Why Bonds Still Belong: Rethinking Fixed Income in Modern Portfolios

2026-02-10

Overview

Despite recent challenges, including muted returns, structural shifts, and evolving correlations, bonds retain unique diversification, planning, and return stacking potential that equity-only alternatives cannot fully replicate.

Key Topics

Bonds, Diversification, , Liability Immunization, Return Stacking

Loading the Elevenlabs Text to Speech AudioNative Player...

Disenchantment with Bonds: Reasonable, but Retrospective

In recent years, the bond market has disappointed many investors. Rising rates and inflation have driven high interest rate volatility, while long-duration bonds have underperformed, dramatically underperforming cash and generating outright negative returns. With a flat term structure, it’s tempting to see duration as uncompensated risk, especially when yields offer little cushion and price sensitivity to interest rate shifts is high.

Figure 1: Rolling 5-Year Annualized Returns of U.S. Bonds

Source: Bloomberg. U.S. Bonds are the Bloomberg U.S. Aggregate Bond Index (LBUSTRUU). Index returns are gross of all fees, taxes, and transaction costs. You cannot invest in an index. Past performance is not indicative of future results.

This skepticism is understandable, but it often misses a crucial point: bond yields act like gravity for long-run returns (a point we’ll address specifically in the third section of this article). While recent returns may look weak, that backward-looking disappointment is largely a reflection of rising yields, which can be good news for future returns. In fact, bond returns often behave inversely to past performance. As yields rise, they drag present returns down but simultaneously lift forward-looking expectations. In many ways, all changing yields really do is push and pull returns across time. Interest rates, after all, govern the time value of money. When yields rise, they simply reallocate return from the present into the future, and when they fall, they pull some of that future return forward.

“Fixed income markets are currently offering forward-looking returns that are not only positive but unusually competitive relative to recent history.”

A natural follow-up to this perspective is, “But what about inflation?” Shouldn’t investors focus on real returns, not just nominal yields? The answer is yes, but with important nuance. Inflation expectations are a key input into bond pricing. While they are neither directly observable nor perfectly stable through time, they are reflected in the shape of the yield curve and in the compensation investors demand for holding nominal bonds. Inflation risk is not eliminated, but it is not ignored either. What matters most going forward is not simply the level of inflation, but whether it surprises relative to what is already implied.

This is where an irony often emerges. Many of the same investors who argue, correctly, that market timing is extraordinarily difficult in equities are often quick to do exactly that in bonds, particularly when it comes to duration. Avoiding longer-term bonds solely because the yield curve is flat or inverted is not a neutral risk-management choice. It implicitly assumes that the bond market is systematically underpricing inflation and interest rate risk, and that future outcomes will be worse than what is already reflected in prices. That may prove true at times, especially during periods of macro transition, but it is a strong and active forecast to make consistently in one of the deepest, most liquid, and most information-rich markets in the world.

“Many of the same investors who argue, correctly, that market timing is extraordinarily difficult in equities are often quick to do exactly that in bonds”

The more constructive takeaway is this. Real yields are currently near decade highs. At the time of writing, investors can access positive real expected returns across much of the Treasury curve, something that was largely absent throughout the 2010s. Even after accounting for inflation uncertainty and the possibility of surprises, today’s starting yields represent a materially improved forward-looking opportunity set. For long-term allocators, this does not eliminate risk, but it does mark a return to being meaningfully compensated for bearing it.

While returns over the last several years may make traditional bonds feel less reliable, the foundations of return stacking remind us that every return stream, even a modest one, can add value when combined thoughtfully. Bonds remain one of the few liquid, low-correlated building blocks available to investors.

Register for our Advisor Center

Tools Center:
Easily backtest & explore different return stacking concepts

Model Portfolios:
Return stacked allocations, commentary and guidance designed
for a range of client risk profiles and goals

Future Thinking:
Receive up-to-date insights into the world of return stacking theory and practice

Register to Access the Portal >

Imperfect Correlation Isn’t a Bug, It’s a Feature

There’s growing talk that bonds have lost their diversification power, especially as stock–bond correlations have turned positive at times in recent years. Still, the concept of diversification doesn’t hinge on perfect negative correlation. It simply requires that assets behave differently, and even today, bonds and equities still do.

Figure 2: Rolling 12-Month Correlation between U.S. Stocks and U.S. Bonds

Source: Bloomberg. U.S. Stocks is the S&P 500 Total Return Index (SPXT). U.S. Bonds is the Bloomberg U.S. Aggregate Bond Index (LBUSTRUU). Index returns are gross of all fees, taxes, and transaction costs. You cannot invest in an index.

More importantly, all else held equal, more diversification is better than less. A portfolio with multiple independent return streams, even if those streams sometimes rise and fall together, is mathematically more stable than one that relies on a single source of return. When advisors abandon bonds entirely, they aren’t just changing correlation profiles: they are reducing the dimensionality of their portfolios, weakening both return potential and risk management over time.

There’s also a common misconception around correlation itself. Two assets can be perfectly correlated and still move in different directions. Correlation captures relative movement, not absolute return. So, while it’s possible that bonds no longer hedge the short-term volatility of stocks, they may still drift differently over time. Combining stocks and bonds, then, may do less to dampen volatility but can still significantly dampen dispersion in expected terminal wealth.

And that drift matters. Fixed income, true to its name, offers a greater degree of certainty in long-run return characteristics. While equities are generally expected to deliver higher returns over time, their outcomes are path-dependent and volatile. Bonds, in contrast, provide investors with clearer expectations, especially when held to maturity. That clarity is critical in portfolio construction, particularly for long-term planning and liability management.

Relying solely on equities not only amplifies volatility, it also degrades certainty. For advisors managing wealth over decades, the removal of bonds narrows the toolkit at precisely the moment when more tools are needed. Replacing a 60/40 with a single-risk equity allocation, even if buffered or hedged, may address near-term drawdowns but introduces long-term vulnerabilities in drift, sequencing, and income reliability.

Imagine another “lost decade” for equities, like the 2000s in the U.S. or the 1990s in Japan. In such a scenario, relying solely on equities as the portfolio’s return engine can leave investors exposed to years of stagnation with little to show for the risk.

Figure 3: The Lost Decade in U.S. Equity Returns (December 31, 1999 to December 31, 2009)

Source: Bloomberg. U.S. Stocks is the S&P 500 Total Return Index (SPXT). U.S. Bonds is the Bloomberg U.S. Aggregate Bond Index (LBUSTRUU). Index returns are gross of all fees, taxes, and transaction costs. You cannot invest in an index. Past performance is not indicative of future results.

Moreover, bonds, even if not negatively correlated with stocks at all times, offer an independent return stream that can continue to compound during equity droughts. This makes bonds not just a volatility diversifier but a vital contributor to long-term return resilience. For advisors building multi-decade plans, ignoring that possibility is not a luxury they can afford.

Bonds as Planning Tools: Immunization Matters

One of bonds’ most enduring virtues lies in planning and liability matching. Whether funding retirement income or meeting a future spending goal, fixed income remains one of the most efficient tools for achieving time-based certainty. This is where bonds do something equities simply cannot.

The concept of immunization, that is, aligning a bond portfolio’s duration with the investor’s time horizon, allows for the cancellation of opposing risks. Specifically, as interest rates move, bond prices and reinvestment opportunities move in opposite directions. If the portfolio is immunized properly, these two forces offset each other, making future wealth more predictable.

This planning power isn’t limited to individual bonds held to maturity. Even bond funds can be used for liability-driven strategies by applying a simple approximation known as the “2× duration minus 1” rule. According to the academic research behind this idea (Leibowitz and Bova, 2015; Lozada, 2016), the optimal investment horizon for a bond fund, meaning the horizon at which price risk and reinvestment risk approximately cancel and investors can expect to receive the starting yield, is roughly equal to: 2 × Duration – 1 year.

So, if an advisor is using a bond fund with an effective duration of 5 years, the approximate investment horizon for which the fund becomes an effective immunizing tool would be: 2 × 5 – 1 = 9 years.

In this case, an investor planning for a known cash need in 9 years could use this bond fund with reasonable confidence that their future wealth would be relatively insensitive to interest rate moves in the interim.

Figure 4: Starting Yield-to-Worst versus Subsequent Duration-Matched Returns for U.S. Bonds

Source: Bloomberg. Graph shows the starting yield-to-worst for theBloomberg U.S. Aggregate Bond Index (LBUSTRUU) versus its subsequent duration-matched return (over a period equal to 2x Modified Duration – 1). Data is from 1976 to 2025. The starting date was chosen due to data availability.

To make things more concrete, a U.S. Treasury ETF targeting the 3–7 year maturity range might currently have a yield-to-worst of 3.8% and an effective duration of 4.3. Using the 2×D – 1 rule, an investor could allocate capital to this fund knowing that if they hold it for 7.6 years, their terminal value would be largely shielded from rate volatility. They would be relying on the fund’s average yield to deliver a stable, predictable return – something close to 3.8% annualized over the period – even as rates fluctuate along the way.

“Shifting into T-bills or cash may reduce short-term volatility, but doing so introduces significant reinvestment (or market timing) risk for anything beyond a 6- or 12-month horizon.”

In contrast, shifting that same allocation into T-bills or cash may reduce short-term volatility, but doing so introduces significant reinvestment (or market timing) risk for anything beyond a 6- or 12-month horizon. Without the term premium and duration exposure that bonds offer, investors are forced to make repeated interest rate bets over time, often at inopportune moments.

Stock/Bond Portfolio Simulator

This simulator generates 10,000 portfolio return paths for three stock-bond correlation assumptions (-0.66, 0, and +0.66) and plots the distribution of realized annualized returns, volatilities, and terminal wealth for each.

As the simulation horizon approaches 2 × Duration - 1, the return and terminal wealth distributions tend to converge across correlation regimes, while the volatility distributions remain distinct. The result suggests that while correlation may affect the bumpiness of the path, a bond allocation’s liability-hedging properties can help provide greater return certainty over certain horizons—specifically, those near the 2 × Duration - 1 convergence point.

For illustrative purposes only. Simulated results do not represent actual performance and are not indicative of future results. See important methodology and assumption disclosures below. Data as of February 23, 2026.

%
%
%
Bond Weight: 40%
Simulating 10,000 paths across 3 correlation scenarios...
10.00 years
Realized Annualized Return Distribution
Realized Volatility Distribution
Terminal Wealth Distribution

Overview

This simulator runs 10,000 Monte Carlo paths of a fixed-mix stock/bond portfolio, rebalanced monthly, across three stock-bond correlation assumptions. Bond returns are derived from a stochastic yield curve model calibrated to historical U.S. Treasury data. Stock returns are modeled as geometric Brownian motion with user-specified parameters, correlated with bond returns via the yield curve's dominant principal component. The simulation horizon is set to twice the Macaulay duration of the selected bond, a holding period over which a constant-maturity bond portfolio's annualized return converges toward its starting yield.

Yield Curve Model

The yield curve simulation uses a two-component model calibrated to daily U.S. Treasury par yield curve rates from October 1993 through the most recent available date. The maturity grid spans 10 points: 3-month, 6-month, 1-, 2-, 3-, 5-, 7-, 10-, 20-, and 30-year.

  • Short rate (Vasicek): The 3-month rate is modeled as an Ornstein-Uhlenbeck process calibrated via maximum likelihood estimation on historical short rates extracted from bootstrapped discount curves. The short rate evolves as dr = κ(θ − r)dt + σdW, where κ (mean-reversion speed), θ (long-run mean), and σ (volatility) are estimated from the historical series. The Vasicek innovation is shared with the first principal component of the forward curve (Z0), linking the short rate and forward curve dynamics through a common level factor.
  • Forward curve (shifted log-normal PCA-HJM): Historical par yields are converted to discount factors via a semiannual coupon bootstrap (with log-linear interpolation at intermediate coupon dates), then to instantaneous forward rates. A 2% displacement is applied before taking logs to handle near-zero rate periods (displaced diffusion). Daily changes in log(f + 0.02) are standardized (zero mean, unit variance per maturity segment) and decomposed via Principal Component Analysis with 3 factors. In the simulation, three independent standard normal draws (Z0, Z1, Z2) are scaled by √(λk × 252 × Δt) to convert daily PCA variance to monthly, then projected back through the eigenvectors and inverse standardization (multiplying by the per-segment standard deviation and adding back the per-segment mean drift). Forward curves evolve as f(t+dt) = (f(t) + 0.02) × exp(shock) - 0.02, where volatility is proportional to the shifted rate level (f + 0.02). This ensures forward rates are bounded below by -2% by construction.
  • Curve reconstruction: At each time step, the full discount curve is built by combining the Vasicek short rate (used for the short end: D(0.25) = 1/(1 + r×0.25)) with the PCA-simulated forward rates (used for longer maturities: D(Ti) = D(Ti-1) × exp(−fi-1 × ΔTi-1)). Discount factors are then converted to par yields using semiannual coupon pricing with log-linear interpolation of discount factors at intermediate coupon dates (simple yield for maturities under 1 year; semiannual coupon pricing for 1 year and above).

Bond Return Calculation

Bond returns simulate a constant-maturity Treasury portfolio. At each monthly step, the portfolio holds a par bond at the selected maturity. The return is computed by pricing the bond after one month of time decay under the new yield curve:

  • The bond's coupon rate is set to the par yield at the selected maturity at the start of each period (the bond is purchased at par).
  • After one month, the bond's remaining maturity is reduced by 1/12 of a year.
  • The bond is repriced at the shifted maturity using the new yield curve, producing a total return including both price change and coupon accrual.
  • For maturities below 6 months, simple discount pricing is used. For longer maturities, semiannual coupon bond pricing is applied.
  • The bond is sold and a new par bond is purchased at the start of each period (monthly rebalance to constant maturity).

Stock Return Model

Stock returns follow a discrete-time geometric Brownian motion with user-specified annualized expected return (μ) and volatility (σ): rstock = μΔt + σ√Δt × Zstock, where Δt = 1/12 (monthly).

Stock-Bond Correlation

The stock-bond correlation is implemented via Cholesky decomposition of a bivariate normal. The first principal component of the yield curve (PC1, which represents parallel level shifts and is the dominant driver of bond returns) provides the bond-side random shock. A positive PC1 shock raises yields, which is negative for bond prices, so the sign is inverted when constructing the stock shock:

Zstock = −ρ × ZPC1 + √(1 − ρ²) × Zindep

where ρ is the user-specified correlation between stock returns and bond returns, ZPC1 is the standard normal shock driving the first principal component of the yield curve, and Zindep is an independent standard normal draw. The negation ensures that a positive ρ means stocks and bond returns (not yields) move together.

To ensure a fair comparison across the three correlation scenarios (ρ = −0.66, 0.00, +0.66), the simulator uses identical bond return paths, identical PC1 draws, and identical independent stock noise draws across all three scenarios. Only the correlation mixing parameter changes.

Portfolio Construction

The portfolio is constructed as a fixed-mix (constant-weight) allocation between stocks and bonds, rebalanced monthly. Each month's portfolio return is the weighted sum of the stock and bond returns: rport = wstock × rstock + wbond × rbond. No transaction costs, taxes, or management fees are modeled.

Simulation Horizon

The yield curve is simulated for a full 20-year horizon. A slider allows the user to explore any holding period from 1 to 20 years without re-running the yield curve simulation. The default horizon is set to twice the Macaulay duration of the selected bond, computed from the current par yield using standard semiannual coupon cash flow weighting. This default is chosen because it represents the approximate holding period over which a constant-maturity bond portfolio's cumulative return converges toward its starting yield-to-maturity, regardless of subsequent interest rate movements. Moving the slider reveals how the convergence of terminal wealth distributions (and divergence of volatility distributions) evolves over time.

Output Statistics

  • Terminal Wealth: The portfolio value at the end of the horizon, starting from $100.
  • CAGR: Compound annual growth rate, computed as (Terminal/100)1/T − 1.
  • Realized Volatility: The annualized standard deviation of monthly log returns over each path, computed as std(log returns) × √12.
  • Distributions: Kernel density estimates using a Gaussian kernel with Silverman's rule-of-thumb bandwidth.

Data Sources

  • Historical yield data: Daily U.S. Treasury par yield curve rates (constant maturity), sourced from the U.S. Department of the Treasury.
  • Current yield curve: Fetched from Treasury.gov at the time of model calibration.

Limitations & Caveats

  • The Vasicek model assumes a single-factor mean-reverting short rate. Real short rates exhibit jumps, regime changes, and time-varying volatility not captured here.
  • The PCA model captures the historically observed covariance structure of yield curve movements but does not impose no-arbitrage drift conditions (unlike a full HJM implementation). The shifted log-normal dynamics ensure forward rates remain above -2%, but do not prevent all arbitrage opportunities.
  • Stock returns are modeled as Gaussian with constant drift and volatility. Real equity returns exhibit fat tails, volatility clustering, and time-varying expected returns.
  • The stock-bond correlation is held constant within each scenario. In practice, stock-bond correlations are time-varying and regime-dependent.
  • No transaction costs, bid-ask spreads, taxes, management fees, or liquidity constraints are modeled.
  • The constant-maturity bond return calculation assumes frictionless monthly rebalancing to a single on-the-run Treasury maturity.
  • Results will vary across runs due to Monte Carlo randomness. With 10,000 paths, summary statistics are generally stable but not perfectly reproducible.

This tool is for educational and illustrative purposes only. It does not constitute investment advice.

Return Stacking Perspective: Bonds Within a Broader Mosaic

In a return stacking framework, bonds do more than contribute returns. They offer a structural component that supports thoughtful portfolio construction. Their lower volatility, predictable income, and defined duration profile make them well suited to serve as one of several foundational exposures in a capital-efficient design.

Return stacking is built on the ability to combine multiple independent return streams within a single portfolio. Bonds help facilitate this by contributing stability, freeing up capital, and offering a differentiated risk profile. Whether paired with equities, alternatives, or systematic strategies, they help maintain balance as additional exposures are layered in through overlays or other tools.

“the result is a compelling combined stream for meeting intermediate-term liquidity needs, with a balance of certainty and opportunity that is difficult to replicate using equities or cash alone.”

As discussed in Section 3, using bonds with a defined duration and known yield-to-maturity allows us to build portfolios around a base we can reasonably expect to deliver a specific return over a known horizon. For example, if we have confidence in a bond allocation returning approximately 3.65 percent annualized over the next 7.6 years, that creates a stable foundation on which to stack a diversified mix of alternative strategies. If we believe that mix of alternatives can generate an additional 300 basis points of excess return, the result is a compelling combined stream for meeting intermediate-term liquidity needs, with a balance of certainty and opportunity that is difficult to replicate using equities or cash alone.

Crucially, bonds diversify across more than just asset class. They bring differences in timing, cash flow, and volatility. Even when expected returns are modest, they can smooth drawdowns, support rebalancing discipline, and create funding flexibility for opportunistic allocations. These characteristics improve the overall shape of portfolio outcomes, not just the average level of return.

In a world that demands resilience, adaptability, and precision, bonds remain a valuable part of the return stack. They are not the only base investors can build upon, but they are among the most versatile and capital-efficient—making them an essential consideration for any stacked approach.

Conclusion: Unfashionable in cycles, indispensable over time.

The case for bonds today is not about nostalgia or tradition. It is about recognizing their evolving role in a modern, diversified, capital-efficient portfolio.

Yes, recent performance has been disappointing. Bonds have delivered volatility without much return, and rising rates have eroded mark-to-market values for many investors. But that backward-looking pain is precisely what resets the forward-looking opportunity. Yields act as gravity for long-term returns. When they rise, they don’t destroy return—they shift it from the past into the future. And with real yields now at decade highs, the prospective compensation for holding duration has meaningfully improved.

Yes, stock-bond correlations have shifted. But diversification does not require perfection. It only requires difference. Bonds still behave differently than equities. They remain one of the few liquid, low-beta, income-producing assets available to investors. In the event of a sustained equity bear market, they may once again prove their worth as a parallel return engine—if not a hedge, then a ballast.

Yes, inflation matters. But unless we assume markets are systematically mispricing risk, those inflation expectations are already embedded in the yield curve. Real rates are what matter most for long-run fixed income returns, and by that measure, bond investors are in a better position now than they have been in years.

And yes, portfolio construction matters. Bonds enable planning with greater precision through immunization strategies, even when using bond funds. They reduce reinvestment risk and support liability-aware frameworks that cash and short-term instruments cannot fully replicate.

But perhaps most importantly, bonds remain a critical enabler of return stacking itself. They provide the stability, cash flow, and risk control that allow other return streams to be layered on top.

For advisors and allocators looking to build resilient portfolios that balance risk, return, and real-world objectives, bonds are still a tool worth using. Not because they always shine, but because they continue to matter.

Bibliography

Leibowitz, Martin & Bova, Anthony & Kogelman, Stanley. (2015). Bond Ladders and Rolling Yield Convergence.
Financial Analysts Journal. 71. 32-46. 10.2469/faj.v71.n2.4.

Lozada, Gabriel. (2016). Constant-Duration Bond Portfolios’ Initial (Rolling) Yield Forecasts Return Best at Twice Duration.
At https://content.csbs.utah.edu/~lozada/Research/IniYld.pdf

Disclosures

This material is for informational and educational purposes only and is intended for use by investment professionals. It is not intended as investment advice or as a recommendation to buy or sell any security or to adopt any investment strategy. The views expressed are those of the authors as of the date of publication and are subject to change without notice.

The information contained herein has been obtained from sources believed to be reliable, but no representation is made as to its accuracy, completeness, or timeliness. This document does not constitute investment research or a research recommendation and is not subject to the requirements applicable to independent investment research.

Investing involves risk, including possible loss of principal. Past performance is not indicative of future results. Diversification and asset allocation do not guarantee profit or protect against loss in declining markets.

References to specific indices, securities, or strategies are for illustrative purposes only and do not constitute a recommendation. Indexes are unmanaged and it is not possible to invest directly in an index.

Nothing in this document should be construed as legal, tax, or accounting advice. You should consult your own legal and tax advisors before implementing any transaction or strategy.

This material may not be reproduced, distributed, or transmitted in whole or in part without prior written permission.