I have several research interests at the intersection of economics and the interdisciplinary science of complex systems, spanning the theoretical-empirical spectrum. Taking a complex systems approach to economics can mean multiple things. Broadly speaking, it is a focus on networks, heterogeneity, non-linear dynamics and bounded rationality, which have now become popular topics in economics. Narrowly speaking, it means building economic models from the bottom-up, without assuming equilibrium, such as Agent-Based Models (ABMs). In this latter sense, the complexity economics approach is still controversial and considered “heterodox” by the majority of economists. I think economics would greatly benefit from embracing non-equilibrium, complex systems models. So, as one of my main research goals, I aim to provide evidence of the usefulness of this approach.
On a purely theoretical level, this means investigating when the equilibrium assumption is reasonable. In the first two chapters of my PhD thesis, I investigate when equilibrium is a valid assumption in games: Do players who learn from repeatedly playing a game converge to equilibrium, or do they chaotically keep changing their behavior? Our results suggest that equilibrium may be an unrealistic assumption in more situations than what might be thought. This provides a theoretical foundation for non-equilibrium models.
On a more applied level, non-equilibrium models can be useful first of all qualitatively. By qualitatively, I mean to mathematically formalize and explain a theoretical mechanism. Many economists would say that for these purposes the standard assumptions of economic theory — optimization and equilibrium — are ideal. Yet, there are situations in which non-equilibrium models provide simpler narratives and allow to more easily incorporate realistic assumptions. For example, we show that an ABM of the housing market accounts with a simple narrative for a wide range of phenomena related to income segregation, inequality and house prices, and that incorporating income heterogeneity is crucial in policy exercises. Spatial equilibrium models have more difficulties in accounting for these phenomena and dealing with heterogeneity.
However, where I think ABMs really have an edge is in out-of-sample forecasting. Being the assumptions in ABMs closer to reality, the ABM should capture more signal of real economic processes, and so do a better job in quantitative prediction. This is where I want to bring my research in the future: comparing the out-of-sample predictive power of equilibrium models and ABMs. This does not mean that I think it is possible to predict economies far in the future with ABMs! I rather view out-of-sample prediction as an overfitting-free measure of realism. If it was possible to show explicitly that ABMs are, say, ten times more realistic than equilibrium models, policy makers would rather use ABMs for counterfactual analysis.
The success of ABMs in forecasting will crucially depend on the quality and granularity of the data. While equilibrium models might be ideal with small datasets, ABMs should be able to leverage big datasets. My expertise in big data comes from my work on housing market. In collaboration with researchers at the Bank of Italy, we analyzed a massive dataset of housing ads, showing that clicks on ads are a good proxy of demand and that big housing data have institutional applications. I am also interested in the growing cross-fertilization between machine learning and econometrics, in particular on whether and how the amount of data could help understand causality.
I am a James S. Mc Donnell Foundation Postdoctoral Fellow at the Institute of Economics and EMbeDS at the Sant’Anna School of Advanced Studies. I completed my DPhil (PhD) at the Mathematical Institute of the University of Oxford, under the supervision of Professor Doyne Farmer, with an affiliation with the Complexity Economics group of INET Oxford. Prior to starting my PhD, I got a BSc in Physics and a MSc in Physics of Complex Systems from the University of Turin. At the same time, I studied Economics at the Scuola di Studi Superiori and Collegio Carlo Alberto in Turin. I prepared my master thesis at the Centre d’analyse et de mathematique sociales (CAMS) in Paris, under the supervision of Professors Jean-Pierre Nadal, Pietro Terna and Annick Vignes.
For more information about my academic background and expertise, have a look at my CV (updated 15/04/2020).