Betwixt and Between? Bringing Information Systems and Accounting Systems Research Together

Abstract

In 2008, the Research and Publications Committee of the Information Systems Section of the American Accounting Association decided to sponsor a special issue of the Journal of Information Systems (JIS) entitled “Reviews of Information Systems Research.” The objective of the special issue is to “publish papers that review a stream of research in information systems (IS) broadly defined.” The Committee intended that submissions would review and integrate the IS (information systems) and AIS (accounting information systems) literatures and suggest future research directions in both disciplines. The special issue followed a previous valiant and groundbreaking effort in IS/AIS research integration for the IS section by Professors Vicky Arnold and Steve Sutton (Arnold and Sutton 2002).As editor of this special issue, I took a somewhat different approach to the task than is normal. First, rather than a regular call for papers, I requested researchers to submit extended abstracts. The objectives of this approach were to ensure that the scope of the proposed article was concomitant with the objective of the special issue and to identify any potential overlaps in subject matter. In this process, I was able to negotiate the amalgamation of several writing teams. I also ensured that where there was commonality in subject matter, the writing teams were introduced to each other and worked to manage the writing process. Second, I had clear views on how the papers should be structured. As an author of one of the chapters in the earlier monograph for the IS section, I was impressed with the systematic approach Dr. Arnold took to ensuring a common approach in the structure of the contributions and the discipline exercised in ensuring that the goals of the monograph were achieved. It is simpler to achieve a common approach in a monograph than it is in separate papers in JIS. My ambition was, then, to strongly suggest directions to authors but not to mandate a single approach. As a consumer of many literature reviews, I realize how easy it is to maroon readers in a Sargasso Sea, not knowing how to navigate their way. Readers need clear navigational markers and a sense of direction. Third, I saw the review process as a mutual exercise among writing teams, reviewers, and myself as editor. Given the scope of this exercise, I deliberately took a more active editorial role than is normal. These objectives probably added somewhat to the time taken for publication but did, I believe, improve the quality of the papers. While the original objective of the Research and Publications Committee was to publish a single forum within an issue of JIS, it quickly became apparent that this was not a feasible option given the differential speed in paper development, reviewing, and revision. The review papers were published in three issues of JIS.Now that this endeavor has concluded, we can take stock, consider the state of AIS research, and place the papers in the nexus between IS and accounting research. The nine review papers can be divided into three groups of research, viz.: public policy, organizational, and socio-technology. In the first group, Boritz and No (2011) and Kauffman et al. (2011) take different approaches to the study of privacy. Boritz and No (2011) explicitly consider privacy in an electronic commerce setting while Kauffman et al. (2011) address consumer privacy from a broader perspective. Along with others that use many tools from companies such as Google including search, maps, Gmail on personal computers, and different Android platforms, I must make conscious (or unconscious?) decisions on the trade-off between functionality and the need to ensure that private information is kept just that. Kauffman et al. (2011) consider a range of policy issues including alternative approaches to privacy around the globe and the choices made by corporations in the establishment of privacy policies. Boritz and No (2011) take advantage of their focus on e-commerce privacy to discuss targeted issues such as the role of privacy seals and other Internet technologies for privacy as well as the relationship of privacy to online consumer behavior. Both Kauffman et al. (2011) and Boritz and No (2011) assess the role of regulation and voluntary privacy codes. Masli et al. (2011) turn to a vital but difficult question: How do we determine if investments in information technology generate positive returns for the enterprise? We know that enterprises and agencies continue to invest heavily in information technology. Yet, evidence on the efficacy of that investment is mixed. Interest in determining the effect of IT investment on firm performance was boosted by Brynjolfsson (1993) whose seminal paper pointed to the “mystery” of a seeming paradox between significant spending on IT and productivity in the U.S. economy. Masli et al. (2011) attack this problem by coupling our rich tradition of accounting research into the relationship between key performance metrics and capital markets and IT situational variables.Turning to the enterprise as the unit of analysis, we have four tightly coupled papers that address IT governance (Wilkin and Chenhall 2010), enterprise resource planning system adoption and use (Grabski et al. 2011), outsourcing (Blaskovich and Mintchik 2011), and data and information quality (Neely and Cook 2011). Taken together, these papers address core concerns for the AIS research community. Outsourcing is a key value driver and source of both risk reduction and increased risk. IT governance, or more strictly correct, the enterprise governance of information technology, is a long-standing concern of the accounting and assurance community. The CobiT IT governance framework, for example, arose from the work practices of IT auditors and their need to standardize audit practices and improve underlying management practices at their organizations. Wilkin and Chenhall (2010) align their discussion of IT governance with areas of long-standing research in the IS domain including strategic alignment, risk management, resource management, performance measurement, and value delivery. Each of these areas has been of concern to the AIS community as well and, importantly, to sister sub-disciplines. For example, the balanced scorecard (BSC) framework is widely used and understood within the management accounting community. There are many opportunities for collaborative research with colleagues from IS and management accounting on BSC and other performance measurement systems. The IT governance research area is also productive as there are well-established frameworks that can provide the researcher with necessary structure and a lingua franca in which to communicate with board members, IT and operational managers, and auditors. These frameworks include CobiT, ValIT, RiskIT, and ISO/IEC 38500.Blaskovich and Mintchik (2011) address outsourcing from a variety of perspectives. Outsourcing has been an increasingly important feature of IT corporate environments over the last 20 years. The development of so-called cloud computing from its origins as Software as a Service (SaaS) in the 1990s through to its current rapid adoption should give additional impetus to the study of outsourcing. An interesting and important aspect of Blaskovich and Mintchik's work is the way that it explicitly links their discussion of outsourcing not just to accounting information systems, including information assurance, but also to other sub-disciplines, including management accounting. While we as researchers might be tempted to think that outsourcing is an area of research that has nearly run its full course, Blaskovich and Mintchik point to a number of significant research questions that are unanswered or where the results of earlier research are now challenged by changes in the technological and business environment. While outsourcing is a worthy area of research in its own right, the fundamental importance of outsourcing for the maintenance of internal control over financial reporting makes outsourcing of key concern for AIS researchers. While outsourcing is a source of risk and potential control issue within the context of internal control over financial reporting (ICFR), outsourcing can also bring immediate gains in process maturity and management of control. How do we build a software development team certified at CMMI Level 4 or 5? Take a few years and a few million dollars and invest a great deal of management time and effort in building the current coding shop. Alternatively, outsource to a specialist development contractor. Mapping and understanding these inherent tensions will present a number of interesting research projects for AIS researchers.While we are talking of research areas that would, at first glance, seem to be beyond their half-life, Grabski et al. (2011) confront the question of Enterprising Resource Planning (ERP) adoption and use. ERP investment received a considerable boost with the need for enterprises to replace legacy systems with new integrated systems in the face of the Y2K challenge of the middle-to-late 1990s. There was a rush of studies that covered adoption challenges of ERP that followed this considerable investment in enterprise systems. I believe that Grabski et al. (2011) conclusively demonstrate that the study of enterprise systems remains of vital importance for the AIS community and has the ability to positively impact on research in IS. Adoption by enterprises of ERP is for the long-term and has the potential to be organizationally transformative. Technology changes continue to impact the ERP environment in a variety of ways. For example, as Grabski et al. (2011) note, the requirement for enterprises to report performance data in eXtensible Business Reporting Language (XBRL) presents major challenges to ensure that there is an appropriately strong theoretical and practical foundation to any technical solutions to these externally mandated requirements. At the same time, these requirements have the potential to significantly improve the reach, reliability, and robustness of both internal and external reporting processes. I am sure the authors will not mind me saying that as editor I had considerable difficulty ensuring that this paper was kept to a reasonable length. This was not just because the existing literature is broad—it certainly is—but also because there were so many interesting future research areas it was difficult to prioritize and prune.Data and information quality (DIQ) is of central importance to all aspects of the accounting function. Neely and Cook (2011) take a systematic approach in their study of data and information quality. Accountants are both consumers of DIQ—we rely on high quality underlying transactional data and other enterprise information as the foundation for the creation of management and external reporting—and contributors to the level of DIQ within enterprises. Accountants both create and transform data and information. Further, as Neely and Cook (2011) note, DIQ concepts are core components of IT governance frameworks including CobiT. Neely and Cook (2011) point to several areas of research that are significantly under-researched. They identify a wide range of research opportunities in areas such as data governance, impact of operational processes on DIQ, and the effect of DIQ on decision making. There are clear links between the study of DIQ and other areas covered within the papers in JIS and elsewhere in the AIS domain. For example, Neely and Cook (2011) discuss the impact of DIQ on privacy. This is one of the major challenges for compliance officers, particularly in sensitive areas such as health care that is impacted by the explicit privacy data quality regulations under the Health Insurance Portability and Accountability Act (HIPAA) legislation. A second example is the impact of XBRL on DIQ. The relatively recent adoption of XBRL for financial reporting has given rise to a number of questions about the data quality of XBRL-tagged financial statements (Debreceny et al. 2010). As, in the Securities and Exchange Commission (SEC) context in the United States, we move to detailed tagging of notes and additional disclosures, there will be many more questions of data quality that are at the intersection of research in financial accounting standard setting, taxonomy design, AIS information generation, assurance, and XBRL. We are in a transitional stage at present, with XBRL reporting being bolted on to pre-existing modes of financial reporting with unknown impact on DIQ. As Trites (2011) in a recent study for the Canadian Institute of Chartered Accountants points out, as we move tagging more to the transactional or report level using XBRL GL (global ledger) or other semantic Web technologies, data quality should improve (Garbellotto 2010; ISACA/IFAC 2011; van Hilvoorde and Garbellotto 2007). How such a process is to be achieved and the effect of tagging further up the information value chain are largely un-researched and unknown. In a similar manner, explicitly including tags that represent the state of internal control as attributes to XBRL GL enabled information flows may radically change the nature of maintaining internal control.Finally, in this review of the papers published in JIS, we turn to two papers that cover different aspects of human interaction with systems, particularly accounting information systems. Dilla et al. (2010) focus on interactive data and information representations. Kelton et al. (2010) address the impact of more established information presentation formats such as multimedia, Chernoff faces, graphics, and tables on judgment and decision making. We can see these papers as a complete representation of the research questions that face the AIS research community when it comes to how decision makers interact with computerized systems or the output from such systems on paper and in other forms. We must be concerned by how the product of our information processing systems is used and the benefit and limitations of alternative modes of representation. This involves the intersection of technology, task, and decision maker. Dilla et al. (2010) and Kelton et al. (2010) each remind us, within their own area of concentration, of the importance not just of the technologies involved but also of the effect of cognitive limitations faced by decision makers as they interact with differing representations of information and of the importance of the tasks involved. We exist in a technological environment where vast amounts of data can be stored and made available to a wide array of devices anywhere the Internet is available—which is pretty much everywhere on the planet. Personal computers, laptops, tablets, and smartphones have very powerful graphical interfaces that allow decision makers to interact with the information in new and interesting ways. Yet, even the most casual empiricism around the accounting function of most organizations shows that a very high proportion of reports are produced by today's accountants using spreadsheets and presented in a tabular format. It will be important to not only understand how information is produced and presented in the accounting function, but also the incentives for preventing and barriers that prevent adoption of more productive communication modes.A principal objective of the special issue was to analyze research from both the IS and AIS research domains and to consider future research in both domains. The papers presented in the special issue provide a prism through which we can view the current state of AIS research. In this subsection, I review what that view tells us about AIS research and look to the areas of research not addressed in the special issue. I then compare and contrast the state of research in both the IS and AIS domains, as revealed by the debates that are underway in those domains.The nine papers published in the special issue encompass a high proportion of AIS research activity. Recall that this was not a command and control exercise. Rather there was an open call for papers, and the papers in the special issue are the result of a community exercise. Yet, there are some areas that are, in my view, clearly missing from this survey and should be addressed in future exercises. I now provide four examples of research areas that could warrant future literature reviews in JIS or elsewhere. First, information security has received only the scantiest attention from AIS researchers (Calderon et al. 2006; Chandra and Calderon 2003; Wallace et al. 2011). Yet, information security is an acute risk factor in any AIS environment. Security is a matter of concern to all those concerned with general IT controls. At the same time, security is closely linked to application controls that are fundamental to the reliability of accounting information systems (Anderson 2009; IT Governance Institute 2009).Second, an active area of research continues to be the impact of the Internet on accounting and business, sustainability reporting. Research on Internet financial reporting (IFR) continues to be active. Recently, more attention has been given to the concomitant study of XBRL, narrowly, and semantic representation of accounting and performance and compliance reporting, more broadly. The initial focus of this study is on the relatively short-term impacts on disclosure caused by compliance requirements. As I discuss above, pushing XBRL and other semantic tagging technologies up the information value chain has considerable potential for enhancement of the effectiveness and efficiency of accounting information systems, broadly defined. Yet, we have little evidence of the effects of such an integrative process. Equally, we know very little of the effect of pushing XBRL down the information value chain. How will stakeholders consume information in XBRL form? How will XBRL interoperate with other semantic representation standards such as SDMX, RDF, and OWL? This area of research is growing rapidly, and some structure and guidance are necessary.A third area of research that has had a long tradition in the AIS community is artificial intelligence (AI) and knowledge based systems (KBS). Unfortunately, interest in AI/KBS has slipped away in recent years. This diminution in attention to AI comes at a time when the power of even low-end computers now matches the resources required for even complex AI tasks. The range and breadth of data sources to feed AI tools dramatically exceed those available a decade ago. These data sources include corporate datamarts, email and document depositories, ERP systems, and broader Internet-based datasets. There are also many recent advances in AI technologies that have not yet informed research in the AIS domain. For example, progress has been rapid in the application of the Bayesian networks, that we are so familiar with in accounting and auditing, to AI (Darwiche 2010; Koller and Friedman 2009). Bayesian algorithms are used, for example, in a number of crowdsourcing solutions (Doan et al. 2011; Gloor 2006; Gloor and Cooper 2007). An interesting related area that applies Bayesian networks, is topic modeling (also sometimes referred to as opinion mining), where key matters of interest are extracted from large volumes of textual information, such as emails, legal cases, and news reports (Anthes 2010; Chen and Zimbra 2010). Even further into the future is the potential for automated agents to negotiate successfully with humans (Lin and Kraus 2010). Each of these techniques would seem to have considerable application in the accounting and auditing domain, particularly forensic auditing and continuous assurance, and deserve added attention.The fourth area of interest is continuous reporting and assurance (Alles et al. 2002; Gibbins and Pomeroy 2007). The overarching idea in this literature is that advantage can be taken of modern information systems, notably integrated ERP systems, to report performance information on a continuous or frequent basis and provide frequent or recurrent assurance on key corporate information. Research approaches include modeling studies (Calderon et al. 2006; Vasarhelyi et al. 2004), design science (Alles et al. 2006; Alles et al. 2008), and think pieces (Vasarhelyi et al. 2010). The growth in the literature and adoption of continuous assurance within corporations coupled with the availability of software solutions mean that it is now time for a thorough analysis of the nature and future place of continuous reporting and assurance.The papers in the special issue provide an opportunity to compare and contrast the state of debate on research in both the AIS and IS communities. The differences are marked. Recent debates in the IS domain have primarily been about the future of IS research and, closely aligned to this fundamental problem, the role of research method. There is only limited discussion of what should be researched (e.g., Davis et al. 2010). Scholars in IS are concerned about the standing of the domain vis-à-vis other cognate disciplines. Is IS a distinct or an applied discipline? To what extent should IS research contribute to practice (Agarwal and Lucas 2005; Benbasat and Zmud 2003; Straub and Ang 2011)? Much of the discussion is about how research is to be conducted. In particular, there is a long-standing debate about the appropriate balance between natural and design science approaches (Hevner et al. 2004; Holmström et al. 2009; March and Smith 1995). The debate has now reached the state of a religious war (Baskerville et al. 2011; Österle et al. 2011) that may well last as long as the Thirty Years War (Wilson 2009).In the AIS domain, we have only echoes of these debates. Rather, we are more concerned with the place of AIS research in the broader research ecosphere, in particular the IS and accounting domains. Over the last decade, the editors of the major journals in our sub-discipline have given attention in their editorials to the state of research in accounting information systems (Steinbart 2009; Stone 2002; Sutton 2000, 2004, 2008, 2010; Vasarhelyi 2006). Steinbart (2009) asserts that AIS has its own “unique identity” but must also contribute to both IS and accounting and auditing research. This followed a number of calls for improving the manner in which AIS researchers interact with and leverage IS research. Sutton (2004) even called for the AIS research community to “rediscover” our “IS roots.” Steinbart (2009) notes that research contributions need not be to both of these disciplines at the same time. Indeed, it is difficult to conceive of many studies that could simultaneously contribute to both the IS and accounting domains. On a similar note, Sutton (2010) asserts that “The goal should not be to build a separate and isolated discipline, but to recognize we [AIS researchers] are a viable part of accounting and integrate with the larger discipline.” This search for an appropriate place for AIS research is the result of being the sub-discipline of two applied disciplines (a second derivative-applied discipline?). Perhaps because of our relatively insecure place between the two applied disciplines and the relatively small size of the AIS community, we are agnostic to method (although personally I would like to see more design science research in collaboration with colleagues in computer science, to which I will return shortly). More important is how far we research beyond the intersection of accounting and auditing, on the one hand, and technology, on the other (Sutton 2010, 294). I believe it is this intersection that generates comparative advantage for the AIS research community. Strengthening the way we integrate with the IS research community and contribute to that community is clearly desirable but is, by no means, the end of the story. We must better exploit our comparative advantage in bringing together our knowledge of accounting and technology and our understanding of how each interacts and influences the other.An important aspect of this intersection is to communicate to others in the academy how technology impacts the various aspects of accounting and auditing. It is interesting that the special issue that is the focus of this editorial was to strengthen the link between IS and AIS. It would have been equally interesting to have a call for papers that merely changed “information systems” to “accounting and auditing.” This is particularly important given the Balkanization of our research caused, in part, by the existence of specialist technology sections within the American Accounting Association (information systems and strategic and emerging technologies) and associated journals (JIS, JETA, and IJAIS). Technology papers are published, for the most part, in the specialist technology journals that only AIS researchers read. There are many interesting questions that need research integration papers. For example: What would financial reporting look like if engineered with current, rather than eighteenth-century, communication technologies? How would auditing change if the norm were that every substantive test were on the client's complete data set rather than a sample drawn in essentially the same manner as five decades ago? How does management accounting change in a world of massive corporate datasets and computing power? We need to put the A back into AIS.There is a final issue that comes from this special issue. It is interesting that the issue was designed to integrate IS and AIS research. What happened to computer science in this picture? For much of the research that I undertake, I rely equally as much or more on computer science research as I do IS research. I employ computer science theory and practice in areas such as datamining, databases, ontologies, and human-computer interaction (HCI). As an AIS community, we should reconsider how we draw from computer science research and collaborate with computer science researchers on design science research projects. It is unlikely that AIS researchers will be able to contribute to computer science research, as we may well be able to do for IS research. A mutually rewarding collaboration may, however, further research in both disciplines.