Neural Networks: A New Tool for Predicting Thrift Failures*

A neural network model that processes input data consisting of financial ratios is developed to predict the financial health of thrift institutions. The network's ability to discriminate between healthy and failed institutions is compared to a traditional statistical model. The differences and similarities in the two modelling approaches are discussed. The neural network, which uses the same financial data, requires fewer assumptions, achieves a higher degree of prediction accuracy, and is more robust.     

What to Learn from Near Misses: An Inductive Learning Approach to Credit Risk Assessment

This paper presents a new dimension of inductive learning for credit risk analysis based on the specific impact of Type I and Type II credit errors on the accuracy of the learning process. A Dynamic Updating Process is proposed to refine the credit granting decision over time and therefore improve the accuracy of the learning process. The new dimension is tested on credit files of small Belgian businesses. Results indicate an improvement of the learning process in terms of predictive accuracy, stability, and conceptual validity of the final decision tree.     

Preference-Order Recursion for Finding Relevant Pure,Admissible and Optimal Statistical Decision Functions*

A preference-order recursion algorithm for obtaining a relevant subset of pure, admissible (non-dominated, efficient) decision functions which converges towards an optimal solution in statistical decision problems is proposed. The procedure permits a decision maker to interactively express strong binary preferences for partial decision functions at each stage of the recursion, from which an imprecise probability and/or utility function is imputed and used as one of several pruning mechanisms to obtain a reduced relevant subset of admissible decision functions or to converge on an optimal one. The computational and measurement burden is thereby mitigated significantly, for example, by not requiring explicit or full probability and utility information from the decision maker. The algorithm is applicable to both linear and nonlinear utility functions. The results of behavioral and computational experimentation show that the approach is viable, efficient, and robust.     

Automating Knowledge Acquisition and Refinement for Decision Support: A Connectionist Inductive Inference Model*

An important application of expert systems technology is to provide support for nonstructured decision making. Usually, nonstructured decision making is characterized by heavy reliance on heuristic knowledge, which is very difficult to articulate or document, and therefore traditional knowledge acquisition approaches are not very successful. The quality and effectiveness of an expert system supporting unstructured decision making is affected when traditional knowledge acquisition approaches are used. To alleviate this problem a model is proposed that combines inductive inference and neural network computing, and an example is presented that illustrates the potential of this model in unstructured decision support.     

Structure of Uncertainty and Decision Making: An Experimental Investigation*

Decisions in the real world usually involve imprecise information or uncertainty about the precesses by which outcomes may be determined. This research reports the results of a laboratory experiment which examined whether the structure of uncertainty, namely, both the center and the range of the probability distribution describing the uncertainty, is an important determinant of choice. Specifically, it examines how the uncertainty of audit by the Internal Revenue Service of income tax returns affects taxpayers' decisions about intentional noncompliance. The context is relevant as almost nothing is known about how taxpayers assess detection risks using the probability information they have. The study focuses on intentional noncompliance. The factors affecting it are distinct and separate from those affecting unintentional noncompliance. Other factors that affect intentional tax noncompliance, such as risk, tax rates, and penalty rates, were controlled in the experiment. It was hypothesized that the lower the mean and the lesser the range (ambiguity) of the perceived audit probability, the greater the international noncompliance. As hypothesized, the analysis indicates that both the mean and the range of the perceived audit probability rate affect intentional noncompliance, though the effect of ambiguity is greater at a relatively higher level of mean. This result suggests that the strength of the information describing an uncertain event is captured better by both the mean and the range of the uncertainty than either of those components singly.     

A Stochastic,Psychometric Valuation Model for Personnel Selection Systems*

This paper describes a method for the valuation of personnel selection systems that is based on contributions from accounting, utility theory, psychometric techniques, Markov processes, labor economics, and present-value analysis. The model developed here can be used by managers to assess the usefulness of proposed selection procedures. It extends previous work in the area by including both the stochastic nature of the employment process and the time value of the associated costs and benefits.     

A Decision Model for the Robot Selection Problem Using Robust Regression*

Industrial robots are increasingly used by many manufacturing firms. The number of robot manufacturers has also increased with many of these firms now offering a wide range of models. A potential user is thus faced with many options in both performance and cost. This paper proposes a decision model for the robot selection problem. The proposed model uses robust regression to identify, based on manufacturers' specifications, the robots that are the better performers for a given cost. Robust regression is used because it identifies and is resistant to the effects of outlying observations, key components in the proposed model. The robots selected by the model become candidates for testing to verify manufacturers' specifications. The model is tested on a real data set and an example is presented.     

Formulation Tactics and the Success of Organizational Decision Making*

One hundred and seventy-seven cases profiling organizational decisions were analyzed to determine how managers carry out formulation as they initiate a decision-making process. Analysis revealed that formulation was carried out by applying problem, idea, target, and reframing tactics. Decision adoption rates, decision value, and time to carry out the decision-making process were used to determine the success of these tactics, controlling for situational effects of importance, urgency, resources, and management level. Reframing was found to be the most successful tactic under all conditions, but was the least frequently used by decision makers. Problem and idea tactics were the least successful and success did not improve with additional resources, but these tactics were used more often than the other tactics. Problem tactics were even less successful when applied to urgent and important decisions. Target tactics were surprisingly effective for crises and important decisions. Decision makers seem prone to use ineffective formulation tactics and give little consideration to the opportunities or constraints imposed by the situation as a tactic is selected.     

A Note on Consistency in Utility Assessment*

A methodology for determining a von Neumann-Morgenstern utility function is outlined based on the axioms crucial to such a function. Reconciliation of inconsistent judgments is facilitated using the theory of reciprocal matrices. Numerical measures of the collective divergence of a set of judgments from perfect consistency or coherency are provided.     

A Computerized Approach to Decision Process Tracing for Decision Support System Design*

We propose the use of computerized process tracing (CPT) tools as an appropriate approach for monitoring the information acquisition and evaluation phase of specific decision processes. CPT tools are unobtrusive and seem particularly relevant for evaluating certain decision tasks that may be supported by decision support systems (DSS). CPT tools can be an important component of DSS development. An information systems research taxonomy developed by previous researchers [29] [36] is used to position research work involving the methodology of CPT. Using a critique suggested by Libby [28], CPT tools are evaluated and compared to alternative process tracing tools. A brief empirical example using CPT is provided, and future uses relative to DSS are suggested. The appendix includes an example of a specific CPT tool.     

Strategic Assessment Model (SAM): A Multiple Criteria Decision Support System for Evaluation of Strategic Alternatives*

The evaluation of strategic alternatives is a particularly difficult task. This difficulty is due to the complexities inherent in the evaluation process and the lack of structured information. The evaluation process must consider a multitude of relevant information from both the internal and external environments of the organization. Various analytical and normative models have helped decision makers utilize large volumes of information in strategic evaluation; however, most of these models have some limitations. We present a multiple criteria decision support system, called strategic assessment model (SAM), that addresses some of the limitations inherent in the existing models. SAM captures the decision maker's beliefs through a series of sequential, rational, and analytical processes. The environmental forces—decomposed into internal, task, general opportunities, and threats—are used along with the analytic hierarchy process (AHP), subjective probabilities, the entropy concept, and utility theory to enhance the decision maker's intuition in evaluating a set of strategic alternatives.     

Coordinating Service‐Sensitive Demand and Capacity by Adaptive Decision Making: An Application of Case‐Based Decision Theory*

The subject of this article is the simultaneous choice of product price and manufacturing capacity if demand is stochastic and service‐level sensitive. In this setting, capacity as well as price have an impact on demand because several aspects of service level depend on capacity. For example, delivery time will be reduced if capacity is increased given a constant demand rate. We illustrate the relationship between service level, capacity, and demand reaction by a stylized application problem from the after‐sales services industry. The reaction of customers to variations in service level and price is represented by a kinked price‐demand‐rate function. We first derive the optimal price‐capacity combination for the resulting decision problem under full information. Subsequently, we focus on a decision maker (DM) who lacks complete knowledge of the demand function. Hence the DM is unable to anticipate the service level and consequently cannot identify the optimal solution. However, the DM will acquire additional information during the sales process and use it in subsequent revisions of the price‐capacity decision. Thus, this decision making is adaptive and based on experience. In contrast to the literature, which assumes certain repetitive procedures somewhat ad hoc, we develop an adaptive decision process based on case‐based decision theory (CBDT) for the price‐capacity problem. Finally, we show that a CBDT DM in our setting eventually finds the optimal solution, if the DM sets the price based on absorption costs and adequately adjusts the capacity with respect to the observed demand.     

A Comprehensive Model for Managing Credit Risk on Home Mortgage Portfolios*

Managing credit risk in financial institutions requires the ability to forecast aggregate losses on existing loans, predict the length of time that loans will be on the books before prepayment or default, analyze the expected performance of particular segments in the existing portfolio, and project payment patterns of new loans. Described in this paper are tools created for these functions in a large California financial institution. A forecasting model with Markovian structure and nonstationary transition probabilities is used to model the life of a mortgage. Logistic and regression models are used to estimate severity of losses. These models are integrated into a system that allows analysts and managers to depict the expected performance of individual loans and portfolio segments under different economic scenarios. With this information, analysts and managers can establish appropriate loss reserves, suggest pricing differentials to compensate for risk, and make strategic lending decisions.     

A PENSION DECISION MODEL FOR THE EMPLOYER*

Recently, the Financial Accounting Standards Board (FASB) reevaluated accounting for pension plans. The issue is emotional and highly political in nature. The FASB attempted to justify its approach on the basis of measuring economic activity, but it failed to provide much in the way of analytical support. This paper provides a managerial decision model and an economic basis for the existence of pension plans. A pension plan is described as a cost-saving, risk-sharing, incentive contract. The analysis is developed using agency theory. The model presented here meets three suggested objectives of an employer: 1. Maximization of utility through the maximization of profit 2. Ability to conform the risk characteristics of an employment contract to the risk characteristics of the employer 3. Diversification of the risk inherent in the employment contract Profit is maximized by producing cost savings associated with employee tenure and loyalty. Sharing cost savings with employees (i.e., offering a pension plan) meets the above objectives. The employer determines the optimal sharing rate for the expected cost savings. An examination of the employer's underlying decision process reveals implications for pension plan accounting which generally are consistent with and support the FASB's Statement of Financial Accounting Standards No. 87 [5].     

Stochastic and Intransitive Behavior in a State-Preference Model of Asset Choice*

Most models of investor behavior assume a time-state independent utility function and result in a deterministic solution where a given set of inputs uniquely specifies the decision. In contrast, a state preference model using a time-state dependent utility function is derived in this paper. The model allows the investment choice decision to be analyzed in a game theoretic context. The general solution is a mixed strategy which allows for a probabilistic interpretation of the decision. The approach presented in this paper can accommodate anomalies such as intransitivity of preference and satisficing as rational behavior. An example of a possible implementation is given along with interpretations of the outcomes.     

A Multiperiod Audit Staff Planning Model Using Multiple Objectives: Development and Evaluation*

The audit staff planning problem, a specific type of manpower planning problem, has been modeled using goal programming and, more recently, multiple objective linear programming. Prior studies developed single-period models and did not go beyond the model building stage. This study develops a multiperiod audit staff planning model and evaluates the model using a test application involving actual decision makers (partners in public accounting firms). The multiperiod model includes seven objectives to be optimized: profit (to be maximized), late completion of work, work declined, staff augmentation, staff reduction, underutilization of the work force, and shortfall in meeting professional development targets (all to be minimized). Over a four-quarter planning horizon with one “busy season,” the model is subject to constraints with respect to the projected audit work load, ability to substitute personnel and to perform interim audit work, available staff hours (including overtime limitations), supervisory requirements, and professional development targets. Results of the test application showed that the model was capable of producing a range of values for each objective. The participants were exposed to much of that range when making their decisions. The results also showed that all objectives were important and that participants were consistent in choosing their preferred level of each objective over several runs of the model. These results and the reactions of the participants demonstrate that the model is usable by actual decision makers and has potential for a number of specific applications.     

Notes and Communications “A DECISION THEORY EXAMPLE IN FOOTBALL”: A COMMENT*

Janssen and Daniel analyzed the choice between a one- or a two-point conversion for a particular game situation in college football. Their decision criteria was maximum expected utility based on a von Neumann-Morgenstern utility function defined over the games outcomes. An alternative approach based on a stochastic dominance criterion is presented that does not rely on knowledge of the relative importance of tying vs. winning; rather, it relies on a notion of consistency in the sequential problem.     

Strict Uncertainty: A Criterion for Moderately Pessimistic Decision Makers*

This paper proposes a decision rule to rank actions under strict uncertainty, the available information being limited to the states of nature, the set of alternative rows, and the consequence of choosing every row if a given state occurs. This rule is suitable to moderately pessimistic individuals and social groups, these agents being neither maximax nor maximin decision makers but people who assume that the best outcome from the action will not occur. For these decision makers the paper shows the existence of a consistent weight system in which one and only one weight is attached to each state of the world under plausible conditions of domination. Most of the traditional axioms are satisfied by the proposed ranking approach. In the frame of disappointment (measured by ranges of column dispersion), the meaning of some controversial postulates used in the literature is explained. The proposed criterion is a departure from Laplace's (1825) rule and from the remaining standard criteria. Only in the special case of equal column dispersion do both Laplace's rule and the proposed weights lead to the same solution.     

Planning Forward by Looking Backward: Retrospective Thinking in Strategic Decision Making*

Retrospective thinking occurs whenever one remembers something from the past, but one can also think retrospectively about hypothetical future events, by imagining that the event has already transpired and then working backward in the mind from the future toward the present. Designers, artists, planners, and other creative individuals utilize retrospective thinking when they mentally envision the object they wish to create and then thinkabout how it might be constructed. Some theorists and some previous research have suggested that forward and backward thinking are different cognitive processes, but there has been little investigation on the nature of these differences. This article examines the literature and describes a laboratory study designed to investigate retrospective thinking as a technique for increased idea generation. Subjects (n=64) were professional planners and managers with an average business experience of over 20 years. In a planning task, significantly more ideas were generated in the retrospective thinking mode than in the prospective mode, but the ideas were not found to be of greater quality. Evidence emerged that a subset of subjects were better able to make use of the retrospective technique, suggesting that selection along with training may be appropriate for the technique's mosteffective use.     

Generating Explanations for Goal-Based Decision Making*

As information systems make the transition from decision support to decision making, there will be a concomitant need for the programs to explain or justify their actions. Without such explanations, humans will not readily cede authority to a machine. This paper describes an automated decision-making program, VOTE, which generates natural language explanations for its decisions in both English and French. The program domain is Congressional roll call voting. VOTE simulates voting decisions of specific members of the United States House of Representatives on given bills. VOTE's natural language generation facility is also used by the underlying databases to interpret the knowledge representations. These underlying knowledge representations are described and applications of this decision making model to the general business domains of strategic planning, investment, and marketing are suggested.