Global Trade

Key Concepts

• Global Trade Data Sources: UN Comtrade, BACI, Gravity Trade Database, MBR World Trade Database, IMF Direction of Trade Statistics (DOTS).
• Global Value Chains: World Input-Output Database (WIOD), Eora Global Supply Chain Database, OECD Trade in Value Added (TiVA) database.
• Trade Policy Data: UN TRAINS database, World Bank WITS database.
• Trade Models: Armington model, Eaton-Kortum model, Krugman model, Melitz model, Constant Elasticity of Substitution (CES) trade models.
• Gravity Equation: Bilateral trade flows are related to importer and exporter fixed effects and bilateral frictions.
• Exact Algebra Counterfactuals: A method for solving trade models by expressing changes relative to an initial equilibrium.
• Linearization: Approximating a non-linear model with a linear one to simplify analysis.
• Friends and Enemies Matrix: A matrix representing the impact of productivity growth in one country on the welfare of another.
• Hub and Authority Scores: Network measures indicating a country's importance as a source of shocks (authority) or sensitivity to shocks (hub).
• Dynamic General Equilibrium (DGE) Model: A macroeconomic model with intertemporal optimization and general equilibrium effects.
• Ramsey Model: A neoclassical growth model with endogenous savings and investment decisions.
• Capital Holdings: Bilateral holdings of capital between countries.
• Decoupling: Reduction in trade and capital flows between countries.

1. Data Sources for Global Trade Analysis

• UN Comtrade: Comprehensive global trade data since 1960, reported by Harmonized System (HS) six-digit product codes (around 10,000 categories). Individual countries may report at more disaggregated levels (e.g., US at 10-digit).
• BACI (SEPY): Cleaned-up version of UN Comtrade data, mapping goods between HS classifications from 1995-2023.
• Gravity Trade Database (SEPY): Aggregate country-level trade data back to 1948, combined with standard gravity equation variables (distance, common border, etc.).
• MBR World Trade Database: Data from 1962-2000 at a disaggregated product level.
• IMF DOTS: Long-term data on bilateral trade between countries, dating back to 1948.
• WIOD: Input-output matrices for 43 countries and the rest of the world, reported for 56 sectors.
• Eora Global Supply Chain Database: Intersectoral flows across 1,900 sectors in 190 countries (1990-2022), but with a significant portion of imputed data.
• OECD TiVA: Data on trade in value added, attempting to correct gross flows for global value chains. Relies on a proportionality assumption that a country uses the same input-output structure for all its export partners.
• UN TRAINS: Tariff data by country, partner, and product at the HS six-digit level, including most favored nation (MFN) and preferential tariffs.
• WITS (World Bank): Interface for accessing TRAINS data, but may have issues with missing preferential tariff data.
• Peter Schott's Website: Concordances for different versions of the HS classification and between NAICS and HS.
• Forum for Research in Empirical International Trade (FREIT): Website with concordances and links to data.
• Federico Tena World Trade Historical Database: Focuses on the period 1800-1938.
• TRA HIS (SEPY): Data on a consistent basis for 188 years from 1827 through to 2014.
• Ricardo Trade Database: Data for a historical period.
• Linked Firm Trade Transactions Database (Census Bureau): Firm-level trade data for the US, accessible through a Census Research Data Center.
• Bills of Lading Data (Datamyne, Panjiva): Data on shippers, recipients, products, and ports, but trade values are often imputed.

2. Modeling Global Value Chains and Trade Policy

• Limitations of Trade Data: Trade data are reported on a gross basis, not breaking out the value-added chain.
• Addressing Global Value Chains:
  • Use gross flows data with a model that explicitly models input-output linkages (e.g., WIOD).
  • Transform the data to construct trade in value added (e.g., OECD TiVA).
• Challenges in Measuring Non-Tariff Barriers: Difficult to measure. TRAINS database has some information on the number or presence of non-tariff barriers. Coverage ratios can be calculated.
• Importance of Data Validation: Theodora Teti's work highlights issues with imputed tariff data in the WITS database, leading to potentially incorrect trade elasticity estimates.

3. Classic Trade Models and General Equilibrium Analysis

• Constant Elasticity Trade Models: A class of models that imply a gravity equation with a constant elasticity of trade flows with respect to trade costs. Examples include Armington, Eaton-Kortum, Krugman, and Melitz models.
• Armington Model: Goods are differentiated by country of origin.
• Gravity Equation: Bilateral trade flows depend on importer and exporter fixed effects and bilateral frictions (e.g., distance).
• Multilateral Resistance: Trade costs with all other nations affect bilateral trade flows.
• General Equilibrium System: The equilibrium of these models reduces to a system of equations for the wages in each country.
• Exact Algebra Counterfactuals: A method for solving trade models by expressing changes relative to an initial equilibrium. This approach avoids estimating unobserved fundamentals.
• Linearization: Approximating the model with a linear system to simplify analysis and gain intuition.
• Friends and Enemies Matrix: A matrix representing the impact of productivity growth in one country on the welfare of another. This matrix can be computed from a single matrix inversion in the linearized model.
• Market Size Effect: Changes in wages in each market affect income, depending on the share of income from that market.
• Cross Substitution Effect: When a country becomes more productive, it affects the competitiveness of other countries in each market.

4. International Friends and Enemies: A Case Study

• Motivation: To understand the relationship between economic dependence and political alignment.
• Economic Friend: A country whose growth raises another country's income.
• Economic Enemy: A country whose growth reduces another country's income.
• Political Alignment: Measured using voting in the UN, strategic rivalries, and formal alliances.
• Model: Countries undertake political actions that raise the productivity of their trade partners but incur utility costs.
• Instruments:
  • China's emergence into the global economy.
  • The invention of the long-distance airplane, which reduced trade costs differentially for countries with long sea distances.
• Findings: Increases in economic dependence on a trade partner predict political realignment. Theory-based measures of economic exposure are consequential.
• Hub and Authority Scores: Network measures indicating a country's importance as a source of shocks (authority) or sensitivity to shocks (hub).

5. Combining Trade and Macro Models

• Motivation: To connect trade and capital flows in a dynamic general equilibrium framework.
• Model: A Ramsey model with Armington trade, bilateral capital holdings, and idiosyncratic shocks to investment productivity.
• Capital Allocation Decision: Representative agents allocate wealth across countries based on rental rates, productivity draws, and capital market frictions.
• Extreme Value Distribution: Idiosyncratic shocks are drawn from a Fréchet distribution, leading to a gravity equation for capital holdings and a common expected return across all host countries.
• Equilibrium: Characterized by time-invariant fundamentals and steady-state wealth stocks.
• Dynamic Exact Algebra: Can be used to solve for the transition path in the nonlinear model.
• Linearization: The model can be linearized to obtain a closed-form solution for the transition path.
• Impact Matrix (R): Determines the initial impact of shocks on the state variables.
• Transition Matrix (P): Governs the updating of the state variables.
• Findings:
  • Opening up the Ramsey model slows down the rate of convergence to steady state.
  • The model generates different predictions for deglobalization compared to models without capital flows.
  • Capital market frictions and trade frictions interact in systematic ways.

6. Key Arguments and Perspectives

• Importance of General Equilibrium: Partial equilibrium measures of trade dependence are insufficient. A general equilibrium approach is needed to account for outside options and the choice set.
• Causality Challenges: Reverse causality and omitted variables are a concern when studying the relationship between economic dependence and political alignment. Model-based instruments can help address these challenges.
• Value of Linearization: Linearization provides intuition for the mechanisms at play in trade models and allows for the computation of the full matrix of welfare exposure from a single matrix inversion.
• Trade and Macro Integration: Combining trade and macro models provides new insights into the growth process and the impact of trade and capital market frictions.
• Limitations of Constant Elasticity Models: The constant elasticity assumption imposes strong cross-substitution restrictions.

7. Notable Quotes

• "A country is an economic friend of its trade partner if growth in that trade partner raises the country's income."
• "A country is going to be an economic enemy if growth reduces its income."

8. Technical Terms and Concepts

• Harmonized System (HS): A standardized system of names and numbers for classifying traded products.
• Most Favored Nation (MFN): A principle of non-discrimination in trade policy.
• Preferential Tariff: A tariff rate that is lower than the MFN rate, granted to specific trade partners.
• Gravity Equation: A model that predicts bilateral trade flows based on importer and exporter characteristics and bilateral frictions.
• Structural Gravity: The idea that multilateral resistance terms (trade costs with all other nations) are important determinants of bilateral trade flows.
• Exact Algebra Counterfactuals: A method for solving trade models by expressing changes relative to an initial equilibrium.
• Linearization: Approximating a non-linear model with a linear one to simplify analysis.
• Fréchet Distribution: An extreme value distribution used to model idiosyncratic shocks.
• Home Bias: The tendency for countries to trade and invest disproportionately with themselves.
• Feldstein-Horioka Puzzle: The high correlation between domestic saving and domestic investment.

9. Logical Connections

• The lecture begins by introducing various data sources for global trade analysis, emphasizing their strengths and limitations.
• It then transitions to a discussion of classic trade models, focusing on the constant elasticity of substitution (CES) framework.
• The concept of exact algebra counterfactuals is introduced as a method for solving these models.
• Linearization is presented as an alternative approach that provides intuition and simplifies computation.
• The "International Friends and Enemies" paper is presented as a case study that applies these techniques to analyze the relationship between economic dependence and political alignment.
• The lecture concludes by discussing how to combine trade and macro models to gain new insights into the growth process and the impact of trade and capital market frictions.

10. Synthesis/Conclusion

The lecture provides a comprehensive overview of data sources, trade models, and analytical techniques used in international trade analysis. It highlights the importance of general equilibrium thinking, the challenges of causality, and the value of combining trade and macro perspectives. The "International Friends and Enemies" case study and the discussion of the dynamic general equilibrium model demonstrate how these tools can be applied to address important policy questions. The lecture emphasizes the ongoing evolution of the field and the potential for further integration of trade and macro approaches.