Compensation systems also have important theoretical implications. While economic theory proposes that compensation follows from economically rational behavior, sociological theory proposes that compensation must also be a social system which responds to social forces in the organization. For instance, employees share beliefs about which jobs are equivalent and which are not and about how much difference should exist among jobs. Consequently, compensation systems embody a fundamental conflict: How can organization compensation systems respond both to social considerations and to economic forces? This question underlies the conflict between economic and sociological theories, and it is also a very practical one for compensation systems since economic and social forces are very real considerations in organizations. Organizations must find ways to make their compensation systems respond to these conflicting forces.
The study of an organization's compensation system can help in identifying the way that these theoretical issues are manifested in practice. Furthermore, a reform effort to construct a more rational compensation system was implemented in the corporation studied, and this reform can illustrate another, possibly superior, reconciliation of the conflict and also indicate what aspects of the conflict are so enduring that they are not easily disposed of by reform efforts.
Strangely, economists and sociologists have ignored the compensation literature. In part, this may be because this literature is mostly devoted to claims about the virtues of the various systems and the procedural details of how to implement them. Theoretical implications are not discussed. Thus, social scientists assume that this literature is of little theoretical interest.
However, this reaction is seriously mistaken. The compensation literature explicitly reveals how organizations address the difficult task of determining the value of jobs. Economic and sociological theories offer diverse interpretations of the influence of jobs on compensation, but these various interpretations are speculations, and they are rarely compared with actual procedures. For instance, human capital theory considers the value of a job to be derived from the employee qualifications required to do the job's tasks. Sociological theory considers the value of jobs to be derived from social normative processes. The detailed analysis of job compensation systems can help settle the conflicts by showing exactly what kinds of considerations go into job compensation.
The organization under study provides a particularly interesting case for studying the social processes affecting job compensation because, during the period studied, it shifts from a traditional status system to a rational job evaluation system. This affords an opportunity to study the operation of each of these systems and the transition between them.
Moreover, job evaluation is a particularly interesting compensation procedure to consider because it is an explicit attempt to move compensation practices from traditional status considerations to economic rationality. Job evaluation is also of theoretical interest, for it concretely illustrates the conflict and tension between economic and social forces which operate in compensation systems.
The study of a job compensation system must consider how jobs affect earnings. Few studies have been able to do this because the survey data generally available to researchers rarely have good indicators of jobs. This is a particularly glaring omission for sociological research, which stresses the importance of job structures. Sociologists have studied the ways that occupational status categories affect earnings, but it is generally recognized that jobs are the units that determine earnings, and sociologists would study jobs if it were possible to identify jobs in their data (Featherman 1973; Kelley 1973a; Baron and Bielby 1980). The data available here offer an unusual opportunity to study the properties of jobs and to analyze their influence on earnings.
Although jobs are considered the basic units of attainment, here is really very little known about them. A typical model of organizations assumes that organizations are constructed of a fairly fixed set of jobs and that these jobs have relatively consistent ranking and value, but these assumptions are rarely tested. Consequently, the analyses that follow address very simple, basic questions. Is an organization composed of a constant set of jobs? To what extent do new jobs appear and old jobs disappear? Are jobs stably valued? To what extent does the value of jobs vary over changing social and economic circumstances?
Second, since jobs are defined by a fairly fixed set of tasks, the common expectation is that they will tend to require the same kinds of individual education, skills, and experience over long periods of time. To what extent are jobs filled by the same kinds of individuals over time?
Third, little empirical analysis has been done to investigate the determinants of job salaries. By ignoring social structural features of employment, human capital theory implies that salary differences among jobs can largely be explained by the qualifications of the individuals in the jobs, that is, the human capital in jobs. In contrast, structural theory contends that a job's structural position has large effects on the job's value independent of its human capital composition (Thurow 1975). In addition, as shall be elaborated subsequently, the two theories also posit different relationships between individual attributes and job salaries, and they posit differential responsiveness to economic forces. The present analyses test these conflicting contentions.
Finally, the job evaluation system studied here, implemented as part of a large affirmative action effort, had the Analyses are done to determine to what extent job salaries are affected by the gender composition of the job holders, independent of the jobs' human capital composition, and how this changes over time. These analyses also investigate the extent to which the job evaluation system improves the compensation of predominantly female jobs. The results have implications for assessing the potential benefits and pitfalls of job evaluation systems for reducing the structural sources of sex bias (Treiman and Hartmann 1981).
