On the other hand, promotions are also one of the most important rewards in an organization, and, as such, they must be allocated in a way that gives hope and motivation to the maximum number of employees (Dreyfuss 1938; Stinchcombe 1965; Edwards 1979). There are two reasons for this. First, as just noted, promotion chances are likely to be an effective way of controlling employees, offering the possibility of material rewards and symbolic status to a far larger number of people than can possibly receive the actual promotions. Therefore, organizations are likely to benefit in terms of employee compliance and commitment to the extent that promotion chances are extended over as large a group as possible.
Second, U.S. norms about opportunity tend to encourage employees to believe that they retain the chance to improve their status throughout their careers. Turner (1960, p. 860), calling this the "contest mobility norm" writes that it encourages delaying "the final award as long as practicable [in order] to permit a fair race." Although Turner's discussion is primarily applied to schooling, studies suggest that workers and managers often retain expectations of promotions at later ages (Chinoy 1955; Sofer 1970; Kanter 1977). By extending the possibility of promotions to the largest number of employees, an organization avoids the appearance of unfair exclusion of some groups from continuing opportunity.
The efficiency and motivating functions may be contradictory if an organization clearly identifies one group of employees as potentially able to contribute much more than others. For instance, it is a common belief in organizations that young employees can contribute more because of the greater amount of years they may spend in the work force. If such organizations decided promotions solely on efficiency considerations, young employees would be the only ones receiving promotions, in much the way Martin and Strauss describe. However, this would totally neglect the reward aspect of promotions, and it would encourage widespread disengagement by middle-aged and older employees. Consequently, organizations may attempt to reconcile these two conflicting considerations in order to attain a reasonable balance of efficient promotions along with some distribution of promotions by age.
Although the specific arrangement may vary from organization to organization, some general features of a plausible distribution of promotions may be inferred. First, the common belief in the greater potential contribution of youth suggests that organizations will show a marked preference for promotion of youth. Second, motivational and fairness considerations will encourage organizations to allocate some promotions to older age groups, although only a small number may be considered necessary to stimulate motivation and feelings of opportunity. Third, motivational and fairness considerations also encourage organizations to avoid sharp discontinuities in promotion chances between adjacent age groups so that no age group suddenly perceives itself disproportionately deprived relative to its immediately younger cohort (Runciman 1966; Pettigrew 1967; Rainwater 1974). These features may be restated as hypotheses: (1) youth-preference hypothesis- young employees will be given preference over older ones for available promotions; (2) contest mobility hypothesis-older employees will retain some chance of promotion, albeit small; and (3) gradual-decline hypothesis-promotion rates will tend to decline gradually, with no sharp cutoff points beyond which promotions decline precipitously.
Interestingly, one specific form these three hypotheses could take is described by an exponential-decline function. Consistent with these hypotheses, an exponential-decline/unction portrays a curve on which the highest promotion chances occur at the outset, declines are always a fixed proportion of one's current chances, and promotion chances become increasingly rare without becoming impossible (i.e., the curve is asymptotic to the X axis). Of course, proportional declines are only one form of gradual decline, and the exponential-decline function is only one specific form that these hypotheses could take. However, proportional declines may be a particularly appropriate form of gradual decline, for research suggests that deprivations are perceived in proportional, rather than absolute, terms (Rainwater 1974).
However, the efficiency-motivation model is unclear about a number of points. First, it contradicts the stationarity assumption, which posits that mobility rates are independent of time, and it partially contradicts the human capital model, which may be interpreted to predict a modest increase in promotions in early career when individuals make initial investments in their on-the-job training, defer immediate rewards, and subsequently receive increased rewards. Second, it is not clear how employee attributes or organizational levels affect this model. Does this model operate in the same way for college-educated and less-educated employees? Does it operate in the same way at different levels in the organizational hierarchy? Third, it is not clear how this model is affected by changes in organizational growth. During increasing growth, are youth given even more preference, do all groups benefit equally, or are all qualified young employees already promoted so that the additional promotions "spill over" to older employees? During declining growth, are the promotions of older employees preserved or are they sacrificed to maintain the promotions of young employees? These questions are addressed in the following analyses.
Age versus Tenure
After Mincer's (1974) impressive demonstration those years of experience is the more important determinant of earnings than age, some comment is required on the use of age in this analysis. Although Mincer's evidence is compelling, two features of his analysis distinguish it from the present one. He studied earnings, not promotions; and he studied large samples in external labor markets, not internal labor markets. In particular, within the primary segment of the internal labor market, employees tend to remain with firms, and firms' investments in employees will tend to be based on this assumption.
This model contradicts some of the models summarized previously. It contradicts Dalton's (1951) conclusion of a constant promotion rate, and it contradicts the clear cutoff point and precipitous decline predicted by Martin and Strauss's career timetable. However, like Dalton's findings, it suggests that older employees receive some promotions, although at a much lower rate than Oalton suggests; and, like Martin and Strauss's (1959) findings, it suggests that young employees get preference and that substantial promotion declines occur with age, but not to the extent suggested by the career-timetable description. Although it contradicts each model in some respects, this model predicts some of the central elements of each study, suggests the relationship between them, and even provides a theoretical rationale for the Dalton and Martin-Strauss findings.
Analysis of invest training and promotions in employees who have more years to stay in the firm, which is calculated by the number of years until age 65 (discounted by an adjustment for exit rates). Consequently, even human capital theory predicts that age will be an important determinant of promotions in the primary segment of the internal labor market.
The more important theoretical consideration for the present analysis is the sociological analysis of age statuses. Sociologists have amassed a large literature on the status properties of age which tend to affect individuals' social conditions at different times in their lives (Foner 1974,1980; Neugarten 1973; Riley 1977; Kallebergand Loscocco 1983). This viewpoint can be extended by contending that age stratification affects organizations' promotion decisions. In any case, empirically it matters little whether age or tenure was chosen as primary, since the two are very highly correlated and, in-deed, they are inseparable for college-educated employees in these data. Although economic theory does have a great deal invested in whether age or tenure is chosen, it is not clear that sociologists do since tenure could possess status properties.
