## CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC

## Source details

**Canonical URL:** [CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC](https://www.imf.org/-/media/files/publications/weo/2022/april/english/ch4.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/weo/2022/april/english/ch4.pdf.md)
- [Structured JSON version](/-/media/files/publications/weo/2022/april/english/ch4.pdf.json)

---

### Pandemic impact on trade: patterns and timing
- At its trough in the second quarter of 2020:
  - Volume of global trade in goods fell 12.2 percent (relative to last quarter of 2019).
  - Trade in services fell 21.4 percent (relative to last quarter of 2019).
- Recovery and sectoral patterns:
  - Trade in goods had recovered to pre-pandemic levels by October 2021.
  - Trade in services remained sluggish, driven mainly by the collapse of travel; transport services appear to be recovering while disruptions in seaborne trade remained elevated.
  - Trade in other services (notably telecommunication services) was more robust.
- Product- and industry-specific effects:
  - Exports of GVC-intensive goods fell 30 percent between January and April 2020; exports of other goods fell about 18 percent.
  - Initial drop was relatively more severe in some industries like automobiles, amid disruptions to key inputs such as semiconductors.

### Research questions and empirical approach
- Main questions:
  1. How well a standard model of demand and prices accounts for pandemic trade patterns, compared with previous large recessions.
  2. Which pandemic-specific factors determined trade patterns.
  3. What international spillover effects were generated by mobility restrictions.
- Empirical strategy:
  - Import demand model linking real import growth of goods and services to growth in demand and relative price of imports for a sample of 127 countries over 1985–2019.
  - Granular bilateral monthly trade data and a gravity model (Santos Silva and Tenreyro 2006 specification) on monthly imports at the six-digit product level from Trade Data Monitor to estimate international spillovers.
  - Oxford COVID-19 Government Response Stringency Index (0–100) used as a key measure (highly correlated with workplace closing component).

### Key empirical findings: model fit and forecast errors
- Demand and relative prices alone do not explain pandemic trade patterns:
  - Import demand model fits historical data up to 2019 but produces large forecast errors for 2020 when goods and services are considered separately.
  - For services: model predicts about –8 percent growth for 2020; actual services trade fell by 25 percent.
  - For goods: model predicts a 10 percent decline; observed fall was 6 percent.
- Estimated coefficients:
  - Coefficients on import-adjusted demand are positive for most countries and greater than 1.
  - Coefficients on relative price average between –0.2 and –0.3.

### Pandemic-specific determinants of trade deviations
- Excess import demand for goods:
  - Countries with more severe pandemic outbreaks, more stringent containment measures, or larger declines in mobility showed “excess import demand” for goods (fall in goods imports smaller than predicted).
  - Forecast error for goods imports was 3 percentage points more positive for countries in the third quartile of COVID-19 cases than for those in the first quartile.
- Services overprediction:
  - Overprediction for services was most pronounced in countries where travel services accounted for a large share of total service imports.
- Health preparedness:
  - Better health-preparedness of importers’ trade partners (measured by the Global Health Security Index) was associated with less negative deviations in goods imports.

### International spillovers from lockdowns (timing, magnitude, robustness)
- Aggregate spillover magnitudes and timing:
  - Lockdowns in a country’s trade partners on average accounted for up to 60 percent of the observed decline in imports in the first half of 2020.
  - Spillover effects first materialized in February 2020, grew in March and April, began declining in May, and by June were indistinguishable from zero as goods imports rebounded.
- Regression evidence:
  - Estimated coefficient from a weighted regression of the change in imports between 2020:Q2 and 2019:Q4 against partner-country Oxford Stringency Index is –0.015 (t-stat = –2.44).
- Robustness:
  - Spillover estimates robust to controlling for exporter COVID-19 cases and deaths per capita, export restrictions, and fiscal policy responses in trade partners.

### Heterogeneity of spillovers
- Industry and position effects:
  - Spillovers larger in GVC-intensive industries than in non-GVC-intensive industries; especially large in electronics.
  - Spillovers larger in downstream industries (close to final user) than in upstream industries (inputs).
- Teleworkability and upstreamness:
  - Spillover effects were more than twice as strong for countries whose exporting partners were less able to rely on remote working (teleworkability mitigated spillovers).
  - A one-standard-deviation increase in the upstreamness index reduces the spillover supply effect of the lockdown by almost one-third.
- Temporal attenuation:
  - Spillover effects diminished over time: imports fell by much less in response to lockdowns in partner countries in 2021 than in 2020, indicating global supply chains adjusted.

### GVC resilience and market-share shifts
- Overall resilience:
  - Goods trade, including trade in GVC-intensive goods, was resilient due to demand rotation toward goods, short-lived spillovers, and adaptation of GVC networks.
- Regional market-share changes (GVC-intensive products):
  - By June 2020, “Factory Asia” countries increased their market share in GVC-intensive industries by 4.6 percentage points in “Factory Europe” and by 2.3 percentage points in “Factory North America.”
  - Factory Europe lost the most market share during the first phase of the crisis.
- Dynamics through mid-2021:
  - Initial gains for Factory Asia and losses for Factory Europe were pared back by June 2021, suggesting changes may be temporary.
  - Factory North America continued to lose market share, predominantly within its own domestic markets.
- Historical context:
  - Asia’s market-share gains by mid-2020 were large and quick relative to changes since 2000 but appear to be reversing rapidly.

### Ongoing disruptions and risks to GVCs
- Persistent frictions:
  - Some industries (e.g., automobiles) faced large supply disruptions.
  - Shipping costs remained elevated along some routes (though down from peaks), and port congestion persisted.
- Other risks:
  - International or civil conflicts, cyberattacks, and extreme weather events associated with climate change could pose additional challenges to GVCs.

### Model-based policy scenarios: diversification and substitutability
- Model framework:
  - Extension of Bonadio and others (2021) general equilibrium model of global production networks and trade.
  - Captures trade in intermediate goods and services across 64 countries and 33 sectors.
  - Model does not feature endogenous input–output linkages and does not include inventory management.
- Definitions used:
  - Diversification: across countries (not products), of intermediate goods and services (not final goods), and of use of intermediate inputs (not their production or export). Modeled by shifting domestically sourced share toward roughly half of its observed value via averaging actual distribution with equal-weight sourcing from each country.
  - Substitutability: modeled by increasing elasticity of substitution between intermediate inputs from different countries from 0.5 to 2.0.

### Scenario findings: single large-supplier shock and multicountry shocks
- Single large-supplier shock (25 percent labor supply contraction calibrated to closely match China):
  - Average economy’s GDP falls by 0.8 percent under baseline diversification; decline is reduced by almost half in high-diversification scenario.
  - GDP-weighted average across countries: loss of 3.2 percent under baseline diversification (China contributing 2.7 percentage points) and 2.6 percent in high-diversification world (China contributing 2.4 percentage points).
- Diversification effects:
  - Diversification reduces volatility of GDP growth under correlated multicountry shocks: volatility of GDP growth in the average country is reduced by 5 percent.
  - Diversification provides little protection against exceptionally highly correlated shocks (example: first four months of the COVID-19 pandemic yields same world GDP fall under high diversification as under observed diversification).
- Substitutability effects:
  - With greater substitutability—even though it amplifies the shock in the source country—all countries other than the source country benefit; their GDP losses are reduced by about four-fifths relative to the baseline.

### Trade costs, concentration, and policy levers
- Trade-cost reduction effects:
  - A one-quarter reduction in costs of trading in intermediates lowers the Herfindahl index of geographic concentration in sourcing intermediates by 4 percentage points from 60 percent as observed.
- Nontariff barriers:
  - Scope to reduce nontariff barriers, particularly in emerging markets and low-income developing countries.
- Policy levers emphasized:
  - Reduce nontariff barriers, reduce trade policy uncertainty, and provide an open, stable, rules-based trade policy regime to support greater diversification.

### Empirical facts on diversification and sectoral scope
- Home bias and concentration:
  - Firms in the Western Hemisphere source 82 percent of their intermediates domestically on average.
  - Benchmark concentration reflecting world production concentration for these intermediates is 31 percent.
- Import-side diversification:
  - Limited room to diversify further among inputs already sourced from abroad except in the Western Hemisphere; main scope for diversification is reducing domestic sourcing.
- Sectoral scope:
  - Greatest room to diversify in services industries such as hospitality, finance, and health care.
- Policy implication on reshoring:
  - Reshoring would lower diversification further and increase concentration risk, arguing against reshoring as a resilience strategy in simple terms.

### Firm-level trade-offs and examples
- Firm-level trade-offs not captured by model:
  - Costs of holding larger inventories, fixed costs of establishing new supplier relationships, and efficiency gains from fewer suppliers can reduce incentives for diversification.
- Automotive sector example:
  - Automobile manufacturers on average have about 250 Tier 1 suppliers, rising to 18,000 suppliers in the full value chain.
  - Toyota’s post-Tohoku adaptations: standardize components across models; build comprehensive supplier and inventory databases; regionalize supply chains; ask single-source suppliers to disperse production or hold extra inventory.
- Mechanization and automation:
  - Firms may adopt greater mechanization to gain resilience against labor-supply shocks.

### Supply chain pressures, sectoral impacts, and inflation signals
- Time profile of pressures:
  - Supply chain pressures increased to unprecedented levels at pandemic onset, eased in second half of 2020, then accelerated again to a new peak by end of 2021.
  - Shipping costs steadily increased until September 2021, then started a moderate decline.
  - Delivery times lengthened in 2021; indices of future delivery times indicate persistent disruptions.
- Trade and inflation signals:
  - Import unit values and import volumes diverged in 2021 (flat import volumes and rising unit values), suggesting supply disruptions contributed to inflationary pressures.
- Firm-level indicators (United States, high-frequency data as of January 20, 2022):
  - Share of firms reporting foreign supplier delays: increased from 9 percent in October 2020 to 20 percent in December 2021.
  - Share of firms reporting production delays: reached 14 percent in December 2021.
  - Share of firms reporting delays in delivery/shipping to customers: reached 26 percent in December 2021.
  - Growing share of small businesses reported difficulties locating alternative foreign suppliers.
- Automotive sector dynamics:
  - Trade and sales collapsed in spring 2020 and rebounded in second half of year without reaching pre-pandemic levels; shortage of automotive chips (semiconductors) was a key factor.

### Box 4.1 — Semiconductor shortage and high-frequency bilateral evidence
- Semiconductor shortage and autos:
  - Car producers curtailed semiconductor orders during downturn; semiconductor production reallocated to consumer electronics.
  - Pent-up car demand in second half of 2020 hit reallocated semiconductor capacity, constraining automotive recovery and raising prices.
  - Trade tensions and domestic shocks (example: drought in Taiwan Province of China) aggravated the shortage.
- High-frequency bilateral seaborne trade empirical approach:
  - Uses daily bilateral seaborne trade volumes (Automatic Identification System data) and seven-day moving averages of year-over-year growth rates relative to 2017–19 averages.
  - Lockdown stringency (LSjt) measured 0–100.
  - Estimated import equation preserved exactly:
    - ˆM_{ij,t+h} = γ_{it} + α_{ij} + β LS_{jt} + X_{jt}′ δ + ∑_{k=1}^{7} ˆM_{ij,t−k} + ε_{ij,t+h}
    - ˆM_{ij,t+h}: bilateral import growth (seven-day moving average of year-over-year growth rates with respect to pre-pandemic (2017–19) averages).
    - γ_{it}: importer-time fixed effects.
    - α_{ij}: bilateral pair fixed effect.
    - LS_{jt}: lockdown stringency (0–100) of the exporter country.
    - X_{jt}′: controls (ratio of new COVID-19 cases to population and aggregate measure of exporters’ exposure to foreign lockdowns).
  - Lockdown measures are lagged to account for delivery lags in shipping.
- High-frequency results:
  - Over full 2020–21 sample, exporter lockdowns have a large and statistically significant impact on bilateral trade volumes.
  - As stringency ranges 0–100, point estimates of around 5 imply that a change in stringency of just 20 points can temporarily halt bilateral trade (numeric interpretation preserved).
  - Lockdowns have no statistically significant effect on trade volumes in 2021 (finding: “lockdowns have no statistically significant effect on trade volumes in 2021”).
  - Panel evidence: full-sample declines up to around –16 percent at short horizons; 2021 effects statistically indistinguishable from zero.
- French Customs firm-level evidence:
  - Adjustment mainly via intensive margin (volumes); extensive margin contributed marginally.
  - Heterogeneous effects:
    - Impact of importing-country lockdowns on exports of firms selling final consumer goods (downstream) was nearly nine times larger than for firms selling intermediate inputs (upstream).
    - Impact of lockdowns and COVID-19 deaths on exports was almost 67 percent larger for less automated firms.
    - Imports of firms in industries holding the lowest stocks of inventories fell more than twice as much as among firms in industries with average inventory intensity; firms in industries with highest inventory intensity increased imports.
    - Exporters in more inventory-intensive industries experienced a smaller drop in sales.
  - Heterogeneity evaluated by interacting stringency and deaths with industry-level upstreamness, firm-level imports of industrial robots (automation proxy), and industry-level inventory intensity.

### Policy implications and recommendations
- Public health and mobility:
  - Vaccinating widely across countries is important to minimize supply disruption spillovers onto partner countries.
  - Strengthening health systems and investing in digital infrastructure helps mitigate transmission of shocks in future pandemics or variants.
  - Facilitating full return of mobility is important to boost services demand back to pre-pandemic trends; quicker-than-expected easing of mobility restrictions could pose an upside risk to global trade projections.
- Infrastructure and information:
  - Upgrade and modernize port infrastructure on key shipping routes to reduce choke points and lower costs of diversification and substitutability.
  - Governments can resolve informational externalities (e.g., digitalized tax and document filings to map interfirm transactions and supply chain networks) to support stress-testing and identify supply chain weaknesses.
- Trade policy:
  - Target reductions in nontariff barriers, reduce trade policy uncertainty, and provide an open, stable, rules-based trade policy regime to support greater diversification.
- Firm-level measures:
  - Consider automation, inventory policies, supplier standardization, regionalization, and supplier dispersal as tools to raise resilience.

*Source: CHAPTER 4, GLOBAL TRADE AND VALUE CHAINS DURING THE PANDEMIC, International Monetary Fund | April 2022.*

### Introduction

### Introduction

### Pandemic impact on trade: patterns and timing
- At its trough in the second quarter of 2020, the volume of global trade in goods fell 12.2 percent, and trade in services fell 21.4 percent, compared with the last quarter of 2019.
- Trade in goods had recovered to pre-pandemic levels by October 2021.
- Trade in services remained sluggish, driven mainly by the collapse of travel; transport services appear to be recovering while disruptions in seaborne trade remained elevated.
- Trade in other services (notably telecommunication services) was more robust.
- Exports of GVC-intensive goods fell 30 percent between January and April 2020, while exports of other goods fell by about 18 percent.
- The initial drop was relatively more severe in some industries like automobiles, amid disruptions to key inputs such as semiconductors.

### Research questions and empirical approach
- The chapter examines:
  1. How well a standard model of demand and prices accounts for pandemic trade patterns, compared with previous large recessions.
  2. Which pandemic-specific factors determined trade patterns.
  3. What international spillover effects were generated by mobility restrictions.
- Methodology highlights:
  - Uses an import demand model linking real import growth of goods and services to growth in demand and relative price of imports for a sample of 127 countries over 1985–2019.
  - Uses granular bilateral monthly trade data and a gravity model (Santos Silva and Tenreyro 2006 specification) on monthly imports at the six-digit product level from Trade Data Monitor to estimate international spillovers.
  - The Oxford COVID-19 Government Response Stringency Index is used (ranges from 0 to 100; highly correlated with workplace closing component).

### Key empirical findings
- Demand and relative prices alone do not explain pandemic trade patterns:
  - The import demand model fits historical data up to 2019 but produces large forecast errors for 2020 when goods and services are considered separately.
  - The model underpredicts the large observed decline in services trade: the model predicts a growth rate of about –8 percent for services in 2020, while services trade actually fell by 25 percent.
  - The model overpredicts the fall in goods trade: predicting a 10 percent decline, against the 6 percent observed fall.
  - Estimated coefficients on import-adjusted demand are positive for most countries and greater than 1; coefficients on relative price average between –0.2 and –0.3.
- Pandemic-specific factors mattered:
  - Countries with more severe pandemic outbreaks, more stringent containment measures, or larger declines in mobility showed “excess import demand” for goods (the fall in goods imports was smaller than predicted).
  - The forecast error for goods imports was 3 percentage points more positive for countries in the third quartile of the distribution of the number of COVID-19 cases than for those in the first quartile.
  - For services, overprediction was most pronounced in countries where travel services accounted for a large share of total service imports.
  - Better health-preparedness of importers’ trade partners (measured by the Global Health Security Index) was associated with less negative deviations in goods imports.
- International spillovers from lockdowns were substantial:
  - Lockdowns in a country’s trade partners on average accounted for up to 60 percent of the observed decline in imports in the first half of 2020.
  - The estimated coefficient from a weighted regression of the change in imports between 2020:Q2 and 2019:Q4 against partner-country Oxford Stringency Index is –0.015 (t-stat = –2.44).
  - Spillover effects first materialized in February 2020, grew in March and April, began declining in May, and by June were indistinguishable from zero as goods imports rebounded.
  - Spillover estimates are robust to controlling for exporter COVID-19 cases and deaths per capita, export restrictions, and fiscal policy responses in trade partners.
- Heterogeneity in spillovers:
  - Spillovers were larger in GVC-intensive industries than in non-GVC-intensive industries.
  - Spillovers were larger in downstream (close to final user) industries than in upstream (input) industries.
  - Spillover effects were more than twice as strong for countries whose exporting partners were less able to rely on remote working (teleworkability mitigated spillovers).
  - Spillover effects diminished over time, indicating that global supply chains were able to adjust.

### Conclusions on GVCs and resilience
- GVCs were able to adjust to the asynchronous development of the pandemic, as reflected in changes in market shares among GVC regions.
- Building resilience:
  - There is substantial room to diversify away from domestic inputs to enhance resilience.
  - Gains in resilience can be achieved by (1) increasing geographic diversification of input sourcing across countries and (2) increasing substitutability of inputs across source countries.
  - Diversification substantially reduces global GDP losses in response to shocks in key upstream suppliers and reduces GDP volatility following correlated productivity shocks observed in historical data over the past 25 years.
  - Reducing diversification increases volatility; greater input substitutability across source countries reduces GDP losses from shocks in individual countries.
- Policy implication:
  - Containing the pandemic domestically is important not only for domestic activity but because outbreaks and lockdowns can have negative spillovers onto trade partners.
  - The observed reduction of spillovers over time, including for GVC-intensive goods, suggests that global supply chains adjusted; this cautions against policies seeking to effect permanent changes in the structure of global production and trade without accounting for adjustment dynamics.

*Source: CHAPTER 4, GLOBAL TRADE AND VALUE CHAINS DURING THE PANDEMIC, International Monetary Fund | April 2022.*

### CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC

### CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC

### Spillovers from lockdowns, teleworkability, and industry position
- Empirical evidence from granular bilateral trade data, controlling for demand in importing countries, shows statistically significant negative spillovers from lockdowns in partner countries.
- Spillovers were:
  - Larger in GVC-intensive industries, and especially in electronics, than in non-GVC-intensive ones.
  - Stronger among partner countries less able to rely on teleworking (teleworkability measured using Dingel and Neiman (2020) cross-country data).
  - Larger for downstream industries (such as transportation and textiles) than for more upstream industries (such as metals and minerals).
- Quantitative magnitudes and specific results:
  - A one-standard-deviation increase in the upstreamness index reduces the spillover supply effect of the lockdown by almost one-third.
  - Spillovers tended to be short-lived and were mitigated when partner countries could use telework.
  - The spillover effects waned over time: imports fell by much less in response to lockdowns in partner countries in 2021 than in 2020.

### GVC resilience and market-share shifts during the pandemic
- Overall finding: goods trade, including trade in GVC-intensive goods, was resilient during the crisis due to demand rotation toward goods and short-lived spillovers, and because GVC networks adapted.
- Regional market-share changes (GVC-intensive products):
  - By June 2020, “Factory Asia” countries increased their market share in GVC-intensive industries by 4.6 percentage points in “Factory Europe” and by 2.3 percentage points in “Factory North America.”
  - Factory Europe lost the most market share during the first phase of the crisis.
- Dynamics through mid-2021:
  - Initial gains in market share for Factory Asia and initial losses for Factory Europe were pared back by June 2021, suggesting the changes may be temporary.
  - Factory North America continued to lose market share, predominantly within its own domestic markets.
- Historical context:
  - Asia’s market-share gains by mid-2020 were large and quick relative to historical changes (shown since 2000) but appear to be reversing rapidly.

### Ongoing disruptions and risks to GVCs
- Notwithstanding resilience, some industries—such as automobiles—faced large supply disruptions.
- Shipping costs remain elevated along some routes (though down from peaks), and port congestion persists, contributing to continuing supply chain disruptions.
- Other shock types that could pose challenges to GVCs include international or civil conflicts, cyberattacks, and extreme weather events associated with climate change.

### Policy options analyzed with a model-based framework
- Model framework:
  - Extension of Bonadio and others (2021) general equilibrium model of global production networks and trade.
  - Captures trade in intermediate goods and services across 64 countries and 33 sectors (model calibration described in Online Annex 4.4).
  - The model does not feature endogenous input–output linkages and does not include inventory management; therefore cannot evaluate inventory-based risk mitigation.
- Scenarios analyzed:
  - Supply disruption in a single large input supplier country.
  - Supply shocks to multiple countries.
  - Outcomes compared under observed levels versus higher levels of:
    - Diversification of intermediate-input sourcing across countries.
    - Substitutability of inputs from different country suppliers.

### Definitions and mechanisms
- Diversification (as used in this chapter):
  - Diversification (1) across countries (not across products); (2) of intermediate goods and services (not final goods); and (3) of the use of intermediate inputs (not their production or export).
  - Modeled as shifting the domestically sourced share toward roughly half of its observed value by averaging actual distribution with equal-weight sourcing from each country.
  - Potential benefits: reduce reliance on a single country and establish alternative supplier relationships.
  - Potential downsides: exposure to more volatile supplier countries; empirical evidence on benefits is mixed.
- Substitutability:
  - Ease with which producers can switch inputs from suppliers in one country to suppliers in another.
  - Can arise from greater production flexibility (e.g., Tesla rewriting software to use alternative semiconductors) or from standardization of inputs (e.g., General Motors reducing unique semiconductor chip families by 95 percent).
  - Modeled by increasing the elasticity of substitution between intermediate inputs from different countries from 0.5 to 2.0 (range similar to Feenstra and others (2018)).

### Room to diversify: empirical facts and implications
- Home bias and concentration:
  - On average, firms in the Western Hemisphere source 82 percent of their intermediates domestically.
  - Benchmark concentration that reflects world production concentration for these intermediates is 31 percent.
  - This indicates substantial “home bias” in sourcing intermediates and suggests sizeable room to diversify away from domestic sourcing toward foreign suppliers.
- Import-side diversification:
  - There is limited room to diversify further among inputs already sourced from abroad except in the Western Hemisphere; main scope for diversification is reducing domestic sourcing.
- Sectoral scope:
  - Sectors with the greatest room to diversify are services industries such as hospitality, finance, and health care.
- Policy implication:
  - Reshoring would lower diversification further and increase concentration risk, arguing against reshoring as a resilience strategy in simple terms; fuller analyses show reshoring can lead to more volatile economic activity even after structural adjustments.

*Source: CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC (ch4 - CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC), April 2022.*

### CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC

### CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC

### Diversification and Substitutability: scenario findings
- Scenario: a 25 percent labor supply contraction in a single large global supplier of intermediate inputs (calibrated to closely match China).  
- Average economy’s GDP falls by 0.8 percent under the baseline level of diversification; in the high-diversification scenario the decline in GDP is reduced by almost half.  
- GDP-weighted average across countries: loss of 3.2 percent under baseline levels of diversification (with China contributing 2.7 percentage points of that loss) and 2.6 percent in the high-diversification world (with China contributing 2.4 percentage points).  
- Diversification reduces the volatility of GDP growth under correlated multicountry shocks: the volatility of GDP growth in the average country is reduced by 5 percent.  
- Diversification provides little protection against exceptionally highly correlated shocks (example: first four months of the COVID-19 pandemic scenario yields the same world GDP fall under high diversification as under observed diversification).  
- Substitutability effects: with greater substitutability—even though it amplifies the shock in the source country—all countries other than the source country benefit; their GDP losses are reduced by about four-fifths relative to the baseline.  
- Modeling assumption on elasticity: baseline elasticity of substitution = 0.5; higher elasticity of substitution = 2.0 in counterfactuals discussed.

### Trade costs, concentration, and policy levers
- A one-quarter reduction in the costs of trading in intermediates lowers the Herfindahl index of geographic concentration in sourcing intermediates by 4 percentage points from 60 percent as observed in actual data.  
- Nontariff barriers: there is scope to reduce nontariff barriers, particularly in emerging markets and low-income developing countries (Figure 4.14 illustrates higher simple-average indices for these groups relative to advanced economies).  
- Conventional policy tools to reduce trade costs include tariff and nontariff barrier reductions; with tariff barriers low globally, reducing nontariff barriers is highlighted as having substantial medium-term growth benefits.

### Firm-level trade-offs, evidence, and examples
- The model does not capture all firm-level trade-offs: costs of holding larger inventories, fixed costs of establishing new supply relationships, and efficiency gains from dealing with a smaller number of suppliers can reduce incentives for diversification.  
- Empirical note: automobile manufacturers on average have about 250 Tier 1 suppliers, rising to 18,000 suppliers in the full value chain.  
- Toyota’s post-Tohoku adaptations illustrating practical measures to increase resilience:
  - standardize some components across vehicle models to enable global sharing of inventory and production flexibility;
  - build a comprehensive database of suppliers and parts held in inventory;
  - regionalize supply chains to avoid dependence on a single location;
  - ask single-source suppliers to disperse production of parts to multiple locations or hold extra inventory.
- Firms may also adopt greater mechanization to gain resilience against shocks to labor supply.

### Policy implications and recommendations
- Public health and mobility:
  - Vaccinating widely across countries is important to minimize supply disruption spillovers onto partner countries.  
  - Strengthening health systems and investing in digital infrastructure would help mitigate transmission of shocks in future scenarios (including COVID-19 variants or other pandemics).  
  - Facilitating the full return of mobility is an important element in boosting services demand back to pre-pandemic trends; a quicker-than-expected easing of mobility restrictions could pose an upside risk to global trade projections.  
- Enhancing infrastructure:
  - Upgrade and modernize port infrastructure on key global shipping routes to reduce global choke points and lower costs of diversification and substitutability.  
- Closing information gaps:
  - Governments can resolve informational externalities to help firms make strategic sourcing decisions.  
  - Examples: using tax and document filings (digitalized) to map interfirm transactions and supply chain networks to support stress-testing and identify supply chain weaknesses.  
- Reducing trade costs:
  - Target reductions in nontariff barriers, reduce trade policy uncertainty, and provide an open, stable, rules-based trade policy regime to support greater diversification.

### Supply chain pressures and sectoral impacts
- Evolution of pressures:
  - Supply chain pressures increased to unprecedented levels at the onset of the COVID-19 pandemic, eased in the second half of 2020, then accelerated again to reach a new peak by the end of 2021.  
  - Shipping costs steadily increased until September 2021, when they started a moderate decline.  
  - Delivery times lengthened in 2021; indices of future delivery times indicate persistent supply chain disruptions.  
- Trade and inflation signals:
  - Import unit values and import volumes diverged in 2021 (flat import volumes and rising unit values), suggesting supply disruptions contributed to inflationary pressures.  
- Firm-level indicators (United States, high-frequency data as of January 20, 2022):
  - Share of firms reporting foreign supplier delays increased from 9 percent in October 2020 to 20 percent in December 2021.  
  - Share of firms reporting production delays reached 14 percent in December 2021.  
  - Share of firms reporting delays in delivery/shipping to customers reached 26 percent in December 2021.  
  - A growing share of small businesses reported difficulties locating alternative foreign suppliers.  
- Sectoral example:
  - Automotive industry: trade and sales collapsed in spring 2020 and rebounded in the second half of the year without reaching pre-pandemic levels; the shortage of automotive chips (semiconductors) was a key factor, driven in part by a pandemic-induced shift in semiconductor demand (increase in demand for electronics supporting remote work while demand for cars fell).

*Source: CHAPTER 4 GLOBaL TRaDE aND vaLUE ChaINS DURING ThE PaNDEMIC (IMF, April 2022).*

### Box 4.1. Effects of Global Supply Disruptions during the Pandemic

### Box 4.1. Effects of Global Supply Disruptions during the Pandemic

### Semiconductor shortage and the automotive sector
- Car producers curtailed orders for semiconductors during the pandemic-related downturn, shifting semiconductor industry production toward sectors such as consumer electronics.
- When pent-up demand for cars accelerated in the second half of 2020, semiconductor production capacity had already been reallocated, constraining the automotive-sector recovery and resulting in higher prices.
- Trade tensions and domestic shocks (such as a drought in Taiwan Province of China) aggravated the semiconductor shortage.
- The shortage highlighted vulnerabilities of global value chains and spurred calls for reshoring and increasing supply chain resilience.
- Figure context: Trade in Automobiles and Semiconductors (Index, January 2018 = 100) is used to illustrate divergent trade movements for “Automobiles” (Harmonized System six-digit codes for manufactured intermediate inputs and final goods (vehicles)) and “Semiconductors” (Harmonized System six-digit codes 854150 and 854190).

### Empirical approach: high-frequency bilateral seaborne trade data
- Data and measurement:
  - Uses a unique data set of daily bilateral seaborne trade volumes (Automatic Identification System data collected by Marine Traffic) and constructs seven-day moving averages of year-over-year growth rates relative to pre-pandemic (2017–19) averages.
  - Lockdown stringency (LSjt) is measured on a 0–100 scale (Hale and others 2020).
- Estimated import equation (daily frequency, horizon h):
  - ˆM_{ij,t+h} = γ_{it} + α_{ij} + β LS_{jt} + X_{jt}′ δ + ∑_{k=1}^{7} ˆM_{ij,t−k} + ε_{ij,t+h}
  - Components preserved exactly as stated:
    - ˆM_{ij,t+h}: bilateral import growth from exporter j to importer i (seven-day moving average of year-over-year growth rates with respect to pre-pandemic (2017–19) averages).
    - γ_{it}: importer-time fixed effects.
    - α_{ij}: bilateral pair fixed effect.
    - LS_{jt}: lockdown stringency (0–100) of the exporter country.
    - X_{jt}′: vector of control variables (the ratio of new COVID-19 cases to the population and an aggregate measure of exporters’ exposure to foreign lockdowns).
  - Lockdown measures are lagged to account for delivery lags in shipping (example given: if voyages take three days, stringency measures are lagged by three days).
- Empirical caveat:
  - The bilateral specification captures lockdown-induced trade disruptions at the bilateral level but does not rule out substitution (sourcing from different countries). An alternative aggregate approach is noted (Cerdeiro and Komaromi 2020).

### Results: lockdown effects on bilateral trade
- Over the full 2020–21 sample, exporter lockdowns have a large and statistically significant impact on bilateral trade volumes.
- Numeric interpretation provided in the source:
  - “As the stringency variable has a range of 0–100, the point estimates of around 5 imply that less than a full lockdown (a change in stringency of just 20 points) can temporarily halt bilateral trade.”
- Temporal dynamics:
  - Lockdowns show statistically significant negative effects in the full sample (2020–21).
  - Lockdowns have no statistically significant effect on trade volumes in 2021 (finding: “lockdowns have no statistically significant effect on trade volumes in 2021”).
  - Interpretation: activity became less susceptible to lockdowns as economies adapted to the pandemic, underscoring resilience of global value chains.
- Visual summary (as described in source):
  - Figure 4.2.1 displays the response of bilateral import growth to exporter lockdowns with 95 percent confidence bands; panel 1 for full sample shows declines up to around –16 percent at short horizons, panel 2 for 2021 sample shows effects statistically indistinguishable from zero (axes and confidence bands described in source text).

### Firm-level evidence from French Customs data
- Data and scope:
  - Monthly French Customs data on firms’ imports and exports for 2019 and 2020.
- Main margin of adjustment:
  - Adjustment occurred mainly along the intensive margin (volumes); the extensive margin (varieties dropping out of France’s trade basket) contributed marginally, indicating the temporary nature of the shock.
- Heterogeneous firm and industry effects (key quantitative findings preserved):
  - Downstream vs upstream:
    - “The average impact of importing-country lockdowns on exports of firms selling final consumer goods (downstream firms) was nearly nine times larger than that for firms selling intermediate inputs (upstream firms).”
  - Automation:
    - “The impact of lockdowns and the spread of the virus (measured by COVID-19 deaths) on exports was almost 67 percent larger for firms that are less automated.”
  - Inventory intensity:
    - “Imports of firms in industries holding the lowest stocks of inventories fell more than twice as much as those among firms in industries with average inventory intensity.”
    - “Firms in industries with the highest inventory intensity increased imports.”
    - “Exporters in more inventory-intensive industries also experienced a smaller drop in sales,” suggesting inventories play a shock-absorbing role.
- Methodological notes:
  - Heterogeneity is evaluated by interacting stringency and deaths with industry-level upstreamness (Antràs and others 2012), firm-level imports of industrial robots (proxy for automation), and industry-level inventory intensity (ratio of inventory to sales).
  - Results on inventory intensity are sensitive to the measure of industry-average inventory-to-sales ratios.
- Figure context: Figure 4.3.1 (“Impact of Supply Chain Upstreamness, Automation, and Inventories on Trade Adjustment”) presents effects on export and import growth across Downstream / Midstream / Upstream groups and automation/inventory intensity categories.

### Policy-relevant implications (as reflected in the analysis)
- Vulnerabilities exposed by the semiconductor shortage and pandemic-induced lockdowns motivate:
  - Calls for reshoring and measures to increase supply chain resilience.
  - Consideration of automation and inventory policies as mechanisms to increase firm-level resilience to lockdown shocks.
- Evidence on adaptation:
  - The loss of statistical significance of lockdown effects in 2021 implies that adaptation (operational changes, sourcing adjustments, inventories, automation) reduced sensitivity of trade to lockdowns over time.

*Italic: Source — Box 4.1. Effects of Global Supply Disruptions during the Pandemic (chapter 4), WORLD ECONOMIC OUTLOOK: WaR SETS BaCK ThE GLOBaL RECOvERy, International Monetary Fund | April 2022*

---


_Source: https://www.imf.org/-/media/files/publications/weo/2022/april/english/ch4.pdf_