An Ontological Model for Accounting Information Systems

Jian Guan, Andrew Cobb, Alan S. LevitanJournal of the Association for Information Systems 2006

Abstract

This research proposes a rigorous ontological model for accounting information systems (AIS) by mapping the Resource-Event-Agent (REA) model onto the Bunge-Wand-Weber (BWW) ontological framework. While REA is a well-established domain ontology for AIS, it suffers from representational deficiencies, including construct overload, a lack of dynamic modeling, and an inability to handle multiple levels of abstraction. By applying BWW constructs and utilizing UML as a modeling grammar, this research addresses these limitations to provide a more precise and complete foundation for AIS design.

Introduction

Accounting Information Systems (AIS) are critical for managing complex economic activities, yet the models used to design them often lack formal ontological precision. The Resource-Event-Agent (REA) model has long served as the primary framework for capturing economic activities. However, as AIS grow in complexity, the limitations of REA—particularly its inability to represent complex business rules and system dynamics—become more apparent. This paper proposes a theoretical overhaul of REA using the Bunge-Wand-Weber (BWW) ontological model to create a more robust foundation for AIS.

The REA Model and its Limitations

The REA model, illustrated in Figure 1, focuses on the core primitives of economic resources, events, and agents.

The traditional Resource-Event-Agent (REA) model for accounting information systems.

Despite its widespread use, REA faces several "ontological bugs" when treated as a formal domain ontology:

  1. Construct Overload: Different concepts (e.g., an asset vs. a business process) are often represented by the same modeling construct (an "Entity"), creating ambiguity.
  2. Construct Deficit: REA lacks formal mechanisms to represent accounting rules and constraints within the model itself, often requiring separate text documentation.
  3. Static Nature: The model excels at showing structure but struggles to capture the behavior or state changes of a system over time.
  4. Lack of Decomposition: It is difficult to manage different levels of detail or "zoom" within a standard REA model, which is necessary for complex enterprise systems.

Proposed Methodology: Mapping REA to BWW

To resolve these issues, the authors propose mapping REA primitives to the Bunge-Wand-Weber (BWW) ontological framework. BWW is a highly rigorous framework designed to ensure that information system models accurately reflect real-world phenomena.

The methodology involves:

  • Mapping Primitives: Redefining Economic Resources, Events, and Agents using BWW constructs to ensure ontological clarity.
  • Grammar Selection: Adopting the Unified Modeling Language (UML) as the visual grammar. UML is better suited than traditional Entity-Relationship diagrams for handling system decomposition and behavioral modeling.
  • Metamodel Creation: Developing a formal metamodel that bakes domain semantics, rules, and dynamics directly into the model's structure.

Key Insights

A primary insight of this research is that ontological clarity is essential for preventing implementation errors. For instance, treating a "Cash Receipt" (an event) identically to "Cash" (a resource) ignores the fact that events represent immutable historical facts. Furthermore, the BWW-based approach integrates structure and behavior, allowing the model to reflect how economic events cause state changes in resources. Finally, hierarchical decomposition allows designers to define high-level accounting procedures that propagate consistently down to lower-level enterprise system details.

Conclusion and Future Work

This research identifies specific modeling inadequacies in the current REA framework and provides a roadmap for a new AIS domain ontology. While the full metamodel is still under development, the work establishes UML as a necessary grammar for representing complex ontological relationships in accounting. Future work will focus on formalizing the metamodel to the point where it can automatically generate accounting conclusions, such as revenue or profit calculations, and validating the model against complex real-world scenarios.

Summary updated

Abstract

This research proposes an ontological model for accounting information systems (AIS) based on the Bunge-Wand-Weber (BWW) modeling constructs.The proposed model is derived from the Resource-Event-Agent (REA) accounting model.REA is a well-known model for AIS and has recently been proposed as an AIS domain ontology.REA is grounded on event-based accounting theories and its sound underlying concepts make it a strong candidate as a domain ontology.However, there are several deficiencies in the current representational scheme used in REA that limit its value as a domain ontology.First the use of entity relationship modeling constructs has resulted in a lack of precision and completeness in the current model.The current model also fails to capture the dynamic aspects of the domain.Finally the current model lacks the ability to represent the different levels of abstraction required to capture the requirements of as complex a domain as accounting information systems.This research argues that the above deficiencies can be addressed through application of the BWW-based modeling constructs and presents an ontological model for the domain of AIS where the above mentioned deficiencies are addressed.
BibTeX
@article{guan2006ontological,
  author = {Jian Guan and Andrew Cobb and Alan S. Levitan},
  title = {{An Ontological Model for Accounting Information Systems}},
  journal = {Journal of the Association for Information Systems},
  year = {2006},
}
Figure