Key Points
When long-term Treasury yields are exceptionally low, investors may be better off foregoing some long-duration exposure and accepting a limited amount of equity risk instead.
The result holds in both the 2006–2025 and 1996–2025 simulations, but the longer period is a more demanding test because it includes many years when 30-year Treasury yields were high enough to justify buying the long bond.
Simply requiring a 5% yield before buying the 30-year Treasury produced approximately the same terminal wealth as buying it regardless of yield, while generally reducing risk.
Risk depends on how it is measured. Strategy 4 produced much higher return-to-volatility ratios, but it also experienced deeper maximum drawdowns. Sharpe ratios therefore tell only part of the risk story.
The large improvement in returns came not merely from avoiding low-yield long bonds, but from investing the foregone allocation in equities. The next question is whether other equity portfolios, including higher-dividend strategies, can retain much of that return advantage with less downside risk.
A fair criticism of this paper is that it may be six years too late. As I argued in Long-Duration Bond Funds Still Haven’t Recovered From COVID, the greatest danger came when investors were buying very long-duration securities at exceptionally low yields.
Today is different. Current long-term yields make the 30-year Treasury considerably more defensible, although I would still approach the purchase of long duration assets cautiously as discussed in Is the 30-Year Treasury Yield Really That Attractive?.
A 30-year Treasury yielding 6 percent and one yielding 2 percent are both obligations of the United States government. But they are not economically equivalent investments. The question is whether investors should assume the same amount of duration risk regardless of the yield they are being paid.
I test four strategies over two periods: 2006–2025 and 1996–2025. Each portfolio receives $10,000 every six months, at the beginning of January and July.
Strategy 1 — Buy All Maturities. Divide each contribution equally among 3-month, 6-month, 1-year, 2-year, 5-year, 10-year and 30-year Treasuries whenever those maturities are available.
Strategy 2 — Buy the 30-Year Only at 5% or More. If the 30-year yield is below 5%, redistribute its allocation among the six shorter Treasury maturities.
Strategy 3 — Buy at 5%, Sell Below 3%. Follow Strategy 2 but check the 30-year yield at every month-end. If it falls below 3% and the portfolio owns 30-year Treasuries, sell those positions and reinvest the proceeds among the shorter maturities.
Strategy 4 — Use the S&P 500 Instead. When the 30-year yield is at least 5%, invest one-seventh of the contribution in it. Otherwise, invest that $1,428.57 in SPY, the SPDR S&P 500 ETF Trust, while the other six-sevenths remain in the six shorter Treasury maturities.
I compare the strategies using ending wealth, money-weighted annual return, annualized volatility, maximum drawdown, and the Sharpe ratio. The Sharpe ratio combines return and volatility into a conventional measure of risk-adjusted performance:
Sharpe ratio = annualized return above the 3-month Treasury-bill rate ÷ annualized volatility.
A higher Sharpe ratio is better: it indicates more excess return per unit of volatility.
2006–2025
There are 40 contributions totaling $400,000. The 30-year Treasury met the 5% threshold on only 2 contribution dates, was available but below 5% on 37, and was unavailable on 1. In Strategy 4, the one-seventh allocation that otherwise would have gone to the 30-year Treasury was invested in SPY on the 38 nonqualifying or unavailable dates.
Return. Strategies 1–3 produced almost identical ending wealth: $497,902, $498,215 and $497,645, respectively. Strategy 4 finished at $670,138, with a money-weighted annual return of 4.79%, compared with about 2.09% for the Treasury-only strategies.
Risk. The 5% purchase rule substantially reduced conventional volatility: annualized volatility fell from 3.42% in Strategy 1 to 2.37% in Strategy 2. Strategy 4’s volatility was also 3.42%, surprisingly close to Strategy 1. But maximum drawdown tells a different story. Strategy 4 suffered a 7.45% maximum decline, compared with 2.58%for Strategy 2. Both are legitimate measures of risk, but for an investor concerned about how far a portfolio can fall during a bad period, the deeper trough in Strategy 4 may be more revealing than its relatively modest average variability.
Risk-adjusted return. The Sharpe ratio—excess return per unit of volatility—was 0.24 for Strategy 1, 0.27 for Strategy 2, 0.27 for Strategy 3 and 0.79 for Strategy 4. On the conventional volatility-based measure, Strategy 4 therefore provided much greater compensation for risk, although its larger maximum drawdown is an important qualification.
Accounting check on Strategy 4. The SPY investments made in place of the 30-year Treasury grew to approximately $237,000 by the end of 2025. The corresponding allocations in Strategy 2, which were instead invested in shorter Treasuries, grew to roughly $65,000. The difference—about $172,000—closely matches the $171,923 difference in terminal wealth between Strategies 4 and 2. That reconciliation provides a useful independent check on the Strategy 4 result.
The Longer Test: 1996–2025
The 2006 sample contains an unusually long period of very low interest rates. Going back to 1996 provides a more demanding test because long-term yields were substantially higher during much of the earlier period.
There are 60 contributions totaling $600,000. The 30-year Treasury met the 5% threshold on 15 contribution dates, was available but below 5% on 37, and was unavailable on 8. Unlike the shorter sample, the rule therefore repeatedly permitted purchases of the 30-year Treasury during the higher-yield years.
Return. Strategies 1 and 2 again produced virtually identical ending wealth: $889,650 and $889,963, respectively. Strategy 3 finished somewhat lower at $884,806. Strategy 4 reached $1,142,691, with a money-weighted annual return of 3.94%, compared with about 2.47% for Strategies 1 and 2.
Risk. Strategy 2 again reduced conventional variability: annualized volatility fell from 2.66% in Strategy 1 to 2.47%, while Strategy 4’s volatility was only moderately higher at 2.84%. Maximum drawdown tells a less favorable story for Strategy 4. Its largest peak-to-trough decline was 6.03%, compared with only 2.16% for Strategy 2. Both measures are valid, but the substantially deeper trough may be the more meaningful warning about Strategy 4’s risk for an investor concerned about losses during severe market declines.
Risk-adjusted return. The Sharpe ratio—excess return per unit of volatility—was 0.28 for Strategy 1, 0.30 for Strategy 2, 0.29 for Strategy 3 and 0.61 for Strategy 4. Thus Strategy 4 provided substantially more return per unit of conventional volatility, although that statistic does not fully capture its larger maximum drawdown.
Accounting check on Strategy 4. The SPY investments made when the 30-year Treasury either failed the 5% test or was unavailable grew to approximately $333,600 by the end of 2025. The corresponding allocations in Strategy 2, which instead went into shorter Treasuries, grew to approximately $80,900. The difference—about $252,700—almost exactly matches the $252,728 difference in ending wealth between Strategies 4 and 2. As in the shorter sample, this provides a useful independent accounting check on the Strategy 4 result.
The longer period therefore reinforces the basic result. Requiring a 5% yield before buying the 30-year Treasury reduced risk without sacrificing meaningful terminal wealth. The automatic sale rule added little. Redirecting the rejected long-bond allocation into equities produced substantially higher returns. The additional return was substantial relative to the increase in conventional volatility, although maximum drawdown was considerably larger.
Conclusion
These results reinforce the argument in my recent papers, Is the 30-Year Treasury Yield Really That Attractive? and Long-Duration Bond Funds Still Haven’t Recovered From COVID. Long-term interest rates have varied enormously over the past two decades—and even more over longer periods. Buying a high-duration asset when yields are exceptionally low can leave an investor accepting substantial interest-rate risk for very little compensation.
The Treasury results here are consistent with that concern. Requiring a 5% yield before buying the 30-year Treasury produced about the same wealth as buying it mechanically, but generally with less risk. The additional 3% sale rule added little.
If investors are going to assume substantial long-term risk, they should ask whether they are being adequately compensated for that risk rather than buying duration mechanically.
The largest improvement in return and in the conventional return-to-risk ratio came from Strategy 4, which invested the rejected 30-year allocation in equities. That result should not be interpreted as proof that the S&P 500 is always the right alternative. It instead raises the next question: if long-duration bonds are unattractive, what should replace them?
Next week I plan to extend the experiment by comparing the S&P 500 with other equity alternatives, including higher-dividend stock portfolios, to see whether some of the additional return can be retained with less downside risk.
Verification Notes and Limitations
This is a historical simulation, not a record of returns from actual Treasury securities purchased on these dates. Treasury yields are drawn from the U.S. Treasury’s historical daily par-yield-curve data. These constant-maturity rates provide a consistent historical measure of yields across maturities, but they are not transaction prices for particular Treasury CUSIPs. The results should therefore be interpreted as estimates of how the strategies would have performed rather than as precisely replicable real-world investment returns.
For Treasury maturities of one year or longer, each simulated purchase is treated as a par coupon bond whose coupon rate equals the constant-maturity yield at purchase. Coupons are paid semiannually. Coupon payments remain in the portfolio and are reinvested rather than treated as withdrawals. When a Treasury reaches maturity, its principal is reinvested in the same maturity at the then-prevailing yield. Short-term Treasury bills are modeled using their corresponding constant-maturity rates.
Each Treasury position is tracked separately and normally held to maturity. Strategy 3 is the exception: qualifying 30-year positions are sold when its explicit sale rule is triggered. Unsold securities remaining at the end of the experiment are marked to market using the December 31, 2025, Treasury yield curve, rather than simply valued at face value.
SPY is used for the equity alternative because its history begins in 1993 and therefore covers both samples. Adjusted closing prices are used so that dividends and stock splits are incorporated into the equity return series.
Returns exclude the external $10,000 semiannual contributions. Annualized volatility is calculated from six-month portfolio returns, while maximum drawdown is calculated from the time-weighted portfolio path so that contributions are not mistaken for investment gains. The Sharpe ratio uses the 3-month Treasury rate as the risk-free benchmark and is calculated consistently across all four strategies.
The simulation necessarily abstracts from some real-world considerations. It does not model specific Treasury issues, bid-ask spreads, commissions, taxes, execution timing differences, or other transaction costs. Constant-maturity yields are also interpolated measures rather than directly observed prices for a security held by an investor. These limitations make the exercise better suited to comparing the relative behavior of the four strategies than to claiming that an investor would have earned exactly the dollar amounts reported.
Author’s Note
Economic and Political Insights covers a fairly wide range of subjects, from investing and personal finance to economic policy and politics. Recent pieces include my forecast for control of the House in 2026, If I Were Speaker: A Tax Reconciliation Agenda, and the first installment of my new Race of the Day series, beginning with Washington’s 3rd Congressional District.
There will be a lot more election analysis and financial research between now and the end of the year. Most of my work will remain free, although some more detailed research and analysis may be reserved for paid subscribers.
If you enjoy the work, please at least become a free subscriber. And for readers who want access to the paid material while also supporting the research, I am offering 20% off a paid subscription.