Why did traditional economic models miss the 2008 crash by a factor of 20?

Can Agent-Based Modeling predict the next financial crisis better than the Fed?

Discover why Doyne Farmer argues that conventional economics fails in complex scenarios. Learn how Agent-Based Modeling and Complex Systems Thinking offer a more accurate way to forecast global economic shifts. Stop relying on outdated predictive models—read on to understand how Complex Systems Thinking can safeguard your investments against the next economic shock.

Recommendation

Traditional economic theory tends to function adequately under fairly simple scenarios. However, as complexity in a system increases, the efficacy of conventional models deteriorates. In a recent Boston Consulting Group-Henderson Institute podcast, moderator Martin Reeves interviews Doyne Farmer, a leader in complex systems who brings a new perspective to economics. According to Farmer, Complex Systems Thinking — and in particular Agent-Based Modeling — provide far more accurate forecasting and analytics. Economists, business owners, and investors interested in a rigorous examination of emerging economic reasoning will find this a robust analysis.

Take-Aways

• A new paradigm, Complex Systems Thinking, is replacing conventional economic models in assessing complicated situations.
• Traditional economic models failed to predict the magnitude of the 2008 global financial crisis.
• Agent-Based Modeling provides an alternative economic tool that relies on a data cycle to assess how stakeholder decisions affect the overall economy.

Summary

A new paradigm, Complex Systems Thinking, is replacing conventional economic models in assessing complicated situations.

Three principles — “the utility function, belief systems, and equilibrium” — have formed the foundation of economic models and predictive analysis. According to this foundation, individuals make decisions based on boosting their utility within a structure that balances supply and demand with competing stakeholders. Economists have relied on these assumptions to construct economic frameworks designed to forecast future growth, inflation cycles, and fiscal and monetary policy.

“So conventional mainstream economics is built on the idea that agents individually, selfishly maximize their utility subject to their beliefs and they assume equilibrium and then they solve for everybody’s decisions and therefore arrive at what the economy is going to do.”

While the three-principle model remains useful in evaluating relatively simple scenarios, the template becomes less reliable as the complexity of situations and events increases. To deal with more complicated and difficult economic situations, academics are exploring Complex Systems Thinking as an alternative solution. Complex Systems Thinking focuses on outcomes that are radically different than the root variable inputs that underly the final “phenomenon.” A noneconomic example involves the evolution of human consciousness: It results from billions of neurons — which on their own are simple cells — connecting. Linked, they form an entirely new complex development, indistinguishable from its formative elements.

“Complex thinking spans a lot of different disciplines, and the core concept is that there are many phenomena in the world, emergent phenomena where the building blocks are fundamentally different than the phenomenon itself.”

Conventional economic models start to lose their efficacy in predictive analysis when “independent variables” roughly exceed 12 in number within a system. Rather than depending on utility, belief, and equilibrium as the rationales for individual decisions, Complex Systems Thinking relies on stakeholders making choices by surveying the entire system, while assessing “the information at hand, using simple cognitive rules.”

Traditional economic models failed to predict the magnitude of the 2008 global financial crisis.

The 2008 Great Recession provides a compelling example of the dangers of conventional economic forecasting. An exogenous event did not initiate the contagion that swept through the international economy. Rather, the collapse resulted from a complex and interconnected banking system.

“It was the financial system itself that generated the crisis…virtually no one predicted that the results would be so huge or even close.”

Even the Federal Reserve underestimated the catastrophic effects on the global economy due to the banking and housing crash “by a factor of 20 times.”

Agent-Based Modeling provides an alternative economic tool that relies on a data cycle to assess how stakeholder decisions affect the overall economy.

The Complex Systems Thinking architecture relies on the principles of Agent-Based Modeling. This uses simulations of stakeholder decisions and how those choices affect the economy as a whole. Each simulation produces data, which economists feed back into additional simulations that ultimately deliver a forward-looking predictive assessment.

“And then you run a simulation where you put in some information, the agents make decisions, the agent’s decisions affect the economy, that generates more information, and then you repeat the loop.”

Experts used the Agent-Based Modeling formula to accurately predict the effect of COVID-19 lockdowns on GDP in the United Kingdom. The model predicted a decline of 21.5%, and the actual tally was 22.1%. Conventional models proved far more inaccurate in their analysis.

“Companies model their warehousing inventory systems. Southwest Airlines models their flight system that way. So it’s [Agent-Based Modeling] widespread outside of economics.”

Businesses can deploy Agent-Based Modeling to more precisely predict sales, inventories, volumes, and myriad other metrics inside the ecosystem and value chain of a company and an industry. With continued advances in computational power, AI, and machine learning, businesses, government agencies, and institutions will turn increasingly to Complex Systems Thinking and Agent-Based Modeling as the most effective tools for predictive forecasting.

About the Podcast

Doyne Farmer is a complex systems scientist at the University of Oxford and the Santa Fe Institute. Martin Reeves is chair of the BCG Henderson Institute.