Microfounded Tax Revenue Forecast Model with Heterogeneous Population and Genetic Algorithm Approach

The ability of governments to accurately forecast tax revenues is essential for the successful implementation of fiscal programs. However, forecasting state government tax revenues using only aggregate economic variables is subject to Lucas’s critique, which is left not fully answered as classical methods do not consider the complex feedback dynamics between heterogeneous consumers, businesses, and the government. In this study we present an agent-based model with a heterogeneous population and genetic algorithm-based decision-making to model and simulate an economy with taxation policy dynamics. The model focuses on assessing state tax revenues obtained from regions or cities within countries while introducing consumers and businesses, each with unique attributes and a decision-making mechanism driven by an adaptive genetic algorithm. We demonstrate the efficacy of the proposed method on a small village, resulting in a mean relative error of $$5.44\% \pm 2.45\%$$ 5.44 % ± 2.45 % from the recorded taxes over 4 years and $$4.08\% \pm 1.21$$ 4.08 % ± 1.21 for the following year’s assessment. Moreover, we demonstrate the model’s ability to evaluate the effect of different taxation policies on economic activity and tax revenues.