vulnerable to disruptive commodity prices, unexpected protective trade policies, and overall decline in external demand from a shrinking global economy.
  - In the long run, all countries stand to lose from geoeconomic fragmentation.

*Source: wpiea2025129-print-pdf - Introduction (IMF).*

### Introduction ...........................................................................................................

### Introduction

### Overview and research question
- Since the advent of the US-China trade tensions, there has been a growing debate on the impact of rising barriers to trade on global and bilateral trade patterns.
- Geoeconomic fragmentation poses losses to all countries in the medium to long term, but may lead to different trade dynamics in the short-term across countries with different structural characteristics.
- The paper focuses on disentangling two channels by which so-called “connector” countries might relate to changing US–China trade patterns:
  - Trade reallocation: connector countries produce more goods domestically to export to the US—with rising domestic value added—as the US shifts away from China, likely using intermediate inputs imported from China.
  - Trade rerouting: connector countries serve as a one-stop place for Chinese exports to the US with minimal to no domestic value added (in the extreme case through transshipment) to circumvent trade barriers, especially for products identified as having strategic importance (e.g., electronics and chemicals).

### Scope and data
- The paper zooms in on six Asian emerging markets: India, Indonesia, Malaysia, Philippines, Thailand, and Vietnam.
- It leverages granular trade and FDI data and a multi-region input-output database to examine value-added components of exports and the role of greenfield FDI.

### Analytical approach
- Construct two value-added measures in a country’s exports to the US using a multi-region input-output database:
  - Share of domestic value added in a country’s exports to the US.
  - Share of Chinese value added in a country’s exports to the US.
- Use a synthetic control approach to estimate the effect of US–China trade tensions on these measures, with particular focus on strategic sectors including electrical and machinery and petroleum, chemical and non-metallic mineral products (mapped to the database’s 26 sectors following International Monetary Fund (2023) guidance).

### Key empirical distinctions
- A higher share of domestic value added versus a synthetic counterfactual indicates evidence consistent with trade reallocation.
- A lower share of Chinese value added versus a synthetic counterfactual indicates no evidence of trade rerouting on a significant scale; conversely, an increase would be consistent with rerouting (Chinese exports re-exported to the US with minimal domestic value added).

---

### Main findings (aggregate and country-level)
- The positive correlation between these countries’ imports from China and exports to the US can reflect alternative forces (growing domestic markets, Chinese supply chain reconfiguration) and thus does not necessarily imply connector roles bridging US and China.
- Several countries (including India, Malaysia, and Vietnam) have seen an increase in their shares in Chinese imports and an increase in their shares in Chinese abroad-absorbed value added, pointing to:
  - Chinese firms relocating some production abroad (e.g., lower labor costs).
  - Growing domestic markets in connector countries enabling Chinese market penetration.
- Vietnam emerges as a clear case of trade reallocation, not trade rerouting, during 2018-2022:
  - For strategic sectors, Vietnam experienced statistically significant increases in the share of domestic value added in its exports to the US:
    - 6 and 7 percentage points higher than its synthetic counterfactual in 2018 for electrical and machinery sector and petroleum, chemical and non-metallic mineral products sector, respectively.
    - 10 and 12 percentage points higher in 2022 for those sectors, respectively.
  - Vietnam saw a statistically significant decline in the share of Chinese value added in its strategic sector exports to the US, suggesting no large-scale rerouting from China to the US to circumvent tariffs.
  - Vietnam expanded domestic content not only to the US but globally; global effects are slightly smaller and take longer to reach statistical significance.
- FDI dynamics in Vietnam:
  - Vietnam attracted more cumulative greenfield FDI projects committing Chinese capital in strategic sectors after 2018, likely supporting the scaling up of domestic production.
  - FDI inflows to Vietnam from other countries (including the US, Japan, and Korea) also significantly increased over the same period.
- Other five Asian countries:
  - Evidence of trade reallocation is elusive.
  - Malaysia: positive but not statistically significant effect on the share of domestic value added in its strategic sector exports to the US.
  - Other countries: negative and not statistically significant effects on domestic value added share in strategic sectors (except Thailand).
  - On trade rerouting (share of Chinese value added in exports to the US for strategic sectors):
    - Malaysia: estimated negative (but not statistically significant) effect.
    - Other countries: small positive and statistically significant effects; interpreted as increased supply chain integration with China (backward linkages) rather than conclusive evidence of rerouting.

---

### Interpretation and implications
- Distinguishing trade reallocation from trade rerouting is crucial:
  - Trade reallocation may boost domestic production and attract inward FDI, benefiting the connector country at least in the short term.
  - Trade rerouting involves minimal domestic production, raising vulnerability to countervailing trade restrictions and offering limited domestic gains.
- Aggregate trade-flow correlations (e.g., rising imports from China together with rising exports to the US) can overstate rerouting if value-added activities within product codes are unobserved.
- Using multi-region input-output linkages to construct value-added measures combined with synthetic control counterfactuals allows causal inference separating reallocation from rerouting.

---

### Contribution to literature
- Distinguishes trade reallocation versus trade rerouting empirically using value-added decomposition—offering novel evidence relative to prior work that largely relied on aggregate trade flows or correlations.
- Enriches the connector-country discussion by accounting for China’s production relocation and domestic market growth in candidate connector countries.
- Provides country-level and sector-level evidence—especially for Vietnam—complementing recent literature on labor markets, structural change, FDI flows, and environmental effects tied to US–China trade tensions.

---

### Structure of the paper
- Section 2 documents stylized facts related to geoeconomic fragmentation and connector countries, and analyzes potential forces behind increasing imports from China.
- Section 3 presents the empirical methodology and data sources.
- Section 4 presents empirical results and validation analysis.
- Section 5 concludes.

*Source: wpiea2025129-print-pdf - Introduction (IMF).*

### 2016.  Since 2017, it has started to increase again, not only in strategic sectors, but also

### wpiea2025129-print-pdf - 2016.  Since 2017, it has started to increase again, not only in strategic sectors, but also

### Trade patterns and connector countries
- Since 2017, Chinese exports have started to increase again not only in strategic sectors but also in non-strategic sectors and for the overall manufacturing sector.
- Two alternative interpretations for the broad-based increase:
  - Evidence of Chinese exports being channeled through other countries to the US to circumvent tariffs for strategic sectors.
  - Chinese firms’ production relocation and supply chain reconfiguration: supplying intermediate products and relocating low-value added production (assembly and packaging) abroad to countries with lower labor costs as labor costs in China increase and China moves up the value chain.
- Empirical observation (2013-17 vs 2018-22) for non-aligned countries:
  - India, Malaysia, Thailand, and Vietnam have seen both an increase in their shares in Chinese exports and an increase in their shares in US imports.
  - India, Indonesia, Malaysia, and Vietnam have also seen an increase in their shares in Chinese imports.
  - All six Asian countries of interest have seen an increase in their shares in Chinese abroad-absorbed value added (indicator of domestic market absorption of Chinese exported value added).
- Interpretation: increases in these countries’ shares in Chinese exports and US imports may reflect:
  - Production relocation and import-back of processed products by Chinese firms; and/or
  - Growing domestic markets that absorb more Chinese exported value added.
- Conclusion from trade flows: no direct evidence that Chinese exports are rerouted through other countries to the US. Connector roles may reflect other forces beyond trade rerouting.

### Analytical approach and hypotheses
- Methodology: synthetic control approach (Abadie et al. (2010)), treatment period set as years from 2018 onwards. Focus countries: India, Indonesia, Malaysia, Thailand, Philippines, and Vietnam.
- Value-added measures used (from a multi-region input-output database):
  - (i) share of domestic value added in a country-sector’s exports to the US;
  - (ii) share of Chinese value added in a country-sector’s exports to the US.
- Hypothesis I (trade reallocation):
  - There has been trade reallocation resulting from the US-China trade tensions, leading the country to produce more goods domestically to export to the US as the US shifts away from China, i.e.,
  - Domestic V alue Added_{t,outcome} / Total exports to US_{t,outcome} > Domestic V alue Added_{t,counterf actual} / Total exports to US_{t,counterf actual}
  - Operational test: a higher share of domestic value added in a country’s strategic sector exports to the US relative to the synthetic control (with statistical significance) implies trade reallocation.
- Hypothesis II (trade rerouting):
  - There has been significant trade rerouting resulting from the US-China trade tensions, i.e., importing Chinese goods and re-exporting them to the US with minimal to no domestic value added.
  - Define Y_{counterf actual t} ≡ Chinese V alue Added_{t,counterf actual} / Total exports to US_{t,counterf actual}
  - Define Y_{outcome t} ≡ (Chinese V alue Added_{t,counterf actual} + Δ_t) / (Total exports to US_{t,counterf actual} + Δ_t + ε_t)
  - If ε_t is zero or sufficiently small compared to Δ_t, then Y_{outcome t} > Y_{counterf actual t}. A lower share of Chinese value added in the outcome relative to the synthetic control (statistically significant) would reject the inequality and thus reject significant trade rerouting.
  - Caveat: a higher share of Chinese value added can also reflect deeper supply chain integration (backward linkages) rather than rerouting.

### Data
- Primary MRIO database: Eora26 Global Supply Chain database covering 189 countries and 26 sectors spanning 1990 to 2022.
- Adjustment: to account for a visible seam between 2015 and 2016 in Eora due to a structural break, value added measures from 2016 onwards are adjusted so that the adjusted 2016 value equals the linearly-extrapolated value based on 2010–2015 while keeping the post-2016 trend unchanged.
- Other data:
  - Orbis Cross-Border Investment database (greenfield FDI projects announced 2003–2022; focus on projects with latest status either completed or announced).
  - Matching characteristics in synthetic control: value-added measures 2010–2017 (pretreatment trend), nominal GDP in US dollar based on purchasing power parity in 2010 and 2015, exports of goods as a share of GDP in 2010 and 2015 (from IMF World Economic Outlook Database).
- Statistical inference: placebo tests following Abadie and Gardeazabal (2003), Bertrand et al. (2004), Abadie et al. (2010).
  - Construct a 10th−90th percentile range from the placebo distribution excluding countries with pretreatment MSPE more than five times the MSPE of the country at hand.
  - An estimated effect is statistically significant at the 90 percent confidence level if it lies outside the constructed 10th−90th percentile range.

### Empirical results: trade reallocation (strategic sectors)
- Strategic sectors mapped to MRIO sectors following IMF 2023 April WEO: (i) electrical and machinery; and (ii) petroleum, chemical and non-metallic mineral products.
- Electrical and machinery sector results:
  - Malaysia and Vietnam have seen higher shares of domestic value added in their exports to the US after 2018 compared to synthetic counterfactuals.
  - Vietnam’s share of domestic value added increased since 2018, rising above 40 percent, while the counterfactual shows a gradual decline.
  - India, Indonesia, Philippines, and Thailand exhibited lower shares of domestic value added than their synthetic counterfactuals.
  - Statistical significance (10th−90th percentile placebo range):
    - Vietnam: positive and statistically significant effects since 2018.
      - Vietnam has seen its share of domestic value added increase by 6 percentage points in 2018, accumulating to more than 10 percentage points in 2022.
    - Malaysia: positive effect but not statistically significant at the 90 percent confidence level.
    - Thailand: among countries with negative effects, Thailand’s negative effect is statistically significant.
- Petroleum, chemical and non-metallic mineral products sector results:
  - Malaysia and Vietnam have positive effects on the domestic content of their exports to the US versus synthetic counterfactuals.
  - Statistical significance:
    - Vietnam: positive and statistically significant.
      - Vietnam has seen its share of domestic value added increase by 7 percentage points in 2018, accumulating to almost 12 percentage points in 2022.
    - Other countries: negative but not statistically significant effects.
- Overall interpretation:
  - Vietnam has benefited from the trade reallocation effect for the strategic sectors examined (electrical and machinery; petroleum, chemical and non-metallic mineral products), showing statistically significant increases in the domestic content of its exports to the US.
  - Other countries in the sample do not exhibit statistically significant increases in domestic content of exports to the US for these strategic sectors.

### Findings and implications
- Trade reallocation evidence:
  - Strongest for Vietnam in strategic sectors: statistically significant increases in domestic value added embedded in exports to the US since 2018.
  - Malaysia shows positive but not statistically significant changes; India, Indonesia, Philippines, and Thailand do not show evidence of benefiting via increased domestic content (with Thailand showing a statistically significant negative effect in one instance).
- Trade rerouting evidence:
  - No trade-flow-based evidence supporting large-scale rerouting of Chinese exports through these countries to the US.
  - Apparent connector-country patterns can reflect production relocation, import-back of processed products, and growing domestic markets absorbing Chinese exported value added rather than tariff-circumvention rerouting alone.

*Source: IMF working paper content (Eora26 Global Supply Chain database analyses, synthetic control approach, placebo tests).*

### 4.2    Has there been trade rerouting?

### 4.2    Has there been trade rerouting?

### Method and interpretation
- The analysis tests whether there has been significant trade rerouting from China by examining the share of Chinese value added in a country’s exports to the US for strategic sectors.
- Comparison is made between actual shares and a synthetic counterfactual constructed as a weighted average of other countries’ shares (excluding the six countries of interest).
- Interpretation rules:
  - A lower share of Chinese value added, with statistical significance relative to the synthetic counterfactual, would provide evidence to reject the hypothesis that trade rerouting has occurred on a significant scale.
  - A higher share of Chinese value added, even if statistically significant, does not necessarily indicate trade rerouting; it could reflect increasing supply chain integration with China (particularly through backward linkages), or relocation of assembly/packaging that sources inputs from China but exports to the US.

### Findings for electrical and machinery sector (Figures 5 and 6)
- Country-level observed vs. synthetic trends:
  - Malaysia and Vietnam have seen lower share of Chinese value added in their exports to the US since 2018, compared to their synthetic counterfactuals.
  - India, Indonesia, Thailand, and Philippines have seen higher share of Chinese value added in their exports to the US.
- Estimated effects and statistical significance:
  - Malaysia and Vietnam have seen negative effects on the share of Chinese value added in their exports to the US.
  - Vietnam’s negative effects are statistically significant at 90 percent confidence level from 2019 onwards.
  - Positive effects are statistically significant in:
    - India, Thailand, and Philippines during 2018-22.
    - Indonesia during 2018-19 only.
- Petroleum, chemical and non-metallic mineral products:
  - Vietnam has seen a negative and statistically significant effect on the share of Chinese value added in its exports to the US.
  - Other countries except India have seen negative but not statistically significant effects (see Appendix D reference in source).

### Interpretation and overall conclusions from value-added measures
- Evidence rejects the null hypothesis of significant trade rerouting through Vietnam:
  - Vietnam has seen statistically significant negative effects on Chinese value added in its strategic sector exports to the US.
  - Conclusion: trade rerouting has not happened as a result of the US-China trade tensions through Vietnam on a significant scale.
- For other countries:
  - Failure to reject the null hypothesis does not imply evidence of trade rerouting.
  - Positive effects on the share of Chinese value added could reflect increased supply chain integration with China (e.g., relocation of assembly/packaging that sources inputs from China but exports to the US — example cited: Foxconn began assembling top-end Apple iPhones in India in 2019 (Reuters, 2018)).
- Synthesis:
  - Results from two value added–based measures suggest that, among the countries of interest, Vietnam has experienced a trade reallocation effect as the US shifts away from China and reconfigures supply chains towards Vietnam, rather than trade rerouting to circumvent tariffs.
  - Vietnam benefited from trade reallocation which boosted its domestic production and increased domestic content embedded in its exports to the US.
  - Effects (either trade reallocation or trade rerouting) in other countries remain elusive.

*Source: IMF working paper content (section 4.2 and related Figures discussed in the source).*

### 4.4    Discussions

### 4.4    Discussions

### Threats to identification and confounding factors
- Potential confounders that affect both strategic and non-strategic sectors could threaten identification (example: lockdowns during the Covid-19 pandemic; economy-wide structural reforms).
- Replication for non-strategic manufacturing sectors: average share of domestic value added in a country’s exports to the US across all non-strategic manufacturing sectors was calculated and examined for effects.
- Result: positive effects on Vietnam’s non-manufacturing sectors are more muted and "not statistically significant until 2022", indicating no statistically significant effects on the domestic content of its non-strategic manufacturing exports to the US.
- Contrast: strategic sectors show positive and statistically significant effects throughout the period of "2018-22".
- Implication: the difference between strategic and non-strategic sectors helps address potential confounding factors that affect all manufacturing sectors and supports attributing positive and statistically significant effects on strategic sectors to the US-China trade tensions.

### Robustness: pandemic timing, synthetic control group, and placebo tests
- Two research-design features mitigate concerns about pandemic confounding:
  - The Covid-19 pandemic spans only the later half of the post-treatment period, with the bulk of significant evidence already captured in the unaffected years "2018-2019".
  - If the pandemic was a simultaneous global shock, pandemic-related supply chain reconfiguration would affect countries in the synthetic control group in a similar manner, reducing confounding on relative distribution of value added.
- Placebo/synthetic-control construction:
  - Appendix E presents top countries—with the largest weights that add up to "70 percent"—used in constructing synthetic counterfactuals for each country.
  - Observation: frequently mentioned connector countries such as Mexico and Poland do not receive large weights in constructing synthetic counterfactuals.
  - All countries (except the six Asian countries in the sample, and China and the US) are included in constructing the "10th−90th percentile range" of the placebo test distribution.
  - Finding: the effect on Vietnam remains statistically significant even when including some other countries in the synthetic control group that may have been affected by US tariffs.

### Structural factors and interpretation of value-added measures
- Concern: other structural factors (for example, a relative increase in labor costs) may drive changes in domestic value added or Chinese value added around the time of the US-China trade tensions.
  - Example mechanism: an increase in domestic input costs in China may incentivize rerouting by Chinese firms; similar increases in connector countries could be reflected as an increasing share of domestic value added in their exports.
- Counterarguments presented:
  - The share of value added (domestic or Chinese) is slow-moving.
  - Pre-treatment matching characteristics include dynamics of the variable of interest over almost one decade as well as economy and export sizes; therefore, significant increases in domestic input costs or productivity growth should already be accounted for in constructing synthetic counterfactuals.
  - Failure to reject the trade rerouting hypothesis could still reflect Chinese firms’ incentive to relocate production to connector countries to benefit from favorable structural characteristics rather than mere transshipment.
  - Some structural drivers in sample and counterfactual countries predate the US-China trade tensions as part of a slower-moving structural transformation toward manufacturing and services.

### Implications for FDI and sectoral effects
- Positive and statistically significant effects on FDI are observed in strategic sectors only, not in non-strategic manufacturing sectors (see Appendix G for sector-specific results on petroleum, chemical and non-metallic mineral products sector and non-strategic manufacturing sectors).
- Synthesis of evidence for Vietnam (expanded in Conclusions):
  - Vietnam appears to have benefited from trade reallocation as a result of the US-China trade tensions during "2018-2022".
  - Vietnam’s domestic production of strategic sector exports to the US has scaled up, reflected in an increasing share of domestic value added embedded in its exports to the US for strategic sectors.
  - No evidence that Vietnam facilitated trade rerouting/transshipment of Chinese exports on a significant scale: Vietnam has seen a lower share of Chinese value added in its strategic sector exports to the US.
  - Positive effects on domestic value added also found in Vietnam’s strategic sectors’ exports to the rest of the world.
  - Vietnam has seen an increase in committed inward FDI from China in its strategic sectors.
  - Interpretation: Vietnam’s connector role has manifested through trade reallocation and Chinese production shifting to Vietnam (via FDI and factory setup), contributing to scale-up of domestic production of strategic sector products and more domestic content embedded in Vietnam’s exports, to the US and to the rest of the world; motives include tariffs and "lower labor costs".

### Broader implications and risks of fragmentation
- Benefits of rising trade and FDI inflows to the domestic sector remain unclear:
  - Potential benefits include higher employment and wages, technological spillovers, and deeper domestic firms’ global value chain integration.
  - Vietnam’s “dual economy” (disconnect between domestic and FDI sectors) may weigh on technological diffusion and productivity spillovers.
- Fragmentation risks:
  - In relative and short-term terms, non-aligned countries that maintain open trade to all partners may suffer the least from geoeconomic fragmentation.
  - Such export-oriented, globally integrated countries remain vulnerable to adverse shocks from intensified fragmentation: disruptive commodity prices, unexpected protective trade policies, and overall decline in external demand from a shrinking global economy.
  - In the long run, all countries stand to lose from geoeconomic fragmentation.

*Source: 4.4 Discussions (chapter excerpt) from the provided IMF working paper content.*

### Appendix C    Additional Empirical Results on the Share

### wpiea2025129-print-pdf - Appendix C    Additional Empirical Results on the Share

### Appendix C — Share of Domestic Value Added in Exports: scope and methods
- Sectors analyzed: petroleum, chemical and non-metallic mineral products; non-strategic manufacturing products.
- Two empirical objects presented for each sector and country:
  - Dynamics: actual values (blue lines) and synthetic counterfactuals (grey lines) of the share of domestic value added in a country’s exports to the US (In Percent).
  - Estimated effects: plotted as the estimated effect (blue lines) with the 10th−90th percentile range of the placebo test distribution (In Percentage Points).
- Synthetic counterfactual construction:
  - Based on a weighted average of the share of domestic value added in other countries’ exports to the US for the same sector, excluding the six countries of interest in the paper.
- Key methodological adjustments and markers:
  - Dashed vertical indicates the year 2018 when the US-China trade tensions started.
  - To account for a visible seam between 2015 and 2016 in the Eora Global Supply Chain Database due to a structural break, value added measures from 2016 onwards are adjusted so that the adjusted 2016 value is equal to the linearly-extrapolated value based on data during 2010 and 2015 while the post-2016 trend is kept unchanged.
  - Placebo distribution excludes countries that have pretreatment (before 2018) mean squared prediction error (MSPE) of more than five times the MSPE of the country in study.

### Appendix D — Share of Chinese Value Added in Exports: scope and methods
- Sectors analyzed: petroleum, chemical and non-metallic mineral products; non-strategic manufacturing products.
- Presentation mirrors Appendix C:
  - Dynamics: actual and synthetic counterfactual shares of Chinese value added in a country’s exports to the US (In Percent).
  - Estimated effects: plotted with the 10th−90th percentile range of placebo tests (In Percentage Points).
- Synthetic counterfactuals constructed analogously (weighted average excluding the six countries of interest).
- Same methodological adjustments as Appendix C:
  - Dashed vertical at 2018; adjustment of 2016 value to match linear extrapolation from 2010–2015 while keeping post-2016 trend unchanged.
  - Exclusion in placebo tests for pretreatment MSPE > five times the treated country’s MSPE.

### Appendix E — Countries used for constructing synthetic counterfactuals (structure and key property)
- Presents lists of top countries used to construct synthetic counterfactuals for:
  - The share of domestic value added (Tables 3–5).
  - The share of Chinese value added (Tables 6–8).
- Common specification across tables:
  - Each column lists the countries used for a given treated country, with the largest weights that add up to 70 percent.
- Table coverage highlights (as presented):
  - Table 3: Top countries for Electrical and Machinery (columns for Indonesia, India, Malaysia, Philippines, Thailand, Vietnam) with ranks shown up to at least Rank 40 in some columns.
  - Table 4: Top countries for Petroleum, Chemical and Non-Metallic Mineral Products (columns for Indonesia, India, Malaysia, Philippines, Thailand, Vietnam) with ranks shown up to at least Rank 44 in some columns.
  - Table 5: Top countries for Non-Strategic Manufacturing Products (columns for Indonesia, India, Malaysia, Philippines, Thailand, Vietnam) with ranks shown up to at least Rank 9 in the printed listing.
  - Table 6–8: Analogous listings for constructing synthetic counterfactuals for the share of Chinese value added across the same sectoral groupings (Electrical and Machinery; Petroleum, Chemical and Non-Metallic Mineral Products; Non-Strategic Manufacturing Products).
- Notes: All tables specify “Each column lists the countries used in constructing the synthetic counterfactual for each country in the sample, with the largest weights that add up to 70 percent.”

### Appendix F — Vietnam: Domestic Value Added in Exports to the World
- Focus: Vietnam’s share of domestic value added in exports to the world for:
  - Petroleum, chemical and non-metallic mineral products.
  - Non-strategic manufacturing products.
- For each sector the appendix presents:
  - Panel (a): Trend — actual (blue line) versus synthetic counterfactual (grey line) (In Percent).
  - Panel (b): Estimated effect — blue lines with the 10th−90th percentile placebo range; placebo excludes countries with pretreatment MSPE > five times Vietnam’s MSPE.
- Methodological markers repeated:
  - Dashed vertical indicates the year 2018 when the US-China trade tensions started.
  - Adjustment of 2016 values to equal the linear extrapolation from 2010–2015 while keeping post-2016 trend unchanged (Eora database seam correction).

### Appendix G — Cumulative number of Greenfield FDI projects that commit Chinese capital to Vietnam
- Sectors covered: petroleum, chemical and non-metallic mineral products; non-strategic manufacturing sector; and references to electrical and machinery sector in captions.
- For each sector the appendix presents:
  - Panel (a): Trend — actual (blue line) versus synthetic counterfactual (grey line) of the cumulative number of greenfield FDI projects that commit Chinese capital to Vietnam since 2003.
  - Panel (b): Estimated effect — blue lines with the 10th−90th percentile placebo range; placebo excludes countries with pretreatment MSPE > five times Vietnam’s MSPE.
- Synthetic counterfactual construction:
  - Based on a weighted average of the cumulative number of greenfield FDI projects that commit Chinese capital to other countries for the same sector since 2023, excluding Vietnam (as specified in figure notes).
- Data source for FDI figures: The Orbis Cross-border Investment Database.

*Sources: The Eora Global Supply Chain Database (Lenzen et al., 2012, Lenzen et al., 2013), Aslam et al. (2017), The Orbis Cross-border Investment Database, and authors’ calculations.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025129-print-pdf.pdf_