Episode 350: Scott Cederburg: A Critical Assessment of Lifecycle Investment Advice

Episode Summary

Executive Summary: The episode revisits Scott Cederberg’s updated lifecycle investing paper, which uses 39 countries of real return data and block bootstrapping to test optimal asset allocation across life. The central result remains strikingly stable: a 100% equity, globally diversified portfolio dominates common target-date and 60/40 strategies in utility, retirement income, and ruin risk, with only small, context-dependent deviations from all-equity near retirement, under high valuations, leverage, or different spending rules.

Main Topics: Lifecycle asset allocation and the paper’s core setup (Priority: 5/5): Scott explains the paper’s goal: determine how investors should save and invest across working life and retirement using real, long-horizon data from developed countries, with stocks, international stocks, bonds, and bills. Why the updated version matters: robustness checks and what-if scenarios (Priority: 5/5): The revision adds many alternatives requested by critics: different valuations, leverage, correlations, spending rules, age-varying allocations, country exclusions, U.S.-special assumptions, and other sensitivity tests. All-equity global diversification versus target-date and 60/40 benchmarks (Priority: 5/5): The study finds the optimal fixed portfolio is all equity—domestic plus international—while standard target-date funds and 60/40 allocations are materially worse on utility and retirement outcomes. The role of bonds, sequence risk, and retirement spending rules (Priority: 4/5): Bonds appear attractive at monthly horizons, but long-horizon simulations show they become riskier and less diversifying over time; limited bond use near retirement mainly reflects sequence-of-returns risk under the 4% rule. Valuations, U.S. specialness, and domestic vs. international tilts (Priority: 4/5): Conditioning on high or low domestic valuations shifts weights modestly, mostly between domestic and international stocks; the paper also tests whether the U.S. being ‘special’ would justify higher U.S. weights. Leverage, borrowing costs, and why it usually doesn’t change the main conclusion (Priority: 4/5): When leverage is allowed, optimal portfolios sometimes lever the same all-equity mix, but results are highly sensitive to borrowing costs and practical constraints, and the authors caution against literal implementation. Resilience of the results and reactions from academia/practice (Priority: 3/5): Despite criticisms about sample choice, Germany, World War II data, small countries, and correlations, the main findings remain stable; Scott says the most surprising result is the robustness across specifications.

Key Arguments: Global developed-country data provide a better basis for long-horizon retirement analysis than U.S.-only samples, which are too short to cover 75-year horizons. Using a block bootstrap preserves important return features—volatility clustering, mean reversion, skewness, and fat tails—without forcing strong parametric assumptions. Bonds look good at short horizons, but over long horizons they can become riskier and more correlated with stocks, reducing their diversification value. An all-equity globally diversified portfolio outperforms target-date and 60/40 benchmarks not just on average wealth, but also on worst-case retirement outcomes and ruin risk. The utility advantage of the optimal all-equity strategy is large enough that benchmark investors would need to save much more to match it. The small bond allocation near retirement in the base case is driven mainly by sequence-of-returns risk under rigid 4% withdrawals, not by an enduring advantage of bonds. Allowing age-varying allocations changes little: the optimization is nearly identical to the fixed-weight solution except for a short-lived retirement-period bond/bill blip. Valuation timing mostly changes the split between domestic and international stocks; bonds matter only at the extremes of valuation. If investors believe the U.S. is special, they need more than a 50% belief to justify meaningful home bias; otherwise global diversification remains favored. Leverage can improve utility if borrowing is cheap enough, but the practical and crisis-time risks make it a fragile recommendation rather than a broad policy prescription.

Data Points: Countries in sample: 39 - Developed countries used in the dataset Time span: 1890 to 2023 - Historical return sample period Country-years: Over 2,650 - Observation count across countries and years Asset classes modeled: 4 - Domestic stocks, international stocks, bonds, and bills Average real bond return: 1% per year - Sample average cited for bonds Average real international stock return: 7% per year - Sample average cited for international equities One-year horizon variance vs. 30-year horizon for bonds: 2.3x higher on a per-period basis - Long-horizon bond risk increases relative to short horizon One-year horizon variance vs. 30-year horizon for stocks: 0.75x on a per-period basis - Stocks become relatively safer over long horizons due to mean reversion Target-date fund default in employer plans: 98% - Vanguard plans default statistic mentioned Vanguard investors with at least one TDF: 83% - Adoption statistic cited in the discussion Vanguard investors with all money in one TDF: 58% - Concentration in a single target-date fund Optimal fixed-weight portfolio: 33% domestic stocks / 67% international stocks / 0% bonds / 0% bills - Base-case optimal allocation Equivalent savings rate for optimal strategy: 10% actual savings rate baseline - Base-case comparison setup Equivalent savings rate for 60/40 strategy: 19% - Needed to match utility of 10% saved in optimal strategy Equivalent savings rate for TDF: 16.1% - Needed to match utility of 10% saved in optimal strategy TDF ruin probability in retirement: Almost 20% - 4% withdrawal rule with target-date fund All-stock ruin probability in retirement: About 7% - 4% withdrawal rule with globally diversified equities Age-varying optimal portfolio: About 99% equity on average - Dynamic optimization result Retirement-age temporary bill allocation: 27% in bills at age 65 - Short-lived bond tent under the age-varying strategy Back to >90% stocks: By age 68 - Age-varying allocation quickly reverts toward equities Near-100% stocks again: By age 70 - Age-varying strategy after retirement blip Equivalent savings rate for age-varying strategy: 9.93% - Compared with 10% for fixed-weight optimal strategy Lowest valuation quintile domestic allocation: 65% domestic / 35% international - When domestic prices are low Highest valuation quintile allocation: 16% domestic / 75% international / 9% bonds - When domestic prices are high Equivalent savings rate with valuation conditioning: 9.7% - Improvement versus fixed-weight strategy Share of valuation-conditioning gains from domestic/international tilts: About 80% - Most of the benefit comes from shifting between domestic and international stocks Small-country exclusion result: 26% domestic / 73% international / 0% bonds / 0% bills - When excluding countries smaller than 0.5% of world market or similar screens Germany excluded result: 34% domestic / 66% international / 0% bonds / 0% bills - Robustness to removing Germany Pre-WWII data excluded result: 31% domestic / 69% international / 0% bonds / 0% bills - Robustness to removing pre-World War II observations Exclude U.S. from international pool: 45% domestic / 55% international / 0% bonds / 0% bills - Effect of removing U.S. ex post winner from sample Correlation effect on domestic stock weight: -3% domestic weight per +0.1 correlation - Higher labor-income/stock correlation shifts portfolio away from domestic stocks Mid leverage cost scenario: 1.4% spread above Treasuries - Lowest plausible retail borrowing rate discussed Low leverage cost scenario: 0.37% above Treasury bills - Futures-implicit borrowing cost estimate High leverage cost scenario: 12% margin rate - Retail margin cost example from E-Trade Optimal leverage at 1.4% spread: 55% leverage - Portfolio remains 34% domestic / 66% international / 0% bonds / 0% bills Optimal leverage at 0.37% spread: 100% leverage - Capped leverage case with some bonds (15%)

Pivotal Quotes: "you should never own bonds, you just invest in stocks internationally diversified throughout your entire lifetime." — Scott Cederberg: Scott’s headline takeaway on the optimal lifecycle portfolio "We were mostly surprised by how resilient the results are to running all of these alternative specifications or different sets of assumptions or whatever." — Benjamin Felix: Discussion of robustness after extensive sensitivity analysis "The all-stock strategy is still safer than something like a TDF." — Benjamin Felix: Summarizing the retirement ruin-risk comparison between equities and target-date funds

Implications: For DIY investors and plan sponsors, the paper challenges bond-heavy defaults and supports globally diversified equity as the baseline. It also suggests most objections are robustness questions, not refutations. The bigger issue is implementation risk: leverage, timing, and withdrawal rules matter a lot.