Economists Rethink Equilibrium as AI and Climate Transitions Expose Limits of Traditional Models
Artificial Intelligence & Machine Learning News
A growing debate in economic research is challenging one of the field’s core tools: equilibrium. As artificial intelligence reshapes labor markets and the global push toward net zero accelerates structural change, economists are increasingly questioning whether traditional equilibrium-based models can adequately capture today’s dynamics.
Professor R. Maria del Rio-Chanona, Assistant Professor at UCL, frames the issue directly. “When is equilibrium useful, and when do we need nonequilibrium models? If you work on AI transitions or net-zero policy, the answer is key for deciding modelling tools.” Her question reflects a broader shift underway across both academic and policy circles.
Researcher Marco Pangallo(below) reframes the issue with a practical distinction.
Equilibrium models solve systems of equations to determine outcomes. Nonequilibrium approaches, by contrast, map how economic variables evolve over time based on current conditions. The distinction, summarized as equations versus maps, is gaining traction among policymakers and quantitative researchers.
Equilibrium models have long dominated macroeconomics due to their simplicity and clarity. By focusing on end states that satisfy certain conditions, they avoid the need to model how an economy actually gets there. This makes them effective near steady states, where systems behave predictably.
However, critics argue that today’s most important economic transformations do not resemble steady states. The rapid adoption of AI is creating uneven adjustments across firms, workers, and institutions. At the same time, the transition to a low-carbon economy is producing cascading effects across supply chains, capital allocation, and regulatory frameworks. These processes are inherently path-dependent and unfold over time, often far from equilibrium.
Nonequilibrium models attempt to address these gaps by explicitly modeling trajectories. They allow for heterogeneity across agents, nonlinear feedback loops, and network effects, all of which are central to understanding modern economic systems. Approaches such as agent-based models and advanced heterogeneous-agent frameworks are increasingly positioned along this maps side of the spectrum.
Jean-Philippe Bouchaud, chairman of Capital Fund Management and a member of the Académie des Sciences, has been among the most vocal proponents of this shift.
Bouchaud(below) argues that traditional equilibrium models struggle to explain real-world phenomena such as excess market volatility, boom-and-bust cycles, and persistent inflation dynamics. These features, he notes, are better understood as emergent outcomes of complex, out-of-equilibrium systems.
The implications extend beyond academia. For investors, regulators, and policymakers, the choice of modeling framework can shape how risks are identified and managed. Equilibrium models may still provide useful benchmarks, but reliance on them alone could obscure the transitional dynamics that define today’s economy.
As AI adoption accelerates and climate policies reshape industries, the question is no longer whether equilibrium models are useful, but when they are insufficient. Increasingly, economists are concluding that understanding the path of change, not just its endpoint, is what matters most.
Join us: Cornell Financial Engineering Manhattan 2026 Future of Finance & Ai Conference : September 11 2026
