Current rules restrict Series I savings bonds to TreasuryDirect accounts with a $10,000 annual cap, locking them out of tax-advantaged 401(k)s and IRAs where most households accumulate wealth. Using a 100,000-path simulation, this analysis shows that allowing a modest I-bond allocation inside retirement accounts provides meaningful downside protection against sequence-of-returns risk during severe inflation shocks by eliminating the duration risk inherent in traditional bond funds.
This analysis considers changes to the rules governing purchases of Series I savings bonds and the portfolio decisions those changes could make possible.
Currently, individuals purchase I bonds through TreasuryDirect, subject to a $10,000 annual purchase limit per person. I bonds are not a standard investment option within IRAs or 401(k) plans. While some plans offer TIPS funds, pooled funds carry perpetual duration risk and lack a fixed maturity date, exposing retirees to mark-to-market losses during rate spikes. Individual Treasury laddering is rarely available in standard plan menus. For a deeper examination of how I bonds compare to TIPS and their unique structural advantages, see this prior analysis on Series I Bonds and Practical Guidance.
The proposed policy options are to allow Series I savings bonds inside tax-advantaged retirement accounts, increase allowable purchases, or do both. Allowing purchases within retirement accounts is the substantive change: it gives households that save primarily through retirement plans access to a non-marketable, zero-floor inflation hedge without interest-rate or duration risk. Increasing purchase limits alone leaves this access barrier intact.
The analysis examines how three portfolios affect retirement wealth and risk under ordinary inflation conditions and conditions with larger inflation shocks. The all-stock portfolio provides a benchmark for assessing the protection offered by traditional bonds; comparing the two balanced portfolios isolates the effect of replacing some traditional bonds with I bonds.
Initial wealth: $1,000,000 for each retiree.
Withdrawals: $40,000 in year one, paid monthly at approximately $3,333.33. Retirement starts in February; payments rise 3% each subsequent February. Total scheduled withdrawals over ten years are $458,555.
Portfolio one: 100% S&P 500.
Portfolio two: 60% S&P 500 and 40% broad U.S. investment-grade bonds.
Portfolio three: 60% S&P 500, 30% broad U.S. investment-grade bonds, and 10% Series I savings bonds.
Base inflation scenario: annual inflation averages approximately 3%, with a standard deviation of 1.5 percentage points. Inflation is persistent and positively skewed, allowing occasional large increases.
Larger inflation shocks scenario: the inflation-shock scale doubles, increasing the annual inflation standard deviation to approximately 3 percentage points while average inflation remains near 3%. Large positive inflation surprises impose disproportionately greater losses on traditional bonds.
Rebalancing: compare resetting each portfolio to its original weights every February with allowing weights to drift. In the base inflation scenario, rebalancing reduces the spread of outcomes for both balanced portfolios, raising median and 10th-percentile wealth while lowering 90th-percentile wealth. It has no effect on the 100% S&P 500 portfolio.
Outcomes: report the worst simulated ending balance and the 1st, 10th, 25th, 50th, 75th, and 90th percentiles for every portfolio. These results show both protection against low ending balances and wealth forgone when investment outcomes are favorable.
Key Findings
The largest reduction in downside risk comes from moving from an all-stock portfolio to either balanced portfolio, at the cost of substantial upside. Replacing some traditional bonds with non-marketable I bonds provides additional protection at the lowest percentiles—particularly during rate spikes that hurt traditional bond funds—with little change at the 25th percentile and generally lower median and upper-percentile wealth.
All-stock benchmark: with annual rebalancing under base inflation, 10th-percentile wealth is approximately $493,000 for all stocks, compared with $677,000 for 60/40 and $680,000 for 60/30/10. All stocks produce higher median wealth—$1,233,000 versus $1,169,000 and $1,162,000—and much higher 90th-percentile wealth: $2,758,000 versus $1,959,000 and $1,937,000.
Incremental effect of I bonds: relative to 60/40, the rebalanced I-bond portfolio adds approximately $5,400 at the 1st percentile and $2,600 at the 10th, but reduces median wealth by $6,300 and 90th-percentile wealth by $22,300.
Larger inflation shocks: the rebalanced I-bond portfolio’s advantage over 60/40 increases to approximately $26,900 at the 1st percentile and $17,700 at the 10th. Median wealth is approximately unchanged, while 90th-percentile wealth remains $27,400 lower.
Rebalancing: in the base case, annual rebalancing reduces the spread of outcomes for both balanced portfolios, raising median and 10th-percentile wealth while lowering 90th-percentile wealth. By definition, rebalancing has no effect on the all-stock portfolio.
Methods
The analysis uses 100,000 simulated market paths, each covering 120 months. Withdrawals occur at the beginning of each month and increase by 3% each February. All portfolios experience the same market conditions. Taxes and management fees are excluded to isolate core risk dynamics; holding I bonds inside tax-deferred accounts eliminates annual interest tax drag relative to taxable holdings.
In the base case, annual stock returns have an assumed arithmetic mean of 8% and standard deviation of 18%. Traditional bond returns have an assumed arithmetic mean of 4.5% and standard deviation of 6%. These are illustrative assumptions rather than historical estimates.
Inflation is modeled using a mixture of gamma shocks with shape parameter 2 and normal shocks. The gamma component has a weight of 0.85, and annual persistence is 0.40. The base process averages 3% inflation with a standard deviation of 1.5 percentage points and simulated skewness of approximately 0.77. The larger-inflation-shocks scenario doubles the scale of inflation fluctuations, raising the standard deviation to approximately 3 percentage points while average inflation remains near 3%. This scenario tests greater inflation uncertainty and more severe inflation surprises, rather than persistently higher average inflation.
Positive inflation surprises hurt traditional bonds more than stocks, with an additional bond-return penalty for large positive surprises. In the larger-shocks scenario, these effects reduce the simulated average traditional bond return to approximately 4.1% and increase its standard deviation to approximately 8.6%.
I bonds have an assumed fixed component of 0.9%, combined with lagged six-month inflation adjustments and a zero floor on the composite interest rate. Unlike TIPS or market bonds, I bonds cannot trade below par when yields spike. A separate sensitivity check sets the fixed component to zero.
Withdrawals come proportionately from all holdings. Each portfolio is evaluated with annual February rebalancing to its original weights and without rebalancing, allowing weights to drift. The principal simulations assume that the proposed rule changes permit both the initial I-bond allocation and subsequent purchases needed for rebalancing.
Results
Holding bonds substantially improves protection against low ending balances compared with an all-stock portfolio, while sacrificing considerable wealth when markets perform well. Replacing some traditional bonds with I bonds provides a smaller additional cushion in the base case and stronger protection when inflation shocks are larger.
The severe downside outcomes at the lowest percentiles are driven by sequence of returns risk: when market downturns or inflation spikes strike during the first few years of retirement, forced withdrawals lock in permanent capital losses. Non-marketable I bonds buffer this vulnerable window because their underlying principal cannot suffer mark-to-market impairment.
Base Inflation Scenario
Under base inflation, the principal benefit of I bonds appears at the lowest percentiles. Median and upper-percentile wealth are slightly lower than for 60/40, while the 25th percentile is nearly unchanged with annual rebalancing.
Worst simulated outcome: with annual rebalancing, the all-stock portfolio reaches zero, compared with minimum balances of approximately $126,000 for 60/40 and $134,000 for 60/30/10. The all-stock portfolio exhausts its assets in 28 of 100,000 paths. These are model-dependent results, and the minimum for each portfolio need not occur on the same path.
1st and 10th percentiles: all stocks leave approximately $186,000 and $493,000; 60/40 leaves $419,000 and $677,000; and 60/30/10 leaves $425,000 and $680,000. Most of the protection comes from holding bonds, with I bonds providing a modest additional benefit.
25th percentile: all stocks leave approximately $781,000, compared with approximately $882,000 for either balanced portfolio. The difference between the balanced portfolios is less than $400.
Median: all stocks leave approximately $1,233,000, compared with $1,169,000 for 60/40 and $1,162,000 for 60/30/10. Replacing traditional bonds with I bonds reduces median wealth by approximately $6,300.
75th and 90th percentiles: all stocks leave approximately $1,904,000 and $2,758,000; 60/40 leaves $1,543,000 and $1,959,000; and 60/30/10 leaves $1,528,000 and $1,937,000. The I-bond allocation reduces wealth relative to 60/40 by approximately $15,000 and $22,300.
Without rebalancing: relative to 60/40, I bonds add approximately $15,000 at the 1st percentile and $2,400 at the 10th. They reduce wealth by approximately $3,200 at the 25th percentile, $8,500 at the median, $11,900 at the 75th percentile, and $10,900 at the 90th.
Effect of rebalancing: annual rebalancing reduces the spread of outcomes for both balanced portfolios. It raises median and 10th-percentile wealth but lowers 90th-percentile wealth. For 60/30/10, median wealth rises from approximately $1,133,000 without rebalancing to $1,162,000 with rebalancing, while 90th-percentile wealth falls from $2,033,000 to $1,937,000. Rebalancing has no effect on the all-stock portfolio.
Larger Inflation Shocks Scenario
Greater inflation uncertainty weakens the protection provided by traditional bonds—and pooled bond funds susceptible to duration losses—making the I-bond allocation more valuable at the low end. Average inflation remains near 3%; the change is greater variability and more severe inflation surprises.
Worst and 1st-percentile outcomes: with annual rebalancing, all stocks leave $0 and approximately $185,000; 60/40 leaves $42,000 and $373,000; and 60/30/10 leaves $119,000 and $400,000. The I-bond portfolio provides stronger protection against very low ending balances.
10th and 25th percentiles: all stocks leave approximately $493,000 and $780,000; 60/40 leaves $641,000 and $851,000; and 60/30/10 leaves $659,000 and $862,000. I bonds improve these outcomes relative to 60/40 by approximately $17,700 and $10,900.
Median: all stocks leave approximately $1,233,000. The balanced portfolios have effectively equal medians: approximately $1,146,000 for 60/40 and $1,147,000 for 60/30/10.
75th and 90th percentiles: all stocks leave approximately $1,907,000 and $2,763,000; 60/40 leaves $1,526,000 and $1,958,000; and 60/30/10 leaves $1,516,000 and $1,931,000. The I-bond portfolio still gives up approximately $9,900 and $27,400 relative to 60/40.
Without rebalancing: I bonds improve the 1st, 10th, and 25th percentiles relative to 60/40 by approximately $45,800, $19,400, and $6,400. They reduce median wealth by approximately $5,400 and the 75th and 90th percentiles by $10,900 and $14,200.
Overall stability: I bonds reduce the standard deviation of ending wealth relative to 60/40 by approximately 3.5% with annual rebalancing and 1.7% without rebalancing. The corresponding reductions under base inflation are approximately 1.9% and 0.7%. Their stabilizing contribution is therefore greater when inflation shocks are larger.
Conclusion
The current retirement architecture forces a strange compromise on American savers: households can save hundreds of thousands of dollars tax-free in employer-sponsored plans, but the asset best engineered to protect their baseline standard of living against inflation spikes remains trapped behind an arbitrary $10,000 annual purchase cap on an isolated Treasury website.
As this analysis demonstrates, incorporating non-marketable Series I savings bonds into retirement accounts provides targeted insurance against worst-case retirement outcomes. By eliminating duration risk and guaranteeing a floor on nominal principal, a modest 10% allocation meaningfully cushions sequence-of-returns risk during severe inflation shocks—precisely when traditional market-traded bonds fail to provide safety.
Simply raising the individual TreasuryDirect cap offers little help to the vast majority of Americans whose investable wealth resides inside 401(k)s and IRAs. A true modern retirement framework should align public policy with household security by opening plan access to non-marketable Treasury inflation hedges, giving retirees the option to buy tailored downside stability when they need it most.
Authors Note: Some tables in Excel are behind a paywall. The annual subscription with this coupon in $30.
Some tables behind a paywall.
