Abstract

Do traditional two-state worklife estimates need adjustment for unemployment? To answer, an augmented three-state model classifies individuals as either 1) employed; 2) unemployed; or 3) inactive but not marginally attached. Periods of unemployment may reduce worklives; however, removal of those marginally attached or discouraged from the inactive state raises worklives. The three-state model results are compared to worklife estimates from the same initial data using the traditional two-state model. Results show that in many cases, the two-state model results are a good proxy for the three-state results that control for unemployment.