J. Efrim Boritz

Effectiveness of Neural Network Types for Prediction of Business Failure

The study examines the effectiveness of different neural networks in predicting bankruptcy filing. Two approaches for training neural networks, Back-Propagation and Optimal Estimation Theory, are considered. Within the back-propagation training method, four different models (Back-Propagation, Functional Link Back-Propagation With Sines, Pruned Back-Propagation, and Cumulative Predictive Back-Propagation) are tested. The neural networks are compared against traditional bankruptcy prediction techniques such as discriminant analysis, logit, and probit. The results show that the level of Type I and Type II errors varies greatly across techniques. The Optimal Estimation Theory neural network has the lowest level of Type I error and the highest level of Type II error while the traditional statistical techniques have the reverse relationship (i.e., high Type I error and low Type II error). The back-propagation neural networks have intermediate levels of Type I and Type II error. We demonstrate that the performance of the neural networks tested is sensitive to the choice of variables selected and that the networks cannot be relied upon to “sift through” variables and focus on the most important variables (network performance based on the combined set of Ohlson and Altman data was frequently worse than their performance with one of the subsets). It is also important to note that the results are quite sensitive to sampling error. The significant variations across replications for some of the models indicate the sensitivity of the models to variations in the data.

Assurance on XBRL-Related Documents: The Case of United Technologies Corporation

ABSTRACT: The eXtensible Business Reporting Language (XBRL) was developed to provide financial information users with a standardized method to prepare, publish, and exchange business information in digital format. XBRL is being used around the world for financial reporting and government e‐filings. Although there has been growing awareness about assurance issues related to the use of XBRL, current audit practices and standards fall short of providing the needed guidance for the provision of assurance on XBRL‐Related Documents. In this paper, we report on a mock assurance engagement that we conducted on the XBRL‐Related Documents of United Technologies Corporations 10‐Q for the third quarter of 2005 and repeated on its 10‐Q for the third quarter of 2008 to identify the issues that companies and auditors might encounter if they are requested to provide assurance on XBRL‐Related Documents. We describe the assurance framework applied in the mock assurance engagement, present the findings from the examination...

Competency‐Based Education and Assessment for the Accounting Profession: A Critical Review*

In recent years many professional accounting associations have become interested in establishing competency-based professional requirements and assessment methods for certifying accounting professionals. A competency-based approach to qualification specifies expectations in terms of outcomes, or what an individual can accomplish, rather than in terms of knowledge or other capabilities possessed by the individual. This idea has an obvious appeal to many practitioners and administrators of professional qualification programs. However, there is limited knowledge about competency-based approaches in the accounting profession and among accounting academics, constraining discussion about the value of these approaches and about the strengths and weaknesses of the different competency models that have sprung up in various jurisdictions. In this paper we review and synthesize the literature on competency-based approaches. We identify a number of theoretical benefits of competency-based approaches. However, we also find many alternative definitions and philosophies underlying competency-based approaches, and a variety of visions of how competencies should be determined and assessed. We note that there is limited evidence supporting many competency-based approaches and we identify 14 research questions that could be investigated to help policy makers to more effectively address policy matters related to competency-based education and assessment.

Assurance Reporting for XML-Based Information Services: XARL (Extensible Assurance Reporting Language)

Extensible Business Reporting Language (XBRL) is an XML-based method for financial reporting. XBRL was developed to provide users with an efficient and effective means of preparing and exchanging financial information over the Internet. However, like other unprotected data coded in XML, XBRL (document) files (henceforth, documents) are vulnerable to threats against their integrity. Anyone can easily create and manipulate an XBRL document without authorization. In addition, business and financial information in XBRL can be misinterpreted, or used without the organizations consent or knowledge. Extensible Assurance Reporting Language (XARL) was initially developed by Boritz and No (2003) to enable assurance providers to report on the integrity of XBRL documents distributed over the Internet. Providing assurance on XBRL documents using XARL could help users and companies reduce the uncertainty about the integrity of those documents and provide users with trustworthy information that they could place warranted reliance upon. A limitation of the initial conception of XARL was its tight linkage with the XBRL document and the comparatively primitive approach to codifying the XARL taxonomy. In this paper, we have reconceptualized the idea of XARL as a standalone service for providing assurance on potentially any XML-based information being shared over the Internet. While our illustrative application in this paper continues to be XBRL-coded financial information, the code that underlies this version of XARL is a significant revision of our earlier implementation of XARL, is compatible with the latest version of XBRL, and moves XARL into the Web services arena.

Adaptive fraud detection using benford's law

Adaptive Benfords Law [1] is a digital analysis technique that specifies the probabilistic distribution of digits for many commonly occurring phenomena, even for incomplete data records. We combine this digital analysis technique with a reinforcement learning technique to create a new fraud discovery approach. When applied to records of naturally occurring phenomena, our adaptive fraud detection method uses deviations from the expected Benfords Law distributions as an indicators of anomalous behaviour that are strong indicators of fraud. Through the exploration component of our reinforcement learning method we search for the underlying attributes producing the anomalous behaviour. In a blind test of our approach, using real health and auto insurance data, our Adaptive Fraud Detection method successfully identified actual fraudsters among the test data.

SEC's XBRL Voluntary Program on Edgar: The Case for Quality Assurance

XBRL (eXtensible Business Reporting Language) was developed to provide users with an efficient and effective means of preparing and exchanging financial information over the Internet. After years of development, XBRL is now in the implementation stage, with many companies, governments, regulators, and stock exchanges around the world implementing or planning to adopt XBRL for electronic filing of financial statements. In this paper, we examine the XBRL fillings in the SECs XBRL Voluntary Filing Program (VFP) on EDGAR from its inception to December 31, 2007 and report findings from our observations and validation tests. We identify persistent and increasing quality control and assurance issues pertaining to the filings under the VFP and discuss potential countermeasures needed to ensure that XBRL filings are reliable and gain user confidence and acceptance.

Detecting fraud in health insurance data: learning to model incomplete benford's law distributions

Benfords Law [1] specifies the probabilistic distribution of digits for many commonly occurring phenomena, ideally when we have complete data of the phenomena. We enhance this digital analysis technique with an unsupervised learning method to handle situations where data is incomplete. We apply this method to the detection of fraud and abuse in health insurance claims using real health insurance data. We demonstrate improved precision over the traditional Benford approach in detecting anomalous data indicative of fraud and illustrate some of the challenges to the analysis of healthcare claims fraud.

Internet Privacy in E-Commerce: Framework, Review, and Opportunities for Future Research

Increased Internet traffic and the sophistication of companies in tracking that traffic have made privacy a critical issue in electronic commerce (e-commerce). This has spawned a number of research works addressing Internet privacy from the perspectives of three main stakeholders - customers, companies and governments, as well as the interactions among them. The purpose of this paper is to analyze the extant studies and develop an understanding of the relationships among them. Accordingly, we review the research on Internet privacy in e-commerce that has been conducted in the fields of information systems, business, and marketing. We develop a framework for classifying the studies, review key findings, and identify opportunities for future research.

Internet Privacy Research: Framework, Review and Opportunities

Increased Internet traffic and the sophistication of companies in tracking that traffic have made privacy as a critical issue in electronic commerce (e-commerce), and in turn spawned a number of research works in the literature. Despite this, what is lacking is an effort to understand the relationships among the various studies. The purpose of this paper is to consider the fields of information systems, business and marketing, and provide a framework for the research works that have dealt with three main stakeholders, namely customer, company and government, as well as the interaction arising among them. We review the literature and identify opportunities for future research.

Firm-Specific Characteristics of the Participants in the SEC's XBRL Voluntary Filing Program

ABSTRACT: A number of papers have attempted to study firm-specific characteristics of the participants in the SEC-administered XBRL Voluntary Filing Program (VFP). However, to date, their findings have been conflicting—contrary to the underlying theory or inconclusive due to methodological limitations. Some of these limitations include the use of limited subsets of VFP data, the use of portfolio matching designs containing matching weaknesses, and omission of key explanatory variables. This paper attempts to overcome some of these limitations by using a more comprehensive sample, employing a more effective matching procedure, and a more complete set of variables suggested by both voluntary disclosure and organization theories. Consistent with the theory, higher voluntary disclosure propensity, stronger corporate governance, and better profitability are found to be robustly significant factors associated with voluntary XBRL adoption in the U.S. Innovativeness is a distinguishing characteristic for non-high...