Individual behaviors, interactions and collective efficiency. What economists can learn from network science and complex-system theory.

Mainstream economics has always employs simple, linear models to study how prices emerge and scarce resources are allocated in societies. In standard economic models, agents are assumed to behave as fully rational individuals who do not interact among them. Furthermore, macro-economic models typically suppose that the average behavior of the economy can be nicely proxied by the average of the behaviors of the individuals that populate societies, I.e. that to study how economies evolve is sufficient to understand how the “representative” firm, consumer, bank, behave. This provides a simple-enough mathematical model that allows for nice static equilibria where markets are shown to work smoothly and efficiently. Only shocks exogenously hitting to the system can displace such an idyllic state, but standard economic theory suggests that in general the economy should go shortly back to another nice equilibrium state. 

The Global Financial Crisis (GFC, 2007-2008) has put in a deep crisis the mainstream theoretical approach to the study of economic systems. In this respect, the GFC was not only an economic crisis, but also a crisis for economists, as the profession was essentially not able to spot early-warning signals, predict the crisis, understand its deep causes, and design appropriate policy instruments to prevent global crisis in the future.

In these lectures, I will discuss how network science and complex-system theory can provide an alternative and more effective theoretical framework for economists. This new approach, knows as Agent-Based Computational Economics (ACE) is based on the idea that economies are evolving complex systems, populated by boundedly-rational agents that interact through complex networks . Aggregate economic dynamics results from aggregation of micro-economic behaviors and must be interpreted as a sequence of meta-stable states, endogenously emerging from micro-level behaviors and interactions. Employing a number of real-world examples, I will discuss how ACE models are able to describe and explain our economies much better than mainstream models.