## wpiea2019143

## Source details

**Canonical URL:** [wpiea2019143](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019143.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019143.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019143.pdf.json)

---

### I. Introduction — context, objectives, and high-level findings
- Context and motivation
  - After five decades of negotiations culminating with the creation of the WTO in 1995, low tariffs in advanced economies became a core pillar of the international trading system.
  - The landscape changed abruptly in 2018 when the U.S. administration imposed, or announced they might impose, higher tariffs targeting specific sectors or trading partners.
  - U.S. tariffs are described as responding to concerns about market-distorting government interventions, technology transfer, subsidies to state-owned enterprises, and, in some cases, U.S. national security interests.
- Objectives and approach
  - Use a multi-sector computable general equilibrium model (Caliendo, Feenstra, Romalis and Taylor, 2017; CFRT) to estimate effects of three potential U.S. trade policy actions with systemic importance.
  - Model leverages input-output data covering 165 countries and 17 sectors, and detailed applied-tariff data including pre-tension periods.
- Scenarios analyzed (high level)
  - Auto and auto-parts tariffs stemming from a Section 232 investigation.
  - Two U.S.–China outcomes associated with a Section 301 investigation: a “transactional deal” (managed trade halving the bilateral trade deficit) and an escalation with higher bilateral tariffs.
- Key high-level findings
  - Aggregate GDP effects are relatively small but generally negative for most countries, including the U.S.
  - Protected sectors typically do not expand sufficiently to offset losses in competitiveness elsewhere.
  - Spillovers to third countries are generally linked to direct gross trade exposures; indirect supply-chain exposures can be first-order.
  - Wide sectoral variation implies high adjustment costs in moving to new equilibria.
  - Heterogeneous regional and distributional effects within the U.S.: a transactional deal benefits U.S. states with high shares of primary-sector activity (examples: Arkansas, North Dakota, Oklahoma, Wyoming); these same states suffer most under the auto tariffs scenario.
  - Richer U.S. states may benefit from certain trade scenarios; poorer states seldom benefit.
  - Impacts depend crucially on assumptions such as the extent of U.S. re-shoring in affected sectors.
- Limitations and scope
  - Modeling choices reflect uncertainties in potential negotiation outcomes and retaliation forms.
  - Some policy changes (investment regimes, intellectual property, or rejection of rules-based trading system) are not well captured and may have first-order effects.
  - The employed model is static, describing steady-state implications and ignoring uncertainty effects on investment and activity.

### II. Model, data, and calibration
- Modeling framework
  - Core model: CFRT (Caliendo et al, 2017) — a multi-sector computable general equilibrium model featuring firm heterogeneity (Melitz-style) and cross-border input-output trade (Caliendo and Parro, 2015).
  - CFRT features: tariffs modeled as revenue shifters (ad-valorem tariffs apply to final price); avoids some corner solutions found in other imperfect-competition setups.
- Shortcomings of CFRT (explicit)
  - No dynamics: static model cannot analyze investment responses or trade-policy uncertainty effects on investment and short-run macro dynamics.
  - Investment omission: cannot capture changes in capital accumulation or productivity effects over medium/long term.
  - Fixed overall trade balances to close the model.
- Databases and aggregation
  - Input-output data: Eora global supply chain database (2015 table used).
  - Tariff data: 2015 applied-tariff data from UNCTAD accessed through WITS.
  - Resulting merged database: 17 sectors and 165 countries.
- Calibration of elasticities of substitution
  - Two sector-level elasticities central to CFRT:
    - σs: elasticities of substitution across foreign or domestic intermediate inputs (assumed larger).
    - ωs: elasticities of substitution across composite foreign and domestic inputs (assumed smaller).
  - Rationale: switching between foreign suppliers is empirically and technically easier than switching between foreign and domestic suppliers.
  - Estimation strategy: sectoral trade elasticities Γ_s estimated from a log-linear trade equation over samples excluding small flows (preferred sample excludes 10 percent of the smallest flows).
  - Relation used to calibrate σ_s:
    - σ_s = (Γ_s + 1)/1.5
  - ω_s calibrated by mapping Feenstra et al (2018; FLOR) estimates: the σ_s/ω_s ratio from FLOR is used; where sectors could not be matched the simple average ratio 1.76 is applied.

- Sector-level trade elasticity Γ_s (preferred calibration versus CP (2015))
  - 1 Agriculture, Hunting, Forestry and Fishing: CP (2015) 9.1 — Preferred 9.1
  - 2 Extractive Industries: CP (2015) 13.5 — Preferred 19.0
  - 3 Food, Beverages and Tobacco: CP (2015) 2.6 — Preferred 2.5
  - 4 Textiles, Textile Products, leather and footwear: CP (2015) 8.1 — Preferred 3.0
  - 5 Wood, Paper, Printing and Publishing: CP (2015) 14.0 (note 1/) — Preferred 5.4
  - 6 Petroleum, Chemical, Non-Metallic Mineral Products: CP (2015) 18.1 (note 1/) — Preferred 12.1
  - 7 Metal Products: CP (2015) 7.0 — Preferred 13.9
  - 8 Electrical equipment and Machinery: CP (2015) 4.8 (note 1/) — Preferred 2.6
  - 9 Transport equipment: CP (2015) 8.2 — Preferred 8.0 (note 3/)
  - 10 Manufacturing, Nec; Recycling: CP (2015) 4.0 — Preferred 4.0 (note 2/)
  - Notes:
    - 1/ Simple-averaged elasticities of more disaggregated sectors.
    - 2/ Using the same as CP.
    - 3/ Simple-average of elasticities across goods producing sectors.
  - Additional calibration details:
    - For sectors 1 and 10 CP values used because estimates were not robust across samples.
    - For sector 9 mean elasticity across all other goods-producing sectors is used following CP.
    - The Γ_s estimates correspond to the first set of numbers in Table 1 of CFRT.

### III. Scenarios analyzed (precise descriptions)
- Baseline: tariffs in effect at end-2018 except for Scenario 3 where specified tariff changes are excluded from the baseline.
- Scenario 1 — U.S. tariffs on autos and auto-parts
  - U.S. introduces a 25 percent tariff on all imports of vehicles and auto parts, independently of origin.
  - Retaliation: all trading partners directly affected impose an equivalent 25 percent tariff on imports of vehicles and auto-parts from the U.S.
  - Additional uniform retaliation: some partners with bilateral auto surpluses impose an additional uniform tariff Z% on all other U.S. goods such that:
    - 25% × value of U.S. auto & auto‑part imports from country i = 25% × value of country i’s auto & auto‑part imports from the U.S. + Z% × value of country i’s all other imported goods from the U.S.
    - where the tariff rate Z ≥ 0 is chosen so that the equality holds.
  - Mapping: all tariffs on autos and auto parts mapped to the “transport equipment goods” sector using 2017 UN/Comtrade shares.
  - USMCA side-letter caveats and selective-retaliation variants considered.
- Scenario 2 — Transactional U.S.–China deal (halving bilateral deficit)
  - Objective: reduce the bilateral deficit between U.S. and China by half.
  - Two modeled components:
    1. Increase Chinese purchases of U.S. goods by simultaneous reduction in Chinese non-tariff import barriers vis-à-vis the U.S. and an increase in non-U.S. barriers; each change calibrated to achieve one-quarter of deficit reduction.
       - Relative change scaled by inverse of sectoral elasticities, except for agriculture, extractive industries and transport equipment where especially large changes are introduced.
    2. Reduction in Chinese exports of electronics and other manufactured goods to the U.S. via export tariffs; sectoral export reductions inversely proportional to elasticities of substitution; calibrated to halve the bilateral deficit.
  - Rationale: inverse-elasticity scaling avoids unreasonable quantity changes and provides a Ramsey-like lower bound on distortions.
- Scenario 3 — Escalation (uniform 25 percent tariffs)
  - U.S. imposes a 25 percent uniform tariff on all goods imported from China and China retaliates with the same 25 percent uniform tariff on all U.S. goods.
  - Abstract capture of higher bilateral tariffs; other policy changes not quantified.

### IV. Results — effects on the U.S. and sectoral/regional heterogeneity
- Aggregate GDP effects (U.S.)
  - U.S. GDP falls under all three scenarios ― by around 0.2-0.3 percent of GDP ― following the introduction of tariffs.
- Sectoral heterogeneity (U.S.)
  - Sectoral value-added responses are heterogeneous and scenario-dependent.
  - Example: U.S. agricultural sector
    - Benefits noticeably under the “transactional deal” where China purchases more U.S. goods.
    - Contracts under the auto tariffs scenario as several trading partners retaliate by increasing tariffs on that sector to match U.S. tariff intensity.
  - Protected sectors do not expand meaningfully because they rely on Chinese inputs; protected sectors face higher input costs and foreign inputs are not easily substitutable.
- Regional and distributional implications within the U.S.
  - Heterogeneous sectoral responses imply diverse regional outcomes across U.S. states due to geographical specialization.
  - A “transactional deal” benefits states with a high share of the primary sector (mining, hydrocarbons, and agriculture) — examples: Arkansas, North Dakota, Oklahoma, Wyoming — which would suffer under the auto tariffs scenario.
  - Escalation with China or a deal where China restrains exports to the U.S. would have a more homogeneous negative effect across states.
  - Distributional implication: richer states (higher median incomes and a higher proportion of the rich) would benefit relative to poorer ones under the auto tariff scenario and under intensification of the U.S.–China trade war; poorer states benefit only when China purchases more primary products from the U.S.
- Figures referenced (IMF staff calculations)
  - Figure 3: Real GDP impact on the U.S. and selected trading partners (percent change).
  - Figure 4: Impact on U.S. Sectors (percent change and changes in sectoral value added in billions of USD; constant prices).

### V. Box 1 — Trade policy and effects on U.S. states (detailed findings)
- Database and mapping
  - Economic activity broken into 17 sectors for each of the 50 states and the District of Columbia.
  - State-level impacts mapped outside the model using sectoral shares in state output.
- Scenario-specific state findings
  - Auto tariffs (uniform retaliation)
    - Primary good sectors (agriculture, mining, hydrocarbon extraction) suffer large falls in value-added.
    - States with large shares of these sectors experience larger falls in GDP: Arkansas, North Dakota, Oklahoma, West Virginia, Wyoming.
    - These states do not reap benefits from any increase in U.S. auto output under this scenario.
  - Auto tariffs (selective retaliation)
    - When retaliation targets “emblematic” products (e.g., Harley Davidson motorcycles, blue jeans, bourbon whiskey) GDP losses of heavily primary-sector states are largely reduced.
  - Trade war with China (both impose 25 percent tariffs across all products)
    - Most sectors fall in output; most U.S. states exhibit negative GDP changes.
    - Effects are relatively less heterogeneous across states than under other scenarios.
  - Trade deal: China purchases more selected U.S. goods
    - States exposed to agriculture and hydrocarbon products, and less diversified states, tend to benefit.
  - Trade deal: China restrains exports to the U.S.
    - Such a deal leads to losses in all states.
- Socio-economic heterogeneity
  - Positive changes in state-level GDP under the auto tariff scenario or higher bilateral tariffs with China are positively associated with states having higher median income, higher mean income, and a higher proportion of high-income households, and negatively associated with higher proportions of low-income households.
  - A trade deal where China purchases more U.S. goods tends to favor states with higher shares of primary goods and somewhat benefit states with lower average income.
- Quantitative correlations with income measures (values preserved exactly)
  - Correlation of estimated state-level GDP changes with state-level income measures:
    - Auto tariffs, uniform retaliation: 0.3** 0.4*** -0.1 0.4***
    - Auto tariffs, selected retaliation: 0.5*** 0.6*** -0.2 0.6***
    - Intensification of trade war with China: 0.5*** 0.6*** -0.3* 0.6***
    - Trade deal: restraint on China’s exports to U.S.: 0.2 0.3** -0.1 0.3**
    - Trade deal: China purchases of U.S. goods: -0.1 -0.2* 0.0 -0.3*
  - Income categories related to columns in table:
    - States with higher median income
    - States with higher mean income
    - Higher proportion of low-income household (annual income < $15,000)
    - Higher proportion of high-income households (annual income > $150,000)

- USMCA exemptions and quotas (numeric details)
  - If ratified, USMCA side-letters stipulate:
    - Canada allowed to ship up to 2.6 million cars to the U.S. under USMCA preferences (versus 1.8 million shipped in 2018).
    - Canada allowed $32.4 billion worth of auto parts without getting hit by tariffs.
    - Mexico’s side-letter: quota for auto parts is $108 billion.
  - Counterfactuals assuming full exemption for Canada and Mexico in Scenario 1:
    - Exports of auto-parts would likely remain within USMCA quotas.
    - Passenger vehicle quotas: simulated expansion implies the quotas would become binding; tariffs would kick in on U.S. vehicle imports from Canada and Mexico unless exports remain within quotas.
    - Additional note: USMCA quotas are already binding for current levels of U.S. vehicle imports from Mexico.

### VI. Role of elasticities, robustness checks, and spillovers
- Elasticities matter: substitution elasticities between foreign suppliers and between foreign and domestic suppliers drive magnitudes and signs of effects.
- Calibration specifics
  - σ_s computed from Γ_s via σ_s = (Γ_s + 1)/1.5.
  - ω_s obtained by mapping FLOR estimates; where unmatched use average σ_s/ω_s ratio 1.76.
- Robustness check: raise foreign-domestic elasticities to equal foreign-foreign ones (operationally set foreign-domestic elasticities equal to foreign-domestic elasticities divided by 1.01 to avoid corner solutions).
  - Intuition: makes re-shoring production easier (easier to substitute foreign goods with domestic production).
- Findings from sensitivity
  - If re-shoring is easier for the U.S., potential losses would be larger for most other countries; most countries (other than the U.S.) lie below the 45-degree line in Figure 9.
  - In baseline calibration it is relatively easier to find another foreign supplier than a domestic one (Feenstra et al. (2018) calibration).
  - If substituting with domestic production were equally easy, trade diversion would be muted and adverse spillovers to third countries would be much larger.
  - Countries heavily involved in Asian or American value chains—such as Canada, Korea, and Mexico—would be hurt if re-shoring were easier; in Korea gains may even turn into losses.
  - Robustness check results reported only for scenarios where China introduces export restraints and U.S.–China escalation; unable to find a solution with higher foreign-domestic elasticities in the auto tariff scenario.
- Mechanisms producing spillovers
  - Difficulty of substituting certain foreign goods (electronics and other manufacturing) with domestic production leads to replacement by additional imports from other countries rather than domestic re-shoring.
  - Indirect exposures through global supply chains can be first-order determinants of outcomes (e.g., Korean value-added in Chinese electronics exports).

### VII. Conclusions and policy implications
- Overarching quantitative result
  - GDP in the U.S. falls under most scenarios analyzed, even when China voluntarily limits exports of selected goods to the U.S.
  - Some production sectors and U.S. regions exhibit substantial changes; which sectors gain and which lose depends on the scenario.
- Policy takeaways
  - Results underscore the value of adhering to existing levels of liberalization and highlight risks associated with fragmentation or breakdown of the trading system.
  - The U.S. and trading partners could work constructively to address distortions in the trading system and tap trade policy areas with high growth potential.

*Source: IMF staff calculations and discussion in wpiea2019143 (References section and main text excerpts).*

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

### wpiea2019143 - References

### I. Introduction
- Context and motivation
  - After five decades of negotiations culminating with the creation of the WTO in 1995, low tariffs in advanced economies became a core pillar of the international trading system.
  - The landscape changed abruptly in 2018 when the U.S. administration imposed, or announced they might impose, higher tariffs targeting specific sectors or trading partners.
  - U.S. tariffs are described as responding to concerns about market-distorting government interventions, technology transfer, subsidies to state-owned enterprises, and, in some cases, U.S. national security interests.
- Objectives and approach
  - Use a multi-sector computable general equilibrium model (Caliendo, Feenstra, Romalis and Taylor, 2017; CFRT) to estimate effects of three potential U.S. trade policy actions with systemic importance.
  - Model leverages input-output data covering 165 countries and 17 sectors, and detailed applied-tariff data including pre-tension periods.
- Scenarios analyzed
  - Auto and auto-parts tariffs stemming from a Section 232 (national security) investigation of the sector.
  - Two U.S.-China outcomes associated with a Section 301 (technology transfer) investigation:
    - A “transactional deal” resembling managed trade: calibrated so each side closes half of the bilateral trade deficit.
    - An escalation with higher bilateral tariffs (no transactional deal).
- Key high-level findings (from scenario results preview)
  - Aggregate GDP effects are relatively small but generally negative for most countries, including the U.S.
  - Protected sectors typically do not expand sufficiently to offset losses in competitiveness elsewhere.
  - Spillovers to third countries are generally linked to direct gross trade exposures, though indirect supply-chain exposures can be first-order.
  - Wide sectoral variation implies high adjustment costs in moving to new equilibria.
  - Heterogeneous regional and distributional effects within the U.S.: a transactional deal benefits U.S. states with high shares of primary-sector activity (examples listed: Arkansas, North Dakota, Oklahoma, Wyoming); these same states would suffer most under the auto tariffs scenario.
  - Richer U.S. states may benefit from certain trade scenarios; poorer states seldom benefit.
  - Impacts depend crucially on assumptions such as the extent of U.S. re-shoring in affected sectors.
- Limitations and scope
  - Modeling choices reflect uncertainties in potential negotiation outcomes and retaliation forms.
  - Some policy changes (investment regimes, intellectual property, or rejection of rules-based trading system) are not well captured and may have first-order effects.
  - The employed model is static, describing steady-state implications and ignoring uncertainty effects on investment and activity.

### II. Model, Data and Calibration
- Modeling framework
  - Core model: CFRT (Caliendo et al, 2017) — a multi-sector computable general equilibrium model featuring:
    - Firm heterogeneity within sectors (Melitz-style).
    - Cross-border input-output (intermediate-input) trade (Caliendo and Parro, 2015).
  - Advantages of CFRT relative to related frameworks (e.g., Costinot and Rodriguez-Clare, 2014; CRC):
    - Tariffs modeled as revenue shifters (ad-valorem tariffs apply to final price), implying tariff reductions are analogous to joint reductions in bilateral iceberg costs and fixed exporting costs — potentially amplifying benefits.
    - Avoids non-generic corner solutions that can arise in some imperfect-competition setups with certain elasticity calibrations.
- Shortcomings of the CFRT framework (noted explicitly)
  - No dynamics: the model is static and cannot analyze investment responses or trade-policy uncertainty effects on investment and short-run macro dynamics.
  - Investment omission: cannot capture changes in capital accumulation, productivity effects over medium/long term.
  - Fixed overall trade balances to close the model — standard in the literature but limits intertemporal trade-balance dynamics.
  - Despite fixed trade-balance assumption, authors note that macro fundamentals typically drive saving-investment balances more than sectoral/bilateral policy; dynamic models have found muted overall deficit responses even under large tariff changes.
- Databases and sectoral aggregation
  - Input-output data source: Eora global supply chain database (multi-region input-output table; Lenzen et al, 2013).
    - Original Eora: 26 sectors across 190 countries, annual series 1990-2015.
    - Study uses latest (2015) Eora table.
  - Tariff data: 2015 applied-tariff data from UNCTAD’s Trade Analysis Information System accessed through WITS.
  - Intersection yields a combined I-O database with 17 sectors and 165 countries (Table 1 lists sectors; model calibrated to 17 sectors and 165 countries).
  - The database enables sectoral bilateral exposure analysis across the 165 countries.
- Calibration of elasticities of substitution
  - Two sets of sector-level elasticities are central to CFRT:
    - σs: elasticities of substitution across foreign or domestic intermediate inputs (assumed larger).
    - ωs: elasticities of substitution across composite foreign and domestic inputs (assumed smaller).
  - Rationale: switching between foreign suppliers is empirically and technically easier than switching between foreign and domestic suppliers.
  - Estimation strategy:
    - Elasticities across intermediate inputs (σs) are newly estimated in this paper following steps after Table 1 in CFRT: starting with estimating Гs in equation (1) using the triangular trade procedure.
- Comparative notes and illustrative sensitivities reported
  - The text documents that certain elasticity calibrations can produce extreme results in other models: for example, in a 28-sector and 15-country application of CRC-style models with standard elasticity estimates, a 25 percent increase in U.S. auto tariffs leads to a 100-percent drop in total Canadian auto exports (authors note these results are available upon request).
  - The paper discusses sensitivity of results to parameter calibrations, including elasticities of substitution.

### III. Comparative Role of Trade vs. Macroeconomic Models
- Complementarity with macro models
  - CFRT trade model sheds light on long-term allocative inefficiencies and supply-chain disruptions across 17 sectors and 165 countries.
  - Macroeconomic models (e.g., IMF’s GIMF) can provide insights on short-term demand effects, rigidities, and macro distortions.
  - Both approaches are useful: trade models capture long-run reallocation across sectors and countries; macro models capture short-run dynamics and demand-driven channels.
- Important caveats reiterated
  - CFRT’s static nature means it does not capture uncertainty-driven investment responses and short-run business-cycle effects.
  - Existing literature documents that trade-policy uncertainty can significantly affect investment and activity (see Handley and Limão, 2017; IMF Scenario Box references).

### IV. Organizational Roadmap of the Paper (as stated)
- Section II: model, databases, and elasticity estimation.
- Section III: description of scenarios and results.
- Section IV: sensitivity analysis to key parameters including elasticities of substitution.
- Section V: conclusions.

*Source: IMF staff calculations and discussion in wpiea2019143 (References section and main text excerpts).*

### Section 4 of Caliendo and Parro (CP 2015) for the goods-producing sectors in Eora. Data on

### Section 4 of Caliendo and Parro (CP 2015) for the goods-producing sectors in Eora

### Data and estimation
- Tariffs for each sector, s, between any two countries, j and k, (τ^s_kj) and corresponding trade flows (X^s_kj) for the year 2016 are from UNCTAD Trade Analysis Information System and sourced through WITS.
- The estimating equation (shown as Equation (1) in the source) is estimated over three samples:
  - all bilateral trade data;
  - excluding 5 percent of the smallest flows;
  - excluding 10 percent of the smallest flows (preferred sample; excludes small outliers that unduly tilt the estimation).
- Equation (1) (log-linear form involving sectoral trade flows and tariffs) is the basis for estimating sectoral trade elasticities Γ_s.

### Sector-level trade elasticities (Table 2)
- Table 2 compares CP (2015) estimates and the preferred calibration for Γ_s:
  - 1 Agriculture, Hunting, Forestry and Fishing: CP (2015) 9.1 — Preferred 9.1
  - 2 Extractive Industries: CP (2015) 13.5 — Preferred 19.0
  - 3 Food, Beverages and Tobacco: CP (2015) 2.6 — Preferred 2.5
  - 4 Textiles, Textile Products, leather and footwear: CP (2015) 8.1 — Preferred 3.0
  - 5 Wood, Paper, Printing and Publishing: CP (2015) 14.0 (note 1/) — Preferred 5.4
  - 6 Petroleum, Chemical, Non-Metallic Mineral Products: CP (2015) 18.1 (note 1/) — Preferred 12.1
  - 7 Metal Products: CP (2015) 7.0 — Preferred 13.9
  - 8 Electrical equipment and Machinery: CP (2015) 4.8 (note 1/) — Preferred 2.6
  - 9 Transport equipment: CP (2015) 8.2 — Preferred 8.0 (note 3/)
  - 10 Manufacturing, Nec; Recycling: CP (2015) 4.0 — Preferred 4.0 (note 2/)
- Notes:
  - 1/ Simple-averaged elasticities of more disaggregated sectors.
  - 2/ Using the same as CP.
  - 3/ Simple-average of elasticities across goods producing sectors.
- Additional calibration details:
  - For sectors 1 and 10 the values from CP are used because estimates were not robust across samples.
  - For sector 9 estimates were not robust; following CP the mean elasticity across all other goods-producing sectors is used.
  - The Γ_s estimates correspond to the first set of numbers in Table 1 of CFRT.

### Calibration of σ_s and foreign-domestic elasticities ω_s
- σ_s is calibrated based on the relation between sectoral elasticities of substitution and the Pareto shape parameter from Eaton, Kortum and Kramarz (2011):
  - σ_s = (Γ_s + 1)/1.5
- To calibrate foreign-domestic elasticities of substitution ω_s:
  - Estimates by Feenstra et al (2018; FLOR) are used.
  - Codes from the published version of FLOR were used to estimate σ_s and ω_s under their two-stage least squares procedure consistent with their table 3.
  - The ratio σ_s/ω_s is computed for each sector s; FLOR sectors are matched to Eora-based sectors; ω_s is then computed using own σ_s estimates.
  - For sectors that could not be matched, the simple average of the ratio, which is 1.76, is used.
- The resulting sectoral elasticities of substitution are presented in Figure 2 (source: IMF staff calculations), showing:
  - sectoral elasticities "between foreign suppliers" and "between foreign and domestic suppliers" for the goods-producing sectors listed.

### Scenarios analyzed (Section III.A)
- The paper quantifies the impact of three scenarios. Baseline uses tariffs in effect at end-2018 except for Scenario 3 where specified tariff changes are excluded from the baseline.
- Scenario 1: U.S. tariffs on imports of autos and auto-parts
  - U.S. introduces a 25 percent tariff on all imports of vehicles and auto parts, independently of origin.
  - Retaliation: all trading partners directly affected (those exporting autos or auto-parts to the U.S.) impose an equivalent 25 percent tariff on imports of vehicles and auto-parts from the U.S.
  - Because some trading partners have bilateral trade surpluses in autos/auto-parts with the U.S., these partners impose additional uniform tariffs across all other U.S. goods exports such that their overall static tariff revenue equals that of the U.S. Mathematically the equality:
    - 25% × value of U.S. auto & auto‑part imports from country i = 25% × value of country i'’s auto & auto‑part imports from the U.S. + Z% × value of country i'’s all other imported goods from the U.S.
    - where the tariff rate Z ≥ 0 is chosen so that the equality holds.
  - Mapping: all tariffs on autos and auto parts are mapped to the “transport equipment goods” sector using the share of autos and auto parts in total bilateral flows of auto-related, trains, airplanes and other transport equipment with the U.S. in 2017 from UN/Comtrade.
  - Side note: USMCA side-letters could exempt Canada and Mexico up to quotas; alternative selective retaliation effects on U.S. states considered in Box 1.
- Scenario 2: A “transactional deal” between the U.S. and China
  - Objective: reduce the bilateral deficit between U.S. and China by half.
  - Two modeled aspects:
    1. Increase Chinese purchases of U.S. goods such that the bilateral deficit is halved via:
       - simultaneous reduction in Chinese non-tariff import barriers vis-à-vis the U.S. and an increase in non-tariff import barriers vis-à-vis non-U.S. partners;
       - each change calibrated to achieve half the goal (each reduces the bilateral deficit by one-quarter);
       - relative change in barriers scaled by the inverse of the elasticity of substitution of each good, except for three sectors (agriculture, extractive industries and transport equipment) where especially large changes are introduced.
    2. Reduction in Chinese exports of electronics and other manufactured goods to the U.S. via export tariffs on these sectors’ goods bound for the U.S., with sectoral changes inversely proportional to elasticities of substitution; this reduction is calibrated to also reduce the bilateral deficit by one-half.
  - Notes: rationale for inverse-elasticity scaling includes avoiding unreasonable quantity changes for highly elastic sectors and providing a Ramsey-like lower bound on distortions.
- Scenario 3: Escalation of trade tensions between the U.S. and China
  - U.S. imposes a 25 percent uniform tariff on all goods imported from China and China retaliates with the same 25 percent uniform tariff on all U.S. goods.
  - The escalation scenario abstractly captures higher bilateral tariffs; other policy changes (e.g., restrictions on FDI or technology-transfer policies) are not quantified by the model.

### Results overview and key findings (Section III.B — Effects on the U.S.)
- Aggregate GDP effects:
  - U.S. GDP falls under all three scenarios ― by around 0.2-0.3 percent of GDP ― following the introduction of tariffs.
- Sectoral heterogeneity:
  - Sectoral value-added responses are heterogeneous and scenario-dependent.
  - Example: U.S. agricultural sector:
    - Benefits noticeably under the “transactional deal” where China purchases more U.S. goods (China treats such goods highly preferentially).
    - Contracts under the auto tariffs scenario as several U.S. trading partners retaliate by increasing tariffs on that sector to match U.S. tariff intensity.
  - The reduction in Chinese exports of selected manufacturing goods to the U.S. (part of Scenario 2) does not result in meaningful expansion of protected U.S. sectors because those sectors rely on Chinese inputs; protected sectors face higher input costs as foreign inputs are not easily substitutable with domestic production.
- Regional and distributional implications within the U.S.:
  - Heterogeneous sectoral responses imply diverse regional outcomes across U.S. states due to geographical specialization.
  - A “transactional deal” where China buys more from the U.S. would benefit states with a high share of the primary sector (mining, hydrocarbons, and agriculture) — e.g., Arkansas, North Dakota, Oklahoma, and Wyoming — which would suffer under the auto tariffs scenario.
  - Escalation with China or a deal where China restrains exports to the U.S. would have a more homogeneous negative effect across states.
  - Distributional implication: richer states (higher median incomes and a higher proportion of the rich) would benefit relative to poorer ones under the auto tariff scenario and under intensification of the U.S.–China trade war; poorer states would benefit only when China purchases more primary products from the U.S.
- Figures referenced (IMF staff calculations):
  - Figure 3: Real GDP impact on the U.S. and selected trading partners (percent change) — shows U.S. and partner GDP changes across scenarios.
  - Figure 4: Impact on U.S. Sectors (percent change and changes in sectoral value added in billions of USD; constant prices) — shows sizable heterogeneity in sectoral percent changes and USD-level impacts across scenarios.

*Source: Section 4 of Caliendo and Parro (CP 2015) for the goods-producing sectors in Eora; IMF staff calculations (as presented in the provided content).*

### Box 1: Trade Policy and its Potential Effects on U.S. States

### Box 1: Trade Policy and its Potential Effects on U.S. States

### Scenario descriptions and mapping to U.S. states
- Database and mapping:
  - Economic activity broken down into 17 sectors for each of the 50 states and the District of Columbia.
  - U.S. sectoral impacts from simulations are mapped (outside of the model) to state-level output changes based on the share of each productive sector in state output.
- Scenarios analyzed (short description):
  - Auto tariffs: U.S. introduces a 25 percent tariff on imports of autos and auto-parts across the board; trading partners retaliate with the same tariff rate on U.S. auto goods and a uniform tariff rate across all other U.S. goods.
  - Auto tariffs with selective retaliation: trading partners retaliate mainly targeting some “emblematic” U.S. products (e.g. Harley Davidson motorcycles, blue jeans, bourbon whiskey).
  - Intensified U.S.–China trade tensions (“Trade war”): both countries introduce a uniform 25 percent tariff rate across all products.
  - Transactional U.S.–China “trade deal” variants:
    - China restrains exports to the U.S. (restraining some of China’s exports to the U.S.).
    - China purchases additional selected U.S. goods (including agriculture and hydrocarbon products).

### U.S. state-level effects (findings)
- Auto tariffs (uniform retaliation):
  - Primary good sectors (agriculture, mining, and hydrocarbon extraction) tend to suffer large falls in value-added.
  - States with relatively large shares of these sectors experience larger falls in GDP such as: Arkansas, North Dakota, Oklahoma, West Virginia, and Wyoming.
  - These states do not reap the benefits from the increase in output in the U.S. auto industry under this scenario.
- Auto tariffs (selective retaliation):
  - When retaliation targets “emblematic” products, GDP losses of heavily primary-sector states are largely reduced because their exposure to the targeted sectors is lower.
- Trade war with China (both impose 25 percent tariffs across all products):
  - Most sectors experience falls in output; most U.S. states exhibit negative GDP changes.
  - Effects are relatively less heterogeneous across states than under most other trade scenarios.
- Trade deal: China purchases more selected U.S. goods:
  - States relatively exposed to agriculture and hydrocarbon products, and overall less diversified, tend to benefit.
- Trade deal: China restrains exports to the U.S.:
  - A deal that restrains some of China’s exports to the U.S. would lead to losses in all states.
- Socio-economic heterogeneity:
  - Positive changes in state-level GDP under the auto tariff scenario or higher bilateral tariffs with China are positively associated with states having higher median income levels and a higher proportion of rich individuals (and lower proportion of the poor): these scenarios tend to favor richer states and those with a higher proportion of wealthy people and lower poverty rates.
  - By contrast, a trade deal leading to additional purchases of certain U.S. goods by China tends to favor states with a relatively higher share of primary good sectors and would seem to somewhat benefit states with lower average income, a higher proportion of lower earners, and a lower share of the wealthy.

### Quantitative correlations with income measures (Box Table 1 — correlations preserved exactly)
- Notes: *, **, and *** denote statistical significance at the 10, 5, and 1 percent confidence level. “low-income households” and “high-income households” denote those households with an annual income of less than $15,000 and more than $150,000, respectively.
- Correlation of estimated state-level GDP changes with state-level income measures (values shown as in source):
  - Auto tariffs, uniform retaliation: 0.3** 0.4*** -0.1 0.4***
  - Auto tariffs, selected retaliation: 0.5*** 0.6*** -0.2 0.6***
  - Intensification of trade war with China: 0.5*** 0.6*** -0.3* 0.6***
  - Trade deal: restraint on China’s exports to U.S.: 0.2 0.3** -0.1 0.3**
  - Trade deal: China purchases of U.S. goods: -0.1 -0.2* 0.0 -0.3*
- Income categories related to columns in table:
  - States with higher median income
  - States with higher mean income
  - Higher proportion of low-income household
  - Higher proportion of high-income households

### Impacts and spillovers to other countries
- Heterogeneous country-level responses depend on scenario and initial sectoral exposures to the U.S. and other tariff-imposing countries (e.g. China).
- When the U.S. and China impose bilateral tariffs on each other:
  - Both countries tend to suffer.
  - Canada and Mexico—highly integrated with the U.S.—would tend to benefit in the long run from less-intense competition with China in the U.S. market; however, in Scenario 1 (tariffs on vehicles and auto-parts), Canada and Mexico would suffer the most given their significant exposure to the U.S.
  - Germany and Japan would also be affected, to a lesser extent.
- Sectoral examples:
  - Transportation equipment (autos and auto-parts) in Canada and Mexico exhibit significant losses under the 25 percent U.S. auto tariff scenario.
  - China: manufacturing sectors such as electronics and other manufacturing experience significant losses when the U.S. increases tariffs on all Chinese goods (Scenario 3) and when China reduces exports to the U.S. under a transactional deal (Scenario 2).
- Global coverage: simulations include results for 165 countries; outcomes visualized in maps and figures in the Appendix.

### USMCA exemptions and quotas (key numeric details preserved)
- Side-letters to USMCA stipulate certain exemptions for Canada and Mexico if the U.S. were to impose Section 232 tariffs on automobile imports. If ratified:
  - Canada allowed to ship up to 2.6 million cars to the U.S. under USMCA preferences (versus 1.8 million shipped in 2018).
  - Canada allowed $32.4 billion worth of auto parts without getting hit by tariffs.
  - Mexico’s side-letter: quota for auto parts is $108 billion.
- Results from counterfactuals assuming full exemption for Canada and Mexico in Scenario 1:
  - Exports of auto-parts would likely remain within the USMCA quotas.
  - Passenger vehicle quotas: the combination of stringent quotas with simulated expansion in exports implies the quotas would become binding; tariffs would kick in on U.S. vehicle imports from Canada and Mexico unless exports are limited to remain within the quotas.
- Additional note from source:
  - USMCA quotas are already binding for current levels of U.S. vehicle imports from Mexico.

### Role of elasticity of substitution (robustness and sensitivity)
- Elasticities matter: the elasticity of substitution between foreign suppliers and between foreign and domestic suppliers is a key driver of magnitudes and signs of effects.
- Robustness check implemented:
  - Raise sectoral foreign-domestic elasticities to be equal to foreign-foreign ones (operationally, set foreign-domestic elasticities equal to foreign-domestic elasticities divided by 1.01 to avoid corner solutions).
  - Intuition: this makes re-shoring production easier (easier to substitute foreign goods with domestic production).
- Findings from sensitivity:
  - If re-shoring production is easier for the U.S., potential losses would be larger for most other countries; most countries (other than the U.S.) lie below the 45-degree line in Figure 9.
  - In baseline calibration, upon an increase in barriers with a specific partner it is relatively easier to find another foreign supplier than a domestic one (Feenstra et al. (2018) calibration).
  - If it was equally easy to substitute with domestic production, trade diversion would be muted and adverse spillovers to third countries would be much larger.
  - Countries heavily involved in Asian or American value chains—such as Canada, Korea, and Mexico—would be hurt if re-shoring were easier; in Korea gains may even turn into losses.
- Computational note: robustness check results reported only for scenarios where China introduces export restraints and U.S.–China escalation; unable to find a solution with higher foreign-domestic elasticities in the auto tariff scenario.

### Concluding findings and policy implications
- Overarching quantitative result:
  - GDP in the U.S. falls under most scenarios analyzed, even when China voluntarily limits exports of selected goods to the U.S.
  - Some production sectors and U.S. regions exhibit substantial changes; which sectors gain and which lose depends on the scenario.
- Mechanisms emphasized:
  - Difficulty of substituting certain foreign goods (electronics and other manufacturing) with domestic production leads to replacement by additional imports from other countries rather than domestic re-shoring.
  - Indirect exposures through global supply chains can be first-order determinants of outcomes (e.g., Korean value-added in Chinese electronics exports).
- Policy takeaways:
  - Results underscore the value of adhering to existing levels of liberalization and highlight risks associated with fragmentation or breakdown of the trading system.
  - The U.S. and trading partners could work constructively to address distortions in the trading system and tap trade policy areas with high growth potential.

*Source: Box 1, “Trade Policy and its Potential Effects on U.S. States,” wpiea2019143.*

### REFERENCES

### REFERENCES

### Reference list
- Caliendo, L., Feenstra, R.C., Romalis, J., and Taylor, A.M., 2017, “Tariff Reductions, Entry, and Welfare: Theory and Evidence for the Last Two Decades,” NBER Working Paper 21768 (National Bureau of Economic Research: Cambridge).  
- Caliendo, L., and Parro, F., 2015, “Estimates of the Trade and Welfare Effects of NAFTA,” Review of Economic Studies, 82, 1-44.  
- Costinot, A., and Rodriguez-Clare, A., 2014, “Trade Theory with Numbers: Quantifying the Consequences of Globalization,” Chapter 4 in Handbook of International Economics, 2014 (Vol. 4) pp. 197-261.  
- Eaton, J., S. Kortum and F. Kramarz, 2011, “An Anatomy of International Trade: Evidence From French Firms,” Econometrica, vol. 79 (5), pp. 1453-1498.  
- Eaton, J., Kortum, S., Neiman, B. and Romalis, J., 2016, “Trade and the Global Recession,” American Economic Review, vol. 106 (11), pp. 3401–3438.  
- Feenstra, R.C., Luck, P., Obstfeld, M., Russ, K.R., 2018, “In Search of the Armington Elasticity,” The Review of Economics and Statistics, vol. 100(1), pp. 135-150.  
- Handley, K., and Limão, N., 2017, “Policy Uncertainty, Tarde, and Welfare: Theory and Evidence for China and the United States,” American Economic Review, 107 (9), pp. 2731-2783.  
- International Monetary Fund, 2018a, World Economic Outlook (Washington, October).  
- International Monetary Fund, 2018b, External Sector Report (Washington, July).  
- International Monetary Fund, 2019, World Economic Outlook (Washington, April).  
- International Monetary Fund, World Bank and WTO (IMF-WB-WTO), 2018, Reinvigorating Trade and Inclusive Growth.  
- Kucheryavvy, K., Lyn, G., and Rodriguez-Clare, A., 2016, “Grounded by Gravity: A Well-Behaved Trade Model with Industry-Level Economies of Scale.” NBER Working Paper 22484 (National Bureau of Economic Research: Cambridge).  
- Lenzen, M., Moran, D., Kanemoto, K., and Geschke, A., 2013, “Building EORA: A Global Multi-Region Input-Output Database at High Country and Sector Resolution.” Economic Systems Research, 25 (1), pp. 20–41.  
- Mattoo, A. and R. Staiger, 2019, “Trade Wars: What Do They Mean? Why are They Happening Now? What are the Costs?” NBER Working Paper No. 25762.  
- Melitz, M.J., 2003, “The impact of trade on intra-industry reallocations and aggregate industry productivity,” Econometrica, 71 (6), pp. 1695-1725.  
- Melitz, M.J., and Redding, S.J., 2014, “Heterogeneous Firms and Trade,” Handbook of International Economics, 4, pp. 1-54.  
- Ossa, R., 2016, “Quantitative Models of Commercial Policy,” Chapter 4 in K. Bagwell and R. W. Staiger, Handbook of Commercial Policy, Volume 1A, North Holland.  
- Ravikumar, B., A. M. Santacreu and M. J. Sposi, 2019, “Capital Accumulation and Dynamic Gains from Trade,” Working Paper, Federal Reserve Bank of St. Louis.  
- Reyes-Heroles, R., 2016, “The Role of Trade Costs in the Surge of Trade Imbalances,” mimeo.

*Source: wpiea2019143 - REFERENCES*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019143.pdf_