The federal civil service workforce2014

Planners and policymakers must be able to assess how compensation policy, including pay freezes and unpaid furloughs, affects retention. This study begins to extend the dynamic retention model (DRM), a structural, stochastic, dynamic, discrete-choice model of individual behavior, to federal civil service employment. Models are developed and estimated,using 24 years of data, and then used to simulate the effects of pay freezes and unpaid furloughs. A permanent three-year pay freeze decreases the size of the retained General Service (GS) workforce with at least a baccalaureate degree by 7.3 percent in the steady state. A temporary pay freeze with pay immediately restored has virtually no impact on retention. When pay is restored after ten years, the retained GS workforce falls by 2.8 percent five years after the pay freeze and 3.5 percent ten years after it. An unpaid furlough, similar to the six-day federal furlough in 2013, has no discernible effect on retention. For all subgroups of GS employees for which the model is estimated, the model fit to the actual data is excellent, and all of the model parameter estimates are statistically significant. In future work, the DRM could be extended to provide empirically based simulations of the impact of other policies on retention; to estimate effects on other occupational areas, other pay systems, or specific demographic groups; or to create a "total force" model (military and civilian) of DoD retention dynamics and the effects of compensation on those dynamics.