Accounting State Space as the Minimal Unit for Economic Agent-Based Modeling: Advancing Ripple Effect Analysis in Real-Time Economy

The emergence of Real-Time Economy has highlighted the critical importance of accounting data as the fundamental unit of economic activity. In economic Agent-Based Modeling (ABM), accounting state space represents the minimal unit for capturing economic transactions and their ripple effects. This paper develops an Accounting-State Agent-Based Simulation (AS-ABM) that embeds inventory levels and input constraints into an IO-consistent production network, where agents maintain bookkeeping state spaces and interact in stage-oriented time. We demonstrate that accounting state space serves as the essential foundation for realistic economic modeling, enabling the analysis of dynamic ripple effects with stock-out behavior and replenishment lags. Through three simulations—(i) the Leontief inverse benchmark, (ii) an ABM reproducing the benchmark, and (iii) an ABM with input constraints—we show that stock-out policies govern amplification versus damping of ripple effects. The model reproduces key stylized facts on variability and generates Kitchin-type inventory cycles via endogenous delays, providing a mechanism-based explanation for gaps between IO predictions and observed outcomes in Real-Time Economy contexts.