Economic Integration, Sanctions, and Preventive War: An Experimental Investigation

How to cite

Huang, Lingbo; Dong, Lu. 2026. "Economic Integration, Sanctions, and Preventive War: An Experimental Investigation." AEA RCT Registry. AEARCTR-0018051. https://doi.org/10.1257/rct.18051-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Lingbo Huang* , Lu Dong

* Primary investigator

Registry ID
AEARCTR-0018051
DOI
https://doi.org/10.1257/rct.18051-1.0
First registered
Last updated
Status
In development
Trial dates
2026-03-13 – 2026-04-17
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18051

Abstract

This study investigates how economic integration and the threat of sanctions influence the likelihood of preventive war. We develop a two-stage bargaining model where a declining power (DP) faces a rising power (RP). Integration increases the total economic surplus but also accelerates the power shift towards the RP. The DP can use economic sanctions to mitigate this power shift. We test the model's predictions using a laboratory experiment with four treatments varying the level of integration (Low, Medium, High) and the availability of sanctions. We hypothesize that medium integration leads to preventive war due to a large power shift and weak sanctions, while high integration fosters peace because the credible threat of severe sanctions disciplines the RP. We also test whether subjects endogenously choose high integration to access this sanction mechanism.

Intervention

The lab experiment implements a two-stage bargaining game between two players. The treatments are parameterized versions of the game. There are six between-subjects treatments in total (EXO-Low, EXO-Med, EXO-Med-NS, EXO-High, EXO-High-NS, ENDO). EXO-Low could be viewed as the control treatment. Recruited subjects are randomized assigned to one of the treatments depending on the experimental session they register. All details are provided in PAP.

Experimental design

The experiment implements a two-stage bargaining game between two players.
• Roles: Player A (the proposer) and Player B (the responder). Player A represents a “rising power” whose relative bargaining position strengthens over time, while Player B represents a “declining power” whose position weakens. In Stage 1, Player B holds a
battlefield advantage (60% winning probability); without sanctions, A’s odds improve substantially in Stage 2, reflecting B’s relative decline.
• Integration Level: Determines the size of the available surplus in each stage, as well as the magnitude of the power shift and the costs of conflict.
• “Take Action” Option: In most treatments, Player B can choose to “take action” in Stage 1. This costs B 2 points but improves B’s winning odds if conflict occurs in Stage 2. This option captures the idea of economic sanctions that weaken the opponent’s future
position.

Timeline:
1. Stage 1: A pot of points is available. Player A proposes how to divide it. Player B then chooses:
• Accept: Both players receive their allocated shares and proceed to Stage 2.
• Reject: A computerized lottery determines who wins the entire pot (combining both stages), and both players pay a conflict cost. The game ends.
• Take Action: Player B accepts the allocation but pays an additional 2-point cost. Both players proceed to Stage 2, but B’s winning odds in any Stage 2 conflict are now improved.
2. Stage 2: A second pot of points is available. Player A proposes how to divide it. Player B then chooses:
• Accept: Both players receive their allocated shares from both stages. The game ends peacefully.
• Reject: A lottery determines who wins the combined pot from both stages. Stage 1 allocations are voided. Both players pay a conflict cost. The game ends.

Experimental Treatments
Treatment & Take Action? & Stage 1 Pot & Stage 2 Pot & Conflict Cost & Predicted Outcome \\
\midrule
\textbf{EXO-Low} & Yes & 12 & 18 & 8 & Peace (Standard) \\
\textbf{EXO-Med} & Yes & 13 & 27 & 9 & Conflict \\
\textbf{EXO-Med-NS} & No & 13 & 27 & 9 & Conflict \\
\textbf{EXO-High} & Yes & 20 & 30 & 15 & Peace (Sanction-Disciplined) \\
\textbf{EXO-High-NS} & No & 20 & 30 & 15 & Peace (Standard) \\
\textbf{ENDO} & Yes & Choice & Choice & Choice & Choose High, then Peace (Sanction-Disciplined) \\

Design details

Randomization method
randomization by computer
Randomization unit
Treatment randomization is employed at session level. Pairing is randomized in each round with roles fixed throughout the session.
Clusters
Each treatment has 6 sessions of 10 subjects each. 36 sessions in total
Observations
36 sessions
Arms
60 participants per treatment. 360 in total.
Primary outcomes
Conflict Frequency; Take Action Frequency; Player A Payoff; Player B Payoff; Integration Choice
Secondary outcomes
Stage 1 Offer; Efficiency

Cited by

No citation data available for this registration yet.

Keywords