Bonds cushioned equity drops only 11% of the time since 2020

For four decades, the answer to "what protects my portfolio when stocks fall?" was simple: bonds. Since 2020, that answer has failed nine times out of ten.

Do Bonds Still Hedge Equity Risk? The 89% Failure Number

Bonds cushioned equity drawdowns only about 11% of the time since 2020. According to BlackRock's research on rebuilding 60/40 portfolios, bond market returns were negative in 17 of the 19 months when equities declined by 2% or more . That is an approximate 89% failure rate for the asset class whose entire portfolio role is to offset stock losses. The number matters because it comes from one of the world's largest asset managers, not from crypto advocacy.

Quick Answer: Bonds offset equity losses in only 2 of 19 months since 2020 when stocks fell 2% or more — roughly an 11% success rate, per BlackRock. Inflation shocks now push both legs of a 60/40 portfolio down together, which is why allocators are reconsidering the fixed-income sleeve entirely.

The mechanism behind the breakdown is inflation. When growth shocks drive markets, bonds rally as investors flee to safety and central banks cut rates — the classic negative correlation that makes 60/40 work. When inflation shocks drive markets, yields rise and equity multiples compress at the same time, so both legs fall together. BlackRock identifies March 2026 as the second-weakest month for 60/40 portfolios since the 2022 drawdowns, with exactly that dynamic in play .

This is not an argument that bonds are worthless. They still deliver income, and at current yields that income is meaningful — the US 10-year Treasury sat at 4.97% on September 13, 2026 . The argument is narrower and more specific: the diversification job that justified a 40% fixed-income weight is no longer being done reliably. A portfolio sleeve that pays you to hold it but fails to protect you when you need protection is an income allocation, not a hedge.

Every allocation model discussed in this article is tested against the same baseline. BlackRock puts the traditional 60/40 portfolio's historical performance at 6.7% returns with 11.6% volatility . That is the benchmark any proposed replacement — bitcoin, gold, private markets, or a wider active equity sleeve — has to beat on a risk-adjusted basis, not just on headline return.

The September 2026 flashpoint came from Bitcoin Suisse's Crypto Wealth Management Report 2026, which CoinDesk covered on September 12, 2026 under the framing "Ditching bonds for bitcoin" . But the report's own logic is additive rather than substitutive. The Zug-based firm's position is that AI capital expenditure has concentrated equity risk to the point where portfolios need a third return driver distinct from both stocks and fixed income — not that bitcoin is a superior safe haven to Treasuries.

"The biggest difference in results came from the decision to get off zero — the choice of which bucket to fund the allocation from resulted in very marginal differences," — Chris Kuiper, CFA, Director of Research at Fidelity Digital Assets, "Getting Off Zero: Evaluating Bitcoin in 2026," March 25, 2026.

Kuiper's finding reframes the entire debate. If the funding-source question is close to a rounding error, then the headline "sell bonds for bitcoin" is answering the wrong question. The live disagreement among top-tier allocators — and the one worth a trader's attention — is about sizing and sequencing, not about which sleeve gets raided.

Why AI Capex Concentration Is Breaking the Classic 60/40

AI capital expenditure has concentrated equity risk to the point where the equity sleeve and the bond sleeve of a 60/40 portfolio are no longer two independent bets. BlackRock puts the top 10 S&P 500 names at roughly 37% of total market capitalization, up from 29% in 2020 and 19% in 2010 . When a third of an index moves on one thematic driver, the 40% in fixed income is being asked to offset a far narrower and more correlated exposure than the model was originally built to hedge.

J.P. Morgan Asset Management's mid-year 2026 outlook measured the Magnificent 7 at 34% of S&P 500 market value as of June 10, 2026, and described the index in plain terms: an increasingly directional bet on uninterrupted AI momentum . Two independent institutions, using slightly different baskets, land within three percentage points of each other. The direction of travel is not in dispute.

"The index is an increasingly directional bet on uninterrupted AI momentum," — J.P. Morgan Asset Management, Mid-Year Outlook 2026, Global AI Investment: Opportunities and Risk (source: J.P. Morgan Asset Management, 2026-06).

The capex numbers driving that concentration are large enough that the two main sources disagree on scale, and the gap is worth flagging rather than averaging away:

• Bitcoin Suisse: US hyperscaler AI spending of upwards of $800 billion in 2026, and more than $1 trillion in 2027 .
• J.P. Morgan Asset Management: roughly $700 billion in 2026 hyperscaler capex, a 70% increase over 2025 .
• The spread: roughly $100 billion, or about 14% of the lower estimate — different definitions of what counts as a hyperscaler and what counts as AI-specific spend. Attribute the figure to its source rather than treating either as settled.

What matters for allocation is not which number is correct but that both describe a spending cycle whose growth rate — 70% year over year on J.P. Morgan's measure — now sets the earnings trajectory for the most heavily weighted names in the benchmark. A deceleration in that capex line does not hit a handful of stocks; it hits roughly a third of the index simultaneously. That is the mechanism by which diversification inside the equity sleeve has quietly disappeared.

The pattern is also not a US-only artifact. J.P. Morgan extends the same observation to Asia, where a small group of South Korean and Taiwanese semiconductor names drive local benchmarks . For a trader who assumed geographic diversification was doing real work, that is an uncomfortable finding: buying Korea or Taiwan alongside the S&P 500 is, at the margin, buying more of the same AI supply chain under a different currency. International equity exposure and domestic equity exposure are increasingly two expressions of one trade.

J.P. Morgan's own prescription, worth noting because it is not the same as the one crypto-native research arrives at, is wider active selection across the AI ecosystem — data centers, power infrastructure, cooling, equipment suppliers — plus utilities, private markets and income-generating ballast . Bitcoin does not appear in that list. But the diagnosis behind it is identical to the one Bitcoin Suisse presents via CoinDesk: the equity leg has become a single-factor exposure, and adding more of a second leg that fails to cushion drawdowns does not fix a concentration problem. Whether the third return driver should be infrastructure equity, private credit or bitcoin is the argument. That a third driver is needed is close to consensus.

What the Bitcoin Suisse Model Actually Tested

Bitcoin Suisse's Crypto Wealth Management Report 2026 tested four fixed bitcoin weightings — 1%, 2.5%, 5% and 10% — inside an otherwise conventional portfolio of equities, bonds, gold and money-market instruments, and measured what happened when each slice was carved out of the stock sleeve versus the bond sleeve . The headline output: annualized return rose from 6.2% with no bitcoin to 7.2% at a 1% bond-funded allocation, and to 8.6% at 2.5% bond-funded . That is a 240-basis-point gain in headline return from a 2.5% position — a large effect from a small weight, which is the entire point of the exercise and also the reason the numbers deserve scrutiny.

Three details define what the study can and cannot support:

• The tested range is narrow by design. Nothing above 10% was modeled, so the report offers no evidence about what a 15% or 20% position would do to portfolio behavior. The allocations sit in the same 1–10% band that Fidelity Digital Assets and BlackRock also work within.
• Both absolute and risk-adjusted returns improved across the full tested range — and improved whether the position was funded from stocks or from bonds . The funding source changed the magnitude, not the sign. Anyone citing this report as proof that bonds specifically must be sold is over-reading it.
• The baseline is a four-asset mix, not a clean 60/40. Gold and money-market instruments are already in the control portfolio, so the measured improvement is bitcoin's contribution on top of existing diversifiers — a harder test than adding bitcoin to a bare stock-bond pair.

The framing inside the report is additive, not substitutive. Bitcoin Suisse's argument, as relayed by CoinDesk on September 12, 2026, is that investors need a third return driver distinct from both stocks and fixed income — not that bitcoin is a superior safe haven to Treasuries. The report does not claim bitcoin will hold its value during a liquidity event the way short-duration government paper does. It claims the equity risk premium and the term premium have collapsed into a single macro bet, and that a third, structurally different driver improves the arithmetic. Those are different arguments, and the second survives scrutiny far better than the first.

The macro case Bitcoin Suisse builds around the model is where readers should apply the most caution. The firm cites US hyperscaler AI spending upwards of $800 billion in 2026 and more than $1 trillion in 2027, US federal debt crossing $40 trillion, and Treasury yields back at Global Financial Crisis-era levels . That capex figure is contested: J.P. Morgan Asset Management's mid-year 2026 outlook puts 2026 hyperscaler capex at roughly $700 billion, a 70% increase over 2025 . A $100 billion spread between two credible institutions on the same variable is a reminder that the inputs feeding these allocation models carry real estimation error.

One structural caveat applies to the whole exercise: Bitcoin Suisse is a Zug-based digital-asset firm, and the study is published by a party with a commercial interest in higher bitcoin allocations. That does not make the arithmetic wrong — the direction of the result is corroborated by Bitwise's March 3, 2026 backtest, which found bitcoin added positively to a quarterly-rebalanced 60/40 in 76% of one-year periods and 100% of three-year periods since 2014 . But readers should verify specifics against the first-party report page rather than secondary coverage, and treat the return figures as backward-looking model output over a period in which bitcoin's realized return dominated every other asset in the mix.

Bitcoin vs Bonds vs Gold: Comparing the Diversifiers Side by Side

Bitcoin, bonds and gold diversify a portfolio through three unrelated mechanisms, and the correlation data separates them cleanly. BlackRock's portfolio-diversification work covering January 1, 2022 through Q1 2026 measured bitcoin's correlation to the S&P 500 at 0.53, gold's at 0.19, and bitcoin-to-gold at just 0.10 . That last number is the practically useful one: bitcoin and gold barely track each other, so holding both is not doubling up on the same trade. Bonds, by contrast, have been failing at the specific job of offsetting equity drawdowns .

Return history separates them further, though in a direction that demands care. Fidelity Digital Assets' "Getting Off Zero: Evaluating Bitcoin in 2026," published March 25, 2026 by Chris Kuiper, CFA, measured trailing 10-year CAGRs of roughly 70% for bitcoin, about 14.5% for stocks and about 2% for bonds . Fidelity did not carry that 70% forward. Its forward-looking inputs were deliberately conservative: bitcoin at 25% return and 50% volatility, stocks at 14.5% return and 15.5% volatility, bonds at 2% return and 5% volatility . In other words, the firm modeled bitcoin returning roughly a third of its historical rate while keeping its full historical volatility — the assumption set a trader should adopt when stress-testing a reallocation.

The mechanism behind bitcoin's diversification is liquidity, not duration. Fidelity found a 0.87 r-squared relationship between global M2 money supply and bitcoin's price . Bonds are priced off the policy-rate path and term premium; bitcoin has behaved like a claim on the size of the global monetary base. Those two drivers can move together or apart, which is exactly why the pairing adds information to a portfolio rather than duplicating a bet. Gold sits closer to bitcoin conceptually — both are scarce, non-yielding stores of value — yet the 0.10 realized correlation shows the two have not moved in step over the measured window .

Now the caveat that matters most for position sizing. BlackRock's 0.53 equity correlation is materially higher than the sub-0.3 full-sample figures that crypto-native research tends to cite, because the 2022-onward window captures the rate-shock and deleveraging regime rather than bitcoin's quieter early years . Conditional correlation is worse still: in forced-deleveraging episodes, bitcoin tends to trade with risk assets, not against them. Practical implications for a trader:

• Do not underwrite bitcoin as a crisis hedge. A 0.53 average correlation that rises under stress is a return diversifier, not portfolio insurance.
• Bitcoin and gold are complements, not substitutes. At 0.10 correlation to each other, replacing one with the other changes the portfolio's risk profile rather than trimming redundancy .
• Size against 50% volatility, not 70% returns. Fidelity's own forward inputs assume the historical CAGR does not repeat .
• Watch liquidity data, not the rate curve. If the M2 relationship is the transmission channel, global money supply is the variable to track .

One operational note from the same Fidelity work: rebalancing method changed outcomes measurably. Threshold-based rebalancing bands produced returns 1.76 percentage points below simple quarterly rebalancing . With a sleeve this volatile, the rebalancing rule is part of the allocation decision, not an afterthought.

Fund From Bonds or Stocks? Where BlackRock and Bitcoin Suisse Disagree

The funding-source question — whether a bitcoin sleeve should be carved out of bonds or out of equities — is the sharpest live disagreement among institutional allocators covering this topic in 2026, and the three most-cited studies reach three different answers. Bitcoin Suisse's Crypto Wealth Management Report 2026 models allocations funded from bonds, showing annualized returns rising from 6.2% with no bitcoin to 7.2% at a 1% bond-funded allocation and 8.6% at 2.5% . BlackRock recommends the opposite. Fidelity says the choice barely registers. For a trader sizing a position, that disagreement is the most useful part of the literature, because it defines the boundaries of what the data can actually support.

The bond-funded case rests on a logical chain rather than a claim about bitcoin's safe-haven properties. If bond returns were negative in 17 of the 19 months since 2020 when equities fell 2% or more — an 89% failure rate measured by BlackRock — then the fixed-income leg is already failing at its designated job, and reallocating a slice of it costs less diversification than the label suggests. Bitcoin Suisse pairs that with a macro read: US federal debt crossing $40 trillion and Treasury yields back at Global Financial Crisis-era levels . Worth noting: the Zug-based firm's own framing is additive, not substitutive. Its argument is that concentrated AI equity risk requires a third return driver, not that bitcoin replaces Treasuries as ballast .

BlackRock's guidance runs directly against that. In diversification work covering January 1, 2022 through Q1 2026, the firm measured bitcoin's correlation to the S&P 500 at 0.53, against 0.19 for gold and 0.10 between bitcoin and gold . Its conclusion: because of bitcoin's volatility profile, fund a bitcoin allocation from equities while funding most other alternatives from fixed income.

"A little bitcoin can go a long way," — BlackRock portfolio diversification research, on sizing and funding bitcoin allocations (source: BlackRock, 2026)

The mechanical reasoning is risk-budget arithmetic. Bitcoin carries equity-like-or-higher volatility and a 0.53 correlation to the S&P 500 over that window, which makes it closer to a high-beta risk asset than to a duration substitute. Swap it against bonds and total portfolio volatility rises on both sides of the trade — you add a volatile asset and simultaneously remove the low-volatility sleeve. Swap it against equities and the volatility added is at least partially offset by the equity exposure removed, while the bond leg continues to hedge whatever it still hedges. Traders should also note that BlackRock's 0.53 figure is materially higher than the sub-0.3 full-sample correlations crypto-native sources tend to cite, and that conditional correlations spike during deleveraging episodes — precisely when a hedge is needed.

Fidelity Digital Assets takes the argument apart from a third angle. "Getting Off Zero: Evaluating Bitcoin in 2026," published March 25, 2026 by Chris Kuiper, CFA, tested 1%, 3%, 5% and 10% allocations against a 60/40 base and found results nearly identical whether the sleeve was funded from equities, bonds or a split . Its stated conclusion: "The biggest difference in results came from the decision to get off zero — the choice of which bucket to fund the allocation from resulted in very marginal differences" .

One nuance softens but does not reverse that. Fidelity did observe that bond-funded versions edged out equity-funded ones historically, a direct consequence of equities compounding faster than bonds over the trailing decade — roughly 14.5% annualized for stocks against 2% for bonds . That is a backward-looking artifact of which leg performed worse, not evidence that bond-funding is structurally superior. Reading the three studies together, the defensible position is narrower than either headline: the sizing decision dominates, the funding source is second-order, and if a trader's priority is holding total portfolio volatility flat rather than maximizing backtested return, BlackRock's equity-funded route is the more conservative construction.

How Much Bitcoin Should a Trader Actually Allocate? A Decision Framework

The sizing evidence points to a band, not a single number: Fidelity Digital Assets' maximum-Sharpe optimization landed on a 9.4% bitcoin allocation, but the largest risk-adjusted improvement per unit added came far lower, at 1–3% . That gap is the whole framework. The first percentage point does the heaviest lifting for diversification; everything above roughly 3% is a return bet financed with volatility, and should be sized as such rather than justified by an optimizer output.

Fidelity's study, published March 25, 2026 by Chris Kuiper, CFA, tested 1%, 3%, 5% and 10% allocations against a 60/40 base and deliberately used conservative forward assumptions — bitcoin at 25% return and 50% volatility, stocks at 14.5% return and 15.5% volatility, bonds at 2% and 5% . Those inputs matter: the 9.4% optimum was not produced by extrapolating bitcoin's trailing 10-year CAGR near 70%, which would have pushed the number considerably higher. Traders using a lower forward return assumption than 25% should expect the optimal weight to fall accordingly.

Consistency across holding periods is the second input. Bitwise's March 3, 2026 study found that adding bitcoin to a 60/40 portfolio with quarterly rebalancing contributed positively to returns in 76% of one-year periods, 94% of two-year periods and 100% of three-year periods since 2014 , as documented by Bitwise. Read the numbers the right way round: roughly one in four one-year windows was a drag. A position sized so that a losing 12-month stretch forces liquidation is mis-sized regardless of what the three-year figure says.

Rebalancing discipline is the input most traders skip, and it carries a measurable cost. In Fidelity's testing, wide rebalancing bands produced returns 1.76 percentage points below a straightforward quarterly rebalance . That penalty is larger than the entire return improvement Bitcoin Suisse attributed to moving from a zero allocation to 1% funded from bonds — 6.2% to 7.2% annualized . A 1% allocation rebalanced on a calendar beats a 2.5% allocation left to drift, which reframes the question from "how much" to "how much can I actually maintain."

A sizing framework by risk profile

The table below maps the three institutional positions onto practical profiles. Funding source follows the disagreement already established: BlackRock prefers equity-funded for volatility control, Fidelity found the choice marginal, and Bitcoin Suisse modeled bond-funded allocations.

Three constraints apply across every row. First, treat 9.4% as a ceiling derived from one set of assumptions rather than a target — it is the point where Fidelity's optimizer stopped, not a recommendation. Second, the quarterly cadence is part of the allocation, not an administrative afterthought; the 1.76-percentage-point drift penalty is the cost of skipping it . Third, size against the one-in-four odds of a negative one-year window implied by the Bitwise data, not against the 100% three-year figure, because the three-year number is only reachable by a trader who survives the one-year outcomes.

For a trader moving from zero, the operationally honest starting point is the smallest allocation that still changes portfolio behavior — 1% — funded from equities, rebalanced quarterly, held through at least one full drawdown before any increase is considered. Scaling toward 2.5% or 5% should follow demonstrated behavioral tolerance rather than a fresh backtest, because the failure mode in this asset class is almost never an optimizer that was slightly wrong. It is a position that was correctly sized on a spreadsheet and abandoned during a 30% week.

Current Market Context: Bitcoin at $77,300 Meets a Rate Headwind

Bitcoin traded near $77,300 on September 13, 2026, up 0.1% on the day but down more than 3% on the week, while the US 10-year Treasury yield sat at 4.97% . That pairing is the practical test of every allocation argument above: the case for a third return driver arrived in the same week that the asset it recommends was falling and the asset it would be funded from was paying nearly 5% in nominal yield. A reallocation decision made on model output alone ignores that the competing instrument is currently at its most attractive entry point in years.

The mechanism is straightforward. A 4.97% risk-free-adjacent yield raises the hurdle rate for every non-yielding asset in the portfolio. Bitcoin produces no coupon, so its entire contribution has to come from price appreciation and diversification benefit. When the bond leg pays close to 5%, the opportunity cost of funding a bitcoin sleeve from fixed income rises in direct proportion — which is precisely the friction the Bitcoin Suisse framing has to overcome, and which its backtested 6.2%-to-7.2% return improvement does not price in at today's yield level .

The weakness was not bitcoin-specific. In the same session, ether fell 0.5% to $2,509 and XRP declined 0.4% to $1.3575 . A broad-based move across majors is more consistent with a rate-driven repricing of long-duration risk assets than with an idiosyncratic crypto event — the same condition under which conditional correlations tend to rise rather than fall. That matters for anyone treating a small bitcoin sleeve as drawdown insurance: the diversification benefit measured across a full sample is not the diversification benefit available on the specific days when rates are doing the driving.

One widely circulated figure deserves explicit caution. Reported spot bitcoin ETF outflows of roughly $462 million were cited by secondary outlets during the period, but that number was not verified against a primary issuer disclosure and should be treated as unconfirmed rather than as an input to any sizing decision. Flow data is the most frequently mis-cited category in crypto coverage, and a single week of outflows — even if accurate — is a poor basis for changing a multi-year allocation.

The reasonable reading is that current price action neither validates nor refutes the structural argument. The AI concentration problem and the bond-hedge failure rate are multi-year conditions; a 3% weekly drawdown and a 4.97% yield are a single snapshot. What the snapshot does change is entry discipline: staging a bond-funded allocation over several quarters, rather than executing it at once, is the cheaper way to be wrong about timing while still being right about structure.

Risks and Caveats Before Reallocating Away From Bonds

The strongest argument against a bond-funded bitcoin allocation is that bitcoin's diversification benefit is regime-dependent: the correlation that looks low on a full-sample chart rises during forced deleveraging, which is precisely when a portfolio needs a hedge to work. BlackRock measured bitcoin's correlation to the S&P 500 at 0.53 from January 1, 2022 through Q1 2026, against gold's 0.19 . That figure is materially higher than the sub-0.3 readings crypto-native sources tend to quote, and conditional correlations spike higher still in liquidation events. A bond sleeve that fails to cushion 89% of equity down-months is a poor hedge; an asset that co-moves with equities when leverage unwinds is not automatically a better one.

The macro premise underneath the thesis is also less settled than it appears. Bitcoin Suisse's case rests on US hyperscaler AI spending of upwards of $800 billion in 2026 and more than $1 trillion in 2027 , while J.P. Morgan Asset Management's mid-year 2026 outlook puts the same figure near $700 billion, a 70% increase over 2025 . A roughly $100 billion gap between two credible institutions is a reminder to treat capex as a range, not a settled input. Position sizing that only works at the top of that range is fragile.

Three further caveats belong in any reallocation decision:

• Nobody modeled a full swap. Bitcoin Suisse tested 1%, 2.5%, 5% and 10% ; Fidelity Digital Assets tested 1%, 3%, 5% and 10% against a 60/40 base in its March 25, 2026 study ; Bitwise's March 3, 2026 work likewise studied bitcoin added to a 60/40 sleeve . The tested range tops out at 10%. "Ditching bonds" is a headline, not a modeled scenario.
• A serious allocator reaches a different conclusion entirely. J.P. Morgan's prescription for AI concentration is wider active selection across the AI ecosystem — data centers, power infrastructure, cooling, equipment suppliers — plus utilities, private markets and income-generating ballast, not bitcoin . If the underlying problem is concentration rather than fixed income itself, that route addresses it without adding a 50%-volatility asset.
• Implementation drag is measurable. Fidelity found rebalancing bands produced returns 1.76 percentage points below quarterly rebalancing . The discipline matters as much as the sizing decision.

The concrete takeaway: treat this as a small, rules-based sleeve, not a bond replacement. A 1-3% position — where Fidelity located the most significant risk-adjusted improvement — funded gradually, rebalanced on a fixed quarterly schedule, and sized so a 50% drawdown in the sleeve stays survivable, captures most of the documented benefit while leaving the bond leg intact for the scenarios where duration still works.

Frequently asked questions

Does bitcoin actually replace bonds in a portfolio?

No. None of the institutional research reviewed here tested full replacement of a fixed income sleeve with bitcoin. Even Bitcoin Suisse, the firm pushing hardest on the idea in its Crypto Wealth Management Report 2026, frames the allocation as additive — a third return driver distinct from both equities and fixed income — rather than as a safe-haven substitute for Treasuries . The tested allocations topped out at 10% , which leaves the bond leg overwhelmingly intact. Bonds still carry duration exposure, contractual income and a liquidity profile bitcoin does not offer; the argument is about the marginal 1-10% of a portfolio, not the core.

Should I fund a bitcoin allocation from stocks or from bonds?

This is the sharpest live disagreement among top-tier allocators, and there is no consensus answer as of September 2026. BlackRock recommends funding bitcoin from equities — given its volatility profile — while funding most other alternatives from fixed income, summarized as "a little bitcoin can go a long way" . Bitcoin Suisse's modeling leans the other way, reporting annualized returns rising from 6.2% with no bitcoin to 7.2% at a 1% bond-funded allocation and 8.6% at 2.5% . Fidelity Digital Assets splits the difference: its March 25, 2026 study found "the biggest difference in results came from the decision to get off zero," with funding source producing "very marginal differences" .

What percentage of a portfolio should go to bitcoin in 2026?

Institutional models cluster between 1% and 10%. Bitcoin Suisse tested 1%, 2.5%, 5% and 10% ; Fidelity tested 1%, 3%, 5% and 10% against a 60/40 base . Fidelity's maximum-Sharpe optimization pointed to 9.4%, but that figure is an output of backward-looking optimization and sits far above what most allocators implement; the study located the most significant risk-adjusted improvement at 1-3% . The practical read: diminishing marginal benefit per unit of volatility above roughly 3%, so size the sleeve to what survives a 50% drawdown without forcing a sale.

Is bitcoin's correlation to stocks too high to call it a diversifier?

It is higher than crypto-native sources usually acknowledge, but not high enough to disqualify it. BlackRock measured bitcoin's correlation to the S&P 500 at 0.53 over January 1, 2022 through Q1 2026, against gold's 0.19 and a bitcoin-to-gold correlation of 0.10 . That is materially above the sub-0.3 full-sample figures often quoted. The more important caveat is conditional: correlations spike during deleveraging events — precisely when a hedge is supposed to work. A 0.53 average correlation with episodic convergence toward 1.0 describes a return diversifier, not a crisis hedge, which is why the research frames bitcoin as a third return driver rather than a replacement for defensive assets.

Why are bonds failing to hedge equities right now?

Because the shocks driving markets since 2020 have been inflationary, and inflation pushes both legs down together. BlackRock found bond market returns were negative in 17 of the 19 months since 2020 when equities fell 2% or more — roughly an 89% failure rate to cushion equity drawdowns . March 2026 was the second-weakest month for 60/40 portfolios since the 2022 drawdowns for exactly this reason . Rising yields compound the problem: the US 10-year traded at 4.97% on September 13, 2026, competing directly for the same allocation dollars .

Enjoyed this article? Subscribe to get new stories by email whenever they're published.

Subscribe