Retirement Glide Paths: Fund the Liability, Not the Birthday

Age and target date are useful defaults, but funded status, human capital, spending flexibility, pensions, taxes, fees, and behavior determine a defensible equity path.

Age alone cannot identify a safe equity allocation

What most investors get wrong is treating a glide path as an age-based risk score. A glide path is a predeclared schedule or rule for changing asset allocation through accumulation and retirement. Age is a useful proxy for remaining earning years, but it does not measure risk capacity by itself. The same 60-year-old can be underfunded with fixed spending, fully funded with a pension, still earning substantial labor income, supporting dependants, facing a near-term house purchase, or planning flexible work. Those households should not receive the same stock weight from age alone. Sequence risk also differs by phase. During accumulation, an early loss can be offset by future contributions when those contributions are large relative to the portfolio; a late loss matters more when the portfolio dominates new saving. In retirement, withdrawals after a loss sell more units and can impair recovery. ^[6] But a lower stock allocation does not mechanically make the plan safer: it can reduce drawdown while increasing inflation, longevity, and underfunding risk. Define the liability first—real spending floor, flexible spending, date range, horizon, bequest, tax, liquidity, and currency—then measure funded ratio, contribution capacity, guaranteed income, and tolerance for loss. The equity path is a response to those constraints, not a universal statement about what each birthday permits.

Two funds with the same target year can carry materially different risks

A target-date fund combines a target year with a particular series, share class, glide path, underlying funds, active or index implementation, fees, and assumptions about withdrawals. A 'to' path reaches its most conservative point near the target date; a 'through' path continues changing afterward. The Department of Labor warns that funds with the same year can differ materially in strategies, allocations, risks, and expenses. ^[1] Therefore do not publish timeless claims that Vanguard, Fidelity, or BlackRock always holds a fixed percentage at retirement. Product families and series differ, managers can update paths, and the correct evidence is the current prospectus and dated allocation for the exact fund available in the plan. Inspect current and terminal stock, nominal bonds, inflation-linked bonds, cash, credit, real assets, international exposure, rebalancing, tactical discretion, derivatives, securities lending, expenses, and when the path stops changing. Also aggregate assets outside the fund. Holding a target-date fund beside a stock fund creates a household path different from the label. The target year is an implementation default, not a personalized finding that the investor can bear its loss.

The claimed 10,000-path AIBROKER simulation is not reproducible

The old article states that AIBROKER ran 10,000 resampled paths using 1926–2024 stock and bond returns and then reports indexed wealth and late-career drawdown probabilities for four policies. No input files or versions, return definitions, proxy identifiers, glide-path weights by year, contribution schedule, random seed, resampling code, bond duration, inflation treatment, rebalancing convention, output artifact, or independent reproduction is provided. The numbers—such as a 34% late-career drawdown probability and median wealth index of 100—therefore fail the publication gate and must not be described as AIBROKER analysis. A reproducible study would freeze point-in-time total-return series; specify real contributions and dates; define every annual or monthly weight; preserve serial dependence through justified blocks or model it explicitly; use a recorded seed; deduct fees, tax, and implementation costs; and publish code, configuration, path count, uncertainty, and output hash. Annual returns sampled independently erase sequence structure and do not create a forecast merely because the sample starts in 1926. Compare probability and magnitude of spending shortfall, real terminal wealth, maximum drawdown, recovery time, turnover, and outcomes by predeclared scenarios. Until that artifact exists, the honest conclusion is qualitative: more equity generally raises upside and downside exposure.

Withdrawal research cannot prescribe an accumulation glide path

Bengen and Trinity studied withdrawals under historical US return sequences and specified portfolio and spending rules; they did not establish a permanent stock target for every accumulator. ^[4][7] Their historical success rates are conditional on data, horizon, rebalancing, inflation adjustment, fees, tax omissions, and the definition of success. Four percent is not a guarantee, and a range such as 50%–75% equity is not proof that every point in that range is suitable now. Withdrawal research can illuminate sequence risk, but translating it into a glide path requires current yields, valuation uncertainty, international assets, guaranteed income, annuity choices, spending flexibility, health costs, taxes, and survivor needs. Rising-equity, static, declining-equity, liability-matching, and guardrail policies answer different questions. Evaluate them with the same household cash flows and adverse scenarios; do not choose the rule after observing which historical path wins. Separate plan solvency from behavioral tolerance. A mathematically sustainable allocation that the investor abandons during a decline is not implementable, while discomfort alone does not justify locking in a shortfall.

Three naive de-risking rules fail for different reasons

Failure one is calendar-only de-risking: automatically selling stocks at a birthday can crystallize a bad market, ignore funded status, and reduce expected growth when future saving or pension income already absorbs risk. Failure two is retirement-date cliff risk: remaining heavily exposed until one year before retirement makes the spending start hostage to a single market window, especially when the retirement date and initial withdrawals are inflexible. Failure three is emotional optimization: changing the path after gains or losses turns a policy into performance chasing and makes every backtest irrelevant. None supports the blanket claim that moving to bonds in one's 40s is too early or that a 25%–30% decline is the correct universal tolerance test. Translate the plan into dollars: stress the loss at current balance, continued contributions, job loss correlated with markets, inflation, bond drawdown, and the first five years of withdrawals. Precommit a rebalancing band, maximum equity change per review, cash floor, spending response, and exceptional-change authority. Review after a material liability or human-capital change, not because a headline changed.

A four-option decision tree starts with funded status

Decision tree: first determine whether a low-cost target-date fund's dated prospectus approximately matches retirement window, withdrawal pattern, outside assets, and tolerance. If yes and behavior is the main risk, one fund can be a strong default; verify fees and avoid overlapping allocations. If the household has pensions, concentrated equity, unusual taxes, multiple currencies, a large near-term liability, or a retirement date range, a custom deterministic glide path may fit better. If spending and funded ratio can be measured reliably, a guardrail or liability-aware policy can adjust risk within predeclared bounds. A valuation, volatility, or regime-driven dynamic path belongs only after a point-in-time, out-of-sample, costed test with model-risk limits; sophistication is not evidence. For every option, name decision owner, review date, benchmark, allowed range, turnover cap, tax-lot rule, and rollback. Reject any rule whose benefit depends on an unpublished simulation or frequent judgment calls. The simplest implementable path that funds the objective under stress is preferable to the most elaborate path selected in hindsight.

A 45-year-old with $300,000 still lacks nine required inputs

The uncomfortable implication is that a 45-year-old with $300,000, a 15% savings rate, and a goal to retire at 65 still lacks at least nine decision inputs: income in dollars and stability; annual contribution and employer match; desired real spending; Social Security and pensions; assets and debts outside the account; tax and account types; retirement date flexibility; health, longevity, and bequest needs; and the maximum loss that would change spending or behavior. Fifteen percent of an unspecified income cannot be converted into a contribution, and the balance alone cannot support claims that 70%–75%, 60%, or 50% stock is best. Build a household balance sheet and annual real cash-flow timeline. Calculate funded status under conservative real returns, then stress equity and bond losses, job interruption, inflation, long life, and early retirement. Compare exact candidate funds from current disclosures with a custom path under the same assumptions. Document why a selected equity percentage is affordable and what observable event—not a market forecast—would change it. Practical takeaway: choose the path the household can finance, execute, and retain through stress; do not manufacture precision from age and one balance. Repeat the calculation annually with updated real spending, benefit estimates, account balances, contribution capacity, and the exact fund disclosure. Record whether a change comes from the liability, funded status, or implementation—not from recent performance. If none of those changed, a market move alone is not evidence that the long-term path was wrong.

Related analysis

• Related AIBROKER guide 1
• Related AIBROKER guide 2
• Related AIBROKER guide 3
• Related AIBROKER guide 4
• Related AIBROKER guide 5
• Related AIBROKER guide 6
• Related AIBROKER guide 7
• Related AIBROKER guide 8
• Related AIBROKER guide 9
• Related AIBROKER guide 10
• Related AIBROKER guide 11
• Related AIBROKER guide 12
• Related AIBROKER guide 13
• Related AIBROKER guide 14
• Related AIBROKER guide 15