تنظیمات گارانتی مطلوب برای طراحی سیستم های اطلاعات حسابداری

This paper investigates optimal price/warranty arrangements for the design of accounting information systems (AISs) when the user of the system has alternative sources of information. A mathematical model is developed that focuses on two systems design scenarios commonly encountered in practice: (1) enhancement and (2) replacement. In an enhancement scenario the AIS user hires the AIS designer to enhance the value of the current AIS by incorporating additional features into the original system. Systems replacement, on the other hand, implies that the designer produces a new system that makes the original system obsolete. Under a replacement system, a simple limited warranty implements the user's optimal decision rule so that no losses occur due to unobservability of the user's decision. For the enhanced AIS design however, the user's action may depend upon the signal from the user's preexisting AIS. Therefore, when the user's decision is not observable, the designer does not know if the user's decision is a result of the systems enhancement or a result of the original system. Due to incomplete contracts, a more complex and costly warranty arrangement is required. Hence, losses occur under enhancement AIS although quality can be maintained with a limited warranty. Since it is difficult to assess the quality of the information system prior to purchase, the potential for a market failure exists where high-quality systems are driven from the market by low quality systems. These findings show that a market failure can be prevented in the design of information systems through contractual warranty arrangements. These warranty arrangements, however, must be tailored to the specific design task.

This paper examines the issues inherent in contracting for the design of information systems when the system user has alternative informational sources. We develop a principal–agent model where the system user hires a system designer to develop an accounting information system (AIS) to aid in the user's decision-making. Due to unobservability of the system development, the quality of the information system can only be detected by observing an outcome affected by the user's decisions. However, the user's decisions may be a result of information from the alternative sources rather than the system provided by the developer, providing additional complication in the contracting process. The intangible nature of information provides for unique contracting problems.1Cushing (1990) noted that the information system development process is characterized by the existence of some degree of friction between system users and system designers. If these two parties do not have perfectly congruent goals, a contract is developed to specify their relationship. The inability to monitor the designer directly leads to the use of performance measures to represent the designer's actions to the user. Unless such performance measures can be developed, high quality may not be sustained in the market for information systems. As with many products, the user cannot readily discern the quality of the information system prior to using it and, at the same time, increased quality may prove costly to the designer. Applying the theory of the market for lemons (Akerloff, 1970), users without additional information on system quality will tend to underprice higher-quality systems while overpricing lower-quality systems. Therefore, as designers of high-quality systems do not recoup their investment, low quality systems will push high-quality systems from the marketplace. Such a market failure can be avoided if the designers of higher-quality information systems can credibly signal that quality to the marketplace (see, for example, Spence, 1974). One such signaling process found in the market for many products is a limited warranty (see Cooper and Ross, 1985 and Mann and Wissink, 1988). With a limited warranty, the designer promises to share the risk of product failure between him and the user. In the particular case of AISs the measurement of product failure is determined by the specific purpose of the AIS. AISs normally support two broad functions in an organization. The first of these functions performs structured financial transaction processing and success may be measured by the accuracy and timelessness by which reliable reports are generated. This function reports on past events rather than predicting future events. The second function is to aid in managerial decision-making. In this case, the AIS aids the user in making decisions under uncertainty to achieve some risky outcome (for example, profitability). The ultimate measure of the AIS's success is its ability to predict future random events. The user brings to the problem a level of expertise that presents itself in the form of using the AIS to aid in the choice of possible decision alternatives. This existence of expertise may cause the final outcome of the decision process to be the only observable measure of system success. Since the final outcome is a function of both the AIS and the user's expertise, the designer may be reluctant to provide a warranty arrangement. This reluctance could lead, in turn, to a market failure. It should be noted that the same AIS normally supports both functions in an organization. The purpose of this paper is to investigate the managerial decision-making function of an AIS. The research focuses on two systems design scenarios commonly encountered in practice: (1) enhancement and (2) replacement. In an enhancement scenario the user hires the designer to enhance the value of the current IS by incorporating additional features into the original system. A firm that recently has adopted a standard cost system approach to aid in controlling internal costs, for example, might hire a designer to add functionality to their existing information system (which provides information on actual manufacturing costs) by incorporating standard cost information along with related efficiency and price variance reports. The resulting new AIS consists of both the original AIS and the enhancement features provided by the designer. Systems replacement, on the other hand, implies that the designer produces a new system that makes the original system obsolete. For example, a firm that has decided to change from a standard inventory control system, which relied on batch processing, to a just-in-time based inventory control system that relies on on-line processing, might hire a designer to completely redesign the system. A second example encountered in many firms is to replace their traditional AIS with an enterprise-wide system, such as SAP, Oracle Financials, etc. The model developed in this paper has many of the elements of standard principal–agent problems such as in Holmström (1979). In those models the final outcome is a function of the agent's (AIS designer) effort. In the model developed in this paper, however, the final outcome is a function of both the user's decision and the system's success in predicting the random state of nature. Therefore, we cannot merely examine only the designer's effort. Since the user's decision is not mutually observable, the results in our model stem from elements of “blocked communication” (see Demski and Sappington, 1986) and of “incomplete contracts” (see Amershi and Datar, 1991 and Holmström and Milgrom, 1990). The model is similar to that of Demski and Sappington (1993). Their model, however, allows communication of private information from the principal (the user) to the agent (the designer) whereas this communication is blocked in our model. Rasch and Yost (1997) extended this research by addressing what they referred to as the “design dilemma.” The design dilemma (paraphrased) is why many IS designs result in poor user satisfaction although everyone involved in the design wanted the IS to be successful. Rasch and Yost note that anecdotal evidence shows that at the beginning of an IS design, the user and the designer are highly motivated to design a successful IS. Over time, however, the user becomes preoccupied with other urgent business-related issues and cannot devote large amounts of time communicating with the designer. The designer, on the other hand, may fall behind schedule and also experience budget problems. The final result of these pressures is that the IS design is completed, but neither party is completely satisfied with the final information system. Rasch and Yost showed that unobservability of the user's decision making does not create a contracting friction if communication is complete between the user and the designer. They also showed, however, that a lack of communication resulted in the “design dilemma.” Their work showed that a lack of communication caused a suboptimal IS design. They suggested that future research be conducted to further investigate this lack of communication situation. Their suggestion provides the impetus for this paper. In this paper, we extend their work to shed more light on the design dilemma situation in AIS design. The results of our model show that under a complete system redesign this lack of communication does not affect the final system quality. Under an enhancement system redesign, however, a lesser warranty is provided leading to a less optimal system choice. The paper is organized as follows. Section 2 describes the model and provides definitions of system replacement and system enhancement. Section 3 provides numerical examples to demonstrate the contracting issues inherent in the replacement and enhancement scenarios. Section 4 provides an analysis of the model results, and Section 5 presents a summary, conclusions, and recommendations for future research.

This paper investigates optimal pricing/warranty arrangements for the design of AISs when the user has alternative sources of information. A mathematical model is developed where the pricing/warranty arrangement is such that the designer credibly commits to putting forth an optimal level of effort when developing the AIS without creating adverse incentives for the user's decision-making efforts. The analysis compares the optimal warranty arrangements when the designer can observe the actions made by the user to the optimal warranty arrangements when the designer cannot observe those actions. The model developed in this paper focuses on two major AIS design scenarios: (1) system enhancement and (2) system replacement. For the enhancement scenario, the user optimally combines signals from the enhanced AIS with signals from the original system. In the systems replacement scenario, the new system replaces the user's original system. Since the signal from the replacement system unambiguously results in a specific action, the payment schedule is the same whether or not the user's actions are observable. Therefore, no losses occur due to unobservability of the user's action under a replacement and a high-quality system is implemented with a simple limited warranty. For the enhanced AIS design, however, the user's action may depend upon the signal from the preexisting original information system. Therefore, when the user's action is not observable, the designer does not know if the user's action is based on the systems enhancement or the original system. Unobservability under an enhancement requires a more complex warranty arrangement. This warranty arrangement is based on the pooling of possible actions taken by the user. Due to risk aversion, the user's expected payment (price net of warranty payments) is higher so that strict losses occur due to unobservability under an enhancement. Nevertheless, a limited warranty does implement quality in the enhanced IS so that the market failure is averted. The research is important from both an academic and practitioner standpoint. From an academic perspective, this paper has shown that a “lemons market” phenomenon can be prevented if the designer provides some form of warranty arrangement. The optimal form of this arrangement is partially dependent on the design task. In a complete redesign, a simple warranty will credibly commit the designer to designing a quality system and the user's moral hazard does not impose additional costs. For an enhancement design, however, through a more complex warranty arrangement, the designer will provide a quality-enhanced system and the lemons market phenomenon can be prevented. This paper extends the findings of Rasch and Yost (1997) by investigating specific design tasks relating to either complete system redesign or system enhancement. Their general finding was that unobservability results in suboptimal contracting between the designer and the user. In the current paper, we refined their analysis and found that in the complete system redesign scenario, unobservability does not result in a suboptimal contract. System enhancement, however, did require a more complex, suboptimal contract. These theoretical findings could provide the motivation for future behavioral research for validation. If our model predictions are correct, we would expect to see a majority of AIS design efforts involving complete system redesign because the designer is more willing to provide an optimal limited warranty. On the other hand, we would expect to see a limited amount of AIS design work involving system enhancements due to the more complex contractual arrangements required. Since our model focused on the managerial decision-making aspects of an AIS, additional research is needed to determine if our findings also extend to the financial transaction processing function of an AIS. This research would entail restructuring the model to determine if the results hold for both AIS functions. From a practitioner perspective, designers should be willing to engage in simple warranty arrangements for complete system redesign. For enhancement scenarios, however, this warranty arrangement becomes more complex and is a function of both the information signal and the user's actions. We note that anecdotal evidence indicates that there are no “standard” warranty arrangements or guidelines that currently exist to aid in structuring warranty contracts for the design of information systems. Further research in needed to investigate the specific form of these warranty arrangements. Practical rules, or guidelines, for the development of these warranty arrangements would help to insure that a market for lemons does not develop in the AIS design industry.