Exploring Income and Wealth Distributions using Agent-Based Modelling: How does Taxation Impact Economic Inequality?

Abstract

Agent-Based Modeling is a rapidly growing field with numerous economic applications, mainly due to its ability to create heterogeneous populations and grow collective structures from the bottom-up, a contrast to mainstream mathematical economics models. “Sugarscape” is one of the most famous agent-based models — theorized by Joshua M. Epstein and Robert Axtell in their book “Growing Artificial Societies”, the model features a 51x51 grid where heterogeneous agents with varying visions and metabolisms move around in hopes of obtaining the most “sugar”. This research attempts to build on the Sugarscape 3 Wealth Distribution model in the Netlogo Models Library, by adding new components to the original model such as taxation, lending with interest rates, trade, and reproduction, studying their impacts on wealth and income inequality within the distribution as well as their general prosperity. This research finds that a flat wealth tax of about 15% is optimal for maximizing marginal tax revenue, and also finds an appropriately linear association between the flat wealth tax rate and the Gini index. Furthermore, the development of four distinct social classes of lenders, borrowers, lenders & borrowers, and agents who are neither is observed after the addition of lending with interest. Limitations of the model include a small carrying capacity which makes validation of the model with real-world data difficult. This research could be expanded on by examining how adding a wage labor system with collective bargaining or different methods of resource replenishment affects inequality.