Professor Doyne Farmer of University of Oxford is leading an ambitious effort to build a Global economic simulation that models every company in the world as a digital agent making realistic, adaptive decisions. The goal is nothing less than a living, evolving model of the entire economy—capable of producing forecasts with unprecedented clarity. Farmer compares the vision to doing for economics what Google Maps did for traffic: providing practical, real-time answers to complex questions about policy, markets, and risk.
Farmer’s credibility comes from a career that blends academic depth with real-world success. In earlier decades, he helped beat casino roulette using early wearable computing and later co-founded an automated trading firm that was eventually sold to UBS. Now, after decades spanning cosmology, chaos theory, and theoretical biology, he believes advances in complexity science and computing power make a full Global economic simulation achievable for roughly $100 million—a modest sum compared to the trillions lost in past crises.
Traditional economic models struggle because they are either too simple to be useful or too mathematically rigid to reflect reality. They typically assume perfectly rational actors and equilibrium conditions, where supply and demand neatly balance. But the real economy is messy, adaptive, and far from equilibrium. Crashes, bubbles, and sudden shifts emerge internally, not just from external “shocks.” Farmer’s complexity-based approach replaces idealized agents with millions of simpler digital actors that learn through trial and error, imitation, and machine learning. These agents evolve their strategies over time, producing booms and busts organically—much closer to how real economies behave.
The stakes are enormous. The 2008 financial crisis alone cost the world about $10 trillion. Farmer argues that even a modestly accurate model could have flagged the danger years earlier. If such foresight reduced losses by just 1%, the investment would have paid for itself a thousand times over. His team has already demonstrated the concept by retrospectively modeling US real estate markets and simulating the UK economy during Covid, showing how policymakers might have better balanced public health and economic outcomes.
Today, Farmer’s primary focus is climate change, where he believes existing economic models are especially misleading. They repeatedly underestimated how quickly renewable energy would scale and how fast costs would fall. As a first step toward a full Global economic simulation, his group is modeling the entire energy sector—30,000 companies and roughly 160,000 assets such as oil rigs and power plants—using 25 years of operational data. Each company is represented by a digital agent, allowing researchers to simulate global energy supply, pricing, and investment decisions in detail. Early results suggest that a rapid clean-energy transition could save trillions of dollars worldwide.
Farmer sees this work as urgent and deeply personal. At 73, he wants to see a functioning Global economic simulation within the next decade—ideally sooner. He believes such a tool could transform decision-making, helping governments and businesses anticipate crises, accelerate decarbonization, and avoid costly mistakes. By embracing complexity rather than oversimplifying it, Farmer hopes to give humanity a clearer map of the economic terrain ahead—and a better chance of navigating climate change and future shocks.