Background In many developed countries, associations have been reported between higher area unemployment and retirement/ receipt of disability pension. While the main explanation for these relationships has been found to depend on unequal distributions of persons across areas, this argument does not take into consideration life course processes linking area of residence, educational and occupational opportunities and retirement outcomes.

Methods Data was used from 2,441 cohort members of the National Survey of Health and Development with retirement status at age 60–64 years (still in main occupation, retired from main occupation but still working, retired for positive reasons or retired for negative reasons). The main exposure of interest was local authority area unemployment data at ages 4, 26 and 53 years (residential addresses linked to census data in 1951, 1971 and 2001). STATA’s Structural Equation Model (SEM) Builder was used to explore the pathways by which area deprivation in childhood could influence inequalities in retirement status outcomes at age 60–64 years. Pathways considered were occupational socioeconomic position (ages 4, 26 and 53 years), educational achievement (26 years), and work status (ages 26 and 53 years)

Results At age 60–64 years, 37% of respondents were in work, 14% were retired-working, 26% retired for positive reasons and 23% retired for negative reasons. A 1-standard deviation increase in area unemployment at age 53 years was associated with a 12% (95% confidence interval: 1.5%, 23.0%) increased odds of retiring for negative reasons by age 60–64 years, compared to still being in main occupation. This relationship was attenuated by adjustment for SEP at age 53 years. However, occupational SEP at 53 years was directly on a pathway leading from area deprivation in childhood to retirement outcomes. Area deprivation in childhood was not directly related to educational achievement at 26 years. But parental SEP was related to both childhood area deprivation and educational achievement at age 26 years, creating a back-door pathway linking childhood socioeconomic circumstances to retirement outcomes through educational and occupational opportunities.

Conclusion Individual and area-level socioeconomic disadvantages are co-related across the life course and contribute to understanding geographic disparities in retirement age workforce exit. Area effects on retirement outcomes should not be dismissed if associations are attenuated after adjustment for individual social class; as area deprivation (particularly during working years) is an essential contributor to the life course pathway.