Poterba Rauh Venti and Wise 2006 — Lifecycle Asset Allocation Strategies and the Distribution of 401(k) Retirement Wealth

Summary

Using Health and Retirement Study (HRS) earnings histories for 1,400 households linked to simulated return sequences drawn from the historical return distribution, Poterba, Rauh, Venti, and Wise (2006) compare retirement wealth distributions under eight portfolio allocation strategies — from all–Treasury Inflation-Protected Securities (TIPS) to all-equity to several lifecycle fund specifications. The central finding is that lifecycle funds produce retirement wealth distributions nearly identical to age-invariant strategies with the same average equity share; the glide path adds little beyond its average allocation. Whether any given strategy dominates depends critically on three parameters: the expected equity premium, the worker's risk aversion, and the amount of non-401(k) wealth at retirement.

Key Claims

• Lifecycle-vs-invariant equivalence: A lifecycle fund averaging 53% equities over the career produces virtually the same mean retirement wealth (~$405k) and nearly the same variance as a fixed 53% age-invariant strategy; the time-varying equity share of the glide path is not the value-relevant feature
• Equity premium dominance (historical returns): All-equity portfolios yield mean 401(k) wealth of ~$812k for high school (HS)-educated workers vs. ~$193k (bonds) and ~$163k (TIPS); even the 10th-percentile all-equity outcome exceeds the mean bond outcome
• Equity dominance is precarious: Reducing assumed equity returns by 300 basis points flips the expected utility ranking; lifecycle and conservative strategies gain substantially relative to all-equity under lower equity premia
• Risk aversion threshold: At low risk aversion (constant relative risk aversion [CRRA] coefficient ≤ 2–3), all-equity dominates under historical returns; at high risk aversion or with substantial non-401(k) wealth that provides background insurance, lifecycle strategies generate comparable or higher utility
• Non-401(k) wealth interaction: Non-401(k) wealth (Social Security, defined benefit [DB] pensions, housing equity, financial assets) functions as a buffer that reduces the marginal utility of avoiding 401(k) downside risk; workers with substantial non-401(k) wealth benefit less from conservative allocations
• Education gradient in 401(k) balances: Large and striking — college-educated workers have mean projected 401(k) wealth ~9× that of less-than-HS workers ($331k vs. $37k for TIPS strategy; $— vs. $— for all-equity strategy), driven entirely by the earnings/contribution gradient, not investment behavior
• Lifecycle fund industry description (2005): 40+ funds across 10+ families; equity share ranges from 88% (35 years to retirement) to 30% (at retirement); weighted average expense ratio 0.6–0.8%

Concepts Introduced or Extended

• Lifecycle Asset Allocation

Entities Mentioned

• James Poterba
• Joshua Rauh
• Steven Venti
• David Wise
• Health and Retirement Study

Quotes

"The distribution of retirement wealth associated with typical lifecycle investment strategies is similar to that from age-invariant asset allocation strategies that set the equity share of the portfolio equal to the average equity share in the lifecycle strategies."

"The expected utility associated with different 401(k) asset allocation strategies, and the ranking of these strategies, is very sensitive to three parameters: the expected return on corporate stock, the worker's relative risk aversion, and the amount of non-401(k) wealth that the worker will have available at retirement."

My Take

The equivalence result — lifecycle funds are no better than fixed allocations with the same average equity share — is the paper's sharpest finding and has direct policy implications for the target-date fund industry. The glide path structure is marketed heavily as managing longevity risk, but the simulations suggest most of the variation in outcomes is driven by the overall level of equity exposure, not its age-tapering. The sensitivity to the equity premium is a significant caveat: the historical 5–7% equity premium may not persist, and with lower expected returns the advantage of equity-heavy strategies collapses. The education gradient finding — 9× difference in projected 401(k) wealth between college and HS-dropout households — reflects the contribution gradient rather than investment behavior, pointing to the fundamental inequality in defined contribution (DC) pension coverage for low-wage workers.