Accounting State Space as a Foundation for Economic Agent-Based Modeling: An Input–Output Benchmark and Inventory Extension
This paper introduces Accounting-State Agent-Based Modeling (AS-ABM), a framework that enforces double-entry bookkeeping consistency at the transaction level to successfully replicate static Leontief input-output benchmarks while dynamically simulating how inventory constraints and replenishment rules influence economic multipliers and fluctuation patterns.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Economics has long relied on a simplified view of how money and goods move through society. Traditional models often treat the economy as a collection of broad averages, where factories instantly produce whatever is needed and supply chains stretch infinitely without ever running out of materials. These frameworks are useful for seeing the big picture, but they struggle to explain what happens when things go wrong, such as when a sudden shortage of a single part halts production across an entire industry. In the real world, businesses do not operate on averages; they operate on ledgers. Every transaction is a specific record of what was bought, what was sold, and what remains in stock, governed by strict rules that ensure every debit has a matching credit. For decades, economists have lacked a way to simulate these detailed, rule-bound records across thousands of interacting companies to see how a small shock might ripple through the system.
A researcher at Chiba University of Commerce has developed a new way to build economic simulations that starts with these accounting records rather than abstract averages. Instead of guessing how a factory might react to a change in demand, this new approach, called Accounting-State Agent-Based Modeling, gives each simulated company its own digital ledger. In this system, every time a company buys or sells something, the model updates its specific books, tracking exactly how much cash it has, how much inventory it holds, and what it owes. The computer follows the same double-entry rules that human accountants use, ensuring that the simulation never violates the fundamental logic of business. By building the economy from the bottom up, using these precise accounting steps, the researcher created a laboratory where the complex dance of supply and demand can be watched in real time, revealing how the rigid rules of bookkeeping shape the flow of the entire economy.
To test this new method, the researcher first built a simple version of the economy using a standard blueprint known as an input-output table. This blueprint maps out how different industries depend on one another, showing which sectors supply raw materials to others. The goal was to see if the new accounting-based simulation could reproduce a well-known economic result called the multiplier effect. In traditional economics, this effect describes how a single dollar of new spending can generate more than a dollar of total economic activity as it passes from one business to another. The simulation was set up so that a sudden burst of demand hit one industry, and the computer tracked how that demand traveled through the network of suppliers. The results were exact: the accounting-based model produced the same multiplier effect as the traditional static formulas. This proved that the new method was not just a different way of drawing the map, but a working engine that could replicate established economic truths while keeping a record of every single transaction.
The real power of the new model, however, emerged when the researcher added a layer of reality that traditional models often ignore: inventory limits. In the real world, a factory cannot produce goods if it has run out of the raw materials it needs, and it cannot sell what it does not have in stock. The researcher introduced these constraints into the simulation, forcing companies to manage their safety stocks and deal with shortages. When a sudden demand shock hit a company with tight inventory limits, the results changed dramatically. The flow of goods did not spread smoothly as the traditional formulas predicted. Instead, the shortage of materials at one stage caused production to stall, which in turn reduced the orders placed with suppliers further up the chain. The multiplier effect, which had been strong in the unconstrained version, shrank significantly or even disappeared entirely when the supply chain was tight. The simulation showed that the ability of an economy to absorb a shock depends heavily on how much buffer stock companies keep on hand.
The study also explored how different inventory policies affected the stability of the system. Companies that kept larger safety stocks were able to maintain production even when demand spiked, allowing the economic multiplier to recover toward its normal strength. However, the rules these companies used to reorder supplies created a different kind of problem. When companies tried to keep their stock levels perfectly balanced, the simulation showed that they sometimes overreacted to small changes, placing large orders that led to a cycle of overproduction followed by a sudden drop in demand. This behavior, known in supply chains as the bullwhip effect, caused wild swings in production volumes. The researcher tested this across nearly one hundred different random economic structures to see if these swings were a universal feature. While the swings were visible in many cases, the study found no statistical proof that they became systematically worse as the economy grew larger. The fluctuations were highly dependent on the specific details of each simulated economy, suggesting that there is no single rule that guarantees stability or chaos for all systems.
Finally, the researcher checked if the simulation behaved like real-world economies by comparing its output to known patterns observed in business data. One such pattern is that the amount of goods a factory produces tends to vary more than the amount it actually sells, because factories try to smooth out their production schedules. Another pattern is that the inventory of raw materials tends to be more volatile than the inventory of finished goods. The new model successfully reproduced these two specific patterns, confirming that the accounting rules it used were capturing the right kind of economic behavior. However, the model did not reproduce a third pattern regarding how much companies order compared to how much they use, a difference the researcher attributed to the specific way the simulation handled the timing of deliveries versus orders. This partial match suggests that while the model captures the core mechanics of inventory management, the specific rules for ordering in the real world are more complex than the current simulation.
The findings suggest that the rigid rules of accounting are not just a way to keep score, but a fundamental driver of how economic shocks travel through society. By treating the economy as a network of ledgers rather than a set of smooth curves, the new approach reveals that the ability of an economy to grow in response to new demand is limited by the physical reality of inventory and supply constraints. The study does not claim to have solved the mystery of economic fluctuations, but it provides a new tool that allows researchers to see exactly how a shortage in one warehouse can ripple through a factory, a supplier, and a bank. As digital records of transactions become more available in the real world, this method of building economic models from the ground up, using the actual language of business, offers a promising path to understanding the dynamic and often fragile nature of modern supply chains.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.