💰 quantitative finance

Cascading Impacts of the USA--China Trade War on Global Oilseed Supply Chain

Original authors: Diksha Gupta, Ritwick Mishra, Achla Marathe, Krista Danielle Yu, Anil Vullikanti

Published 2026-07-01

📖 5 min read🧠 Deep dive

Original authors: Diksha Gupta, Ritwick Mishra, Achla Marathe, Krista Danielle Yu, Anil Vullikanti

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the global economy as a massive, intricate game of Jenga. Every country is a block, and every trade deal is a connection between those blocks. When you pull one block out (a trade disruption), the whole tower wobbles. Sometimes, just one block falls, but often, the vibration travels through the tower, causing other blocks far away to topple too.

This paper is about what happens when you pull a very specific, heavy block out of the tower: the trade of oilseeds (like soybeans) from the United States to China.

Here is the story of the paper, broken down into simple concepts:

1. The Problem: The "Trade War" Wobble

In 2018 and again in recent years, the U.S. and China got into a trade fight. China put high taxes (tariffs) on U.S. soybeans. Suddenly, the flow of soybeans from the U.S. to China stopped or slowed down drastically.

The authors wanted to know: If the U.S. stops selling to China, what happens to the rest of the world?

Does the whole global economy crash?
Does China just buy from someone else?
Does the price go up, or does the supply just run out?

2. The Tool: A "Digital Twin" of the Economy

To answer this, the researchers built a mathematical simulation (a Linear Programming model). Think of this as a super-advanced video game of the global economy.

The Old Way: Previous models were like a static map. They showed where things were, but if you blocked a road, they just said, "Okay, that road is closed," and stopped there. They didn't account for drivers finding new routes or building new roads.
The New Way: This paper's model is like a live traffic simulator. If you block the main highway (U.S. to China), the simulator instantly calculates:
- Can drivers (traders) find a detour (reallocation)?
- Can construction crews build a new lane (production expansion)?
- Where does the traffic jam move to next?

3. The Experiment: Four Scenarios

The researchers ran four different "what-if" scenarios in their simulator:

Scenario A (The Disaster): The U.S. stops selling to China. No one is allowed to switch suppliers or make more beans.
- Result: It's a mess. China loses a huge amount of food supply (14% loss in their specific sector). But here's the twist: smaller countries (like some in Africa) also suffer disproportionately because they rely on the ripple effects of Chinese trade, even though they aren't directly involved in the U.S.-China fight.
Scenario B (The Detour): The U.S. stops selling, but China is allowed to buy 20% more from Brazil (the other big soybean seller).
- Result: This acts like a life raft. The global damage drops significantly. China is much happier. However, the "detour" puts pressure on Brazil. Brazil has to send more beans to China, which means there are fewer beans for other countries that usually buy from Brazil (like Thailand and Norway). The problem didn't disappear; it just moved.
Scenario C (The Local Fix): China buys more from Brazil, and Brazil tries to grow more beans using its own local resources (land, workers, fertilizer).
- Result: Brazil actually makes a tiny profit because they are producing more. But, because Brazil is using up its own local resources to grow more beans, it creates shortages for other things Brazil usually makes. It's like a farmer stopping his vegetable garden to plant only corn; he has more corn, but no vegetables for his family.
Scenario D (The Global Team-Up): China buys more from Brazil, and Brazil grows more beans by importing resources from the whole world (not just its own backyard).
- Result: This spreads the pain out. Instead of just Brazil feeling the stress of growing more beans, the "cost" of growing those extra beans is shared by many countries that supply Brazil with fertilizer, machinery, and labor. It's a fairer distribution of the burden, though the total amount of beans produced is similar to Scenario C.

4. The Big Takeaways (The "So What?")

The "Butterfly Effect" is Real: A fight between two giants (U.S. and China) doesn't just hurt them. It sends shockwaves that can knock over small economies that seem far away.
Trade Reallocation isn't Magic: You can't just say, "Oh, Brazil will just fill the gap." If Brazil fills the gap, it creates a new shortage for everyone else who buys from Brazil. It's a trade-off. You fix one problem, but you create a new pressure point elsewhere.
Production has Limits: You can't just "make more" instantly. If Brazil grows more beans, it needs more workers and fertilizer. If those things are in short supply, the cost of fixing the problem gets passed down the line to other countries.
Small Economies are Fragile: The study found that while China takes the biggest absolute hit (losing the most money), some tiny economies take the biggest percentage hit. They are the most vulnerable to these cascading failures.

Summary Analogy

Imagine a massive potluck dinner where everyone brings a dish.

The Disruption: The person bringing the main course (U.S. soybeans) suddenly leaves.
The Detour: Someone else (Brazil) tries to bring a bigger dish to fill the gap.
The Consequence: That person (Brazil) now has to skip bringing their own side dish, so the people who usually eat that side dish go hungry.
The Lesson: You can fix the main course shortage, but you have to be careful about who else goes hungry in the process. The paper's math helps us see exactly who gets the short end of the stick before we make the decision.

Technical Summary: Cascading Impacts of the USA–China Trade War on Global Oilseed Supply Chain

Problem Statement

Global supply chains are highly interconnected, rendering them vulnerable to cascading disruptions triggered by trade policy shocks. While existing literature often analyzes trade disruptions using Computable General Equilibrium (CGE) models or Agent-Based Models (ABM), these approaches face limitations in capturing physical bottlenecks, capacity constraints, and the specific mechanics of cascading shortages in interconnected networks. CGE models rely on price-mediated adjustments that may obscure physical flow constraints, while ABMs often lack system-level equilibrium characterizations and scale poorly to economy-wide settings.

The specific problem addressed is the inability of current models to simultaneously represent:

Cascading disruptions propagating through production networks.
Endogenous mitigation mechanisms, specifically trade reallocation (shifting flows to alternative suppliers) and production expansion (increasing output subject to capacity limits).
System-level equilibria that account for the trade-offs between minimizing unmet demand and preserving global output.

The authors focus on the U.S.–China trade war as a case study, specifically analyzing the disruption of oilseed (soybean) flows, where China relies heavily on imports from the U.S., Brazil, and Argentina.

Methodology

The authors propose a Linear Programming (LP) formulation of an Input-Output (IO) system to model the global supply chain as a network of countries, sectors, firms, and households.

System Abstraction

Firms: Each firm represents a single sector in a single country.
Production Function: Modeled using a Leontief production function, where output depends on intermediate inputs from other firms and primary inputs (labor, capital, resources).
Flow Conservation: Total output equals the sum of intermediate flows to other firms and final flows to households.

Optimization Framework

The model is formulated as a two-phase Linear Program to compute approximate equilibrium solutions by minimizing total unmet demand:

Phase 1 (Maximize Output): Maximize global system output ( $X$ ) subject to constraints.
Phase 2 (Minimize Slack): Minimize the total unmet global household demand ( $G$ ) while maintaining the maximum output found in Phase 1.

Key Constraints

Leontief Constraints: Ensure intermediate and primary inputs are sufficient for production levels.
Disruption Constraints: Exogenously limit flows from disrupted firms (e.g., U.S. to China) to a fraction of baseline values ( $\Delta$ ).
Reallocation Constraints: Allow non-disrupted firms to increase flows to alternative destinations up to an elasticity factor ( $\alpha$ ).
Production Constraints: Allow specific firms to expand output ( $\eta$ ) and primary inputs ( $\beta$ ) within feasible bounds.
Slack Variable ( $G$ ): Represents the gap between total household supply and demand, allowing the model to handle supply shortages without forcing infeasible solutions.

Scaling Strategy

To handle large-scale MRIO (Multi-Region Input-Output) tables, the authors apply value-based sparsification, removing intermediate flows below a specific threshold ( $\tau$ ) to reduce dimensionality while preserving the overall network structure.

Key Contributions

The paper makes three primary contributions:

Unified LP Formulation: A novel linear programming formulation that captures cascading supply-chain disruptions alongside endogenous trade reallocation and production adjustment under capacity constraints. Unlike prior static IO-LP models, this enables the joint analysis of disruption propagation and mitigation pathways within a single equilibrium framework.
Large-Scale Empirical Application: Application of the framework to the global oilseeds supply chain using GTAP MRIO data (2023) to analyze tariff-induced disruptions from the U.S.–China trade war. The study conducts counterfactual analyses to quantify how reallocation and production expansion differentially mitigate losses.
Insight into Trade Diversion Limits: The analysis demonstrates that trade diversion is conditional. When trade and capacity constraints limit adjustment, reallocation may be insufficient to fully offset disruptions, resulting in unmet final demand. This shifts the analytical focus from price-mediated adjustment (common in CGE models) to the structural limits of global supply chains.

Results

The authors applied the model to four scenarios involving a disruption of U.S. oilseed exports to China:

1. Baseline Disruption (Scenario I)

Impact: A localized 70% disruption to U.S. oilseed flows to China resulted in a 3.27% loss in global output.
Distribution: China experienced a disproportionate loss of 14.02%. However, several smaller economies (particularly in Africa) faced disproportionately high percentage final-demand losses due to downstream exposure, despite lower absolute losses.
Mechanism: Demand-side effects were more immediately sensitive to the shock than production-side cascades.

2. Trade Reallocation (Scenario II)

Strategy: Allowing a 20% reallocation of Brazilian oilseed exports to China.
Outcome: Global output losses were significantly reduced to 1.36%.
Trade-offs: While this alleviated the bottleneck for China, it shifted pressure onto other countries importing from Brazil (e.g., Thailand, Norway), which absorbed the remaining unmet demand. Brazil itself saw a net reduction in loss, suggesting mutual benefits in aggregate.

3. Domestic Production Expansion (Scenario III)

Strategy: Increasing Brazil's oilseed production by 10% using domestic reallocation of resources and primary inputs.
Outcome: Brazil achieved a net production gain of 0.13%.
Trade-offs: This strategy disproportionately shifted losses toward smaller economies in Africa and Asia, which faced increased final-demand losses as domestic resources were redirected to support the expanded oilseed sector.

4. Global Production Expansion (Scenario IV)

Strategy: Increasing Brazil's production by 10% supported by global inputs (international trade partners) rather than just domestic resources.
Outcome: Aggregate global losses remained similar to Scenario III.
Trade-offs: Diversifying input sources redistributed the adjustment costs more broadly across the international network. While aggregate gains were modest, smaller economies experienced larger percentage recoveries compared to the domestic-only strategy.

Significance and Claims

The paper claims that its framework provides a more realistic representation of supply chain resilience by explicitly modeling physical bottlenecks and capacity constraints, which are often smoothed over in price-based equilibrium models.

Conditional Trade Diversion: The study concludes that trade diversion is not automatic; under constrained substitution and production capacity, mitigation strategies (reallocation and expansion) can reduce aggregate losses but often create localized resource stress or shift burdens to vulnerable economies.
Efficiency vs. Equity: The authors note that their LP formulation is driven by efficiency (maximizing global output) rather than welfare weights. Consequently, the results highlight structural inefficiencies and capacity limits rather than prescribing how societies ought to prioritize demand satisfaction based on vulnerability or humanitarian needs.
Policy Relevance: The findings suggest that while mitigation strategies are effective, their success depends heavily on spare capacity, existing trade relationships, and the specific distribution of adjustment costs across the network.

The authors do not propose future applications or specific policy recommendations beyond the scope of the counterfactual analysis presented, maintaining a focus on the methodological capability to quantify system-level effects of large-scale disruptions.