How Trade Reduces the Risk of War
This generates variation in bilateral trade that cannot be explained by political alignment, regime type, or reverse causation. Countries with little or no part in the birth of aviation suddenly saw this new technology dramatically change the distance between them and their partners. The result is reproduced under an entirely different natural experiment: The 1985 liberalization of Soviet airspace, which, by abruptly shortening flight paths between Western Europe and East Asia, produced a sharp, time-dependent rise in trade between affected pairs. Two independent identification strategies, one answer.
The Findings
According to the authors’ estimates, the doubling of bilateral trade due to reduced shipping costs of aviation reduces the probability of militarized conflict between two countries by roughly 30 percent. It also reduces the intensity of conflict when it occurs. Given the economic and human cost of conflict, this dividend of economic integration is enormous.
The largest effects accrued in the East and Southeast Asian theater, where rapid integration into global manufacturing networks coincided with the stabilization of a historically volatile region. China, South Korea, Thailand, Myanmar, and the Philippines led the leaderboard in how much trade reduced their risk of conflict. The postwar stabilization of East Asia is one of the most consequential geopolitical phenomena of the last 70 years, and these trade effects offer a plausible causal mechanism for a key contributor to how this stabilization was achieved. This is the ledger’s off-diagonal at work: Economic integration generating a security return that analysts of the decoupling trend fail to recognize.
Critically, these statistical models establish that trade not only reduces realized conflict but also dampens the perceptions of enmity that pave the way to conflict. Using a strategic rivalry dataset that measures whether governments view one another as competitive and threatening enemies, the authors find that higher bilateral trade significantly reduces the probability of strategic rivalry—the perception of hostility that shapes state behavior long before boots begin to move. Trade soothes tensions before they harden into violence.
Reckoning with the Realists
Despite the strength of the evidence, it is also undeniable that trade creates mutual gains which, even where they temper our worst instincts, can be weaponized as coercive leverage. Albert Hirschman, in National Power and the Structure of Foreign Trade (1945), argued that asymmetric trade dependence creates leverage that one state can use to coerce another. Kenneth Waltz and Robert Gilpin extended the case: Interdependence breeds vulnerability, and vulnerability breeds conflict. The contemporary geoeconomics literature, studies by Cathrin Mohr and Christoph Trebesch as well as Christopher Clayton and coauthors, have documented how trade ties can become instruments of geopolitical influence through sanctions, export controls, and supply chain weaponization.
In 2026, the bite of these arguments is hard to miss. Europe’s reliance on Russian gas before February 2022 is one example; China’s tightening of rare earth export licenses since 2023 is another; the Strait of Hormuz is currently wreaking havoc on global supply chains. The argument that interdependence creates exploitable vulnerability is unequivocally correct. It is the reason why narrow, targeted efforts to build redundancy or limit concentration in certain commercial domains have defensible logic. Whether the broader de-risking or decoupling agenda can claim the same justification is a different question, one that this new evidence brings us one step closer to answering.
The estimates by Feng et al. imply that blanket efforts to limit dependence by reducing integration will, on average, raise the likelihood of conflict—the security cost the decoupling ledger is not pricing in.
Completing the Ledger
Ultimately, security alliances generate trade returns that the burden-sharing argument has overlooked. Trade generates security returns that the decoupling agenda has taken for granted. These are the two diagonals of Washington’s policy ledger. The post-1945 architecture has been producing returns on margins that current policy debates simply do not see, and each omission makes the architecture look costlier than it is.
This does not prejudge any specific policy question. It establishes what serious cost-benefit analysis should now account for: The security dividend that broad fragmentation forfeits is no longer theoretical. The economic costs of trade restrictions are visible in price indices and trade volumes. The security costs are not, but they are no less real, and they are now quantifiable. Accounting for them is the minimum prerequisite for evaluating the merits of any fragmentation policy. What these costs imply for de-risking and decoupling debates is the subject of a forthcoming piece.
What comes next will be designed in capitals and shaped by politicians, policymakers, and pundits working from a ledger whose diagonals they have never read. They no longer have that excuse. The postwar order lasted because its architects, knowingly or not, built something whose parts reinforced each other. The next order will last only if its architects design with their eyes open.
Philip Luck is director of the Economics Program and Scholl Chair in International Business at the Center for Strategic and International Studies in Washington, D.C. Christopher M. Meissner is professor of economics at the University of California, Davis, and a research associate at the National Bureau of Economic Research.