Agent-based modelling and complexity economics
Editor: Mauro Gallegati, Università Politecnica delle Marche, Ancona, Italy, Alberto Russo, Università Politecnica delle Marche, Ancona, Italy
The recent crisis has brought out the weakness of mainstream (NK-DSGE) models that were not able to anticipate it nor to understand its underlying causes. Admittedly, this is not a fair assessment in that even the possibility of an extended crisis has been expelled from dominant models, perhaps based on the idea that the “central problem of depression-prevention has been solved” as maintained by Lucas in 2003. And it is impossible to study the characteristics of a large crisis if this event is not considered in the model by construction. The big mistake then was to think that the “state of macro” was good, as assessed by Blanchard in 2008. However, once the damages of the crisis became undeniable many modifications have been introduced in the dominant paradigm, especially financial frictions of various species, though these may resemble the epicycles of a Ptolemaic system. Now, does economics need a Copernican-like revolution? The answer that the majority of economists would provide is negative, being the consensus that present models can be conveniently improved to include, for instance, heterogeneous agents and income/wealth distribution dynamics. We will see if this incremental effort can be the right answer to the difficulties of economists in anticipating and managing large crises.
A radical alternative is based on agent-based modelling (ABM), a methodology addressed to study the economy as a complex adaptive system in which collective behaviour emerges from the interaction among heterogeneous boundedly rational agents. As also partly recognized by the mainstream, agents’ heterogeneity is an essential feature of the economic system in which individual variables (for instance, personal income or firms’ revenues) are often distributed as power laws. The other fundamental ingredient is interaction, not only that mediated by the price system (that is indirect interaction), but the direct connection in a network of agents’ interlinkages, as in the case of financial contagion and bankruptcy chains.
The last years have seen a flourishing of ABMs which demonstrate that this approach is growing and it is considered as a valid alternative by many and many researchers around the world (including some central banks, as for instance the BoE). ABMs of various types have already demonstrated that this approach is able to reproduce complex dynamics, as technological evolution or business cycles, based on (relatively) simple individual behavioural rules (a first attempt was the Schelling’s segregation model). The integration between ABM and the Stock-Flow Consistent (SFC) modelling has further improved macroeconomic models, highlighting the role of underlying economic theories, with a particular emphasis on the Post-Keynesian approach. The recent attempts to study policy issues and environmental sustainability enrich this approach and contribute to build an effective alternative to mainstream models.
This special issue is aimed at collecting recent contributions in the ABM field, focussed on various topics, thus making an additional step towards a concrete approach to study the economy as a complex system and the role of policy in managing its evolution.
no papers yet