This perspective has been extended in several important ways over the past decade. Among jobs in a hierarchy, it is only vacant jobs which offer advancement opportunities. Consequently, patterns of filling vacancies will affect career advancement options within organizational hierarchies. White (1970a) showed how jobs are linked in distinctive vacancy chains. A high-level vacancy tends to be filled by individuals from a limited set of lower-level jobs, creating a new vacancy which in turn must be filled from a limited set of jobs at the next lower level, and so on.
White (1970a, p. 282) notes the obvious limitation of this approach in ignoring individual attributes, and Bartholomew (1968) suggests more complex models that combine the vacancy approach with the analysis of individual attributes. Stewman and Konda (1983) have used such procedures to model the ways that promotions from any position are affected by organizational growth or contraction, cohort size, or exit rate. These models focus on the relationship of organizational structure to individual demography, and they provide techniques for projecting the distribution of individuals in the organization in the future, given hypothetical conditions (Stewman 1981). Indeed, Stewman and Konda (1983) suggest that the pyramid imperative may not always apply, and they show how the ratio of vacancies to lower-level individuals can increase at one level in the hierarchy under hypothetical circumstances despite its decreases in all other levels of the hierarchy. Although this model is far more sophisticated than the pyramid imperative and comes to somewhat different conclusions, it has the same elegant simplicity to explain important parameters of career systems from structural properties.
However, like the pyramid imperative, the vacancy approach is more suited for analyzing the way that shape, growth, and exits influence careers than it is for analyzing the influence of individuals' attributes, the jobs they occupy, or their job histories. While in principle the vacancy approach can be extended to include such factors, and Stewman and Konda show procedures for analyzing a few individual attributes, the procedure quickly becomes cumbersome when several attributes are considered.
The fundamental insight of the pyramid imperative and the vacancy approach is that organizational shapes and vacancies impose important restrictions on the average employee. However, it isn't clear that organizations have any "average employees." Even a very steep pyramid (which narrows greatly at each successive level) may offer substantial advancement opportunity to a small elite subgroup of employees through the highest levels of the hierarchy, and the shape of the organizational pyramid and its number of vacancies may have very little impact on their careers. A small elite group could continue receiving the same high promotion rates even when there are few vacancies. In contrast, if a large segment of the work force (the "secondary labor market") was identified by policy or practice as having no advancement opportunity, then the particular shape of the organizational hierarchy or the number of vacancies would not affect the advancement opportunity of this segment. Indeed, the present study even finds that some segments of the primary labor market are not helped by increasing growth. Obviously, these possibilities can only be ascertained by analyzing various subgroups of employees.
The impact of organizational shape also fails to specify when employees will be selected for advancement, and time is a crucial dimension of careers. Description of organizational hierarchies and vacancy chains can either ignore career timing or adopt assumptions about it, but the question of career timing is an empirical matter which is determined by organizational policy and practice. Even if there are only a limited number of vacancies available, organizations have great latitude in determining whether promotions are to be distributed in the beginning, middle, or end of employees' careers and what this timing will imply about individuals' later career options. While some employee groups may not expect to receive their first promotions until after their first 5 years in the organization, individuals in other employee groups who have not received a promotion in their first 5 years may be unlikely ever to receive one. For some employee groups, an early promotion may indicate the first of much future advancement; for others, an early promotion may be as much advancement as one can expect for one's entire career.
While the shapes of organizational structures surely affect career advancement chances, organizations are not run as chance lotteries. Personnel selection systems introduce many other kinds of constraints in determining which employees will be considered for advancements, when employees will be considered for advancements, and what their rate of advancement will be. Although Stewman and Konda's models provide a powerful method for understanding certain kinds of structural constraints on careers in organizations, analyses must also consider important components of selection systems, particularly how they differentiate employee subgroups and define the timing of careers.
