The Reverse Engineering of Economic Systems. Tools and Methodology

We will start* highlighting the methodological necessity to use reverse engineering for economic system when there is a strong relation between the whole and the part (Kaneko, 1998). In "complex systems" when there is a complementarity between the whole and parts, we need some logic about why rules appear for a class of economic systems. This will permit a better understanding of the ways in which agents interact, especially in a dynamic socio-economic context. A large part of studies focus on how rules give rise to dynamic behaviour, while in complex system studies, we have to study how rules form from dynamic behaviour. In searching for a universal class of behaviours for reconstructing the real economic behaviour, there are very few studies in this direction. We will use a "constructive economics" approach trying the emergence of syntactic rule from complex behaviour (Langton 1995). We will start considering what meaning Reverse Engineering have for Economics on the base of features of economic systems. This meaning impose that for the adequacy of Top-down and Bottom-up a new approach Chaotic Itinerancy (Kaneko Tsuda 1994) for integrating them will be suggested. Applying RE approach to economics, we need two kinds of tools. Some aggregated time series analysis tools for which we will propose Recurrence Analysis for its ability to work with few data and to detect chaos and chaotic itinerancy (Itoh M., Kimoto M., 1997). A simulation tool that will start from this meta-analysis for obtaining the results that we considered meaningful. Of course, the here proposed RE must make use of a simulation tool that will reproduce economic time series of production, consumption, work as results, starting from the economic choices and behaviour of agents. A new simulation model based on agents is proposed. This model introduces a micro-foundation macroeconomic approach in a model on Automated trading (Gulyás 2001). In the Appendix there are some information about the simulation model.