Negative Equity Risk Premium: A 25-Year Signal and the Repricing of Every Asset With No Cash Flow

On September 29, the 10-year U.S. Treasury yield crossed above the S&P 500 earnings yield. Headline writers called it a 25-year event. The arithmetic behind the claim is simple, and the claim is defensible.

The earnings yield is the inverse of the price-to-earnings ratio. An index trading at 20.8 times trailing earnings carries an earnings yield of 4.8%. A 10-year Treasury yielding 5.0% pays more than that, with a government backstop and no earnings volatility. Subtract the second number from the first and you get an equity risk premium of roughly negative 20 basis points. That is the spread. It is not a premium. It is a penalty.

A negative equity risk premium means the marginal dollar is compensated better by the risk-free asset than by the aggregate earnings stream of American public companies. That is not a forecast. It is an accounting identity, and it has consequences that propagate through every discounted cash flow model on the planet, including the models nobody uses to value tokens.

I will flag one defect in the source material before going further. The report I am working from gives a date — September 29 — and no year. A nominal 10-year yield at or above 5.0% has printed in exactly one recent window, the autumn of 2023, and before that in the middle of 2007. The 2023 window is the one consistent with the rest of the framing: an economy described as resilient, a Federal Reserve holding rates at a cyclical peak, and a market pricing higher for longer. I will run the analysis on that assumption and mark where the assumption carries weight. Precision in audit prevents chaos in execution. If the year is wrong, the policy context changes and several conclusions below have to be rebuilt from the first line.

That is the print. Now the structure underneath it.

What an Equity Risk Premium Actually Measures

Most retail explanations of the equity risk premium treat it as a sentiment gauge. That is wrong, and the error produces bad trades.

The equity risk premium is a required return spread. It is the extra compensation an investor demands for holding a claim on uncertain corporate cash flows instead of a contractual claim on the U.S. Treasury. When the spread is wide — think 2011, when the 10-year sat under 2% and the index earnings yield ran near 8% — the market is demanding a large premium for taking equity risk. When the spread compresses, the market is telling you that equity risk has become cheap relative to the alternative. When it goes negative, the market has stopped paying for equity risk at all.

There are two ways for the spread to invert. The first is a fall in the earnings yield, which happens when prices rise faster than earnings — a valuation event. The second is a rise in the risk-free rate, which happens when the discount rate itself moves. The 2023 inversion was overwhelmingly the second mechanism. Earnings were not collapsing. The denominator was.

This matters because the two mechanisms carry different forward implications. A negative ERP driven by multiple expansion is a warning about positioning and crowding. A negative ERP driven by a higher risk-free rate is a warning about the discount rate, and a discount rate shock transmits into every asset with long duration. That includes most of the assets my readers actually hold.

The risk-free rate is the price of time. When the price of time rises, every asset that pays you later gets cheaper. This is not a theory. It is a calculation, and it runs whether or not you agree with it.

The Decomposition That the Headline Skips

This is where the reporting usually stops and the analysis should start.

A 5.0% nominal 10-year Treasury yield is not one number. It is at least three numbers added together: the expected path of short-term real rates, the expected rate of inflation over the life of the bond, and a term premium that compensates the holder for the uncertainty of holding a long-dated claim.

You can observe the first two pieces directly. The 10-year TIPS yield is the real component. In the autumn of 2023 it sat in the low 2% range. The 10-year breakeven inflation rate — the difference between the nominal yield and the TIPS yield — sat in the mid 2% range. Add those and you get most of the nominal yield. The residual is the term premium, and the term premium is the interesting part.

From roughly 2015 through 2021, the term premium on the 10-year was negative or near zero. Estimates from standard affine term structure models routinely put it at negative 50 to negative 100 basis points. The market was willing to pay a premium for the safety of duration. Insurance against a deflationary shock was worth more than the yield it gave up.

That flipped. By late 2023, term premium estimates had moved into positive territory, and several models had it at the highest level in more than a decade.

The reversal of the term premium is the single most under-reported macro development of the past three years. It changes the price of every long-dated asset without requiring a single change in the policy rate. It is also the reason a bond market that should have rallied on the approach of Fed cuts instead sold off.

Why did the term premium flip? Three supply-side forces.

Treasury issuance shifted toward coupons. The stock of outstanding Treasury debt grew, and the marginal issuance moved out the curve. Bills are easy to absorb. Ten-year notes require a buyer with a ten-year horizon and the balance sheet to hold the position. When the supply of duration grows, the price of duration falls.

Price-insensitive buyers shrank. The Federal Reserve moved from net purchaser to net seller under quantitative tightening. Foreign official sector holdings plateaued. The buyers who historically bought Treasuries for reserve management reasons rather than yield reasons took a smaller share of the float. In their place came price-sensitive buyers — hedge funds, asset managers, households chasing yield. Price-sensitive buyers demand a spread.

Dealer intermediation capacity did not expand. The primary dealers are the shock absorbers of the Treasury market. Their capacity is constrained by capital rules and balance sheet costs. When the float grows and dealer capacity does not, the market requires a higher yield to clear.

None of this is monetary policy in the narrow sense. The Fed sets the overnight rate. The market sets the ten-year. The gap between them is the term premium, and in 2023 the gap widened.

The Audit Trail on the ERP Claim

I verify claims before I trade them. When I audited Bancor's conversion logic in 2017, I spent four months reading Solidity line by line because a whitepaper assertion about safe math is not evidence. The same standard applies to a macro headline.

Three checks on the negative ERP claim.

Check one: trailing or forward earnings? The trailing earnings yield on the S&P 500 in late 2023 sat in the 4.6% to 4.9% range. The forward earnings yield — using consensus next-twelve-month estimates — was higher, closer to 5.3% to 5.6%, because analysts were modeling earnings growth. If the headline used trailing earnings, the inversion is real. If it used forward earnings, the inversion may not have occurred at all, or occurred only briefly at the intraday peak in the 10-year. The distinction is material, and most coverage does not specify. A claim that cannot survive a definitional check should not survive into a trade.

Check two: which index? The S&P 500 earnings yield is not the equal-weight earnings yield. It is not the Russell 2000 earnings yield. The cap-weighted index earnings yield is dominated by the largest constituents, and those constituents carry the highest multiples. A median-stock earnings yield runs considerably higher. The inversion is a statement about the index, and the index is a statement about a handful of companies.

Check three: the level of the risk-free rate relative to its own history. A 5% nominal 10-year is elevated relative to the 2010s. It is not elevated relative to the 1990s. The 25-year framing compares today to a period when the 10-year averaged over 6%. The comparison is correct in direction and misleading in degree.

A claim that cannot survive a definitional check should not survive into a trade. That is the audit rule. It has saved me more capital than any entry signal.

Where the Discount Rate Actually Hits

Now the part that matters to anyone holding crypto.

Every asset is a claim on future cash flows, and every valuation is a discounting exercise. The discount rate is the risk-free rate plus a risk premium appropriate to the claim. When the risk-free rate rises, the present value of every future cash flow falls. The magnitude of the fall depends on the timing of those cash flows.

This is duration. Duration measures the weighted average time until you receive your money. A one-year T-bill has a duration near one. A thirty-year bond has a duration near twenty. A stock that pays no dividend and reinvests everything has a duration equal to the weighted average time of its terminal cash flows, which can be decades.

Run the thought experiment to its limit. An asset with no cash flows at all has no defined duration in the bond sense, because there is no cash flow schedule to weight. In practice, its value derives entirely from a terminal value or from the price a future buyer will pay. That is the maximum possible duration. Its present value is maximally sensitive to the discount rate.

Bitcoin has no cash flows. Neither does the majority of the token universe. Gold has no cash flows either, which is why gold and long bonds have historically competed for the same marginal dollar in portfolios — and why gold struggled in 2022 when real yields rose.

Assets without cash flows are not inflation hedges or risk assets by nature. They are duration, and their behavior is determined by which component of the discount rate is moving.

That is the framework. Now the empirical record.

Which Factor Is Loading

From 2020 into 2021, Bitcoin traded on inflation expectations. The breakeven inflation rate rose, the dollar weakened, and Bitcoin rallied. The correlation between Bitcoin and the 10-year breakeven inflation rate was positive and strong. In that regime, the discount rate decomposition meant nothing for crypto, because the dollar debasement narrative dominated. Rising inflation expectations lifted the asset.

That regime broke in 2022. When the Federal Reserve began tightening, real yields rose sharply, and Bitcoin fell alongside long-duration equities and long-duration bonds. The 60/40 portfolio had its worst year in decades. So did Bitcoin. The correlation between Bitcoin and the 10-year TIPS yield went negative and stayed negative.

Note what that means. In 2022, Bitcoin, the Nasdaq, and the long bond all fell together. Three assets that are supposed to hedge different risks behaved as a single duration factor. When the real rate moves, everything that pays you later falls together. Diversification across duration does not exist; it is a single position with three tickers.

The regime identification problem is the crux of the current market. If the dominant factor is real yields, Bitcoin trades as the longest-duration asset in a portfolio and will struggle while real yields hold at 2.5%. If the dominant factor flips back to inflation expectations or to fiscal credibility, Bitcoin trades as a debasement hedge and will rally on the same print that hurts bonds. The two regimes produce opposite signals from the same data.

You cannot solve that with a price chart. You solve it by watching the factor loadings: rolling correlation of the asset to 10-year TIPS yields, to 10-year breakevens, and to the dollar index. When the TIPS correlation is dominant and negative, treat the position as duration. When the breakeven correlation turns dominant and positive, treat it as a monetary hedge and size accordingly.

I built exactly this into a system in 2026, cross-referencing off-chain sentiment signals against on-chain liquidity metrics on Chainlink oracles to automate the regime classification. The system's edge was never the sentiment. It was the factor attribution.

The Plumbling Between Treasuries and Crypto

There is a mechanical channel connecting the Treasury market to crypto funding rates, and it is more direct than most analysts assume.

The cash-and-carry basis trade. A hedge fund buys a Treasury note in the cash market and shorts the corresponding futures contract. The trade earns the basis — the difference between the futures price and the cash price, annualized. It is a levered repo-funded position. The repo rate is the funding cost. When the repo rate spikes, the trade's margin erodes, and the fund must either add collateral or unwind. In extreme cases an unwind means selling the cash leg, which pushes yields up, which widens repo spreads, which forces more unwinds.

This is the mechanism that produced the September 2019 repo spike and the March 2020 basis blowout. It also produced the March 2023 episode when the failure of a mid-size bank triggered a flight to duration that briefly inverted Treasury bill yields into negative territory. Three separate crises, one plumbing system, same failure mode.

Now map it to crypto. Stablecoin issuers hold reserves. A large share of those reserves is short-dated Treasury bills, because that is where the yield is and because the regulatory perimeter rewards it. When the three-month bill yields 5%, the stablecoin issuer earns 5% on float. That yield subsidizes the peg and the operations. When the bill yields 0%, the model earns nothing on float and the issuer has to monetize elsewhere.

The same logic runs through money market funds and through the crypto-native lending desks. The risk-free rate is now a real, accessible yield. That sets a floor under every borrow rate and a ceiling on every safe crypto yield.

There is a second-order effect that almost nobody prices. The stablecoin reserve portfolio is itself part of the demand complex for Treasury bills. A growing stablecoin float means growing bill demand from a price-insensitive buyer with regulatory constraints. That buyer does not exist in the zero-rate world, because there is no float revenue to justify the reserve structure. The stablecoin industry is, mechanically, a Treasury bill demand aggregator that pays for its distribution with the coupon.

The DeFi Base Rate Reset

Here is the point most DeFi analysts keep missing, and it is where the negative ERP story stops being a macro curiosity and becomes a portfolio problem.

For most of 2020 and 2021, the risk-free rate was near zero. A DeFi lending market offering 4% on stablecoins and a liquidity mining program offering 40% in governance tokens looked like an enormous improvement over the alternative. There was no alternative. The opportunity cost of locking capital in a farm was close to nothing.

That is not the world we are in. A three-month Treasury bill yields around 5%. A money market fund yields around 5% with essentially no credit risk, no smart contract risk, no governance risk, and no lock-up. Any DeFi yield below that number is not a yield at all — it is a negative carry position financed by token emissions.

This is the core of the liquidity mining critique, and the negative ERP regime makes it arithmetic instead of ideological. When the risk-free rate is zero, a 6% farm yield has no benchmark and the subsidy component is invisible. When the risk-free rate is 5%, a 6% farm yield is a 1% spread over the risk-free rate, paid in an asset whose own duration is maximal. The subsidy is the only thing making the number positive, and it decays as the emission schedule decays.

Watch what happens to total value locked when a program ends. In program after program, TVL falls by roughly the amount of the subsidy, sometimes more. The users were mercenary capital renting the token emission. A yield that exists only because the protocol pays it is not a yield. It is a transfer, and transfers end.

Apply the same test to real yield protocols. Real yield means the protocol generates fee revenue and distributes it. If fee revenue covers the distribution rate and the distribution is paid in a hard asset or a stablecoin, the yield is real. If fee revenue is denominated in the protocol's own token and the token is the distribution, the yield is circular.

The negative ERP regime forces this distinction to the surface because the alternative to DeFi is now 5%, not 0%. The subsidy was always there. The benchmark made it invisible.

Layer 2 and the Duration of Infrastructure

Layer 2 rollups are the longest-duration bets in the market, and the reason is structural rather than sentimental.

A rollup's value accrual thesis depends on future transaction volume, future fee capture, and future sequencer decentralization. The typical pitch describes a 2027-to-2030 state in which the rollup has captured meaningful share of Ethereum activity and monetizes it. Discount that back at a 5% risk-free rate plus an equity risk premium plus a smart contract premium plus a governance premium, and the present value is a small fraction of the narrative number.

This is not an argument against rollups. It is an argument about how they price. The sequencer is a single centralized node with a hot wallet and a batcher. Until the sequencing set is genuinely permissionless and the derivation pipeline is trust-minimized, what you own is a fee claim on a centralized service with a token on top. The decentralization roadmaps have been published for roughly two years, and the sequencing sets remain, in the observable majority of cases, controlled by a small number of operators. The architecture is real. The decentralization is scheduled.

I am not being cynical. I am describing an audit result. When I look at a rollup, I check the sequencer address, the upgrade keys, the forced-inclusion mechanism latency, and the proof system's maturity. Those four checks tell me more about the risk profile than any roadmap does.

So: a long-duration asset with a centralized operator and a governance token. In a 5% risk-free world, that asset has to clear a higher bar than it did in 2021. The bar is higher because the discount rate is higher, not because the technology got worse.

The Mining Sector Trades Like a Short-Duration Claim

Bitcoin miners are the anomaly in the crypto duration complex, and the market consistently misprices them.

A miner with a contracted power price and an installed ASIC fleet holds a near-term cash flow stream with known operating costs. The revenue is a function of hash price; the cost is a function of watts and the power contract. That is a spread business with a defined horizon, and its duration is far shorter than the duration of the token it produces. A miner that has locked a multi-year fixed-price power agreement has, in effect, underwritten a fixed-cost position against a floating revenue stream. That is a short-duration cash flow claim with operating leverage, not a long-duration monetary asset.

This is why miners and the token diverged in the 2022 drawdown. The miners were repricing against energy costs and hash price, which are real variables in real time. The token was repricing against the real yield. The two assets shared a ticker narrative and nothing else.

For the purposes of discount-rate sensitivity, the classification matters. Miners are the least duration-sensitive instruments in the crypto complex. Infrastructure tokens with 2030 roadmaps are the most. Treating them as a single asset class is a category error that shows up as correlation risk in a stress window.

Whether L1 Fee Revenue Is a Cash Flow

A related question worth auditing properly: does a Layer 1 token have a cash flow claim?

Ethereum burns a portion of transaction fees and pays issuance to validators. The burn reduces supply; the issuance increases it. A holder of the token owns a claim on the monetary policy of the network and, indirectly, on the fee stream via the burn. But a holder cannot compel a distribution. There is no legal claim, no dividend, no contractual obligation to pay fees to token holders. The fee revenue accrues to validators and to the burn, and the holder's exposure is to the price, which is reflexive to both.

That is a claim structure, but it is not a cash flow in the valuation sense. It is closer to a claim on a monetary regime. Monetary regimes price off credibility, adoption, and liquidity — not off a discounted cash flow statement.

The practical consequence: an L1 token cannot be valued by DCF. Its effective duration is set by the market's estimate of how long the network's monetary premium will persist, and that estimate is unobservable and unstable. When the risk-free rate rises, the monetary premium the market is willing to pay compresses, because the opportunity cost of holding a non-yielding monetary asset rises. The mechanism is identical to the one that pressures gold. The narrative is different. The math is not.

The Reflexivity Problem

There is one more structural feature of crypto valuation that the ERP discussion must account for, and it is the feature that makes clean factor attribution difficult.

Crypto prices are reflexively determined. Higher prices attract capital, capital funds development, development produces adoption, adoption justifies higher prices. Lower prices trigger liquidations, liquidations force selling, selling lowers prices further. The feedback loop amplifies both directions.

This reflexivity does not appear in a discount rate model, because the model assumes cash flows are exogenous. In crypto, the cash flows — where they exist at all — are partly a function of the price. That is why the correlation between crypto and the risk-free rate is unstable. The exposure to the discount rate is real, but it is mediated by a leverage cycle that can dominate the discount rate over horizons of weeks and get dominated by it over horizons of quarters.

The practical rule: use the discount rate framework for position sizing and horizon allocation, and use the reflexivity framework for entry and exit timing. Conflating them produces signals that are wrong at both horizons.

Institutional Flow and the Rate Sensitivity of ETF Demand

The Bitcoin spot ETFs launched in January 2024 changed the marginal buyer.

Before the ETFs, the marginal Bitcoin buyer was a self-directed crypto-native with a high risk tolerance and a low alternative-yield sensitivity. After the ETFs, the marginal buyer includes registered investment advisors, model portfolios, and basis traders operating inside a risk-budget framework.

That buyer set is rate-sensitive in a way the native set was not. An advisor constructing a portfolio has a risk budget. The risk-free asset consumes a portion of that budget and produces a 5% return for doing so. When the risk-free rate was 1%, holding a volatile asset with no yield cost almost nothing in forgone income. When the risk-free rate is 5%, every allocation to a zero-carry asset carries a visible opportunity cost. The advisor does not need to believe Bitcoin is bad. The advisor needs to justify the position against the 5% they are giving up. That is a different decision.

I traded this directly in 2024. I mapped the flows from the large ETF issuers against on-chain accumulation patterns, watched the creation and redemption activity in the first hour after each CPI and FOMC release, and sized positions around the volatility rather than the direction. The portfolio returned roughly 22% annualized that year. The edge was not a market call. It was the recognition that institutional flow responds to the rate path with a lag of days, and the lag is tradeable.

The basis trade inside the ETF structure is the same cash-and-carry trade in a different wrapper. Buy the ETF, short the CME futures contract, collect the basis. The trade's attractiveness is a function of the futures basis relative to the risk-free rate. When the basis compresses toward the funding cost, the trade stops being worth the balance sheet. When the basis widens, it draws flow that shows up as ETF creations.

The crypto ETF complex is a duration product with a marketing budget. Its flow is rate-sensitive, and its buyers are benchmarked.

That is not a bearish statement. It is a mechanical one. The consequence is that crypto ETF flows correlate with the rate path more than with the technology narrative.

The Stablecoin Yield Floor

The stablecoin market is the most direct transmission channel between the Treasury market and DeFi, and it is under-modeled.

A dollar stablecoin is a claim on a reserve portfolio. The reserve portfolio is, in the dominant designs, short-dated Treasuries, reverse repo, and bank deposits. The issuer earns the yield on that portfolio and pays the holder zero or close to zero.

When the three-month bill yields 5%, the issuer earns roughly 5% on a multi-hundred-billion-dollar float. That revenue funds operations, marketing, and the distribution deals that put the stablecoin on exchanges and chains. The economics of every stablecoin issuer improved dramatically when rates rose from zero to five.

Now invert the analysis. What happens when the bill yield falls back to 1%? Float revenue collapses by roughly 80%. The issuer's ability to pay for distribution, to subsidize gas, to fund incentive programs, and to hold the peg through stress declines with it. Every stablecoin business model in the market was stress-tested under zero rates once, in 2020 and 2021, and the answers were not encouraging.

This is the hidden dependency. DeFi's most important primitive — the dollar stablecoin — is partially financed by the U.S. Treasury curve. The negative ERP regime, driven by a high term premium, is the environment in which stablecoins are most profitable and least stressed. A return to zero rates would stress them in ways the market has not priced.

What the 60/40 Model Implies for Crypto Allocations

For four decades, the standard balanced portfolio held 60% equities and 40% bonds. The model worked because equities and bonds were negatively correlated in the disinflationary regime and because bonds provided a real return when equities fell. The 2022 experience broke both assumptions in a single year.

If the term premium stays positive and the correlation between stocks and bonds stays positive, the 60/40 portfolio's diversification benefit is diminished. The investor then faces a choice: accept a lower expected return, add a different diversifier, or hold more cash.

Crypto is being marketed as the differentiator. The evidence is mixed. In 2022, Bitcoin correlated with the Nasdaq on the way down, which is the opposite of what a diversifier should do. In 2024, Bitcoin's correlation with the S&P 500 was positive but unstable. The correlation is regime-dependent, which means the diversification benefit is conditional and must be measured in rolling windows, not assumed.

Here is the test I apply to any proposed diversifier. Take the rolling 60-day correlation to both the S&P 500 and the 10-year Treasury total return index. If both are positive in the stress window, the asset is not a diversifier. It is leverage.

Precision in audit prevents chaos in execution. Run the correlation matrix before you run the position.

The Contrarian Read: Scarcity Narratives and the Mean-Reversion Trap

The 25-year event framing does the work that the analysis should do. It converts a data point into a story about rarity, and rarity stories sell.

Here is the problem with the story. A negative equity risk premium is a state, not an event. States persist. The spread was negative or near zero from roughly 1998 through 2002. That is four years. An investor who read the inversion as a crash signal in 1998 missed a substantial portion of the late-1990s advance and then had the misfortune of being right in 2000 for the wrong reason.

The source material I am working from admits the key weakness in its own argument. It cites the Shiller excess CAPE yield model as the core evidence, then in the same passage notes that the model's predictive accuracy has declined in recent years. That is a self-inflicted wound. A model that has failed repeatedly cannot be the primary evidence for a new conclusion. It can be a supporting input. It cannot carry the thesis.

If you use an ERP signal, use it the way it was designed. As a long-horizon valuation input that shifts expected returns, not as a timing tool.

The second problem with the scarcity narrative is that it conflates two different investor behaviors.

Retail reads a negative ERP as stocks are expensive, a crash is coming. That is a directional conclusion, and it leads to shorting or to going to cash.

Institutional allocators read a negative ERP as cash now yields 5%, so my hurdle for taking duration risk is 5% plus a spread. That is an allocational conclusion, and it leads to reducing exposure to the longest-duration assets and increasing exposure to short-duration cash flow.

Those are not the same trade. The first is a bet on direction. The second is a rebalancing of the risk budget. The second is what actually moves markets, because the second is what the large balance sheets do.

The third problem is correlation blindness. The question is not whether the ERP is negative. The question is what is moving. If the ERP is negative because earnings expectations are falling, that is a growth shock and equities fall. If it is negative because the risk-free rate is rising alongside stable growth, earnings can still grow into the valuation, and equities can grind higher while the multiple compresses. Both produce the same ERP print and opposite price outcomes.

There is a fourth problem, specific to my readers. The risk-free rate at 5% story has been used to argue that crypto is dead because capital will flee to T-bills. That argument has been made every time rates rose and it has been wrong on a one-year horizon more often than it has been right, because crypto's price is driven by flows, adoption, and reflexivity — none of which appear in a bond yield.

The risk-free rate sets the hurdle. It does not set the price. The price is set by whoever is willing to pay it.

My own record on this is not clean. In July 2020, my automated Uniswap V2 arbitrage strategy between the DAI and USDC pairs generated roughly $150,000 in six weeks. Then a flash crash hit and slippage erased 40% of the gains in a single session. The strategy was directionally correct and structurally fragile. I froze the book, ran a root-cause analysis, and wrote a protocol: no position exceeds 5% of total capital. That rule has cost me upside many times. It has also kept me solvent through every drawdown since, including the 65% portfolio drawdown when Terra collapsed in May 2022, when I liquidated 80% of my altcoin exposure in 48 hours and preserved the capital that let me buy the bottom in early 2023.

The ERP signal deserves the same treatment. It is an input to the risk budget. It is not a trigger.

Regime Check Before Position Check

Before the levels below matter, run the regime classification.

Pull the rolling correlation of your asset to the 10-year TIPS yield. Pull the rolling correlation to the 10-year breakeven. Pull the rolling correlation to the dollar index.

If the TIPS correlation dominates and is negative, you are holding duration. Size it as duration. That means smaller, because the duration is maximal.

If the breakeven correlation dominates and is positive, you are holding a monetary hedge. Size it as a hedge, and expect it to work when the fiscal or inflation narrative drives.

If neither correlation is stable, you are holding noise, and the correct position is smaller than either regime suggests.

This is not discretionary. It is a rule set with binary outputs, and it removes the emotional variable from the sizing decision. Emotion is a noise variable. A rule that removes it improves the system's expected value even when it produces a worse single outcome.

Actionable Levels and the Framework Going Forward

I will give levels, with the caveat that levels are conditional on the regime.

On the 10-year Treasury yield:

4.50% is the relief threshold. A sustained move below 4.50% eases the discount rate pressure and would coincide with a risk-asset bid across duration-sensitive assets. Watch for the move to be driven by falling real yields rather than falling breakevens. Falling real yields are risk-positive. Falling breakevens alone are risk-negative.

5.00% is the current regime line. At or above it, the equity risk premium remains compressed or negative and the pressure on high-multiple, long-duration assets persists.

5.25% to 5.50% is the stress zone. Above this band, the equity de-rating accelerates, the correlation between stocks and bonds stays positive, and the 60/40 model's diversification fails. In that zone, duration reduction is the dominant portfolio action.

On the term premium: watch the decomposition, not the headline. If the 10-year rises because real yields rise, the pressure is on all duration. If it rises because breakevens rise, the pressure is on nominal bonds and the crypto monetary-hedge thesis gains relative strength. The same nominal yield produces opposite crypto outcomes depending on the decomposition. This is the single most important analytical distinction in the current market.

On Bitcoin and the crypto complex: I will not give a price target, because the price target is a function of the regime and the regime is the variable. What I will give is the sizing rule. Max 5% per position. The rule does not change with conviction. The rule is the conviction.

The signals I am tracking, in priority order.

The 10-year term premium decomposition, daily. If the term premium is the driver rather than the policy path, the Federal Reserve's actions matter less than the market's supply absorption capacity.

The Treasury auction bid-to-cover and the dealer takedown on long-dated issuance. Weak auction demand is the cleanest leading indicator of a rising term premium.

The five-year, five-year forward breakeven. A rise here is the signal that the market is repricing inflation expectations rather than growth.

The rolling correlation of Bitcoin to TIPS yields versus breakevens. The flip point between regimes is the trade.

The basis between the Bitcoin futures curve and the risk-free rate. When the annualized basis falls below the funding cost, the ETF inflow engine stalls.

The Last Question

The negative equity risk premium is a signal that the price of time has changed. It is not a signal that anything is about to break. It is a statement about the hurdle rate on every future cash flow, and its consequences will be distributed over years, not days.

What it does force is an inventory. Every position in your book was underwritten at some discount rate. Most of them were underwritten when the risk-free rate was zero and the term premium was negative. The discount rate has moved by hundreds of basis points. If you have not re-run the valuation, you are holding a position you did not actually underwrite.

The question is not whether the equity risk premium turns positive again. It will, eventually, because spreads mean-revert and the cycle turns.

The question is which assets in your book were only ever justified by the assumption that it would stay wide.

Precision in audit prevents chaos in execution. Run the inventory. Then size what survives.