|
Article Excerpt
ABSTRACT
This paper investigates the interaction between compensation structures and expert system technology. One hundred twenty-two participants selected a contract (fixed pay or performance-contingent incentives) and performed one of two risk-rating tasks either in the presence or absence of an externally valid expert system. For both tasks, the expert system increased decision accuracy only for participants selecting performance-contingent incentives, and these incentives increased decision accuracy only in the presence of the expert system, consistent with a complementary interaction. The results support the view that expert system technology changes task requirements and, hence, may also change selfselection and effort compensation contracting effects. The results provide evidence about the interdependence of organization, system, and individual factors in determining task performance, and suggest that organizational performance may be improved through the joint implementation of expert systems and per formance-contingent incentives.
Keywords: performance-contingent incentives; complementarity; information technology; expert systems.
Data Availability: Contact the author.
INTRODUCTION
An important use of accounting information is the development of compensation structures with monetary incentives that motivate individuals to improve decisions (Young and Lewis 1995). The usefulness of such systems depends on how effective they are at motivating individuals to increase effort and performance (Awasthi and Pratt 1990). Bonner et al. (2000) find that the effectiveness of incentives depends on individual, task, and environmental variables, and argue that research examining these interactions is important because monetary incentives may be one of the most costly methods for attempting to improve decisions.
This study extends contracting research by investigating the interaction between information technology (IT) and compensation. Bama and Whinston (1998) suggest a complementary interaction, defined as the mutual arrangement of factors that have a value synergy with each other such that when combined correctly they have a significant positive impact on performance. In order to better understand and test for such a complementary interaction, specific changes in task requirements in the presence of IT are used to predict specific changes in contracting effects. An expert system (1) is used to operationalize IT because it is implemented at an individual level consistent with the types of compensation structures utilized in the study. In addition, prior research suggests that expert systems often fail to realize expected benefits (Gill 1996). This study may also have implications for improving expert system implementation.
Using Milgrom and Roberts's (1995) formal requirements for complementarity, this study predicts that the marginal return to the expert system is larger in the presence of performance-contingent rather than fixed pay contracts. Further, to maintain symmetry, the marginal return to performance-contingent contracts is predicted to be larger in the presence of the expert system. Marginal return is defined as decision accuracy. An experiment was conducted to test the predicted complementary interactions. Participants were assigned to perform one of two loan evaluation tasks, either with or without the option of a commercially available expert system designed for the task. After a training session, participants selected either fixed pay or one of several performance-contingent contracts and made four case decisions.
Contracting theory, used to develop specific predictions, suggests that compensation affects the type of individual attracted to an organization (referred to as self-selection), and the amount of effort exerted by the individual once on the job (Wruck 2000). The expert system should impact both. For self-selection effects, the system changes task requirements, prompting more in-depth analysis of self-efficacy, and provides a detailed, structured model of the task, allowing a more accurate assessment of task-related skill. Therefore, the study predicted and found a stronger relationship between contract selection and skill with the expert system.
For effort effects, the expert system changes task requirements by structuring the task to overcome selective information acquisition biases. Prior research finds that individuals exhibit a tendency to decrease effort in the presence of this increased structure (Todd and Benbasat 1992; Gill 1996). The study predicted that incentives in performance-contingent (but not fixed pay) contracts would provide motivation to maintain high effort with the expert system. Results support predicted effort effects for intensity of effort, but not duration of effort.
Combined, these self-selection and effort effects suggest the predicted complementary interaction between expert systems and compensation. Comparing the presence to the absence of the expert system, participants selecting performance-contingent (but not fixed pay) contracts significantly increased accuracy, suggesting that the expert system is more effective in the presence of performance-contingent compensation contracts. Comparing performance-contingent to fixed pay contracts, accuracy significantly increased with, but not without, the expert system, suggesting that performance-contingent contracts are more effective in the presence of the expert system. The combination of these two accuracy effects supports a complementary relationship between expert systems and compensation structure.
The current study also developed new forms of performance-contingent compensation contracts based on inputs to the expert system. Self-selection, effort, and accuracy were comparable to the profit-based performance-contingent contracts in the expert system condition. These results demonstrate that IT may be used to design contract incentives based on new input-based performance measures, not practically available without IT, that elicit high effort and decision accuracy at a lower cost than profit-based contract incentives due to improved risk sharing.
This study contributes to our understanding of the importance of task changes on the effectiveness of incentives. The study also contributes to our understanding of the adoption of expert systems and demonstrates how incentives may be used in concert with technology to improve performance. Finally, the study extends prior research by examining self-selection effects in addition to effort effects. Bonner and Lewis (1990) argue that skill is an important precondition to being able to use effort to improve performance on a task, and Prendergast (1999) concludes that a significant portion of the impact of compensation on performance is from contract selection. Yet, with few exceptions (Waller and Chow 1985), accounting studies have focused on motivation (effort) effects (Awasthi and Pratt 1990; Libby and Lipe 1992; Sprinkle 2000; Stone and Ziebart 1995; Tuttle and Burton 1999). Bonner et al. (2000) suggest that the role of skill in incentives is a major issue that warrants additional research.
The remainder of this paper is organized as follows. The second section reviews prior literature linking IT and compensation and then uses the theory of complementary management choices to develop hypotheses. The third section summarizes the experimental methods. The fourth section presents results, and the final section presents conclusions and limitations.
COMPLEMENTARITIES, COMPENSATION STRUCTURES, AND EXPERT SYSTEMS
Almost 20 years of research examining whether IT investment leads to productivity improvement provides inconclusive results. Brynjolfsson (1993) reviews 18 and Chan (2000) reviews 41 prior studies of IT investment and productivity research. Both find mixed and inconclusive results, documenting both positive and negative associations between measures of IT investment and productivity. Brynjolfsson and Hitt (1998, 55) argue that it is difficult to establish the overall value of IT because of large variations between firms and conclude:
Computerization does not automatically increase productivity, but it is an essential component of a broader system of organizational changes, which does increase productivity. As the impact of computers becomes greater and more pervasive, it is increasingly important to consider these organizational changes as an integral part of the computerization process.
Recent research at a macro level finds that productivity gains are the result of a deliberate alignment between management's organizational and technological choices (e.g., Tallon et al. 2000; Francalanci and Galal 1998; Clark and Stoddard 1996).
Compensation is one of management's organizational choices beginning to be linked to IT performance. Sviokla (1996) suggests that compensation structure influences the adoption of new expert system technology in the insurance industry. Anecdotal evidence in the banking industry suggests that poorly designed compensation plans can seriously impair a firm's ability to achieve optimal returns on technology investments (Price et al. 1996). Hammer (1996) explicitly argues that to achieve high performance with changed processes implemented through IT, compensation must also change, from "pay for showing up to pay for results." Finally, Barua and Whinston (1998) argue that compensation structure is one of six classes of organizational factors that exhibit important complementary relationships with IT.
Milgrom and Roberts (1995) formally apply the neoclassical theory of complementarities in optimization and games to a firm's strategy, its structure, and its managerial processes. Two variables are considered complementary if doing more of each of them increases the returns to doing more of the other. The theory of complementary management choices has been applied to costing systems and incentive compensation structures (Drake et al. 1999), human resource management practices (Ichniowski et al. 1997), and capital budgeting practices and manufacturing technologies (Miller and O'Leary 1997). This section begins by describing expert systems; the IT used in the study. Then, specific self-selection and effort interaction effects between expert systems and compensation structure are predicted, which result in a final hypothesized pattern of decision accuracy consistent with Milgrom and Roberts's (1995) definition of complementarities.
Expert Systems
An expert system is used in this study for two reasons. First, the theory of complementary management choices requires an alignment between specific choices. Since this study examines individual-level compensation, an expert system was chosen because it represents a type of IT that also generally operates at the individual level of analysis. Chan (2000) suggests that research using individual measures could contribute more specific knowledge about when and how IT provides value.
Second, expert systems are commonly implemented in practice (Brown and Phillips 1995) with demonstrated difficulties suggesting that alignment with managerial choices may contribute to an understanding of variations in success. Gill (1996) finds that only 25 percent of expert systems continued to be used for more than five years after they were built, even though technical performance was not a problem. Gill (1996) and Sviokla (1996) specifically suggest that compensation structure contributes to the successful long-term implementation of expert systems, but do not provide details about the exact nature linkages should take. Contracting theory provides a basis for developing specific linkages.
Expert System and Compensation Structure Contract Effects
An organization's compensation structure serves as both a selection and motivation (effort) device (Wruck 2000). Waller and Chow (1985) develop a framework for contracting effects to demonstrate that selection and effort are highly related because they both stem from the interrelated processes of contract design by organizations and contract selection by workers.
Self-Selection Effects
Employment contracts vary by type of rewards, determination of compensation, and measurement of performance. Potential employees evaluate contracts based primarily on their perceived skill level (Waller and Chow 1985). Prior research consistently finds that a compensation structure with strong performance-contingent (PC) incentives attracts individuals who have the skill2 to perform in a way that will allow them to earn the payoff. Conversely, individuals who do not have such skills are attracted to fixed salary (FP) contracts (Wruck 2000; Prendergast 1999; Young and Lewis 1995; Waller and Chow 1985). Waller and Chow (1985) also point out barriers to this optimal contracting relationship. Workers' perceptions of their own task-related skill may be vague or wrong such that their contract choice is not consistent with their actual task-related skill level. Waller and Chow (1985) attribute some turnover phenomena to the revelation of this sort of mismatch after contract choice.
Waller and Chow's (1985) perceived task-related skill construct is conceptually consistent with the psychological construct of self-efficacy, defined as a person's estimate of his/her own capacity to perform a specific task (Gist and Mitchell 1992). In assessing self-efficacy, an individual must analyze the task requirements, analyze his/her own abilities and experiences in performing these specific task requirements, and assess the availability of specific resources and constraints. Prior research finds that individuals of all skill levels often do a poor job of predicting their own performance (Gist and Mitchell 1992).
Bandura (1982) suggests that individuals determine self-efficacy through personal task attainments, verbal persuasion, physiological arousal, and vicarious experience or modeling. The presence of the expert system should improve the accuracy of self-efficacy by providing a model of the task requirements, which gives the individual additional information not available without the expert system. Todd and Benbasat (1999) argue that a decision support system's structuring of a task...
|
