## wpiea2024001-print-pdf

## Source details

**Canonical URL:** [wpiea2024001-print-pdf](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024001-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024001-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024001-print-pdf.pdf.json)

---

### Introduction and motivation
- Industrial policy (IP) has gained increased prominence amid sluggish post-financial crisis growth, the COVID-19 pandemic and associated supply disruptions, intensifying geopolitical tensions and conflicts, and unmet public demands for climate mitigation and adaptation.
- There is a 73.8 percent probability a subsidy for a given product by one major economy is met with a subsidy for the same product by another within one year (Table 1).
- Key questions: Can the resurgence in industrial policy be corroborated with evidence? Which measures are being used and why? What sectors are targeted and what cross-border spillovers are generated?
- Major challenge: lack of high quality and systematic information on government IP measures; no separate inventory of IP measures existed prior to this effort.

### The New Industrial Policy Observatory (NIPO) and data approach
- NIPO records measures announced or implemented since the beginning of 2023; defines industrial policies as targeted government interventions aimed at developing or supporting specific domestic firms, industries, or economic activities to achieve national economic or noneconomic objectives.
- Improvements over the Global Trade Alert (GTA) database:
  - Distinguishes between strategic plans, enacted policies/regulations, and firm-specific interventions.
  - Records the stated motive of a government tagged from official sources.
  - Associates interventions with pre-specified groups of products in strategic sectors: medical, semiconductors, critical minerals, military/civilian dual-use, low carbon technology, and other advanced technology.
  - Expands tracked policy interventions to include several technology-related interventions.
- GTA tracks over 60 types of policy intervention and contains information on over 61,000 distinct interventions.

### Snapshot of NIPO in 2023 — key statistics and patterns
- Number of measures:
  - Over 2,500 new industrial policies (NIPs) worldwide recorded in the first 12 months of data collection.
  - 71 percent of recorded NIPs are trade distorting.
- Trade coverage and impact:
  - Restrictive measures with precise trade coverage (882 measures) impact at least 22 percent of global trade.
- Geographic concentration:
  - China, European Union, and United States account for 48 percent of the measures.
  - Advanced economies (AEs) were more active than emerging markets and developing economies (EMDEs) in the use of industrial policy in 2023.
- Instrumental differences by income group:
  - AEs rely more on direct financial grants, state loans, and state aid.
  - EMDEs rely more on state loans, tax relief, and capital and equity injections.
  - Trade restrictions on imports and exports are used more frequently by EMDEs relative to AEs.
- Stated motivations (where motive information available):
  - Strategic competitiveness: stated objective for over a third of measures.
  - Climate-related motivations: 28 percent.
  - Supply chain resilience: 15 percent.
  - National security and geopolitical tensions combined: around one in five measures.
- Sectoral coverage and evolution:
  - Medical goods sector was most targeted early in 2023; later overtaken by military/civilian dual use products and advanced technology products, including low-carbon technology, semiconductors, and upstream inputs such as critical minerals.
- Pattern: corporate subsidies are the most common type of trade distorting instrument.

### Methodology, coverage, and sample composition
- Recording rules:
  - NIPO records state measures implemented or announced on or after January 1, 2023 and is updated monthly.
  - Each entry refers to a distinct state intervention.
  - Monitoring covers 75 jurisdictions which make up 94 percent of global GDP.
  - NIPO is a live database launched January 2024; data may be subject to revision and recording lags may exist.
- Jurisdiction classification:
  - 45.3 percent are Advanced Economies (AEs).
  - 54.7 percent are Emerging and Developing Economies (EMDEs).
- Regional representation:
  - Asia Pacific: 13 jurisdictions (17.3%)
  - Europe and Central Asia: 31 jurisdictions (41.3%)
  - Latin America and the Caribbean: 9 jurisdictions (12%)
  - Middle East and North Africa: 9 jurisdictions (12%)
  - North America: 2 jurisdictions (2.7%)
  - South Asia: 4 jurisdictions (5.3%)
  - Sub-Saharan Africa: 7 jurisdictions (9.3%)

### Inclusion criteria and motive tagging
- Inclusion criteria: a GTA measure is included in NIPO if it meets at least 1 of 3 criteria:
  1. Associated with a predefined set of motives.
  2. Covers at least one predefined set of products or service categories.
  3. Is an industrial strategy or plan.
- Stated motive categories (from official sources):
  - National security
  - Geopolitical concerns
  - Security of supply (for non-food products)
  - Strategic competitiveness (domestic competitiveness in strategic sectors)
  - Climate change mitigation and other environmental objectives
- Entries can have multiple motives; “Mentions food security”, “Mentions public health concerns”, and “Mentions other” are recorded but are not valid inclusion criteria by themselves.

### Product and sector classification
- Product/service groups included (six-digit HS subheadings and CPC codes):
  - Low-Carbon Technology (examples: wind turbines, solar panels, biomass systems, carbon capture equipment).
  - Dual-use products (civilian and military use; includes HS codes for hydrogen, steel, iron and aluminum though separate categories exist).
  - Critical minerals.
  - Advanced technology products (medical and industrial applications, opto-electronics, electronics, robotics, optical fiber cable, aerospace, nuclear technology).
  - Semiconductors.
  - Medical products (consumables, medicines, vaccines, medical equipment).
  - IT or digital services (CPC codes 623, 831, 834, 839, 841, 842, 843 according to CPC version 2.1).
- HS nomenclature: 2012 edition used.
- Sectoral scope is unrestricted if other inclusion criteria met.
- Monitoring expanded to include computing-related technologies (microelectronics, quantum information systems, artificial intelligence), clean energy technologies (batteries, electric vehicles).

### Levels of intervention and database composition (Table 2 exact figures)
- Plans & strategies: Number of records in NIPO database 98; Percentage of total records 3.80%; Percent Currently implemented n/a
- Policies & regulations: Number of records in NIPO database 1451; Percentage of total records 56.24%; Percent Currently implemented 95.8%
- Firm-specific interventions: Number of records in NIPO database 1031; Percentage of total records 39.96%; Percent Currently implemented 99.2%
- Total: Number of records in NIPO database 2580; Percentage of total records 100%; Percent Currently implemented 93.6%
- Most interventions are policies and regulations, normally implemented soon after announcement.

### Policy instruments covered and examples
- Broad categories captured:
  - Export barriers (e.g., export control measures targeting 30 drone-related products by China on 9 January 2023).
  - Import barriers (e.g., import licensing requirements on laptops, computers, servers by India on 11 January 2023).
  - Domestic subsidies (e.g., financial grant of EUR 6.5 billion to compensate energy-intensive companies in Germany).
  - Export incentives (e.g., BNDES loans supporting Embraer’s export operations in September 2023).
  - Foreign Direct Investment measures (e.g., Russia measure on 17 January 2023 authorizing exclusion of votes of shareholders from “unfriendly” countries).
  - Procurement policies (e.g., US steps in March 2023 to favor domestic firms in circuit board contracts for national security reasons).
  - Localization incentives/requirements (e.g., local content requirements in the US Inflation Reduction Act).
  - Technology-related trade and investment measures (direct restrictions on transfers/sales of technology; restrictions on collaborations, joint ventures, movement of scientific personnel; bans on TikTok by public bodies in multiple jurisdictions).
- Firm-specific actions included (e.g., Korea financing for CS Wind’s offshore wind tower factory in Vietnam).

### Taxonomy of specific policy instruments (selected items from Table 3)
- Export barriers include: Export ban; Export licensing requirement; Export quota; Export tariff quota; Export tax; Local supply requirement for exports; Export-related non-tariff measure, nes.
- Import barriers include: Anti-dumping; Anti-subsidy; Import ban; Import licensing requirement; Import monitoring; Import quota; Import tariff; Import tariff quota; Internal taxation of imports; Import-related non-tariff measure, nes.
- Domestic subsidies include: Capital injection and equity stakes (including bailouts); Financial grant; In-kind grant; Tax or social insurance relief; Production subsidy; Interest payment subsidy; Loan guarantee; Import incentive; Price stabilization; State loan; State aid, nes; State aid, unspecified.
- Export incentives include: Trade finance; Export subsidy; Tax-based export incentive; Financial assistance in foreign market; Other export incentive.
- FDI measures include: FDI: Entry and ownership rule; FDI: Financial incentive; FDI: Treatment and operations, nes.
- Public procurement measures include: Public procurement access; Public procurement, nes.
- Localization content measures include: Local content incentive; Local content requirement; Local operations incentive; Local operations requirement; Local value added incentive; Public procurement localization; Localization, nes.
- Others include: Anti-circumvention; Control on personal transactions; Controls on commercial transactions and investment instruments; Controls on credit operations; Foreign customer limit; Intellectual property protection; Labor market access; Post-migration treatment; Repatriation & surrender requirements; Special safeguard; Trade payment measure.
- Note: “Instrument unclear” category exists (labelled 12. Instrument unclear).

### Other NIPO variables and trade coverage calculations
- Variables recorded: Entry ID and Title; Jurisdiction; Initial assessment (distortive or liberalizing); Announcement, implementation, removal dates; Level of government implementation; Targeted economic activity (HS 6-digit, UN HS 2012; CPC 3-digit UN CPC 2.1); Source (almost always official sources).
- Trade covered and subsidy size:
  - Trade covered (USD millions) computed for each entry where relevant; interpretation depends on policy type (subsidies, export policies, outward/export incentives).
  - Size of subsidy: subsidy value of the NIPO entry (in USD millions) available for import competing and export incentives.

### Aggregate activity, regional and sectoral patterns (2023 snapshot)
- Total new industrial policy interventions in 2023: over 2,500 measures.
- Trade-distorting measures: around 1,800 measures, or 71 percent.
- Distribution of trade-distorting NIPs by income group:
  - AEs accounted for 70.9 percent of trade distorting measures.
  - EMDEs accounted for 29.1 percent of trade distorting measures.
- Concentration: China, the European Union, and the United States account for 47.7 percent of trade distorting measures.
- Most frequently used instruments: subsidies to domestic producers.
- Second most frequent:
  - In AEs: export incentives, followed by localization policies (public procurement and investment controls).
  - In EMDEs: import barriers, followed by localization policies.
- Jurisdictional patterns:
  - Export incentives: Canada, Germany, Japan, Korea most active by number.
  - Localization measures: United States and India rely most on localization.
  - Export curbs: China, India, Russia most frequent.
  - Low-income developing countries: heavy reliance on import barriers.
- Regional instrument highlights:
  - Europe & Central Asia, and North America: comparatively more use of domestic subsidies.
  - Asia Pacific, Latin America & Caribbean, South Asia: greater reliance on import barriers.
  - South Asia and North America: comparatively more use of localization measures.
  - Asia Pacific and Europe & Central Asia: most active users of export curbs.
  - North America: more public procurement measures than other regions.
  - Measures related to FDI: relatively infrequent.

- Sectoral focus in 2023 (cumulative stock; equal weight for multi-sector measures):
  - Military/civilian dual use products: 25.7 percent.
  - Other advanced technologies (including medical products and semiconductors): 20.6 percent.
  - Low carbon technology products: 15.3 percent.
  - Steel and aluminum: 10.1 percent.
  - Critical minerals (upstream inputs): 3.0 percent.
  - Note: medical goods were more prominent earlier during COVID-19.

### Imports covered by distortive measures and stated motive (Table 6 exact figures)
- Total inward measures in NIPO with trade value data: 882 entries (56% of inward distortive measures in NIPO; 882 out of 1,576).
- Entries with trade value (percent): 93.1%
- Imports covered (billions, USD): 3,789.8
- Imports covered (percent of global imports): 21.6%
- Breakdown by motive:
  - Without Stated Motive: 408 entries; 93.6% entries with trade value; Imports covered (billions, USD) 2,143.4; Imports covered (percent) 12.2%
  - With Stated Motive: 474 entries; 92.6% entries with trade value; Imports covered (billions, USD) 2,269.5; Imports covered (percent) 12.9%
  - Strategic Competitiveness: 240 entries; 95.4% with trade value; Imports covered (billions, USD) 1,507.9; Imports covered (percent) 8.6%
  - Climate Change: 201 entries; 98.5% with trade value; Imports covered (billions, USD) 1,010.6; Imports covered (percent) 5.8%
  - Supply Chain Resilience: 161 entries; 98.1% with trade value; Imports covered (billions, USD) 812.9; Imports covered (percent) 4.6%
  - National Security: 46 entries; 76.1% with trade value; Imports covered (billions, USD) 659.7; Imports covered (percent) 3.8%
  - Geopolitical Concerns: 27 entries; 55.6% with trade value; Imports covered (billions, USD) 75.7; Imports covered (percent) 0.4%
- Note: import coverage based on matching HS six-digit products covered by a measure with HS six-digit UN Comtrade data for 2019; only inward measures included.

### Correspondence between income class, motivation, and instrument
- Motivation profiles by income class:
  - Climate, geopolitics, and national security motives: predominantly given by Advanced Economies (AEs), little used by EMDEs.
  - EMDEs: main motives are strategic competitiveness and other motives.
- Instrument-motive patterns:
  - Climate concerns: instrument of choice are domestic subsidies.
  - National security: tools are varied, including trade measures and public procurement.
  - Strategic competitiveness:
    - In AEs: addressed with domestic and export subsidies.
    - In EMDEs: in around half the cases EMDEs use other instruments.

### Empirical approach to determinants of industrial policy use
- Five classes of hypothesized determinants:
  - (i) global market position,
  - (ii) retaliation dynamics,
  - (iii) political economy factors,
  - (iv) structural factors,
  - (v) cyclical factors.
- Product-country panel regression (Equation (1)):
  - Measures_cs = β0 + β1 RCA_cs + β2 #Measures by Others_2022_cs + δ_cs + ε_cs
  - Outcome Measures_cs can be intensity (NIPs count) or extensive margin dummy (Dummy #NIPs>0).
  - PPML estimator used to account for ~84 percent zeros in dependent variable.
- Country-level specification (Equation (2)):
  - Measures_cs = β0 + β1 Political Econ_c + β2 Structural_c + β3 Cyclical_c + δ_s + ε_cs
  - Outcome Measures_cs = number of new industrial policy measures by jurisdiction c affecting product s in 2023.
  - Most stringent product-level (HS 6-digit) fixed effects included.

### Key regression findings (summary of Tables 7 and 8)
- Sector-level correlates:
  - Positive correlation between new industrial policies and Revealed Comparative Advantage (RCA).
  - Positive correlation between new industrial policies and prior IP measures by other countries in the same sectors (tit-for-tat dynamics).
- Country-level correlates:
  - Structural variables increase goodness of fit more markedly than political economy or cyclical variables.
  - Political economy:
    - Upcoming elections: positive correlation with number of NIPs (statistically significant when controls included).
    - Governments leaning further to the right: tend to use more NIPs.
    - Higher government effectiveness (World Bank WGI): correlated with greater likelihood of undertaking industrial policies.
  - Structural variables:
    - High export concentration (Herfindahl–Hirschman Index of exports): positively correlated with number of IP measures.
    - Sovereign debt rating: with full controls, economies with poorer sovereign debt ratings tend to use more measures.
    - Credit to the private sector: negatively related to use of industrial policy measures.
    - Higher GDP correlates with more measures (consistent with AEs being more active).
  - Cyclical variables:
    - Change in GDP from previous year: associated with fewer measures when only cyclical variables included; loses significance with fuller controls.
    - Real effective exchange rate (REER) appreciation: positively and significantly correlated with industrial policy measures.
- Estimation notes:
  - PPML used due to ~84 percent zeros in dependent variable; results generally robust to OLS though OLS may be biased.

### Context, risks, and policy considerations
- Longstanding economic debate: theoretical rationale for IP (infant industry, knowledge spillovers, dynamic scale economies, coordination failures, informational externalities) versus risks (inefficiencies, rent-seeking, resource misallocation).
- International risks: IP can lead to spillovers, trade tensions, and retaliatory dynamics; rules-based trading system may be ill-equipped to manage these dynamics.
- Historical example: wide-body aircraft (Airbus vs. U.S. producers)
  - United States government estimated civil aviation contributed around 5 percent to GDP pre-pandemic.
  - Airbus case resulted in a record $7.5 billion in authorized annual countermeasures and the Boeing case in annual countermeasures of $4.0 billion.
- New objectives for IP increasingly include climate change, pandemics, and national security, implying evaluation criteria should extend beyond traditional economic considerations.
- Policy emphasis: design policies mindful of retaliation and weakening of the trading system; deliberation and dialogue (e.g., within the WTO) to identify shared rules of conduct is emphasized.

### Conclusions and avenues for future research
- NIPO provides a detailed, manually-vetted monitoring exercise for a broad set of IP measures beginning in 2023, addressing a previous data gap.
- Key 2023 empirical patterns:
  - AEs more active and use domestic and export subsidies; EMDEs more frequently use trade restrictions on imports and exports.
  - Strategic competitiveness is the dominant stated motive; climate change and supply chain resilience follow.
  - Combined trade coverage of climate change, supply chain resilience, geopolitical/national security measures equals that of strategic competitiveness.
- Exploratory regression correlates:
  - Sector-level: correlated with RCA and prior measures by others.
  - Domestic: elections, structural factors, and cyclical conditions correlate with IP use.
- Policy implication: governments should exert caution in using industrial policy given potential spillovers and fiscal risks.
- Future research enabled by NIPO:
  - Assess success/failure of NIPs at meeting stated objectives.
  - Assess domestic macroeconomic implications.
  - Measure distortions to trade patterns and resource allocation.
  - Study impacts on trading partners and global general equilibrium welfare effects.
  - Address whether current trade rules and multilateral mechanisms are adequate and how best to update them.

*Source: IMF Working Paper — Introduction (wpiea2024001-print-pdf).*

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

### Introduction

### Overview and motivation
- Industrial policy (IP) has gained increased prominence amid sluggish post-financial crisis growth, the COVID-19 pandemic and associated supply disruptions, intensifying geopolitical tensions and conflicts, and unmet public demands for climate mitigation and adaptation.
- There is a 73.8 percent probability a subsidy for a given product by one major economy is met with a subsidy for the same product by another within one year (Table 1).
- Key questions raised: Can the reported resurgence in industrial policy be corroborated with evidence? Which measures are being used and why? What sectors are targeted and what cross-border spillovers are generated?
- A major challenge is the lack of high quality and systematic information on government IP measures; no separate inventory of IP measures has existed prior to this effort.

### The New Industrial Policy Observatory (NIPO) and data approach
- NIPO records measures announced or implemented since the beginning of 2023, defining industrial policies as targeted government interventions aimed at developing or supporting specific domestic firms, industries, or economic activities to achieve national economic or noneconomic objectives.
- The NIPO improves on the Global Trade Alert (GTA) database in four respects:
  - Distinguishes between strategic plans, enacted policies/regulations, and firm-specific interventions.
  - Records the stated motive of a government tagged from official sources.
  - Associates interventions with pre-specified groups of products in strategic sectors: medical, semiconductors, critical minerals, military/civilian dual-use, low carbon technology, and other advanced technology.
  - Expands tracked policy interventions to include several technology-related interventions.
- The GTA database tracks over 60 types of policy intervention and contains information on over 61,000 distinct interventions (see Annex I for GTA methodology).

### A snapshot of NIPO in 2023 — key statistics and patterns
- Number of measures:
  - Over 2,500 new industrial policies (NIPs) worldwide recorded in the first 12 months of data collection.
  - 71 percent of recorded NIPs are trade distorting.
- Trade coverage and impact:
  - A subset of restrictive measures affecting imports with precise trade coverage (882 measures) impact at least 22 percent of global trade.
- Geographic concentration:
  - China, European Union, and United States account for 48 percent of the measures.
  - Advanced economies (AEs) were more active than emerging markets and developing economies (EMDEs) in the use of industrial policy in 2023.
- Instrumental differences by income group:
  - AEs rely more on direct financial grants, state loans, and state aid.
  - EMDEs rely more on state loans, tax relief, and capital and equity injections.
  - Trade restrictions on imports and exports are used more frequently by EMDEs relative to AEs.
- Motivations stated for measures (where motive information is available):
  - Strategic competitiveness: stated objective for over a third of measures.
  - Climate-related motivations: 28 percent.
  - Supply chain resilience: 15 percent.
  - National security and geopolitical tensions combined: around one in five measures.
- Sectoral coverage and evolution:
  - Medical goods sector was the most targeted at the beginning of 2023, but was soon overtaken by military/civilian dual use products and advanced technology products, including low-carbon technology, semiconductors, and upstream inputs such as critical minerals.
- Patterns suggest IP activity is concentrated among key economies and that corporate subsidies are the most common type of trade distorting instrument.

### Determinants of industrial policy use (preliminary regression evidence)
- Sectoral correlates:
  - Positive correlation between new industrial policies and revealed comparative advantage (RCA): established sectors are more frequent targets.
  - Positive correlation between new industrial policies and past IP measures imposed by other countries in the same sectors: evidence of tit-for-tat dynamics and potential negative spillovers.
- Country-level correlates:
  - Upcoming elections: countries with an upcoming election are more likely to use IP measures, indicating political economy motivations.
  - Export concentration: countries with high export concentration tend to intervene more.
  - Sovereign debt ratings: poorer sovereign debt ratings do not deter policy activism and are positively correlated with IP measures.
  - Real exchange rate appreciation (loss of external competitiveness): associated with greater use of IP measures.
- Caution: Given the short time series (data covering 2023), these findings are preliminary and suggestive rather than definitive about effectiveness.

### Context, risks, and policy considerations
- Longstanding debate: economic arguments for IP (infant industry, knowledge spillovers, dynamic scale economies, coordination failures, informational externalities) versus risks (inefficiencies, rent-seeking, resource misallocation).
- International risks: IP can lead to spillovers, trade tensions, and retaliatory dynamics that the rules-based trading system may be ill-equipped to manage.
- Historical example:
  - Wide-body aircraft industry (Airbus vs. U.S. producers) illustrates strategic IP with large subsidies, protracted negotiations, WTO disputes, and contested impacts on consumer welfare.
  - United States government estimates civil aviation contributed around 5 percent to GDP pre-pandemic.
  - The Airbus case resulted in a record $7.5 billion in authorized annual countermeasures and the Boeing case in annual countermeasures of $4.0 billion.
- New objectives for IP increasingly include climate change, pandemics, and national security considerations, implying evaluation criteria should extend beyond traditional economic considerations.
- Designing policies mindful of retaliation and weakening of the trading system is important to avoid undermining intended objectives; deliberation and dialogue (e.g., within the WTO) to identify shared rules of conduct is emphasized.

### Contribution and next steps
- NIPO provides a detailed, manually-vetted monitoring exercise for a broad set of IP measures beginning in 2023, addressing a previous data gap.
- The data enable initial descriptive analysis and preliminary regression-based assessments of determinants of IP use; further data collection and research are needed to robustly assess economic effects and effectiveness.

*Source: IMF Working Paper — Introduction (wpiea2024001-print-pdf).*

### Section 3 takes a first look at the data and presents some stylized facts on IP measures in 2023, while Section

### wpiea2024001-print-pdf - Section 3 takes a first look at the data and presents some stylized facts on IP measures in 2023, while Section

### Methodology of the NIPO database
- NIPO records state measures that have been implemented or announced on or after January 1, 2023 and is updated monthly.
- Each entry in the NIPO database refers to a distinct state intervention.
- Monitoring covers 75 jurisdictions which make up 94 percent of global GDP.
- The data presented represent a snapshot in time of the first NIPO launched in January 2024 and may be subject to revision; the NIPO is a live database that will be updated monthly.
- Recording lags may exist as information becomes available at different speeds across various jurisdictions.

### Coverage and sample composition
- Jurisdiction classification:
  - 45.3 percent are Advanced Economies (AEs).
  - 54.7 percent are Emerging and Developing Economies (EMDEs).
- Regional representation:
  - 13 jurisdictions in the Asia Pacific (17.3%).
  - 31 jurisdictions in Europe and Central Asia (41.3%).
  - 9 jurisdictions in Latin American and the Caribbean (12%).
  - 9 jurisdictions in the Middle East and North Africa (12%).
  - 2 jurisdictions in North America (2.7%).
  - 4 jurisdictions in South Asia (5.3%).
  - 7 jurisdictions in Sub-Saharan Africa (9.3%).

### Inclusion criteria for NIPO
- A GTA measure is included in NIPO if it meets at least 1 out of 3 criteria:
  1. The measure is associated with a predefined set of motives.
  2. The measure covers at least one of a predefined set of products or service categories.
  3. The measure is an industrial strategy or plan.

### Stated motive (first inclusion criterion)
- NIPO includes GTA measures whose stated motive is any of:
  - National security
  - Geopolitical concerns
  - Security of supply (for non-food products)
  - Strategic competitiveness (domestic competitiveness in strategic sectors)
  - Climate change mitigation concerns
- Entries can be associated with more than one stated motive.
- Motivation identification method: statements from official sources or direct quotes from senior officials are collected and reviewed.
- Government-stated motive categories defined in source:
  - National security: explicit reference to current or future military security or specifically quotes “national security”.
  - Geopolitical concerns: refers to countering risk from a country or class of countries.
  - Resilience/security of supply (non-food): refers to improving stability or security of local supplies of non-food products now or in the future.
  - Domestic competitiveness in strategic sectors: refers to promotion of domestic competitiveness or innovation in a strategic product or sector.
  - Climate change mitigation and other environmental objectives: refers to climate change mitigation or the transition to a low-carbon economy.
- Measures that fulfill inclusion criteria but also mention food security, public health, or other motives are tagged under “Mentions food security”, “Mentions public health concerns”, and “Mentions other”.
  - Note: “Mentions food security”, “Mentions public health concerns”, and “Mentions other” are not valid inclusion criteria by themselves.

### Product or Service Categories (second inclusion criterion)
- NIPO includes GTA measures that relate to the following groups defined by six-digit HS subheadings and CPC codes:
  - Low-Carbon Technology:
    - Examples: wind turbines, solar panels, biomass systems, carbon capture equipment.
  - Dual-use products:
    - Goods, software, and technology usable for both civilian and military purposes.
    - Includes HS codes of hydrogen and steel, iron and aluminum (but separate categories exist for those).
  - Critical minerals:
    - Minerals necessary for producing a broad range of goods used in everyday life and modern technologies.
  - Advanced technology products:
    - Includes medical and industrial applications of advanced scientific discoveries, opto-electronics, electronics, robotics, optical fiber cable and video discs, aerospace products, and nuclear technology.
  - Semiconductors:
    - Semiconductors and materials/products that allow further development and application of semiconductor-related technologies.
  - Medical products:
    - Medical consumables or non-durable products, including medicines and vaccines as defined in the GTA’s Essential Goods Initiative; medical equipment.
  - IT or digital services:
    - Technological research, digital or IT services (identified by CPC codes 623, 831, 834, 839, 841, 842, and 843 according to CPC version 2.1).
- GTA measures without an associated HS code but with a CPC code are classified using a conversion table.
- Additional text-based keyword search categories: Hydrogen; Steel, Iron & Aluminum; Medical products; Critical Minerals.
- The 2012 edition of the HS nomenclature is used.
- The sectoral scope of NIPO is unrestricted if other inclusion criteria are met.
- Monitoring was expanded to include computing-related technologies including microelectronics, quantum information systems, artificial intelligence, clean energy technologies such as batteries and electric vehicles.

### Level of Intervention (third inclusion criterion)
- Industrial strategies or plans: officially adopted strategic policy guidance issued by a government body to accomplish a NIP objective; tend to span multiple years and relate to multiple subsequent policies and interventions.
  - Example: European Commission’s announcement of its European Economic Security Strategy with proposals related to outbound investment and export controls.

### Policy interventions covered in NIPO
- NIPO includes measures across levels from plans and strategies to policies and regulations to firm-specific interventions.
- Table 2 (Levels of Policy Interventions) — exact figures from NIPO database:
  - Plans & strategies: Number of records in NIPO database 98; Percentage of total records 3.80%; Percent Currently implemented n/a
  - Policies & regulations: Number of records in NIPO database 1451; Percentage of total records 56.24%; Percent Currently implemented 95.8%
  - Firm-specific interventions: Number of records in NIPO database 1031; Percentage of total records 39.96%; Percent Currently implemented 99.2%
  - Total: Number of records in NIPO database 2580; Percentage of total records 100%; Percent Currently implemented 93.6%
- Most interventions covered are policies and regulations, normally implemented soon after announcement.
- NIPO includes GTA policy measures and an additional set of technology-related trade or foreign investment restrictions, guided by a list that includes:
  - Direct restrictions on transfers or sales of technology
  - Restrictions on commercial technological collaboration
  - Other technology-related restrictions targeting specific countries or entities
  - Restrictions on joint ventures in technology sectors
  - Restrictions on flows and employment of scientific personnel
  - Restrictions on public procurement in technology sectors

### Firm-specific actions
- NIPO identifies firm-specific actions such as subsidy awards or actions affecting particular firms, including those resulting from broad-based programs like export incentives.
  - Example: Korea’s financing to support CS Wind's offshore wind tower factory in Vietnam.
- Government decisions on specific foreign direct investments are included (e.g., measures to allow/disallow FDI).

### Broad categories of policy instruments captured
- Export barriers:
  - Export bans, tariffs and quotas, export licensing and other export-related trade barriers.
  - Example: export control measures targeting 30 drone-related products implemented by China on 9 January 2023.
- Import barriers:
  - Import bans, tariffs and quotas, import licensing and other import-related trade barriers.
  - Example: import licensing requirements on laptops, computers, and servers by India implemented on 11 January 2023.
- Domestic subsidies:
  - Tax rebates, grants, state loans and loan guarantees, price stabilization measures, production subsidies and other incentives to domestic production.
  - Example: financial grant of EUR 6.5 billion to compensate energy-intensive companies at risk of carbon leakage from higher fuel prices in Germany.
- Export incentives:
  - Tax-based export incentives, unit-based export subsidies, trade financing and other financial export promotion.
  - Example: in September 2023, the Brazilian Development Bank (BNDES) provided two loans to aircraft manufacturer Embraer to support its export operations.
- Foreign Direct Investment measures:
  - Entry and ownership requirements as well as FDI screening decisions.
  - Example: the measure adopted by Russia on 17 January 2023 grants authorization for certain companies to exclude the votes of shareholders from “unfriendly” countries.
- Procurement policies:
  - Changes to public procurement law or practice that may favor local suppliers.
  - Example: steps taken by the United States in March 2023, motivated by national security considerations, to favor domestic firms in government contracts for circuit boards.
- Localization incentives or requirements and public procurement localization measures:
  - Example: local content requirements in the US Inflation Reduction Act.
- Technology-related trade and investment measures given additional attention:
  - Direct restrictions on transfers or sales of technology
  - Restrictions on commercial technological collaborations and joint ventures
  - Restrictions on movement and employment of scientific personnel
  - Restrictions on public procurement in technology sectors
  - Example: bans on the use of TikTok imposed by public bodies in Australia, Belgium, Canada, Denmark, EU, France, Nepal, New Zealand, United Kingdom, and the United States.

### Taxonomy of Specific Policy Instruments (selected list from Table 3)
- Export barriers (enumerated):
  1. Export ban
  2. Export licensing requirement
  3. Export quota
  4. Export tariff quota
  5. Export tax
  6. Local supply requirement for exports
  7. Export-related non-tariff measure, nes
- Import barriers (enumerated):
  1. Anti-dumping
  2. Anti-subsidy
  3. Import ban
  4. Import licensing requirement
  5. Import monitoring
  6. Import quota
  7. Import tariff
  8. Import tariff quota
  9. Internal taxation of imports
  10. Import-related non-tariff measure, nes
- Domestic subsidies (enumerated):
  1. Capital injection and equity stakes (including bailouts)
  2. Financial grant
  3. In-kind grant
  4. Tax or social insurance relief
  5. Production subsidy
  6. Interest payment subsidy
  7. Loan guarantee
  8. Import incentive
  9. Price stabilization
  10. State loan
  11. State aid, nes
  12. State aid, unspecified
- Export incentives (enumerated):
  1. Trade finance
  2. Export subsidy
  3. Tax-based export incentive
  4. Financial assistance in foreign market
  5. Other export incentive
- Foreign Direct Investment measures (enumerated):
  1. FDI: Entry and ownership rule
  2. FDI: Financial incentive
  3. FDI: Treatment and operations, nes
- Public procurement measures (enumerated):
  1. Public procurement access
  2. Public procurement, nes
- Localization content measures (enumerated):
  1. Local content incentive
  2. Local content requirement
  3. Local operations incentive
  4. Local operations requirement
  5. Local value added incentive
  6. Public procurement localization
  7. Localization, nes
- Others (enumerated):
  1. Anti-circumvention
  2. Control on personal transactions
  3. Controls on commercial transactions and investment instruments
  4. Controls on credit operations
  5. Foreign customer limit
  6. Intellectual property protection
  7. Labor market access
  8. Post-migration treatment
  9. Repatriation & surrender requirements
  10. Special safeguard
  11. Trade payment measure

*IMF Working Papers — The Return of Industrial Policy in Data (excerpt from NIPO methodology and taxonomy).*

### 12. Instrument unclear

### 12. Instrument unclear

### Other Variables in the NIPO Database
- Entry ID and Title: unique ID and title from the GTA database.
- Jurisdiction: jurisdiction implementing the policy intervention or proposal.
- Initial assessment (change relative to 1 Jan 2023): direction of policy change assessed as either distortive or liberalizing. Distortive measures generally discriminate against foreign commercial interests by restricting market access or by altering conditions in favor of local firms. Liberalizing measures tend to enhance market access on a non-discriminatory (i.e., most favored nation) basis or improve policy transparency.
- Announcement, implementation, and removal date: issuance date, date entered into force, and date withdrawn or replaced. NIPO includes measures announced or implemented from 1 January 2023.
- Level of government implementation: supra-national, national, and sub-national.
- Targeted economic activity: affected HS codes at the 6-digit level (UN HS version 2012) and CPC sector codes at the 3-digit level (UN CPC 2.1) where available; sector selection uses UN correspondence tables or GTA wording otherwise.
- Source: almost always official sources documenting the intervention.
- Additional variables:
  - Sectoral coverage: measures exclusively targeting agricultural, manufacturing, or service sectors, measures targeting more than one sector, and horizontal measures.
  - Sanctions: trade-related sanctions included given their use for industrial policies motivated by national security and geopolitical objectives.
- Trade covered: amount of goods trade covered (in USD millions) for each entry, where relevant. Interpretation depends on policy type:
  - For subsidies: trade covered = value of imports covered by subsidies to import-competing firms.
  - For export policies: trade covered = imports in the affected country potentially affected due to outward measures by implementing jurisdiction.
  - For outward/export incentives: trade affected = value of exports from third-parties competing with a subsidized rival in a shared export destination.
- Size of subsidy: subsidy value of the NIPO entry (in USD millions) (only for import competing and export incentives).

### A snapshot of the NIPO database in 2023 — Aggregate activity and instrument use
- Total new industrial policy interventions in 2023: over 2,500 measures implemented throughout the year.
- Number and share trade distorting in 2023: around 1,800 measures, or 71 percent.
- Distribution of trade distorting NIPs by income group:
  - AEs accounted for 70.9 percent of trade distorting measures.
  - EMDEs accounted for 29.1 percent of trade distorting measures.19
- Concentration: China, the European Union, and the United States account for 47.7 percent of trade distorting measures in the database.
- Most frequently used policy instruments (both groups): subsidies to domestic producers.
- Second most frequent instruments:
  - In AEs: export incentives, followed by localization policies (public procurement and investment controls).
  - In EMDEs: import barriers, followed by localization policies.
- Jurisdictional examples:
  - Export incentives: Canada, Germany, Japan, and Korea are most active by number of interventions.
  - Localization measures: United States and India rely most on localization measures.
  - Export curbs: China, India, and Russia are the most frequent users.
  - Low-income developing countries: heavy reliance on import barriers.
- Instrumental differences by income group and fiscal capacity hypothesis:
  - AEs tend to rely on direct financial grants, state loans, and other state aid.
  - EMDEs opt for import tariffs, state loans and tax relief, and more trade restrictions on imports and exports—policies less dependent on direct government budget expenditures.
  - These patterns provide tentative support for a role of fiscal space in instrument choice; to be further explored with regression analysis in Section 4.

### Regional and sectoral patterns
- Regional instrument usage highlights:
  - Europe and Central Asia, and North America: comparatively more use of domestic subsidies.
  - Asia Pacific, Latin America and the Caribbean, and South Asia: greater reliance on import barriers.
  - South Asia and North America: comparatively more use of localization measures than other regions.
  - Asia Pacific and Europe and Central Asia: most active users of export curbs.
  - North America: more public procurement measures than other regions.
  - Measures related to FDI: relatively infrequent and no significant regional differences.
- Sectoral focus in 2023 (cumulative stock of measures; for measures covering multiple sectors, each sector given equal weight):
  - Military/civilian dual use products: 25.7 percent.
  - Other advanced technologies (including medical products and semiconductors): 20.6 percent.
  - Low carbon technology products: 15.3 percent.
  - Steel and aluminum: 10.1 percent.
  - Critical minerals (upstream inputs): 3.0 percent.
  - Note: medical goods were more prominent during the early COVID-19 phase; 2023 focus shifted toward the above sectors.

### Stated motivations for NIPs in 2023 (subset with available motives)
- Predominant stated motives (cumulative stock; where multiple motives per measure, each given equal weight):
  - Strategic competitiveness: 37.0 percent.
  - Climate-related concerns: 28.1 percent.
  - Supply chain resilience: 15.2 percent.
  - Geopolitical concerns and national security: 19.7 percent.
- Observations:
  - Only officially provided motives are recorded; no direct evidence of protectionist intent from motives alone.
  - The proliferation of motives raises questions about design efficiency, alternative policies, and accommodation within the world trading system given potential spillover effects.

### Imports covered by distortive measures and stated motive (Table 6 — exact figures)
- Total inward measures in NIPO with trade value data: 882 entries (56% of inward distortive measures in NIPO; 882 out of 1,576).
- Entries with trade value (percent): 93.1%
- Imports covered (billions, USD): 3,789.8
- Imports covered (percent of global imports): 21.6%
- Breakdown:
  - Without Stated Motive: 408 entries; 93.6% entries with trade value; Imports covered (billions, USD) 2,143.4; Imports covered (percent) 12.2%
  - With Stated Motive: 474 entries; 92.6% entries with trade value; Imports covered (billions, USD) 2,269.5; Imports covered (percent) 12.9%
  - Strategic Competitiveness: 240 entries; 95.4% with trade value; Imports covered (billions, USD) 1,507.9; Imports covered (percent) 8.6%
  - Climate Change: 201 entries; 98.5% with trade value; Imports covered (billions, USD) 1,010.6; Imports covered (percent) 5.8%
  - Supply Chain Resilience: 161 entries; 98.1% with trade value; Imports covered (billions, USD) 812.9; Imports covered (percent) 4.6%
  - National Security: 46 entries; 76.1% with trade value; Imports covered (billions, USD) 659.7; Imports covered (percent) 3.8%
  - Geopolitical Concerns: 27 entries; 55.6% with trade value; Imports covered (billions, USD) 75.7; Imports covered (percent) 0.4%
- Note: Import coverage based on matching HS six-digit products covered by a measure with HS six-digit UN Comtrade data for 2019. Only inward measures included (import policies, domestic subsidies, procurement, localization). Total imports covered account for 56% of inward distortive measures in NIPO (882 of 1,576). Entries may have multiple motives.

### Correspondence between income class, motivation, and instrument
- Motivation profiles by income class:
  - Climate, geopolitics, and national security motives: predominantly given by Advanced Economies (AEs), little used by EMDEs.
  - EMDEs: main motives are strategic competitiveness and other motives.
- Instrument-motive patterns:
  - Climate concerns: instrument of choice are domestic subsidies.
  - National security: tools are varied, including trade measures and public procurement.
  - Strategic competitiveness:
    - In AEs: addressed with domestic and export subsidies.
    - In EMDEs: in around half the cases EMDEs use other instruments.

### A first look at determinants of industrial policy use — empirical approach
- Five classes of hypothesized determinants explored:
  - (i) global market position,
  - (ii) retaliation dynamics,
  - (iii) political economy factors,
  - (iv) structural factors,
  - (v) cyclical factors.
- Regression frameworks:
  - Product-country panel regression (Equation (1)):
    - Measures_cs = β0 + β1 RCA_cs + β2 #Measures by Others_2022_cs + δ_cs + ε_cs
    - Outcome Measures_cs can be intensity (NIPs count) or extensive margin dummy (Dummy #NIPs>0).
    - PPML estimator used to account for ~84 percent zeros in dependent variable.
  - Country-level specification (Equation (2)):
    - Measures_cs = β0 + β1 Political Econ_c + β2 Structural_c + β3 Cyclical_c + δ_s + ε_cs
    - Outcome Measures_cs = number of new industrial policy measures by jurisdiction c affecting product s in 2023.
    - Most stringent product-level (HS 6-digit) fixed effects included.

### Key regression findings (Tables 7 and 8 — summary of results)
- RCA and tit-for-tat dynamics (Table 7):
  - Positive correlation between new industrial policies and:
    - Revealed Comparative Advantage (RCA).
    - Prior activity by other countries affecting the same product in the previous year.
  - Interpretation: states tend to target sectors where RCA is already high (established global market presence) and tend to follow sectors others have targeted previously (tit-for-tat dynamic).
- Country-level determinants (Table 8):
  - Structural variables increase goodness of fit (R-squared) more markedly than political economy or cyclical variables.
  - Political economy:
    - Elections in the year of implementation or the following year: positive correlation with number of NIPs (become statistically significant when other sources of variation are controlled for).
    - Governments leaning further to the right: tend to use more NIPs.
    - Government effectiveness (World Bank WGI): higher effectiveness correlated with greater likelihood of undertaking industrial policies.
  - Structural variables:
    - High export concentration (Herfindahl–Hirschman Index of exports): positively correlated with number of IP measures (consistent with motive to diversify the economy).
    - Sovereign debt rating:
      - Without full controls (Column 3): correlation with sovereign debt rating appears positive.
      - With full controls (Column 1): economies with poorer sovereign debt ratings tend to use more measures.
      - Implication: fiscal-space variables from Kose, et al. (2022) do not appear to prevent countries from engaging in IP; results may be concerning as subsidies intended to reduce external fiscal risks could contribute to fiscal risk buildup.
    - Credit to the private sector: negatively related to use of industrial policy measures, suggesting substitutability between market financing access and government intervention.
    - Higher GDP correlates with more measures (consistent with AEs being most active).
  - Cyclical variables:
    - Change in GDP from previous year: associated with fewer measures when only cyclical variables included (Column 4), implying countercyclical motivation; correlation loses significance once political economy and structural variables are controlled for.
    - Real effective exchange rate (REER) appreciation: positively and significantly correlated with industrial policy measures, consistent with use of NIPs to achieve short-term goals rather than long-term structural transformation.
- Estimation notes:
  - PPML used due to large share of zeros (~84 percent).
  - Results generally robust to OLS estimator though OLS may be biased given zeros.

*Source: wpiea2024001-print-pdf - 12. Instrument unclear*

### Conclusions

### Conclusions

### Overview of the NIPO dataset and purpose
- Introduces the New Industrial Policy Observatory (NIPO) dataset based on granular monitoring of the industrial policy landscape.
- Intended as a public resource to enhance transparency, raise awareness among policy makers, and promote further research into implications of the proliferation of new industrial policies.
- Granularity is intended to facilitate:
  - assessment of the impact of NIPs on targeted economic and non-economic outcomes;
  - study of cumulative impacts of industrial policy measures and dynamics of policy reprisals.
- Note: individual measures may often be targeted and will not usually cause movements of macroeconomic indicators.

### Key 2023 empirical patterns from the NIPO data
- Significant differences between AEs and EMDEs in:
  - frequency of measures; and
  - choices of instrument used.
- AEs:
  - more active in implementing new industrial policies;
  - have done so primarily through the use of domestic and export subsidies.
- EMDEs:
  - have more frequently used trade restrictions on imports and exports.
- Motives reported by governments (dominance and relative shares):
  - Strategic competitiveness is the dominant motive governments give for taking action.
  - Climate change and supply chain resilience follow strategic competitiveness.
  - Geopolitical and national security concerns account for a smaller share; however, the combined trade coverage of the latter three sets of measures (climate change, supply chain resilience, geopolitical/national security) is equal to that of the most popular motive (strategic competitiveness).

### Exploratory regression findings and correlates of NIP introduction
- Sector-level correlates:
  - Introduction of NIPs in a given sector is correlated with revealed comparative advantage.
  - Correlated with how established a certain sector is.
  - Correlated with the use of measures by others in the same sector, indicating potential tit-for-tat dynamics.
- Domestic political economy, structural, and cyclical correlates:
  - Domestic political economy factors (e.g., current or upcoming elections) correlate with the use of industrial policy.
  - Structural factors (e.g., past fiscal profligacy) correlate with the use of industrial policy.
  - Cyclical conditions (e.g., loss of external competitiveness through exchange rate appreciation) correlate with the use of industrial policy.
- Policy implication drawn:
  - The relevance of all these factors suggests that governments should exert caution in using industrial policy.

### Avenues for future research enabled by NIPO
- Assess the success or failure of covered industrial policies at meeting their stated objectives (both economic and non-economic).
- Assess domestic macroeconomic implications of industrial policies.
- Use the NIPO dataset as a basis for measuring:
  - distortions to trade patterns;
  - changes in the allocation of resources across sectors;
  - domestic fiscal implications.
- Study impacts on trading partners, including:
  - direct distortions to competition; and
  - global general equilibrium welfare effects.
- Systemic policy questions:
  - Whether current trade rules and multilateral surveillance and enforcement mechanisms are adequate to curtail negative spillovers while allowing sufficient flexibility when needed.
  - How best to update trade rules and mechanisms in areas where they are lacking.

### Annex I — GTA data collection methodology (summary)
- GTA documents credible announcements of meaningful and unilateral changes by governments that affect relative treatment of foreign versus domestic commercial interests.
- Dataset begins in November 2008.
- Foreign commercial interests covered: trade in goods and services, investment, and labor force migration.
- Over 60 different types of commercial policy intervention—including subsidies—are documented.
- GTA does not track changes in Technical Barriers to Trade and Sanitary and Phytosanitary Measures.
- Each GTA entry provides:
  - implementing jurisdiction; direction of change (distortive or liberalizing); announced policy instrument; announcement date and, where available, implementation date; sectors and products covered; and for goods, potentially affected trading partners (identified based on official United Nations trade flow data).
- Each entry is based on official statements where possible and undergoes a two-stage review process.
- As of this writing, over 60,000 policy interventions have been documented since the GTA initiative began.
- Seven conditions for an intervention to warrant a new GTA entry:
  1) Unilateral Action: deliberate action that tilts the playing field to benefit or harm foreign commercial interests; excludes bi-, pluri-, or multilaterally agreed interventions.
  2) Relative Treatment Test: must alter relative treatment of domestic commercial interests vis-à-vis foreign competitors; distortive if it discriminates against foreign commercial entities in favor of at least one rival with operations in the implementing jurisdiction.
  3) Meaningful Change: likely to meaningfully change international commercial flows.
  4) Credible Action: implemented already or its future implementation date is enacted and known.
  5) Absence of uncontested higher motive: excludes multilateral measures with codified goals superior to preservation of seamless international commerce.
  6) One announcement, one entry: interventions with the same announcement are reported in the same GTA entry.
  7) GTA monitoring period: meaningful change announced on or after 1 November 2008.

### Annex II — NIPO coverage
- The NIPO database actively tracks 75 jurisdictions (not all currently have measures recorded).

*IMF WORKING PAPERS The Return of Industrial Policy in Data*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024001-print-pdf.pdf_