Retirement and depressive symptoms: A 10-year cross-lagged analysis

Dikla Segel-Karpas, Liat Ayalon, Margie E. Lachman

Research output: Contribution to journalArticlepeer-review


The effect of retirement on depressive symptoms remains a subject of scientific inquiry, given the fact that previous studies have found mixed results. Moreover, the possible effect of depressive symptoms on the propensity to retire remains relatively understudied. Given the sheer number of retirees, and the significance of depressive symptoms for individuals' well-being and ability to work, as well as for societies at large, we used a large longitudinal dataset to examine the reciprocal effects of retirement on depressive symptoms, and of depressive symptoms on the propensity to retire. Using six waves of the Health and Retirement Study (HRS) data collected over a period of 10 years (N = 6584), we tested cross-lagged models of the reciprocal relationships between retirement and depressive symptoms. The analysis revealed that retirement results in increased depressive symptoms, and that depressive symptoms increase the likelihood of retirement. No sex differences in the lagged associations were found. We conclude that depressive symptoms are a risk factor for retirement, and practitioners should try and identify older workers suffering from depression prior to the retirement transition. Similarly, as retirement increases depressive symptoms, the transition should be treated as an important and sometimes risky milestone, where adequate preparation is required.

Original languageEnglish
Pages (from-to)565-570
Number of pages6
JournalPsychiatry Research
StatePublished - Nov 2018

Bibliographical note

Publisher Copyright:
© 2018


  • Cross-lagged
  • Depressive symptoms
  • Longitudinal
  • Retirement

ASJC Scopus subject areas

  • Psychiatry and Mental health
  • Biological Psychiatry


Dive into the research topics of 'Retirement and depressive symptoms: A 10-year cross-lagged analysis'. Together they form a unique fingerprint.

Cite this