Predictability and Forecasting of Acquisition Careers in the Army

Abstract

A great deal is known about the movement of personnel population within large organizations (manpower). On the other hand, far less is known about how individual careers unfold through the structure of such organizations, with no established methods to forecast the positions individuals will take in so-called internal labor markets. In this paper, based on methods from network science, probability, and data analysis, we provide a new, empirically calibrated modeling framework for forecasting careers in large organizations. We show that, without the use of information that goes beyond the memoryless framework provided by Markov models, it is not possible to understand and forecast career moves in an organization. When memory effects are included, models improve significantly and begin to provide both useful predictions as well as information about the limits of predictability in career forecasting. Our method is applied to the Army acquisition workforce.