This presentation reviews three sets of agent-based economic / finance models developed by Ian Wright, Pascal Seppecher & Isabelle Salle, and Geoff Willis.
These models were developed independently, and from different roots. Ian Wright’s work is founded in statistical-mechanics. The work of Pascal Seppecher and Isabelle Salle comes from the post-Keynesian stock-flow tradition of Wynne Godley. Geoff Willis’s work is based in econophysics.
The models have in common a demand driven, dynamic, stock-flow consistent approach to economic modelling with relatively simple interactions between the agents.
Despite their differences, the models share a wide range of outputs in common. More importantly, these outputs closely match those seen in the real world.
These outputs include power law distributions for wealth, income and company sizes.
All the models show endogenous cycles of boom and recession.
By including simple behavioural decisions, the models of Seppecher & Salle simultaneously produce short term business cycles in parallel with much longer term cycles of financial crises.
In parallel to this, the models of Willis give a clear link between increased levels of consumer debt and increased financial fragility.
It is of particular note that all three models produce an emergent phenomenon of a stable ratio of macroeconomic shares of returns to labour and capital – one of Kaldor’s unexplained stylised facts. In all three models these stable macroeconomic shares are invariant to technology shocks.
The models of Willis explain the origin of this macroeconomic split with a simple algebraic derivation. They also give a driver from macroeconomic shares to micro level changes in inequality.
All three modelling approaches have their strengths and weaknesses, but taken together they suggest a new route forward with simple ABM models that can effectively model real world economies with a meaningful financial sector.