Future methods of political economy: from Hicks’ equation systems to evolutionary macroeconomic simulation

Traditional macroeconomics and agent-based simulation (ABS) seem to be two disjunctive worlds, two different sprachspiele in the sense of Wittgenstein. It is not just the fact that macroeconomics has a long and distinguished history that on top of more than 200 years of discourse has recently adopted a sophisticated dynamic mathematical framework, while ABS is still in its infancy and for outsiders looks more like an intellectual toy than a serious research tool. Both languages are tools and eventually both are aiming at the same object of investigation: political economy. Why they let their object appear differently certainly is due to the intrinsic properties of the two languages. As is the case with every tool, the properties of the tool are to some extent transferred to the results that can be achieved with the respective tool. What aggravates this split of work styles is the fact that two different large research communities are linked to the use of the two languages; and each member of such a community has built already a considerable human capital stock, which consists mainly of elements that belong to exactly one of the two languages. Any expedition into the use of the foreign language runs into danger to make a part of the own toolset look obsolete, and thus to lose hard earned human capital. The incentives for cooperation disappear. To ease the pains of disaggregated research, the aim of this paper is to improve mutual understanding, and to show how far evolutionary macroeconomic simulation can advance political economy by explaining traditional macroeconomics as a (sometimes implausible) special case of its own more general approach. On the other hand ABS researchers often are unaware of the rich interpretative and empirically oriented treasures that classical macroeconomics has in store. What at first sight looks to be easily transferred into an algorithm turns out to be a highly refined argument, which in turn challenges the skills of ABS modelers. The most promising route to follow in the future certainly will be to be versatile in both languages, to walk on both feet. This short paper should provide a modest first step towards this goal.