The effects of labour market reforms upon unemployment and income inequalities: an agent-based model

Agent-based model calibration using machine learning surrogates

“…Our machine-learning surrogate approach is different from kriging, which has been recently applied to ABMs dealing with industrial dynamics (Salle and Yildizoglu, 2014;Dosi et al, 2017c), financial networks (Bargigli et al, 2016) and macroeconomic issues (Dosi et al, 2016(Dosi et al, , 2017b. In particular, apart from the different statistical framework kriging relies on (it assumes a multivariate Gaussian process), the results it delivers once applied to ABMs may suffer from three relevant limitations.…”

“…This constrains the modeller to introduce additional procedures to select, a priori, the subset of parameters to study, while leaving the rest constant (see e.g. Dosi et al, 2016). Second, the machine-learning surrogate approach performs better in out-of-sample testing: the typical kriging-based meta-model is tested on 10-20 points within an extremely large space, while our surrogate is tested on samples with size 10000 in the first set of exercises and 1000000 points in the last exercise.…”

Between Scylla and Charybdis: Income Distribution, Consumer Credit, and Business Cycles

“…For example, Fischer (2012) builds a simple model with heterogeneous households showing that increasing inequality leads to greater financial volatility stemming from the accumulation of net worth by households at the top of the income distribution. These have been investigated in the literature on macroeconomic agent-based models and complex systems, for example, by Dosi et al (2016), Harting, Dawid, and Neugart (2015), and Markey-Towler and Foster (2013). Similarly, Caiani, Russo, and Gallegati (2018) show that more progressive tax systems, as well as measures supporting the wages of low-income workers, counteract inequality, and promote economic development.…”

“…Note that we are concerned with the consequences of rising inequality, which we take as exogenous, rather than its possible causes. These have been investigated in the literature on macroeconomic agent-based models and complex systems, for example, by Dosi et al (2016), Harting, Dawid, and Neugart (2015), and Markey-Towler and Foster (2013). peer effects in consumption, and the housing market.…”

“…This is explained by the rise of the household default rate which leaves a greater fraction of households out of the credit market. One of the major advantages of heterogeneous agent models-from more traditional versions, such as Aiyagari (1994) and, more recently, Kaplan, Moll, and Violante (2018), to the agentbased models (e.g., Cardaci 2018;Dosi et al 2016;Russo, Riccetti, and Gallegati 2016) that are related to this paper-is that they allow to analyze the distribution of key economic variables among heterogeneous individuals, which is often useful in order to shed some light on the microeconomic dynamics behind aggregate changes. In our case, it is interesting to investigate the change in the distribution of desired and actual consumption, in order to have a more comprehensive picture of the factors leading to falling GDP.…”