John J. Maher

Predicting Bond Ratings Using Neural Networks: A Comparison with Logistic Regression

Bond rating agencies examine the financial outlook of a company and the characteristics of a bond issue and assign a rating that indicates an independent assessment of the degree of default risk associated with the firm’s bonds. Predicting this bond rating has been of interest to potential investors as well as to the firm. Prior research in this area has primarily relied upon traditional statistical methods to develop models with reasonably good prediction accuracy. This article utilizes a neural network approach to modeling the bond rating process in an attempt to increase the overall prediction accuracy of the models. A comparison is made to a more traditional logistic regression approach to classification prediction. The results indicate that the neural networks-based model performs significantly better than the logistic regression model for classifying a holdout sample of newly issued bonds in the 1990–92 period. A potential drawback to a neural network approach is a tendency to overfit the data which could negatively affect the model’s generalizability. This study carefully controls for overfitting and obtains significant improvement in bond rating prediction compared to the logistic regression approach.

Framework for Cognitive Skill Acquisition and Spreadsheet Training

It is well documented that electronic spreadsheet models utilized in many professions to enhance decision making frequently contain errors that have negative effects on the ultimate quality of decisions. Limited research has been published that systematically identifies potential reasons for the causes of these errors, and what procedures can be taken to minimize or eliminate them. Our research provides initial evidence concerning this problem area by investigating how several important cognitive skills are affected by formalized spreadsheet training. Results indicate that one cognitive skill, logical reasoning, significantly increases after a six-week training period. Importantly, the greater the increase in logical reasoning skill, the more effectively the subject developed competent spreadsheet models. These findings provide a meaningful step in more perceptibly understanding and defining important cognitive changes that occur in individuals as they undergo formalized spreadsheet development training. Further extensions of this research should more clearly refine our understanding of the cognitive changes that occur in spreadsheet developers and eventually cultivate the development of more efficient and effective training methods for spreadsheet model designers. ABSTRACT INTRODUCTION

The Impact of Auditor Tenure on Initial Bond Ratings

Abstract The desirability of mandated auditor rotation represents an ongoing debate in the accounting profession. Proponents assert that audit quality (through auditor independence) is threatened by extended auditor–client relationships. Opponents assert that mandatory auditor rotation will actually decrease audit quality, primarily due to the time required for auditors to learn the nuances of a clients business processes. Our research contributes to this important debate by providing empirical evidence regarding the capital markets effects of audit tenure. Specifically, we examine newly issued bonds over the period 1990–2002 and find auditor tenure to be positively related to ratings received. This finding remains consistent across all sample issues regardless of investment grade, firm performance, or time period. We find no evidence that extended auditor–client relationships result in a decrease in the perceptions of audit quality.

Home mortgage analysis for cultivating crucial spreadsheet and model development skills

Abstract Accounting practitioners and academics have identified spreadsheet and model development skills as important for accounting graduates to possess. We present an adaptable mortgage analysis project that provides a practical setting in which students can develop critical spreadsheet modeling skills while demonstrating knowledge of several important principles of accounting and finance. The developed spreadsheet model is designed in a structured manner to permit changing the necessary financial terms and facilitates performing sensitivity analysis on critical variables. The analysis requires an informed decision be made between employing a fixed rate mortgage or an adjustable rate mortgage (ARM) for financing a home purchase. Important theoretical skills required to arrive at an acceptable solution include understanding of present value concepts, effective rate method of interest calculation, and amortization of principle balance. The student must also verify proficiency with the necessary theoretical and practical skills of effective spreadsheet design and model development

Perceptions of Postretirement Benefit Obligations by Bond Rating Analysts

This research examines the measurement and impounding of alternative measures of a corporations other postretirement benefits obligation (OPEBs) by an important segment of the capital markets. The Kaplan and Urwitz (1979) model is used as a benchmark from which to assess the importance of an added OPEB variable in the bond rating process. Using the corporate bond rating as the dependent variable, multiple measures of the OPEB obligation are inserted individually as an added independent variable into an N-chotomous probit model. The results for 1987 and 1988 indicate that measures calculated from publicly available information produce highly significant results. The developed postretirement liability measures are found to provide relevant and material information regarding the risk level of a firms bonds as represented by its bond rating. This insight concerning the additional risk represented by a firms postretirement benefits is beyond that supplied by the firms pension information. This suggests that the additional investor default risk attributed to a firms OPEB can be reasonably proxied by data found in the companys annual report footnote disclosures.

Environmental Liabilities, Bond Ratings, and Bond Yields

We examine the relationship that exists among bond ratings, bond yields, and various estimates of a firms contingent environmental remediation liability using a sample of new bond issues. Our results indicate that the largest external EPA-based estimates of the firms environmental obligations are significantly associated with a firms bond rating, providing relevant incremental information beyond that supplied by the environmental accruals presented in the financial statements. Furthermore, while the accrued environmental liability is shown to have a direct association with the bond yield, the external EPA-based estimates provide an indirect relationship with the bond yield through their influence on the bond rating. These results contribute to the extant literature by empirically clarifying the role of various environmental liability estimates in establishing a firms bond rating and further indicating their connection with the pricing of corporate debt.

Defined-Benefit versus Defined-Contribution Pension Plans: How to Compare

This article proposes a framework that will help managers and employees to compare the retirement benefits that accrue from the two major types of pension plans.

Evidence Regarding Elimination of IFRS-U.S. GAAP Reconciliation

The main purpose of this paper is to provide initial evidence regarding the effects of the Securities and Exchange Commission (SEC)s decision to eliminate the International Financial Reporting Standards (IFRS)-U.S. Generally Accepted Accounting Principles (GAAP) reconciliation by examining short-window trading volume reactions for the year of the elimination and the year prior to the elimination. This provides an important partial snapshot regarding the effects of eliminating U.S. GAAP on the U.S. capital markets. Our results indicate that trading volume decreased the year the reconciliation was eliminated which is consistent with a significant decrease in value relevant information available to equity investors related to those firms that eliminated the U.S. GAAP reconciliation. To further validate our findings, we also compare changes in trading volume for our sample firms with a control group of foreign private issuers unaffected by the SECs rule to eliminate the reconciliation. We find the decrease in announcement period volume to be significantly greater for foreign private issuers that eliminate the reconciliation. These results are consistent with the interpretation that value relevant information has been lost to the capital markets for those firms no longer required to reconcile back to U.S. GAAP.