## wpiea2024218-print-pdf

## Source details

**Canonical URL:** [wpiea2024218-print-pdf](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024218-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024218-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024218-print-pdf.pdf.json)

---

### EU climate goals, EV policy targets, and market implications
- EU objective: make the economy climate neutral by the middle of this century; rapid EV transition targets under the “Fit for 55” package.
- Regulatory targets:
  - 55 percent reduction in CO2 emissions for new cars and 50 percent for new vans from 2030 to 2034 compared to 2021 levels.
  - 100 percent CO2 emission reductions for both new cars and vans from 2035.
- Market implications and observed shares:
  - Third‑party estimates: battery EV cars must make up around 65 percent of total new car sales by 2030; by 2035 all new car sales should be fully electric.
  - Fully electric vehicle sales currently account for about 15 percent of total car sales in Europe.
- Drivers of change: climate goals, energy security (reducing fossil fuel import dependence), and changing consumer preferences.

### China’s role in EV production, prices, and exports
- Market shares and volumes:
  - EVs account for about a third of domestic vehicle sales in China by early 2024.
  - China accounts for more than half of global EV sales.
  - 2023 volumes: around 1.5 million pure‑electric vehicles exported; 5.4 million pure‑electric vehicles sold domestically; pure electric car sales globally ≈ 10 million cars.
- Prices and productivity:
  - Sales‑weighted average price of electric cars (before purchase subsidy) in China is already lower than ICE cars.
  - China pure‑electric EV export unit values hover between USD 18,000 and USD 23,000 (12‑month moving average).
  - As of 2023, Chinese EVs in Europe sell for about 20 percent less than similar European models.
  - Example price disparities: Volkswagen’s ID.4 sells for nearly 50 percent more in Europe than in China; BYD’s Seal U comfort costs nearly double in Europe compared to China.
- Trade and FDI:
  - China exported over 4 million cars in 2023, of which about one‑third were electric.
  - EU is China’s largest export market: in 2023, 21.7 percent of battery‑electric car sales in the EU were made in China; Chinese brands accounted for 7.6 percent of battery‑electric car sales in the EU in 2023 (up from 2.9 percent made in China and 2 percent Chinese brands in 2020).
  - Chinese FDI: around 30 percent of 2023 Chinese FDI in the EV sector went to Europe; BYD announced a plant in Hungary in late 2023; Hungary received more than half of total FDI from China in 2022 and 2023 (Rhodium Group).

### Economic importance of the automotive sector in Europe and policy responses
- Sector importance:
  - Transport equipment is Europe’s largest industrial sector.
  - Direct and indirect employment by the automotive sector ≈ 7 percent of total EU employment.
  - Automotive share in GDP: 4–5 percent in Czechia, Slovakia, Hungary, and Germany; around 0.5 in France, Belgium and Bulgaria.
  - CESEE economies supply nontrivial inputs into automobile production of Germany, France and Italy.
- Policy response:
  - EU anti‑subsidy investigation led to provisional tariffs in June 2024 and permanent countervailing duties effective from July for a period of 5 years (vote on October 4).
  - Additional tariffs range between 7.8 to 35.3 percent on top of existing 10 percent EU most‑favored‑nation tariffs; combined peak tariffs can reach about 45 percent for some EV producers based in China.
  - Tariffs were calculated based on estimates of state aid to firms; cooperating firms had duties cut; tariffs apply to Chinese firms and subsidiaries of EU and U.S. firms operating and exporting out of China.

### Modeling approach, scenario design, and calibration
- Models used:
  - GIMF-GVC: multi‑region, micro‑founded, dynamic general equilibrium model for short‑run dynamics (annual, 7 regions, 3 sectors including Automotive).
  - Trade Model (based on Cuñat and Zymek, 2024): dynamic multi‑country, multi‑sector quantitative trade model for long‑run steady‑state effects (68 economies, 40 sectors; Motor Vehicles ISIC C29 + Other Transport Equipment ISIC C30 = Automotive).
- Key model assumptions:
  - Input‑output data contain a single automotive sector; authors assume two unobserved car types—ICE and EV—with broadly similar production structures.
  - Baseline calibration anchored in Japan’s auto rise (1960–1980s): assume a comparable 15 percentage‑point increase in Chinese market penetration, compressed into 5 years.
- EV‑shock calibration:
  - Modeled as two combined shocks:
    - Supply shock: increases total factor productivity (TFP) in China’s automotive sector by 15 percent.
    - Demand shock: policy‑driven shift toward EVs that disproportionately benefits Chinese vehicles, generating a 15 percentage‑point increase in Chinese market share over 5 years.
  - Long‑run (trade model) benchmark: 16.3 percent of EU car spending goes to Chinese vehicles versus 1.3 percent in 2016–19.
- Alternative scenarios:
  - Trade restrictions: simulate additional tariffs of 25 percent or 100 percent on automotive imports from China.
  - FDI scenario: increased Chinese FDI such that 40 percent of Chinese EVs sold in Europe are produced in Europe (analogous to Japan’s late 1980s U.S. experience); implemented via reductions in equity premia, markups and improvements in investment efficiency in European automotive sectors.

### Key quantitative findings and heterogeneous impacts
- EU aggregate impacts:
  - Short‑term output cost of the transition is small: decline of about ¼ percent in real GDP for most countries over 5 years; close to zero over the long term.
- Country heterogeneity (GIMF‑GVC, 5‑year effects):
  - Hungary: GDP decline after 5 years = 1 percent.
  - Czechia: GDP decline after 5 years = 1.5 percent.
  - Germany, France, Italy and Rest of Europe: around 0.2 percent decline after 5 years.
- Steady‑state (Trade model) outcomes:
  - Slovakia (in trade model): GDP losses exceed 1 percent in the long run.
  - Worst affected: smaller, open economies with large auto sectors (CESEE), with losses persisting in steady state for some.
  - Economies without sizable auto sectors benefit from cheaper Chinese EV imports.
- Labor market impacts:
  - Short/medium term (GIMF‑GVC): some employment losses in motor vehicle sector; reallocation to tradable and non‑tradable sectors; wages fall more in affected regions.
  - Steady state (Trade model): labor relocation away from auto sector up to 2.6 percent of the workforce in Slovakia and 1.7 percent in Hungary.
- Mechanisms of loss:
  - Demand shifts away from domestically produced ICE vehicles toward cheaper Chinese EVs; input‑output linkages amplify losses through upstream suppliers; terms‑of‑trade deterioration for affected exporters.

### Protectionist response (tariffs) — simulated effects
- Tariff scenarios: additional 25 percent and 100 percent tariffs simulated on Chinese automotive imports.
- Short/medium and long‑run effects:
  - Tariffs reduce Chinese car imports but worsen GDP outcomes in both short and long run.
  - 25 percent tariff example: instead of a 15 percentage‑point rise in Chinese share, the share rises by only 2.5 percentage points, but higher consumer prices more than offset producer gains, leaving long‑run real GDP lower.
  - Tariff channels of harm:
    - Higher EV consumer prices reduce real disposable income and demand for other goods.
    - Tariffs increase input costs for European auto producers (due to automotive intermediate goods), lowering labor productivity and real wages economy‑wide.
    - Tariffs cause exchange rate appreciation, lower investment, and raise production costs—effects larger for Hungary than for Germany, France, Italy.
- Climate effects of tariffs:
  - Tariffs slow EV adoption but impacts modest:
    - Under a 100 percent tariff, Chinese vehicles < 10 percent of total new cars sold.
    - Full ban yields additional emissions equal to just 4 percent of 2023 EU car sector emissions (assuming unchanged climate policies).
  - Conclusion: tariffs have modest direct emissions impact if EU climate targets are maintained; postponing targets would raise emissions substantially.

### Increased Chinese FDI scenario — simulated effects
- Implementation (models do not allow foreign ownership directly):
  - Lower equity premia and markups in European automotive sector; productivity and investment efficiency improvements; partial reversal of demand shift to reflect local Chinese production in Europe.
- Two allocations:
  - Proportional FDI across EU economies.
  - Hungary‑focused FDI with disproportionate allocation to Hungary.
- Outcomes:
  - Short/medium term (GIMF‑GVC): proportional FDI reduces 5‑year GDP losses (e.g., effects on GDP levels after 5 years ≈ 0.5 percent for Hungary and 0.15 percent for Germany, France, Italy and Rest of Europe); Hungary‑focused FDI can make Hungary a modest net GDP winner after 5 years.
  - Steady state (Trade model): FDI undoes some economic losses for worst‑affected economies; distribution of averted losses depends on which economies attract FDI.
- Long‑run benchmark: scenario assumption of 40 percent of Chinese EVs produced in EU mirrors Japan’s experience where about 40 percent of foreign sales were through affiliates within a decade of protection.

### Climate accounting and EV adoption path (assumptions and numbers)
- EV purchase share path for EU (linear path to 100 percent by 2035; Table A3.1):
  - 2023: 12
  - 2024: 19
  - 2025: 27
  - 2026: 34
  - 2027: 41
  - 2028: 49
  - 2029: 56
  - 2030: 63
  - 2031: 71
  - 2032: 78
  - 2033: 85
  - 2034: 93
  - 2035: 100
- Emissions accounting parameters:
  - η = 0.5 (relative emissions intensity of EVs vs ICE, based on IEA Life Cycle Assessment Calculator).
  - δ = 1/15 (implying average vehicle lifespan of 15 years).
  - Initial condition for EU economies: X_d0^EV = 0.
- Example projection:
  - For a linear transition path, 63 percent of new vehicles sold in 2030 would be EVs (Table A3.1 reference).
- Result: tariffs or bans modestly slow EV adoption trajectories and add limited emissions if climate targets hold; main downside of protectionism is higher economic cost of the EV transition.

### Contribution to literature and relation to other studies
- This analysis provides combined short‑term dynamic (GIMF‑GVC) and long‑run steady‑state (Trade model) assessments of simultaneous EV transition and rising Chinese EV imports.
- Complements descriptive and partial‑equilibrium studies and augments recent trade‑model analyses by examining GDP, employment, and climate‑transition implications using multiple models.
- Related empirical findings cited:
  - Ayerst and Noumon (2023): electrification in Czechia can lead to modest output increases but large employment losses due to lower labor intensity.
  - Attinasi and others (2024): a simulated 50 percent price drop in Chinese EVs leads to large shifts in global EV market shares and a small, 0.1 percent drop in overall EU production.
  - EU automaker R&D: 5.5 percent of net sales (EU Industrial R&D scorecard, 2023) versus 5.3 and 5.1 percent for Chinese and American top automakers.

### Stylized conclusions and policy recommendations
- Aggregate conclusion:
  - EU‑wide GDP cost of increased Chinese EV imports and EV transition likely small in short term and close to zero long term; large heterogeneity across countries.
- Key country findings:
  - Central European producers face substantial short‑term losses (1–1½ percent of real GDP over 5 years) and persistent long‑run impacts for some.
- Policy recommendations:
  - Proactive labor market policies to ease worker re‑skilling and facilitate sectoral transitions.
  - Caution on protectionist tariffs: tariffs magnify GDP costs, reduce purchasing power, raise input costs, and risk retaliation; environmental benefits of tariffs are modest if climate goals remain unchanged.
  - Encourage FDI that ramps up European EV production to mitigate localized losses.
  - Pursue productivity‑enhancing measures in the EU auto sector to offset competitive pressures.
  - Preserve climate goals (e.g., “Fit‑for‑55” 2035 targets); postponement would raise emissions and impair growth.
- Caveats:
  - Scenarios anchored on Japan’s experience but outcomes could differ; model limitations include abstracting from labor reallocation costs and not modeling potential retaliation to tariffs; emissions estimates assume EU climate targets are met.

### Annex: Japan’s historical experience — relevant numeric benchmarks and lessons
- Japan’s U.S. market penetration: share of imported Japanese cars in U.S. sales rose from close to 0 in 1965 to 15 percent in 1985.
- Japanese auto productivity and cost advantages:
  - Fuss and Waverman (1985): Japanese auto industry annual TFP growth in the 1970s = 4.3 percent; Canada = 1.4 percent; U.S. = 1.6 percent.
  - Gomez‑Ibanez and Harrison (1982): Japanese small cars’ cost advantage ≈ between 13.0 and 16.1 percent around 1980.
  - Börsch‑Supan and Romer (1998): by early 1990s Japan’s auto industry TFP = 120 percent of the U.S.; = 171 percent of Germany.
  - Japan’s auto industry productivity growth was 1–3 percentage points higher than in the U.S. during the 1970s and 1980s.
- Protectionism and FDI benchmarks:
  - VER quantitative limits: 1981 = 1.68 million; 1984 = 1.85 million; 1985–1992 = 2.3 million passenger cars annually; VER implied tariff peak = 21 percent.
  - Japan outward FDI: 1982–1989 outward FDI increased about eight times in dollar terms.
  - Historical affiliate production: within a decade of protection, about 40 percent of foreign sales were done through affiliates (U.S. experience).
- Lesson for scenarios:
  - Authors assume a compressed Japan‑like market share rise for China: 15 percentage‑point increase in EU market share over 5 years; investigate trade barriers and FDI reallocation scenarios informed by Japan’s historical patterns.

*Source: wpiea2024218-print-pdf*

### 1. Introduction ........................................................................................................

### 1. Introduction

### EU climate goals, EV policy targets, and market implications
- The EU aims to make its economy climate neutral by the middle of this century and has set rapid EV transition targets under the “Fit for 55” package.
- Current regulations require a 55 percent reduction in CO2 emissions for new cars and 50 percent for new vans from 2030 to 2034 compared to 2021 levels, and 100 percent CO2 emission reductions for both new cars and vans from 2035.
- Third-party estimates suggest sales of battery EV cars in the EU must make up around 65 percent of total new car sales by 2030 and that by 2035 all new car sales should be fully electric.
- Fully electric vehicle sales currently account for about 15 percent of total car sales in Europe.
- Drivers of the shift include climate goals, energy security concerns (reducing dependence on fossil fuel imports), and changing consumer preferences.

### China’s role in EV production and exports
- China increasingly dominates the EV sector, including batteries.
- Within China, EVs already account for about a third of domestic vehicle sales and more than half of global EV sales.
- The sales-weighted average price of electric cars (before purchase subsidy) in China is already lower than that of internal combustion engine (ICE) cars.
- China’s EV exports have more than quadrupled over the last three years; in 2023 China exported over 4 million cars, of which about one-third were electric.
- The EU is China’s largest export market: in 2023, 21.7 percent of battery-electric car sales in the EU were made in China; Chinese brands accounted for 7.6 percent of battery-electric car sales in the EU in 2023, up from 2.9 percent (made in China) and 2 percent (Chinese brands) in 2020.
- In terms of total (ICE+EV) car sales, battery-electric vehicles made in China and by Chinese brands correspond to about 3 and 1 percent, respectively.
- The average purchase price of battery-electric vehicles in the EU averaged 46,000 euros, more than twice the average price of those imported from China.
- Chinese EVs sold in Europe are about 20 percent cheaper than similar models made by European manufacturers.

### Economic importance of the automotive sector in Europe and policy responses
- Transport equipment is Europe’s largest industrial sector.
- Direct and indirect employment by the automotive sector accounts for about 7 percent of total EU employment.
- The automotive sector’s share in GDP ranges from 4-5 percent in Czechia, Slovakia, Hungary, and Germany to around 0.5 in France, Belgium and Bulgaria.
- Central, Eastern, and Southeastern European (CESEE) economies account for a nontrivial share of inputs into automobile production of Germany, France and Italy.
- Policy response: the EU initiated an investigation into subsidized EVs from China that led to provisional tariffs in June 2024 and permanent tariffs for a 5 year period in October 2024; combined with existing most-favored nation tariffs, peak tariffs can reach about 45 percent for some EV producers based in China.
- Tariffs were calculated based on estimates of how much state aid each firm received; companies that cooperated with the EC probe saw their duties cut. Tariffs apply to Chinese firms and subsidiaries of EU and U.S. firms operating and exporting out of China.

### Modeling approach, scenario design, and calibration
- The paper uses two state-of-the-art models:
  - GIMF: a multi-region, micro-founded, dynamic general equilibrium model of the global economy (references: Kumhof and others, 2010; Anderson and others, 2013) to simulate short-run impacts.
  - A dynamic multi-country, multi-sector quantitative trade model based on Cuñat and Zymek (2024) to describe long-run steady state effects at greater granularity.
- Both models have a single representative automotive sector in available input-output data; authors assume the automotive sector contains two unobserved car types—ICE and EVs—with broadly similar production structures.
- Main scenario calibration:
  - Anchored in the historical episode of Japan’s auto rise in the 1960–1980s, when Japanese cars reached around 15 percent market share in the U.S. in the mid-80s.
  - The baseline calibration assumes China achieves a similar 15 percentage-point increase in market penetration but over a shorter time span of 5 years—about twice as fast as the U.S.-Japan experience.
- The “EV-shock” is modeled as two combined shocks:
  - A generic supply shock that increases total factor productivity (TFP) in China’s automotive sector as a whole.
  - A policy-driven demand shift toward EVs that disproportionately benefits demand for Chinese vehicles due to an assumed comparative advantage in the EV car type, generating a 15 percentage-point increase in Chinese market share over 5 years.
- Baseline steady state for comparisons assumes China’s productivity growth does not take place and the EU does not introduce an EV transition target.

### Key quantitative findings and heterogeneous impacts
- For the EU as a whole:
  - The short-term output cost of the transition to EVs is nontrivial but small, with a decline of about ¼ percent in real GDP for most countries over 5 years, and close to zero over the long term.
- Heterogeneity across countries:
  - Losses are more significant for economies heavily reliant on the car sector, mainly in Central Europe.
  - Hungary sees a 1 percent decline in real GDP over 5 years.
  - Czechia sees a 1.5 percent decline in real GDP over 5 years, with the negative GDP impact persisting in the steady-state.
  - EU economies without a sizable auto production sector benefit from cheaper EV imports from China.
- Labor market implications:
  - Both models point to some employment losses in the motor vehicle sector and the need for workers to shift to other sectors, highlighting the need for policies to ease the transition.
- Policy responses and their impacts:
  - Tariffs on Chinese EVs would deepen losses and raise the GDP cost of the EV transition, including for countries heavily reliant on auto production.
  - The impact of tariffs on emissions would be generally modest, provided tariffs do not cause policy targets to be postponed.
  - Increased Chinese FDI that results in 40 percent of Chinese EVs sold in Europe being produced in Europe (analogous to Japan’s late 1980s experience in the U.S.) would help economies receiving the FDI; effects can be particularly large for smaller (Central European) economies if additional FDI flows disproportionately to them.

### Contribution to literature and relation to other studies
- The paper fills a gap by providing a comprehensive model-based analysis capturing both short-term dynamic and steady-state GDP effects of the dual shocks (EV transition and growing penetration of Chinese EV imports).
- It complements descriptive and partial-equilibrium studies by policy institutions and industry associations and augments recent trade-model analyses (e.g., Kiel Institute’s KITE) by examining GDP, employment, and climate-transition implications using multiple models.
- Related findings from the literature cited:
  - Ayerst and Noumon (2023): electrification in Czechia can be challenging, leading to modest output increases but large employment losses due to lower labor intensity.
  - Attinasi and others (2024): a simulated 50 percent price drop in Chinese EVs due to subsidies leads to large shifts in global EV market shares and a small, 0.1 percent drop in overall EU production.
  - Broader green-transition literature (IMF 2022a; IMF, 2022b; Bluedorn and others, 2022) indicates that with the right policies output losses can remain manageable but worker adjustment needs active policies.

*Source: wpiea2024218-print-pdf - 1. Introduction*

### Section 5 concludes.

### Section 5 concludes

### Letting the past and the present inform possible scenarios
- The authors use Japan’s postwar entry into the U.S. car market (1960s–1980s) as a historical benchmark to inform scenarios for China’s rise in electric vehicles (EVs).
- The analysis documents (i) the speed and extent of Japan’s rise using long-run input-output data, (ii) parallels with China’s EV expansion and recent EU policy responses, and (iii) how these comparisons inform model scenario choices.

### 2.1. Japan’s rise in the 1960–1980s — key observations
- Japanese auto industry phases:
  - Rapid domestic expansion during the 1960s.
  - Export-led growth during the 1970s.
  - Regionalization through FDI since the early 1980s.
- U.S. market penetration:
  - Share of imported Japanese cars in U.S. sales rose from close to 0 in 1965 to 15 percent in 1985.
  - Exports to the U.S. rapidly increased in the 1970s and peaked before Japanese producers switched to supplying the U.S. via local production.
- Trade protection and FDI response:
  - U.S. restrictions culminated in voluntary export restraints in the late 1970s/early 1980s.
  - By the 1980s, Japanese firms pursued regionalization via FDI; within a decade of protection, about 40 percent of foreign sales were done through affiliates.
- Historical tariff benchmark:
  - Equivalent peak tariffs during Japan’s episode reached 21 percent.

### 2.2. China’s rise in the 21st century — key statistics and patterns
- EV adoption and sales (China):
  - By early 2024, EVs made up around 32 percent of total sales in China.
    - Pure-electric vehicles: 22 percent of total sales.
    - Plug-in electrics: 10 percent of total sales.
  - 2023 volumes:
    - Around 1.5 million pure-electric vehicles exported.
    - 5.4 million pure-electric vehicles sold domestically.
    - Pure electric car sales globally amounted to about 10 million cars (IEA, 2024).
- Labor productivity (2015–2023):
  - China’s automobile sector labor productivity increased at an annualized rate of 8.8 percent.
  - China: all industries annualized growth: 9.2 percent.
  - Labor productivity growth in China’s auto sector was significantly higher than in the U.S. and Germany.
- Export prices and Europe price gap:
  - China pure-electric EV export unit values hover between USD 18,000 and USD 23,000 (12-month moving average).
  - As of 2023, Chinese EVs in Europe were selling for about 20 percent less than similar models made by European manufacturers.
  - Example price disparities (industry source notes):
    - Volkswagen’s ID.4 model sells for nearly 50 percent more in Europe than in China.
    - BYD’s Seal U comfort costs nearly double in Europe compared to China (Rhodium Group, 2024).
- EU policy response:
  - European Commission announced additional provisional tariffs on Chinese-made EVs in June 2024 following a WTO-compliant anti-subsidy investigation.
  - A vote on October 4 made provisional countervailing duties permanent and effective from July for a period of 5 years.
  - Additional tariffs range between 7.8 to 35.3 percent, on top of existing 10 percent EU most-favored-nation tariffs on car imports.
- Chinese EV-related FDI in EU:
  - Around 30 percent of 2023 Chinese FDI in the EV sector went to Europe (Rhodium Group 2024).
  - BYD announced a car production plant in Hungary in late 2023 (Rhodium Group 2023).
  - Chinese investments in Europe concentrated in Germany, France, the UK and notably Hungary, which received more than half of total FDI from China in 2022 and 2023.
  - Some investments benefitted from European state aid (examples cited: Envision AESC, CATL; subsidized shares noted in source).

### 2.3. Putting the past and present together — scenario-informing comparisons
- Comparative summary (high-level):
  - Japan (1960s–1980s): large shift prompted by oil price shocks and emissions regulations; significant productivity differential favoring Japan; 15 percentage point U.S. market share increase over roughly a decade; protection at peak equivalent to tariffs of 21 percent; within a decade about 40 percent of foreign sales done through affiliates.
  - China (present): very large shift prompted by green transition; very large productivity growth differentials; EU context features additional tariffs of 7.8–35.3 percent; Chinese FDI expanding EV production in the EU.
- Scenario input choices informed by comparisons:
  - The authors assume a similar magnitude market share increase as Japan’s 1970s U.S. experience but compressed in time: a 15 percentage point increase in Chinese market share in the European market occurring over 5 years (compared with Japan’s increase over a bit more than a decade).
  - They explore scenarios with rising trade barriers (consistent with recent EU tariffs) and scenarios where China serves the EU market partly through FDI.
  - Long-run benchmark: 16.3 percent of EU car spending goes to Chinese vehicles in the long run compared with 1.3 percent in the 2016–19 period.

### 3. Model specification — models used and calibration
- Two complementary structural models are used:
  - GIMF-GVC (Global Integrated Monetary and Fiscal Model with Global Value Chains):
    - Purpose: short- to medium-term dynamics.
    - Type: annual multi-region, dynamic general equilibrium model.
    - Regions in application: 7 regions — Hungary, Czechia, Germany-France-Italy, Rest of Europe, United States, China and Rest of the World.
    - Sectors: 3 sectors — Non-tradable, Tradable, Automotive.
    - Automotive sector features: roundabout production, lower elasticities of substitution (reflecting supplier-specific relations), inclusion of trade in intermediate goods.
    - Calibration: OECD’s Inter-Country Input Output tables; baseline steady state calibrated using averages over 2016 to 2019.
    - Limitation: does not distinguish EV and ICE automotive production; assumes changes in total Chinese automotive market shares are caused by EV-sector growth.
  - Trade Model (Dynamic Quantitative Trade Model based on Cuñat and Zymek, 2024):
    - Purpose: long-run, steady-state analysis of trade patterns and incomes.
    - Type: steady-state, multi-region, general equilibrium model.
    - Coverage: 68 individual economies (including 27 from EU, 34 from EUR) and Rest of the World.
    - Sectors: 40 sectors; Motor Vehicles (ISIC C29) + Other Transportation Equipment (ISIC C30) constitute the Automotive sector.
    - Features: sector-specific bilateral iceberg trade barriers; sector-level structural gravity equations; forward-looking agents; endogenous distribution of capital and trade balances.
    - Calibration: uses averages over 2016 to 2019 to avoid COVID-related distortions.
- Key modelling tradeoffs:
  - GIMF-GVC provides rich intertemporal dynamics (transition paths for output, investment, and trade) but limited sectoral and geographic granularity (3 sectors, 7 regions).
  - Trade Model provides detailed sectoral and geographic structure (40 sectors, 68 economies) and steady-state comparative statics.

### 3.2 Scenarios modeled — setup and parameter choices
- EV-Shock scenario:
  - Objective: calibrate a 15 percentage points increase in China’s share of EU spending over 5 years.
  - Shocks combined:
    - Supply shock: increases total factor productivity (TFP) in China’s automotive sector by 15 percent.
      - Rationale: captures strong productivity growth in Chinese EV production that has lowered prices.
      - Note: the TFP shock corresponds approximately to productivity growth differentials observed between Japan and the U.S. in the 1970s (Annex reference).
    - Policy-driven demand shock: shifts consumption shares and inputs towards EVs (in line with rapid electrification required under EU regulation).
  - Modeling details and equivalence:
    - Because the models do not separate EV vs ICE in input-output data, the demand shift towards EVs and China’s productivity advantage is observationally equivalent to a growing “China preference” in the overall car sector by EU consumers.
    - In the long run (trade model steady-state comparison), 16.3 percent of EU car spending goes to Chinese vehicles versus 1.3 percent in 2016–19.
  - Timing:
    - Most of the rise in penetration is assumed to play out over only 5 years; transition paths are modeled in GIMF-GVC and full steady states compared in the trade model.
- Trade restrictions scenario:
  - Setup: EU adopts protectionist policies layered on the EV shock.
  - Tariff assumptions simulated: additional tariffs of 25 percent or 100 percent applied to automotive sector imports from China.
  - Modeling notes:
    - Models allow dynamic capital accumulation responses to trade barriers.
    - Models do not include endogenous effects of trade barriers on sectoral productivities (learning-by-doing, infant industry arguments, etc.); such channels are ignored given limited consensus.
- FDI scenario:
  - Purpose: quantify effects of Chinese EV delivery to EU market via FDI rather than imports.
  - Modeling approach (because models do not allow foreign ownership directly):
    - Scenarios alter investment-related parameters in the automotive sector in Europe to limit Chinese import penetration to 10 percentage points in the long run (compared with 15 percentage points in EV shock).
    - GIMF-GVC implementation:
      - Reduce equity premium in the automotive sector in Europe (cheaper borrowing for investment).
      - Reduce markups to replicate increased competition from new firms.
      - Partially reverse demand and input weight shifts for Chinese EVs to reflect domestic Chinese production in Europe.
      - Two versions:
        - FDI proportional across European regions (equity premium, markups, and demand shifts change equally across four European regions).
        - Hungary-focused FDI where a disproportionate share of additional investment and demand is allocated to Hungary (increasing demand for Hungary automotive production across Europe).
    - Trade model implementation:
      - Introduce an EV-biased EU car-sector productivity catch-up and increase EU investment efficiency.
      - Structural parameter changes calibrated so European imports of Chinese EVs decline from a peak of around 15 percent of total domestic production to 10 percent after around 3 years.
      - This mirrors the Japanese experience where around 40 percent of Japanese cars sold in the U.S. were eventually produced in the U.S.
  - Long-run FDI benchmark:
    - Proposed scenario assumption: 40 percent of Chinese EVs produced in EU (aligning with the historical Japan-to-U.S. affiliate production share).

*Source: wpiea2024218-print-pdf - Section 5 concludes.*

### 4. Simulation Results

### 4. Simulation Results

### THE “EV‑SHOCK”
- Short/medium term (GIMF‑GVC):
  - GDP losses after 5 years: Hungary: 1 percent; Czechia: 1.5 percent; Germany, France, Italy and the rest of Europe: around 0.2 percent.
  - Mechanism: China's productivity increase in EVs yields cheaper cars and inputs, shifting demand away from ICE vehicles produced in Europe toward Chinese EVs; demand shock dominates slightly positive supply/productivity shock.
  - Sectoral vulnerability: automotive sector in CESEE (notably Czechia, Hungary, and Slovakia) is around four times larger relative to other regions in terms of factor and domestic value‑added shares (OECD ICIO). The sector is highly capital‑intensive and in Hungary and Czechia is substantially more productive than the rest of the economy.
  - Labor mobility constraint: GIMF‑GVC allows labor to move across sectors but workers cannot move across countries.

- Steady‑State (Trade model):
  - Long‑run GDP losses: Hungary and Czechia incur some of the largest losses; Slovakia (not included in GIMF‑GVC but in trade model) is highlighted as worst affected with GDP losses exceeding 1 percent in the long run.
  - Mechanism: terms‑of‑trade deterioration from (i) needing to sell cars at lower export prices due to competition from cheap Chinese EVs; (ii) higher supply of other tradable goods lowering export prices via reallocated production; (iii) improvement in import prices from cheaper Chinese EVs partially offset by a long‑run decline in real exchange rates that raises domestic import prices.
  - Distributional outcome: smaller economies and those with exports concentrated in cars suffer larger macroeconomic consequences; economies not reliant on car manufacturing mainly benefit from cheaper EV imports, with overall EU income impact negligible.

- Labor market effects:
  - Short/medium term (GIMF‑GVC): small increase in employment after 5 years but large initial reallocation away from the car sector to tradable and non‑tradable sectors; wages fall (more in more affected regions); labor reallocated to lower productivity sectors in Hungary and Czechia where automotive has among highest labor productivity.
  - Steady‑State (Trade model): labor relocation away from auto sector as much as 2.6 percent of the workforce in Slovakia and 1.7 percent in Hungary; most displaced workers in Slovakia ultimately absorbed by other manufacturing sectors and retail; such adjustments likely entail significant transition costs.

- Productivity offset scenario:
  - The trade model (Table 3) suggests modest productivity gains in the car sector could protect average welfare in EU main car producers; these gains are a fraction of the baseline 15 percent TFP increase assumed for the Chinese automotive sector.
  - Rationale: Chinese EV imports eventually account for 15 to 20 percent of total domestic production, so smaller productivity gains benefit a larger share of workers and firms in Europe.
  - Evidence on innovation: large European automakers’ R&D expenses represent 5.5 percent of net sales—compared to 5.3 and 5.1 percent for Chinese and American top automakers respectively (2023 EU Industrial R&D scorecard). European transport‑related patents during 2020‑22 exceed those filed by China and the U.S.

### PROTECTIONIST RESPONSE SCENARIOS
- Policy variants considered:
  - 25 percent additional tariff on all Chinese automotive imports into Europe.
  - 100 percent additional tariff on all Chinese automotive imports into Europe.

- Main findings:
  - Tariffs reduce car imports from China but result in worse GDP outcomes in both short and long run.
  - Mechanisms of adverse impact:
    - Prevent consumers from benefiting from cheaper EVs, lowering real disposable income and demand for other goods and services.
    - Promote some domestic auto production but insufficient to fully restore European automotive production.
    - Large input‑output links in automotive sector mean tariffs on intermediate goods raise input costs and lower labor productivity in auto sector, reducing real wages economy‑wide.
    - Tariffs cause exchange rate appreciation, lower investment, and higher production costs—effects larger for Hungary (auto production is a larger share of GDP) than for Germany, France, Italy.
  - Short/medium term (GIMF‑GVC): GDP in Hungary and Germany, France, Italy performs worse with tariffs; tariffs support employment and domestic auto production but at expense of other tradable and non‑tradable sectors.
  - Steady‑State (Trade model): a 25 percent tariff would curb most of the increase in Chinese car‑market penetration (instead of rising by 15 percentage points, share of Chinese‑imported cars in EU car‑sector spending rises by only 2.5 percentage points), but higher consumer prices for EVs more than offset gains to car producers, leaving long‑run real GDP lower—especially for economies that were winners from the EV shock.

- Climate implications linked to tariffs (summary from Section 5, referenced here):
  - Tariffs slow EV adoption but impacts are modest; even a full ban reduces share of EVs in new cars by only 1 percentage point at peak.
  - Under a 100 percent tariff, Chinese EV imports amount to less than 10 percent of total new cars sold.
  - A full ban yields additional emissions equal to just 4 percent of 2023 EU car sector emissions (assuming unchanged climate policies).

### INCREASED FDI INFLOWS SCENARIO
- Two alternatives modeled:
  - Proportional Chinese FDI increase across European economies proportional to the size of each economy’s auto sector (structural shocks same size across countries, partial reversal of demand shift toward European production).
  - Hungary‑focused FDI where Hungary receives a disproportionate amount of FDI (European demand for Hungarian EVs increases by more than for other economies).

- Modeling note:
  - Since GIMF‑GVC and the trade model do not allow foreign ownership of firms, the FDI scenario is implemented via declines in equity premium and markups in Europe’s automotive sector, improvements in productivity and investment efficiency, and a partial reversal of the demand shift.

- Short/medium term (GIMF‑GVC):
  - Proportional Chinese FDI reduces GDP losses: effects on GDP levels after 5 years are around 0.5 percent for Hungary, and 0.15 percent for Germany, France, Italy and the rest of Europe.
  - Hungary‑focused FDI can turn Hungary into a modest net GDP winner after 5 years.
  - Higher FDI benefits both CESEE and EA3 block primarily through higher investment; different FDI allocation has limited spillovers to other regions given Hungary’s small economy.

- Steady‑State (Trade model):
  - FDI undoes some economic losses from baseline proportionally in worst‑affected EU economies.
  - Distribution of averted losses depends on which economies succeed in attracting Chinese FDI.

### Climate Implications
- Methodology overview:
  - Under assumptions: (i) demand for ICE and EVs given total car demand is described by a constant elasticity of substitution function; (ii) productivity improvement in China’s automotive sector is entirely attributed to EV production.
  - Using EU EV adoption projections and average vehicle life span estimates, project carbon emissions under alternative tariff scenarios.
  - By construction, share of EVs approaches 1 in the long run in all scenarios (given assumption that “fit‑for‑55” targets are met), but speed of adoption is slower with tariffs.

- Quantitative outcomes:
  - Example projection for 2030 under linear transition path: 63 percent of new vehicles sold would be EVs (Annex Table A3.1 reference).
  - Under a 100 percent tariff, Chinese vehicles constitute less than 10 percent of total new cars sold.
  - Emissions effect of a full ban on Chinese vehicles: additional emissions are just 4 percent of 2023 EU motor vehicles emissions (relative to baseline).
  - Conclusion: tariffs or bans have modest impacts on EV adoption speed and emissions during the transition; main downside of protectionism is increased economic cost of the EV transition. This conclusion depends on unchanged climate policy stance—postponement of EU EV adoption goals would likely raise additional emissions substantially.

*Sources: OECD; PWT; Fontagné et al. (2022); IEA; and IMF staff calculations.*

### 6. Conclusion

### 6. Conclusion

### Structural shifts facing Europe’s automotive sector
- Two large structural shifts: the green transition (rapid transition to EVs) and increased competition in the EV sector from China.
- Scenario analyzed: Chinese share in EU auto purchases rises from about 1 percent in 2016-19 to 16 percent by 2030.
- Scenarios assume the increase in imports of Chinese autos results from relative productivity and demand shifts.

### Aggregate and distributional findings
- For the EU as a whole:
  - The GDP cost of a shift towards greater auto imports from China is likely to be small in the short-term and close to zero over the long-term.
- Country-level heterogeneity:
  - Real GDP losses are significant in the range of 1-1½ percent over 5 years, and persisting in steady state, for a number of economies heavily reliant on the car sector (mainly in Central Europe).
  - Mechanisms for losses in exposed economies: demand for domestically-produced autos and auto-sector inputs shrinks; production shifts away from a high-productivity sector to lower productivity sectors; labor earnings fall.
  - EU economies not heavily exposed to auto production benefit from availability of cheaper EVs due to Chinese imports.
- Labor and sectoral reallocation:
  - Both models point to some employment losses in the motor vehicle sector and the need for employment to shift to other sectors.
  - GIMF-GVC allows reallocation to take time given wage rigidities.

### Drivers of import-led adjustment
- The simulated costs of the transition depend on the driver of the rise in auto imports from China:
  - Costs are smaller if the increase is driven by relatively high productivity growth in producing EVs in China and relatively low price of Chinese vehicles.
  - Costs are larger if the increase is driven by preference shifts towards Chinese autos.

### Trade policy, FDI, and productivity implications
- Tariffs against Chinese autos and auto parts magnify the simulated GDP cost of transitioning towards EVs (assuming 2035 “all EV” goal for new cars can be maintained under tariffs).
  - Tariffs amplify GDP losses because shielding the European automotive sector does not compensate for GDP losses due to:
    - lower purchasing power for EU consumers;
    - losses imposed on other sectors through costlier inputs; and
    - an appreciated exchange rate that lowers exports and investment.
- Mitigating factors:
  - Chinese FDI into EV production in the EU that eases the increase in imports by ramping up European EV production can mitigate income losses for the worst affected EU economies.
  - Seemingly achievable productivity gains in the EU auto sector can also mitigate income losses.

### Caveats and uncertainty
- Large and unprecedented transformations: wide uncertainty on future trends.
  - Scenarios are anchored largely on Japan’s experience in the U.S. market, but outturns could be different; current and prospective trends (productivity growth differentials, policies to encourage EV adoption, consumer preferences) could bring about a faster and larger shift than the 1970s experience.
  - There has recently been significant pushback by some countries against the forthcoming EU combustion engine vehicle ban, adding uncertainty.
- Model limitations:
  - Models abstract from labor reallocation costs (although reallocation takes time in GIMF-GVC given wage rigidities), which may lead to underestimation of economic losses in the short- to medium-term.
  - Tariff scenarios leave out the potential for retaliation and its adverse consequences, which could raise macroeconomic costs of imposing tariffs significantly.
- Emissions and climate policy assumption:
  - Estimates on the emissions implications of the different tariff scenarios assume that the “Fit-for-55” 2035 targets banning EV sales are met, despite higher EV prices.
  - Postponing current climate goals in the EU because of higher EV prices (due to tariffs on imports) could significantly raise emissions and impact growth.

### Policy implications and recommendations
- Proactive policies to ease labor reallocation are critical:
  - Policies to help auto sector workers re-skill and find employment in other sectors are critical to limit the costs of the transition.
- Trade policy caution:
  - Tariffs against Chinese autos and auto parts can magnify GDP costs and reduce purchasing power; potential retaliation and emission consequences warrant caution.
- Support measures to mitigate localized losses:
  - Encourage FDI that ramps up European EV production.
  - Pursue productivity-enhancing measures in the EU auto sector.
- Preserve climate goals:
  - Avoid postponing climate objectives (e.g., the “Fit-for-55” 2035 targets), since delaying could raise emissions and impact growth.

*Source: wpiea2024218-print-pdf - 6. Conclusion*

### ANNEX I. THE JAPANESE EXPERIENCE: LESSONS FROM PAST STUDIES

### ANNEX I. THE JAPANESE EXPERIENCE: LESSONS FROM PAST STUDIES

### (I) The role of supply and demand shifts
- Research consensus: Japan’s productivity rose earlier (from the 1950s) and a productivity edge was established by 1980; a contemporaneous demand shift in the 1970s toward smaller, fuel-efficient cars amplified Japanese market share gains.
- Productivity and process innovations in Japan since the 1950s included: the just-in-time system, concentration of production facilities in a single location, defect reduction by assigning quality control to production workers, and quality circles (Hart, 1992).
- Empirical productivity estimates:
  - Fuss and Waverman (1985): annual TFP growth of the Japanese auto industry in the 1970s = 4.3 percent; Canadian = 1.4 percent; U.S. = 1.6 percent.
  - Gomez-Ibanez and Harrison (1982): cost advantage of Japanese small cars manufactured and shipped to the United States ≈ between 13.0 and 16.1 percent around 1980.
  - Börsch-Supan and Romer (1998): by the early 1990s, Japan’s auto industry TFP = 120 percent of the U.S. and = 171 percent of Germany.
  - Aggregate summary: Japan’s auto industry productivity growth was 1-3 percentage points higher than in the U.S. during the 1970s and 1980s.
- Demand-side shocks:
  - Two oil price shocks in the 1970s prompted a shift in U.S. auto demand toward smaller and more fuel-efficient cars (Bresnahan and Ramey, 1993).
  - In the U.S., imported cars increased from between 15 and 18 percent of domestic sales in the 1970s to 27 percent in 1980, with Japanese imports accounting for 21 percent (Gomez-Ibanez and Harrison, 1982).
  - U.S. real oil prices increased by 28 percent between October 1973 and May 1974 and by 42 percent between February 1979 and March 1980 (Bresnahan and Ramey, 1993).
- Supply constraints and market response:
  - Japanese firms often operated at full capacity (Gomez-Ibanez and Harrison, 1982).
  - Slow catch-up by U.S. producers in the small-car market and a persistent perception of higher quality for Japanese cars (consumer reports showing Japanese cars achieved vastly higher quality standards) further enabled Japanese market penetration.

### (II) Japan’s government support
- Early protection and support:
  - MITI provided high but not prohibitive protection via import quotas, tariffs, and FDI restrictions in the nascent stage (Nishijima, 2012).
  - Pre-1964 support included low-interest loans and tax privileges before Japan joined GATT (Hart, 1992).
- Financial catalysis:
  - Japan Development Bank (JDB) low-cost financing acted as leverage for private finance (Odano and Islam, 1994; Hart, 1992).
  - Odano and Islam (1994) estimate: a one-unit increase in JDB loans is associated with a 14-unit increase in private banks’ loans to the auto industry between 1955–1985.
  - Rapid increases in bank financing preceded auto sector expansion (Figure A1.4 referenced).
- Subsidies and phased liberalization:
  - MITI provided subsidies to car producers and subcontractors.
  - Liberalization timeline cited:
    - Lifted import quotas: 1960 for buses and trucks; 1965 for passenger cars; 1972 for engines.
    - Tariff reductions: from 40 percent to 10 percent in 1971; to 5 percent in 1973; to zero percent in 1978.
    - FDI liberalized in 1971 (Nishijima, 2012).
  - Conclusion: phased reduction of protection allowed time for the industry to attain international competitiveness.

### (III) Protectionist policy responses starting in the 1980s, and their consequences
- U.S. Voluntary Export Restraints (VERs):
  - VERs initiated May 1981 limited Japanese passenger car exports to the U.S. to 1.68 million annually; extended 1984 limit = 1.85 million; from 1985 to 1992 limit = 2.3 million.
  - Berry et al. (1999): VERs did not raise prices significantly initially but had larger price impacts after five years; implied tariffs peaked at 21 percent.
  - Welfare effects per Berry et al. (1999): VERs increased U.S. producers’ profits, induced moderate net welfare losses to the U.S., while Japanese firms’ profits were less affected; had the VERs been tariffs, the U.S. would have experienced a net welfare gain due to tariff revenue offsetting consumer losses.
  - Feenstra (1988): documented substantial quality upgrading of Japanese autos under the VERs.
  - Dinopoulos and Kreinin (1988): European suppliers, excluded from VERs, raised prices by one-third; estimated Europe’s welfare gains in 1984 exceeded Japan’s losses.
- European restrictions and outcomes:
  - Pre-1970 Japanese market shares in Europe were tiny: 0.4 percent of the British market; 0.2 percent of the French market; negligible in German and Italian markets (Mason, 1994).
  - By 1980: Japan attained 11.9 percent of the British market and 10.4 percent of the German market (Mason, 1994).
  - Specific European measures:
    - 1976 Britain–Japan agreement: limit Japanese share to 11 percent.
    - 1977 France: restricted Japanese car imports to under 3 percent of its market.
    - 1981 Germany–Japan: agreed not to raise Japan’s share of the German market by more than 1 percent per year (Denicolò and Garella, 1999).
    - Italy limited Japanese exports to 2,800 cars from 1956 to 1986.
  - Studies generally find VERs had negative welfare impacts on European countries (Turrini, 1999; Walker, 2015).
  - By the late 1980s, many European-owned car makers lagged significantly behind in productivity, except some German luxury brands (Mason, 1994).

### (IV) The Japanese FDI response, and its impact
- Scale and direction of FDI:
  - From 1982 to 1989, Japan’s outward FDI increased about eight times in dollar terms (Bayoumi and Lipworth, 1997); close to half of this FDI went to the U.S.
- Effects on host-country productivity:
  - Japanese FDI improved overall productivity in the U.S. auto industry via increased competition and knowledge spillovers.
  - Chung et al. (2003): overall productivity of U.S. auto component suppliers increased in the 1980s; little evidence of direct technology transfer to suppliers; suppliers to Japanese transplants had lower initial productivity and experienced no greater productivity growth than non-Japanese affiliated suppliers, indicating adverse selection effects that temporarily inhibited productivity growth.
  - Branstetter (2006): FDI served as a channel of knowledge spillovers for Japanese multinationals undertaking direct investments in the United States; spillovers are bidirectional:
    - Japanese-to-U.S. spillovers mainly via R&D and product development facilities.
    - U.S.-to-Japanese spillovers largely through greenfield affiliates.
- Production localization trends:
  - Japanese firms began producing vehicles in North America and Europe in the 1980s (Figure A1.6 referenced).
  - Figure A1.7 referenced: fraction of Japanese cars on the U.S. market produced in the U.S. rose across the 1980s.

### Key numeric facts and calibrated figures cited in the annex
- Productivity and costs:
  - Fuss and Waverman (1985): Japanese auto industry annual TFP growth in the 1970s = 4.3 percent; Canada = 1.4 percent; U.S. = 1.6 percent.
  - Gomez-Ibanez and Harrison (1982): Japanese small cars’ cost advantage ≈ between 13.0 and 16.1 percent around 1980.
  - Börsch-Supan and Romer (1998): by early 1990s Japan’s auto industry TFP = 120 percent of the U.S.; = 171 percent of Germany.
  - Japan’s auto industry productivity growth was 1-3 percentage points higher than in the U.S. during the 1970s and 1980s.
- Oil price changes (U.S. reference):
  - Real oil prices increased by 28 percent between October 1973 and May 1974.
  - Real oil prices increased by 42 percent between February 1979 and March 1980.
- VER quantitative limits:
  - 1981 VER limit = 1.68 million passenger cars annually to the U.S.
  - 1984 VER limit = 1.85 million.
  - 1985–1992 VER limit = 2.3 million.
  - VER implied tariff peak = 21 percent (Berry et al., 1999).
- Japan outward FDI growth:
  - 1982–1989: outward FDI increased about eight times in dollar terms (Bayoumi and Lipworth, 1997).
- European market shares and restrictions:
  - 1970 Japanese market share examples: Britain = 0.4 percent; France = 0.2 percent.
  - 1980 Japanese market share examples: Britain = 11.9 percent; Germany = 10.4 percent.
  - Italy: limit of 2,800 cars from 1956 to 1986.

*Source: ANNEX I. THE JAPANESE EXPERIENCE: LESSONS FROM PAST STUDIES (extracted from the provided PDF content).*

### 1.2 given its growing dominance in global markets.

### 1.2 given its growing dominance in global markets.

### GIMF-GVC Calibration of Markups by Sector and Region
- Table A2.3. GIMF-GVC Calibration of Markups by Sector and Region (values are markups)
  - Regions / Sectors:
    - Hungary: Consumption 1.05; Investment 1.05; Nontradables 1.1; Tradables 1.1; Automotive 1.05
    - Czechia: Consumption 1.05; Investment 1.05; Nontradables 1.1; Tradables 1.1; Automotive 1.05
    - Germany, France, Italy: Consumption 1.1; Investment 1.05; Nontradables 1.1; Tradables 1.1; Automotive 1.05
    - Rest of Europe: Consumption 1.1; Investment 1.05; Nontradables 1.1; Tradables 1.1; Automotive 1.05
    - China: Consumption 1.1; Investment 1.05; Nontradables 1.2; Tradables 1.2; Automotive 1.2
    - United States: Consumption 1.1; Investment 1.05; Nontradables 1.2; Tradables 1.2; Automotive 1.1
    - Rest of the world: Consumption 1.1; Investment 1.05; Nontradables 1.2; Tradables 1.2; Automotive 1.1
- Source: IMF staff calculations.

### The Trade Model — Structure and Key Mechanisms
- Model environment and agents:
  - Arbitrary number of economies, no aggregate uncertainty.
  - Agents endowed with country-specific human capital, supplied inelastically domestically.
  - Constant probability of death; departed agents replaced by newborn cohort; aggregate populations constant.
  - Actuarially fair life insurance available.
  - Agents optimize holdings of two assets: country-specific capital (accumulated with country-specific investment efficiency and depreciates at a common, constant rate) and a freely traded one-period international bond.
  - Common interest rate ensures market-clearing in the international bond market.
  - Rate of time preference may differ across economies; together with demographic assumptions this ensures a unique steady state with a non-degenerate distribution of international assets and trade balances.
- Production and trade structure:
  - Final consumption, investment and intermediate inputs assembled in Cobb-Douglas fashion from an arbitrary number of sectors.
  - In each sector, countries produce differentiated country-specific varieties and trade à la Armington with sector-specific substitution elasticity.
  - All goods markets are perfectly competitive; equilibrium sector-level bilateral trade flows follow the structural gravity equation typical of “gravity class” quantitative trade models (Costinot and Rodríguez-Clare, 2014).
  - Country varieties produced using capital, labor and intermediate inputs; production technology features country-specific labor share and country-sector specific intermediate-input share.
  - Factor markets perfectly competitive; labor and capital move freely between sectors within economies; no cross-border labor mobility; international capital flows possible via international bond.
- Exact-hat logic and dynamics:
  - Static gravity-class quantitative trade models use “exact hat” algebra (Dekle et al., 2008) to describe impacts of parameter changes conditional on calibration and status-quo trade and world output shares.
  - The Cuñat and Zymek (2024) based trade model generalizes to steady-state comparisons in a dynamic setting:
    - Delivers standard exact hat algebra for changes in trade patterns and macro outcomes conditional on given capital stocks or steady-state trade balances.
    - Provides supplementary dynamic “exact hat” equations that characterize endogenous responses of steady-state capital stocks and trade balances to parameter changes (see Cuñat and Zymek, 2024; Appendix A.5).

### Calibration Objectives and Data Sources
- Four sets of calibration objects required:
  1. Countries’ bilateral trade shares, world-GDP shares and trade balances in the initial steady state.
  2. Sectors’ shares in countries economic activity and country-sector specific intermediate input shares.
  3. Country-specific capital shares.
  4. Sector-specific trade elasticities.
- Data and aggregation:
  - First two sets taken from OECD’s Inter-Country Input-Output Tables (ICIO; OECD, 2023), using the average for years 2016–19.
  - ICIO covers global output split into 76 individual economies and the "rest of the world".
  - Due to limited availability of some other required data, some individual economies are merged with the rest-of-the-world group; some largely non-traded services sectors merged if immaterial.
  - Final calibration: 68 individual economies and the rest of the world; 40 individual sectors covering goods and services.
  - Sectors C29 (“motor vehicles”) and C30 (“other transport equipment”) jointly treated as the “automotive sector” for all trade-model scenarios.
- Labor shares, trade elasticities and other parameters:
  - Country-specific labor shares averaged for 2016–19 obtained from Penn World Tables (Feenstra et al., 2015).
  - For goods-producing sectors, use trade elasticities from Fontagné et al. (2022) compatible with OECD ICIO disaggregation.
  - For all services sectors, trade elasticity set to 5, following Costinot and Rodríguez-Clare (2014).
  - Other parameters (e.g., depreciation rate of capital, average rate of time preference) calibrated using standard literature values; reasonable changes do not materially affect findings.
- Merged ICIO economies in calibration:
  - The 8 countries individually covered in the ICIO but merged with the rest of the world in calibration are Bangladesh, Belarus, Brunei, Cambodia, Jordan, Myanmar, Pakistan and Vietnam.
- Note on ICIO data period:
  - The most recent set of ICIO data available covers the 2016–20 period; 2020 excluded due to large trade disruptions from the Covid-19 pandemic.

### TM Sectors and Sectoral Trade Elasticities (Table A2.4)
- Selected sector names, codes and trade elasticities (full list shown in source):
  - Agriculture A01_02: 3.19
  - Fishing A03: 6.99
  - Mining, energy B05_06: 3.16
  - Mining, non-energy B07_08: 8.31
  - Mining services B09: 5.00
  - Food C10T12: 4.09
  - Textiles C13T15: 4.71
  - Wood products C16: 8.68
  - Paper products C17_18: 7.98
  - Coke, petroleum C19: 4.51
  - Chemical products C20: 8.25
  - Pharmaceuticals C21: 8.54
  - Plastics C22: 6.86
  - Other non-metallic products C23: 4.81
  - Basic metals C24: 7.12
  - Fabricated metals C25: 4.33
  - Electronic equipment C26: 5.38
  - Electrical equipment C27: 4.74
  - Machinery C28: 4.38
  - Motor vehicles C29: 8.66
  - Other transport equipment C30: 8.93
  - Manufacturing nec. C31T33: 5.00
  - Electricity, gas D: 5.00
  - Water supply, sewage E: 5.00
  - Construction F: 5.00
  - Wholesale and retail G: 5.00
  - Land transport H49: 5.00
  - Water transport H50: 5.00
  - Air transport H51: 5.00
  - Warehousing H52: 5.00
  - Postal activities H53: 5.00
  - Hospitality I: 5.00
  - Publishing, broadcasting J58T60: 5.00
  - Telecommunications J61: 5.00
  - Information services J62_63: 5.00
  - Finance and insurance K: 5.00
  - Real estate L: 5.00
  - Professional services M: 5.00
  - Administrative services N: 5.00
  - Other services O_T: 5.00
- Sources: Fontagné et al. (2022); OECD (2023); and IMF staff calculations.

### EV Share of Automotive Sales — Identification and Assumptions
- Motivation:
  - Models calibrated to trade and production linkages for general motor vehicles and transport equipment, not EVs specifically, due to lack of comprehensive EV-specific international trade and input-output data.
  - Introduce assumptions to identify bilateral EV trade and characterize EV component of automotive trade and consumption.
- CES-gravity form for automotive spending share:
  - Share of spending by destination d on automotive-sector output of origin o in period t:
    - s_o d t = (τ_o d t / p_o t)^(-θ) / Σ_o (τ_o d t / p_o t)^(-θ)
    - where p_o t is price of cars in origin; τ_o d t > 1 is ad-valorem equivalent trade barrier; θ > 0 is trade elasticity; s_o d t observable and models calibrated to match it.
- Two-variety decomposition (unobserved EV vs ICE):
  - Automotive-sector output composed of two substitutable varieties i ∈ {EV, ICE}.
  - Assumptions:
    - Ad-valorem trade barriers broadly similar for EV and ICE.
    - Origin prices decomposed: p_o t^EV = c_o t γ_o t; p_o t^ICE = c_o t.
      - c_o t: common component representing input costs and mark-ups.
      - γ_o t: EV-specific component representing origin’s relative productivity in assembling EVs.
    - Consumers treat EV and ICE as CES substitutes with substitution elasticity 1+θ and preference weights β_d t^EV and 1−β_d t^EV respectively.
  - Under these assumptions, re-written gravity share:
    - s_o d t = β_o d t (τ_o d t / c_o t)^(-θ) / Σ_o β_o d t (τ_o d t / c_o t)^(-θ)
    - where β_o d t ≡ (γ_o t^θ β_d t^EV + 1 − β_d t^EV).
  - Relation for units purchased x_o d t^i:
    - x_o d t^EV / (x_o d t^EV + x_o d t^ICE) = β_d t^EV γ_o t^(1+θ) / (β_d t^EV γ_o t^(1+θ) + 1 − β_d t^EV).

### Scenario Assumptions for EV Adoption (EU focus)
- EU consumers’ EV preference path:
  - Throughout scenarios, assume fixed path of β_d t^EV for EU consumers chosen such that in the “EV shock” scenario the share of EVs in EU consumers’ total car purchases linearly approaches 1 by 2035 (see Table A.3.1).
  - Table A3.1. EV Share in EU Automotive Purchases (Percent):
    - 2023: 12
    - 2024: 19
    - 2025: 27
    - 2026: 34
    - 2027: 41
    - 2028: 49
    - 2029: 56
    - 2030: 63
    - 2031: 71
    - 2032: 78
    - 2033: 85
    - 2034: 93
    - 2035: 100
- Additional shocks introduced in scenarios:
  - 15 percent TFP shock for China described in Section 3.2 that lowers China’s common automotive sector production cost c_CHN,t and shifts car spending towards China.
  - Path of China’s relative EV productivity γ_CHN,t introduced to achieve a gradual 15-percentage-point increase in China’s penetration of EU car purchases in the “EV shock” scenario.

### EV Adoption Dynamics and Emissions Accounting
- Vehicle-on-road evolution:
  - Let X_d t^i denote stock of vehicles of type i on the road in economy d period t.
  - Assume both vehicle types have constant retirement probability δ.
  - Share of EVs on the road evolves as:
    - X_d t^EV / (X_d t^EV + X_d t^ICE) = (1 − δ) X_d t−1^EV / (X_d t−1^EV + X_d t−1^ICE) + δ Σ_o x_o d t^EV / Σ_o (x_o d t^EV + x_o d t^ICE)
  - Under the EV shock scenario, share of EVs on the road in EU economies converges to 1 as t → ∞; alternative scenarios change the path to 2035 and thus the speed of convergence.
- Emissions formulation:
  - Car emissions E_d t generated by economy d in period t:
    - E_d t ∝ 1 − (1 − η) X_d t^EV / (X_d t^EV + X_d t^ICE)
    - η denotes relative emissions intensity of EVs compared with ICE vehicles.
  - Compare emissions across scenarios by computing Σ_{t=1}^∞ (Ê_d t − E_d t) / E_d0, where hatted emissions derive from alternative EV adoption paths.
- Parameter values used:
  - η = 0.5 (IEA’s EV Life Cycle Assessment Calculator (IEA, 2024) based)
  - δ = 1/15 (implying average vehicle lifespan of 15 years)
  - Initial condition for EU economies: X_d0^EV = 0
- Implementation:
  - Cumulate EV share on the road and corresponding emissions using equations for stock evolution and emissions (equations (6) and (7)) and EU economies’ EV purchase shares derived from the EV shock scenario and alternative protectionist scenarios described in Section 4.2.
  - Output described in Section 5.

*Source: IMF staff calculations, GIMF-GVC and TM model annex material contained in the provided PDF content.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024218-print-pdf.pdf_