When Bonds Yield More Than Stocks: Repricing the Equity Risk Premium After a Two-Decade Reversal

When Bonds Yield More Than Stocks: Repricing the Equity Risk Premium After a Two-Decade Reversal

The crossover that changes the opportunity set

The chart captures a regime break: the yield on the Bloomberg U.S. Aggregate Bond Index has moved above the S&P 500 earnings yield for the first time since the beginning of the century. Its black series, a Shiller-style equity risk premium measured against aggregate bonds, ends near negative 0.523 percentage points. The blue historical outcome series indicates that earlier starting valuations were associated with large swings in subsequent ten-year relative returns. The immediate message is not that bonds must beat stocks, but that investors are no longer paid an obvious yield premium for accepting equity cash-flow risk.

For most of the post-2000 era, especially after the global financial crisis, bond yields were suppressed while equity earnings yields were comparatively high. There is no longer a risk-free or high-quality carry vacuum forcing allocators toward equities. A diversified bond portfolio now offers contractual income, seniority, and lower volatility at a headline yield that exceeds the inverse price-earnings ratio of the equity index.

This comparison is powerful because it expresses opportunity cost in one picture. Every dollar allocated to stocks forgoes a bond yield that can be observed today, while the earnings yield is an accounting flow whose distribution and growth are uncertain. The crossover raises the hurdle that equities must clear through growth, dividends, buybacks, or valuation expansion.

Yet the units are not perfectly comparable. Aggregate bond yield is closer to a promised nominal return before defaults, calls, reinvestment, and price changes; equity earnings yield is neither promised nor fully distributed. The chart should initiate a cash-flow and risk analysis rather than terminate it.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Decomposing the two yields

A bond yield can be decomposed into expected short rates, term premium, credit compensation, option effects, and liquidity. The Bloomberg U.S. Aggregate contains Treasuries, agency securities, mortgage-backed securities, and investment-grade corporates, so its yield is not a pure government rate. Duration creates mark-to-market sensitivity, mortgages add negative convexity, and corporate bonds add default and spread risk. The quoted yield is compensation for this bundle.

The S&P 500 earnings yield, E divided by P, is the reciprocal of a price-earnings multiple. It reflects current or forecast accounting profits relative to market value. Unlike a coupon, earnings can fall, be restated, or be reinvested at uncertain returns. On the other hand, companies can grow nominal earnings, adjust prices, innovate, and repurchase shares. Equity duration is long and state-contingent.

A more complete relative-value identity is expected equity return minus expected bond return equals an equity risk premium plus differences in growth, duration, and tail exposure. The observed yield spread is only one input. If earnings grow rapidly, a low starting earnings yield can still support strong equity returns. If inflation or refinancing pressure erodes margins, the yield can prove illusory.

Investors should therefore reconcile trailing, forward, and cyclically adjusted profits. Peak margins make current earnings yields look safer than they are; recession-depressed earnings make them look worse. Shiller-style smoothing reduces cyclicality but introduces its own lag and sector-composition issues.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Why the reversal occurred

The bond side of the crossover reflects the end of the zero-rate world. Inflation, restrictive monetary policy, heavier Treasury supply, less price-insensitive central-bank demand, and a rebuilt term premium lifted yields across the curve. The Aggregate index repriced not because its underlying cash flows grew, but because discount rates rose and prices fell. That painful reset created better forward carry for new buyers.

The equity side reflects both resilient profits and elevated valuations. Mega-cap profitability, capital-light business models, buybacks, and enthusiasm around artificial intelligence supported prices faster than aggregate earnings yields could rise. Index concentration matters because a small group of highly valued firms can pull down the earnings yield even when the median stock is cheaper.

Fiscal policy connects the two markets. Large deficits can support nominal corporate revenue in the near term while increasing bond supply and term premium. This combination keeps equity earnings firm and bond yields high, producing an ambiguous signal rather than a conventional recession warning. The same deficit that benefits current cash flow can raise the discount rate applied to that cash flow.

The crossover is thus the joint product of normalization in fixed income and exceptional equity pricing. It should not be attributed to one Federal Reserve decision or one earnings season. Regime analysis must incorporate inflation credibility, fiscal supply, productivity, profit margins, and investor positioning.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

What the equity risk premium actually measures

The equity risk premium is the extra expected return investors demand for holding uncertain residual claims rather than safer fixed-income claims. It is not directly observable. Historical excess returns, survey expectations, dividend-discount models, and earnings-yield-minus-bond-yield proxies can produce very different answers. A negative 0.523 reading in the plotted proxy means the selected equity yield is below the selected bond yield; it does not prove the true expected premium is negative.

In a Gordon growth framework, expected equity return is approximately dividend yield plus long-run dividend growth, with valuation change added over finite horizons. Earnings yield enters through payout and reinvestment. Retained earnings create value only when invested above the cost of capital. The bond comparison omits this growth option, so an equity index can rationally trade at a lower current yield.

Conversely, growth cannot be counted without cost. Competition, disruption, taxes, labor bargaining, and capital requirements can absorb profits. When valuations already capitalize optimistic growth, future returns become sensitive to small disappointments. A compressed proxy premium signals a narrow margin for error even if it is not a timing rule.

The correct language is expected compensation, not guaranteed spread. Realized premiums are volatile over decades because discount rates and cash-flow news arrive unpredictably. A single starting metric offers a prior; scenario analysis updates it.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Lessons from the late 1990s and early 2000s

The last comparable crossover occurred around the technology bubble and its aftermath. Equity valuations were extreme, the earnings yield was low, and subsequent relative performance favored bonds for a meaningful period. That history deserves respect because starting valuation constrains long-horizon returns. It also warns against superficial analogy: today’s index leaders are generally profitable, cash generative, and deeply embedded in the economy, unlike many speculative issuers of 2000.

The macro regimes differ as well. The early 2000s moved toward disinflation and rate cuts after an investment boom. The present configuration includes larger public debt, different inflation risk, heavier index concentration, and an economy shaped by digital capital and artificial intelligence. Bonds may offer higher nominal carry today while facing greater term-premium and inflation volatility.

Historical comparisons work best through state variables. Compare valuation, profit margins, leverage, real yields, inflation expectations, fiscal balance, investment intensity, and household allocation rather than calendar labels. Similar yield spreads can generate different outcomes when the surrounding balance sheets differ.

The enduring lesson is asymmetry. When the initial equity yield cushion is thin, good outcomes require sustained growth or favorable discount rates. Bonds begin with more contractual carry. That shifts the burden of proof toward the equity thesis.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Forward returns and the danger of deterministic charts

The blue series in the chart relates starting conditions to subsequent ten-year average annual relative returns of the S&P 500 versus the Aggregate index. Such evidence is useful because valuation usually has more explanatory power at long horizons than at short horizons. However, the final decade of outcomes cannot yet be observed, so the blue line necessarily ends years before the black series. Readers must not interpret the visual gap as missing bad news or as a forecast already realized.

Overlapping ten-year windows also create strong serial correlation. Adjacent observations share most of the same return months, making a smooth historical line look like many independent experiments when it is not. Statistical confidence must account for overlap, regime dependence, and the small number of truly distinct cycles.

Selection choices matter: trailing or forward earnings, Treasury or aggregate yields, nominal or real returns, arithmetic or geometric averages, and start date can all alter results. A robust study tests multiple definitions and reports uncertainty rather than one fitted relationship.

Even so, valuation is not useless. It disciplines expectations. When a low or negative proxy premium coincides with high margins and optimistic growth assumptions, the distribution of equity outcomes becomes more dependent on favorable surprises. Forecasts should widen, not become falsely precise.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Inflation is the hidden common factor

Inflation affects both sides of the comparison but through different channels. Higher inflation raises nominal bond yields and can produce immediate bond losses when rates reset. Once yields have adjusted, bonds offer more carry, but unexpected inflation still erodes real coupons. Equities can pass through some inflation in revenue, yet margins suffer when input costs rise faster than pricing power and higher discount rates compress multiples.

The stock-bond correlation is therefore regime-dependent. In demand-driven disinflation, weak growth can lower yields and hurt earnings, allowing bonds to hedge equities. In supply-driven inflation, yields can rise while profit margins and valuations fall, hurting both. The crossover alone does not restore the negative correlation that made the traditional balanced portfolio work so well after 2000.

Break-even inflation, inflation swaps, commodity curves, wage growth, and pricing-power indicators help identify the regime. Nominal yield advantage is most valuable when inflation is anchored. If inflation risk is structurally higher, inflation-protected securities may provide cleaner real-rate exposure than nominal Aggregate bonds.

Portfolio decisions should separate expected carry from covariance. An asset can have an attractive standalone yield and still fail as a hedge in the state where protection is needed. Both expected return and crisis behavior belong in the allocation problem.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Duration, convexity, and the path of rates

The Aggregate index has meaningful duration, so its attractive yield does not eliminate price risk. A rough first-order approximation is that a one-percentage-point rise in yields produces a price decline approximately equal to duration, partly offset by convexity and subsequent carry. Investors who may need liquidity before maturity must care about the path, not only the yield to maturity.

Mortgage-backed securities complicate the picture. When rates fall, homeowners refinance and shorten cash flows; when rates rise, refinancing slows and duration extends. This negative convexity can make the index underperform a simple Treasury portfolio in large rate moves. Corporate spreads can also widen precisely when equities fall.

Equities possess a different duration. Growth stocks derive a large share of value from distant cash flows and can be highly sensitive to real rates, but corporate managers can alter investment, pricing, and payout. Equity duration is behavioral and nonlinear rather than contractual.

A valid comparison therefore matches horizons and liabilities. An investor funding a known medium-term payment may value bond cash-flow certainty. A perpetual institution seeking real growth may still require equities despite a lower current yield. The same crossover supports different allocations for different balance sheets.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Earnings quality, margins, and index concentration

The earnings yield can be distorted by where profits sit in the cycle. Aggregate margins have benefited from globalization, low taxes, intangible scale, market power, and falling financing costs over long periods. Some supports may persist; others can reverse through competition, regulation, labor costs, tariffs, or refinancing. Normalizing margins is essential before concluding that equities offer a stable cash-flow yield.

Concentration adds a second layer. The S&P 500 is capitalization-weighted, so highly valued mega-cap companies dominate its earnings yield and risk. The median constituent may offer a materially different spread over bonds. Equal-weighted, sector-neutral, value, small-cap, and international universes can show a more generous premium even when the headline index does not.

Concentration is not automatically irrational. Dominant firms may possess superior margins, balance sheets, and reinvestment opportunities. But common exposure to artificial-intelligence capital spending, regulation, and long-duration growth makes index-level diversification less complete than the constituent count suggests.

Investors should decompose the crossover into sectors and factors. It may argue for rotation within equities rather than wholesale exit: toward quality cash generators, reasonable valuations, strong balance sheets, and less crowded sources of earnings.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Credit is not cash

The Aggregate index includes investment-grade credit and securitized assets, so comparing its yield with equities is not equivalent to comparing stocks with Treasury bills. Corporate bondholders bear downgrade, spread, and default risk; mortgage investors bear prepayment and convexity risk. In recession, some of the apparent yield advantage can be offset by widening spreads and weaker liquidity.

Seniority still matters. Bond coupons and principal rank ahead of equity, while shareholders absorb the residual shock. An investment-grade issuer can experience a large equity drawdown without defaulting on bonds. This asymmetry gives fixed income a powerful advantage when growth disappoints modestly rather than catastrophically.

At the same time, credit upside is capped. If productivity surges, bondholders receive contractual payments while shareholders capture residual growth. The relative allocation is therefore a claim on the shape of the macro distribution, not just its average.

A precise framework separates Treasury duration, agency mortgages, investment-grade credit, and equities. The headline Aggregate yield is a useful blended hurdle, but active allocation should decide which component supplies carry, defense, or spread exposure.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Fiscal dominance and the term premium

Large and persistent fiscal deficits can keep bond supply elevated even when monetary policy eases. If investors demand greater compensation to absorb duration, the term premium can remain high. This would preserve attractive bond carry but expose holders to volatility and prevent equity multiples from expanding as they did in the low-rate era.

Fiscal spending can simultaneously support corporate revenue, employment, and nominal growth. That delays recession and benefits earnings, yet it may also crowd out private investment or keep inflation above target. Stocks and bonds then compete under a higher nominal discount-rate regime rather than returning neatly to pre-pandemic relationships.

Debt sustainability depends on the interaction of interest costs, nominal growth, primary balances, and maturity structure. Markets need not expect default to reprice duration; uncertainty about future inflation, taxes, or issuance is enough to raise required yields.

For allocators, fiscal risk argues for diversification across maturity, inflation sensitivity, geography, and real assets. It also cautions against assuming that a recession automatically recreates the enormous duration rallies of prior decades.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

The household and institutional allocation decision

A household can now earn meaningful income from high-quality bonds without relying entirely on dividends or selling appreciated shares. This is especially valuable for near-term spending needs and retirees with limited tolerance for sequence risk. Liability matching converts the bond yield from an abstract comparison into a funded cash-flow plan.

Pensions, insurers, endowments, and sovereign funds face different constraints. Higher yields improve the economics of immunizing liabilities and can reduce the return target placed on equities. Yet perpetual pools still need real growth, and insurers must manage capital charges and asset-liability duration. There is no universal sixty-forty answer.

Tax treatment also changes relative value. Interest, dividends, and capital gains can face different rates, while municipal bonds may dominate taxable bonds for some investors. Account location and turnover matter as much as the headline spread.

The rational response is a policy review, not a reflexive trade. Expected returns, risk budgets, liabilities, liquidity needs, and governance should be updated using today’s yields. Allocations inherited from the zero-rate era may no longer be efficient.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Scenario analysis for the next decade

In a soft-landing scenario, inflation moderates, policy rates decline gradually, bond carry is supplemented by duration gains, and corporate earnings continue to grow. Both assets can perform, but bonds may deliver unusually competitive risk-adjusted returns from their higher starting yield. Equity upside depends on whether growth exceeds what valuations already imply.

In a productivity boom, artificial intelligence and capital deepening raise real growth and margins. Equities can overcome the initial yield disadvantage, while real rates may stay elevated and limit bond price gains. In a renewed inflation scenario, nominal bond returns suffer and equity multiples compress, though pricing-power companies and real assets may fare better.

In recession, earnings fall and credit spreads widen. Treasuries likely provide the cleanest defense, while the Aggregate benefits from duration but absorbs spread and mortgage effects. In fiscal stress without recession, long yields can rise even as equity valuation falls, producing the most difficult joint drawdown.

Probabilities matter less than robustness. A portfolio should avoid ruin in any one branch and preserve the ability to rebalance when relative prices move. The crossover expands the menu of viable defensive assets, but it does not abolish uncertainty.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Implementation: from headline spread to portfolio weights

The first step is to estimate forward returns consistently. For bonds, begin with yield, subtract expected credit losses and fees, and model roll-down, convexity, and reinvestment. For equities, combine payout yield, sustainable real earnings growth, inflation, and plausible valuation change. Use ranges rather than point estimates.

The second step is to model covariance by regime. Historical averages dominated by the post-2000 disinflationary period may understate the chance of positive stock-bond correlation. Stress tests should include rapid inflation, term-premium shocks, recession, and liquidity withdrawal.

The third step is implementation discipline. Laddered Treasuries, core bonds, Treasury inflation-protected securities, active credit, and cash each serve different purposes. Equity exposure can be diversified away from the most expensive index segment without abandoning the asset class. Rebalancing bands prevent a compelling chart from becoming an all-or-nothing market-timing bet.

Finally, governance must match horizon. A ten-year valuation signal can look wrong for several years. Investors who cannot tolerate that path should not size the trade as if convergence were immediate. The best allocation is one that can survive the journey to the expected outcome.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.

Conclusion: a higher hurdle, not a verdict

The yield crossover is genuinely consequential. For the first time since the beginning of the century, the Bloomberg U.S. Aggregate Bond Index offers a yield above the S&P 500 earnings yield, and the plotted Shiller-style proxy ends near negative 0.523 percentage points. Fixed income once again provides contractual carry high enough to compete directly with equity cash-flow yield.

That fact raises the required case for stocks. Future equity outperformance must come from sustainable growth, payouts, or favorable repricing rather than an obvious initial yield advantage. Elevated margins and index concentration make the result more sensitive to disappointment. Bonds, meanwhile, begin with better carry but retain duration, inflation, mortgage, and credit risks.

History supports caution but not determinism. The early-2000s analogy shows that compressed starting premiums can precede strong bond-relative returns, while differences in profitability, inflation, fiscal policy, and technology prevent a simple replay. Overlapping forward-return data and measurement choices further limit precision.

The appropriate investment response is to reset capital-market assumptions, strengthen fixed-income allocations where liabilities justify them, diversify equity valuation exposure, and stress-test stock-bond correlation. The crossover is not a command to sell every stock. It is a message that safe and senior claims have recovered economic gravity—and equities must now earn their place against a credible alternative.

From a valuation perspective, the section requires separate estimates for starting income, sustainable growth, and terminal repricing. Bond return assumptions should deduct expected credit loss, fees, adverse reinvestment, and the effects of duration or convexity; equity assumptions should test margins, payout, reinvestment returns, dilution, and multiple change. Ranges are more honest than point forecasts because small changes in long-run growth or discount rates compound dramatically. The crossover is most useful as a discipline that exposes hidden assumptions in both assets, not as a deterministic forecast generated by one observed spread.

From a portfolio perspective, the same signal behaves differently in a soft landing, a productivity boom, renewed inflation, recession, or fiscal stress. Allocation must consider both average return and crisis covariance, and it should separate Treasury duration, mortgage convexity, corporate credit, inflation protection, and equity factor exposure. Position size should reflect the investor’s liability horizon, liquidity needs, tax status, governance, and capacity to rebalance. A long-horizon view has little value if the portfolio cannot survive the path by which relative valuation eventually normalizes. Rebalancing thresholds should therefore be written into policy before volatility arrives, preventing a patient valuation judgment from becoming improvised performance chasing at the worst moment.