Plenty of Fish in the Sea

A New Risk Environment

Avoiding Trade Restrictions

What political or economic features make a target less attractive for coercive trade restrictions?

• Barriers to trade imposed by one state (sender) on another (target) intended to compel the target to change its behavior
  • Sanctions, quotas, licencing, tariffs, etc.

Current Understanding

The Broad Set-Up

• Trade restrictions disrupt otherwise efficient trade

  • By prohibiting it or making it more costly

• Costly for both the sender and target

  • Sender hopes the target will blink first

• When the target can mitigate costs imposed, the sender faces higher risk of incurring the cost and reaping no benefits

  • What tools, other than sheer economic size, are available to targets in the event of trade restrictions?

Third-Parties to the Rescue

• Can spoil trade restrictions by filling gaps created in the market

• In fact, their presence alone can deter a sender from imposing trade restrictions (Peksen and Peterson 2016)

  • Third-parties that are both willing and able to step in should:
    • Decrease the costs to the target…
    • Increasing the sender’s risk of failure…
    • Reducing the likelihood restrictions will be imposed

Who Is Able to Step-In? A Product-Level Question

Observable Implications

Research design

• US export restrictions from 2017 to 2024

  • Export Administration Regulations: Commerce Control List and Commerce Country Chart

  • Products identified as critical to actions that are a risk to US national security and interests and buyer states that are identified as at-risk of pursuing those actions

  • State-state restrictions (as opposed to state-entity restrictions)

New data set of US export restrictions

• Products identified at the HS six-digit level and at a daily rate
  • Between 2017 and 2024 the CCL was updated 96 times and the CCC was updated 17 times
• Identified over 500 products facing restrictions at a given time
• By mapping to HS codes, we can match restrictions to international trade data

Coding process

Using multiple human–LLM feedback loops, I identified:

1. The item(s) covered by each ECCN

2. Each item’s HS six-digit code(s)

3. The countries facing export restrictions on the item

ECCN descriptions are far from simple

Match to their controls

Match definitions by control to HS codes

• Created stand-alone definitions of items covered by each ECCN by its relevant control using an LLM

• Mapped these to relevant HS six-digit codes using an LLM supplied with their definitions and expert verification

Match countries to those controls

Countries with fewer partners faced more restrictions

Countries with fewer partners faced more sanctions

Summary

Implications

• Third parties play significant role in sanctioning decisions

• When third parties can supply sanctioned goods

  • Targets have reduced compliance incentives

  • Senders lose leverage

Thank You

Appendix