A Deep Case for Shallow Integration -- by John Sturm Becko, Arnaud Costinot

TEXT START: We study international cooperation in a general environment with pecuniary and non-pecuniary externalities across countries.

The Dissection

The paper constructs a formal case that international cooperation need not require deep political integration. Market-access commitments, extended to treat cross-border externalities as tradable and priceable goods, can supposedly implement globally efficient outcomes.

Its real move is to convert political conflict into contract design. Climate damage, geopolitical spillovers, and other non-market effects are translated into prices, commitments, and exchange conditions. That is a powerful theorem inside a functioning institutional order. It is not a theory of whether that order remains socially or politically viable.

The Core Fallacy

The central error under the Discontinuity Thesis is a category error: it treats allocative efficiency as systemic survival.

Even if every externality can be priced and every commitment enforced, pricing does not restore the mass employment–wage–consumption circuit once AI makes cognitive labor economically unnecessary. A globally efficient allocation can coexist with concentrated ownership of AI capital, permanent exclusion from productive work, and a population maintained through transfers rather than participation.

The paper may be mathematically correct within its assumptions. That is precisely its limitation. It optimizes the exchange layer while leaving the ownership layer—and therefore the fate of human labor—untouched.

Hidden Assumptions

• States remain capable of negotiating and enforcing credible long-term commitments.
• Social values for non-pecuniary externalities can be measured, aggregated, and legitimized politically.
• Property rights, markets, and purchasing power remain sufficiently stable for the mechanism to operate.
• Domestic losers can be compensated, and governments retain the capacity to distribute those compensations.
• Strategic rivalry does not override the gains from cooperation.
• “Global efficiency” is an accepted objective rather than a result imposed by whoever controls the valuation and enforcement machinery.
• The transition to AI-dominated production is slow enough for institutions to adapt.
• Preserving consumption and trade is treated as adequate even if productive participation collapses.

The most important assumption is the one the paper does not examine: that the majority will continue to matter as economic agents. Under P1–P3, that assumption is the first thing to die.

Social Function

Classification: partial truth, transition management, and elite self-exoneration, with a layer of ideological anesthetic.

The partial truth is real: shallow commitments may reduce negotiation costs and coordinate states more effectively than attempts at full harmonization. That makes the framework useful as a lag defense and as transition-intermediation machinery.

The anesthetic is the implication that better institutional architecture can domesticate the coming rupture. It reassures policymakers and globally mobile capital that the order can be repaired through more elegant pricing and contracting, while avoiding the harder questions: who owns the productive systems, who controls energy and logistics, who receives the income, and whether humans remain necessary at all.

The Verdict

This is a rigorous coordination theorem, not a rescue plan for post-WWII capitalism. Shallow integration may make a declining system more efficient at pricing externalities and managing interstate friction. It cannot reverse P1, defeat coordination failure at the scale of human displacement, or restore productive participation after AI severs the wage circuit.

Under the Discontinuity Thesis, the paper describes a possible control panel for the transition—not a cure for the disease. The order may become better coordinated while becoming less humanly necessary. That is not survival. It is orderly obsolescence.