Definition
Post-retirement asset drawdown refers to how total household wealth — including housing equity, retirement accounts, and financial assets — evolves after retirement. Contrary to the standard life-cycle hypothesis prediction of smooth drawdown, empirical evidence shows that assets typically grow through retirement for stable households and decline sharply only at family-status shocks (primarily divorce; widowhood has a smaller, often insignificant effect).
Key Ideas
- Stable households accumulate: Continuously married and continuously single households grow total assets throughout the retirement period — +4–10% per biennial interval for Health and Retirement Study (HRS) cohorts aged 51–75.
- Divorce triggers large declines: The 2→1 (divorced) transition is associated with −27% to −48% asset declines, robust across trimmed-mean and median estimators and statistically significant.
- Widowhood effect is ambiguous: 2→1 (widowed) households have lower assets before the transition (55–65% of continuing married households), and the transition itself causes a small, usually statistically insignificant change for younger cohorts; assets decline for the oldest Asset and Health Dynamics Among the Oldest Old (AHEAD) 70–80 cohort.
- Pre-transition selection: Future widows/divorcees already hold lower assets years before the transition, not just at the time of the shock. Failure to condition on initial asset levels confounds level differences with transition effects.
- Latent health gradient dominates: The ratio of total assets between the top and bottom health quintile (principal component analysis (PCA)-based index of 28 indicators) grows from 1.7 to 2.2 over 1992–2006 for HRS 56–61 married households; the ratio for single-person households rises from 2.8 to 4.1 over the same period.
- Cohort effects: HRS cohort households (born ≈1935–45) held far higher real assets at age 70–75 in 2006 than AHEAD cohort households (born ≈1915–25) held at the same age in 1993.
How It Works
Poterba, Venti, and Wise (2010) track households across HRS and AHEAD waves (biennial) and Survey of Income and Program Participation (SIPP) (annual), classifying each biennial interval by the household's family-status transition (2→2, 2→1 widowed, 2→1 divorced, 1→2, 1→1). Generalized least squares (GLS) regressions with interval fixed effects on trimmed data (top/bottom 1% residuals trimmed) produce smoothed asset trajectories. A latent health index is constructed by taking the first principal component of 28 self-reported health indicators accumulated through the beginning of each interval, then grouping households into quintiles. All asset values are in constant 2000 dollars.
Why It Matters
- Challenges simplistic over-saving or under-saving narratives: rising assets through retirement are consistent with rational precautionary saving against uncertain health costs, not with under-spending.
- The health gradient finding motivates the formal pathway analysis in Education and DI Participation and related work: health is not just a determinant of mortality but of wealth accumulation rates throughout retirement.
- Informing policy on retirement adequacy: the typical household is not rapidly drawing down at, say, age 70; the real vulnerabilities are divorce-induced wealth shocks and the health-poor tail.
Open Questions
- How much of the stable/rising asset profiles reflects selection (healthier, wealthier households survive) vs. genuine accumulation?
- To what extent does annuity income (Social Security, defined benefit (DB) pensions) allow households to hold non-annuity assets stable rather than drawing them down?
- Does the self-insurance motive for holding assets weaken once Medicare/Medicaid coverage expands, and if so, by how much?
Related
Sources