[Paper Review] Economics need a scientific revolution

Jean‐Philippe Bouchaud|RePEc: Research Papers in Economics|Oct 29, 2008

Complex Systems and Time Series Analysis4 citations

TL;DR

The paper argues that economics must undergo a scientific revolution by abandoning dogmatic axioms like market efficiency and rational agents, shifting focus to empirical data, order-of-magnitude reasoning, and complexity-based models—particularly those from physics—to better understand and predict systemic financial crises, which classical models fail to anticipate due to flawed assumptions like Gaussian distributions of returns.

ABSTRACT

I argue that the current financial crisis highlights the crucial need of a change of mindset in economics and financial engineering, that should move away from dogmatic axioms and focus more on data, orders of magnitudes, and plausible, albeit non rigorous, arguments.

Motivation & Objective

• To challenge the dominance of untestable axioms in economics, such as rational agents and market efficiency, which undermine empirical validity.
• To highlight the failure of mainstream economic models—like Black-Scholes—in predicting or preventing major financial crises, such as the 1987 crash and the 2008 credit crunch.
• To advocate for a paradigm shift in economics toward methods inspired by physics, emphasizing empirical data, robustness to extreme events, and systemic complexity.
• To call for reform in economic education and financial regulation, including crash testing financial innovations and independent oversight.
• To position behavioral economics and econophysics as essential but underappreciated fields that offer more realistic models of market behavior.

Proposed method

• Replace axiomatic models with data-driven analysis, prioritizing empirical observation over theoretical elegance.
• Adopt complexity theory from physics to model how small perturbations can trigger large-scale market instabilities.
• Use statistical physics concepts—such as non-Gaussian distributions and fat tails—to better represent extreme market events.
• Integrate insights from econophysics and behavioral economics to model collective irrationality, herding, and systemic risk.
• Develop regulatory frameworks that subject financial innovations to stress testing under extreme scenarios, akin to aerospace or nuclear safety protocols.
• Encourage interdisciplinary training in natural sciences within economics curricula to counteract 'Cargo Cult Science' in financial modeling.

Experimental results

Research questions

• RQ1 Why have classical economic models failed to predict or prevent major financial crises despite their mathematical sophistication?
• RQ2 How do axiomatic assumptions like market efficiency and rational behavior impede the scientific progress of economics?
• RQ3 To what extent can concepts from statistical physics and complexity theory improve the modeling of financial markets?
• RQ4 Why do financial models based on Gaussian distributions systematically underestimate the risk of systemic crashes?
• RQ5 How can economic education and regulation be reformed to prioritize empirical data and systemic stability over theoretical dogma?

Key findings

• The Black-Scholes model contributed to the 1987 crash not by predicting it, but by assuming negligible probability of extreme market moves, leading to destabilizing hedging strategies.
• The 2008 financial crisis was exacerbated by models that underestimated the correlation of simultaneous defaults in sub-prime mortgages, ignoring systemic risk.
• Markets are not self-correcting or efficient; they exhibit collective irrationality, herding, and fragility to small shocks, akin to complex systems in physics.
• Classical economics lacks a framework to understand 'wild markets'—a critical gap given that such phenomena are obvious to non-experts.
• The use of models based on incorrect axioms, even when mathematically elegant, leads to real-world financial devastation, as seen in the proliferation of flawed structured products.
• A shift toward pragmatic, data-focused, and complexity-aware models—inspired by physics—could improve crisis prediction and regulatory oversight.