Identifying predictors of persistence and recurrence of depression in individuals with a major depressive episode (MDE) poses a critical challenge for clinicians and researchers. We develop using a nationally representative sample, the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC; N = 34,653), a comprehensive model of the 3-year risk of persistence and recurrence in individuals with MDE at baseline. We used structural equation modeling to examine simultaneously the effects of four broad groups of clinical factors on the risk of MDE persistence and recurrence: 1) severity of depressive illness, 2) severity of mental and physical comorbidity, 3) sociodemographic characteristics and 4) treatment-seeking behavior. Approximately 16% and 21% of the 2587 participants with an MDE at baseline had a persistent MDE and a new MDE during the 3-year follow-up period, respectively. Most independent predictors were common for both persistence and recurrence and included markers for the severity of the depressive illness at baseline (as measured by higher levels on the general depressive symptom dimension, lower mental component summary scores, prior suicide attempts, younger age at onset of depression and greater number of MDEs), the severity of comorbidities (as measured by higher levels on dimensions of psychopathology and lower physical component summary scores) and a failure to seek treatment for MDE at baseline. This population-based model highlights strategies that may improve the course of MDE, including the need to develop interventions that target multiple psychiatric disorders and promotion of treatment seeking to increase access to timely mental health care.