Evolvable accounting information systems: applying design science methodology and normalized systems theory to tackle combinatorial effects of multiple GAAP
21 September 2016
Promotiezaal Grauwzusters - Lange Sint-Annastraat 7 - 2000 Antwerp (route: UAntwerpen, Stadscampus
Prof Jan Verelst
PhD defence Els Vanhoof - Faculty of Applied Economics
Disclosure requirements are a primary regulatory issue for companies. Companies need to externally report financial information for different purposes and according to multiple formats (e.g., taxation, statutory filing, investor relations). In this regard, companies can be obliged to report according to more than one set of Generally Accepted Accounting Principles (GAAP). Moreover, both GAAP and regulatory environment tend to change over time. Companies use Accounting Information Systems (AIS) to facilitate their reporting requirements and these need to cope with changes in/of GAAP. Therefore, evolvability of the AIS, the “ease of further development”, is a major characteristic of interest for AIS management. In this dissertation we study the evolvability of multiple GAAP AIS by using a mixed method approach of design science and case studies. The research is conducted in three phases (problem identification, design and construct).
In the problem identification phase, we define and circumscribe the problem domain: processing of accounting-related events according to multiple GAAP. The lack of literature about evolvability and multiple GAAP makes our work largely exploratory. We conduct five case studies to study multiple GAAP AIS in practice and document the designs of the case-specific AIS. Next, we use Normalized Systems Theory to evaluate each of the case designs with regard to evolvability. The result of the first research phase is the identification of combinatorial effects (violations of evolvability). In the design phase, we use the documented combinatorial effects in order to develop design principles. We theoretically evaluate the design principles by relating them to prior literature and to the case studies. Finally, in the construct phase, we build a prototype that serves as a proof-of-concept for the design principles. We use Normalized Systems Theory to evaluate the prototype with regard to evolvability.
We elaborate and document that evolvability is a major issue in current multiple GAAP AIS and contribute to a solution for the problem by proposing design principles for more evolvable AIS. The prototype that we construct, shows the relevance of the design principles and the feasibility of building an evolvable multiple GAAP AIS according to the principles from Normalized Systems Theory.