An Explanatory analysis of first-term reenlistment modeling using the PerSMART data warehouse

Abstract

The Navy Bureau of Personnel has constructed a data warehouse (PerSMART) for use by Navy manpower planners. PerSMART is currently building a Retention Monitoring Module (RMM) intended to give Navy manpower planners a way of quickly assessing the impact of current and proposed policies on enlisted retention within the Navy. The purpose of this thesis is to examine the structure of PerSMART and identify possible data and models that could be useful in the construction of the Retention Modeling Module (RMM). The first part of this thesis conducts a literature review of studies looking at civilian and military retention and the effects of compensation, in particular Selective Reenlistment Bonuses (SRB), on retention. The second half of this thesis estimates and specifies a model examining the effects of SRBs on the retention behavior of Zone A sailors at the reenlistment decision point during fiscal years 1995-2001. The model shows a positive relationship exists between SRBs and retention. A one-unit increase in the SRB multiple was found to increase the probability of reenlistment by 3.6 percentage points on average. However, the marginal effect varied across rating groups from 0.8 to 10.4 percentage points. Finally, recommendations are given for future model specifications and sources of data.