BofA expects the “Bessent put” to cap but not lower Treasury yields, with gold still the preferred hedge

Overview

The report examines whether the U.S. government can stop the “Anything But Bonds” trade. Its central judgment is that Treasury buybacks of long-term government bonds, foreign-exchange intervention, and Federal Reserve cooperation have created a succession of “Bessent puts.” These policies may hold the 30-year Treasury yield near 5%, but are insufficient to reduce the cost of capital significantly. The report therefore also proposes contrarian allocations to gold, duration assets, Hong Kong real estate, and the AI financing chain, while using fund flows and the BofA Bull & Bear Indicator to assess market crowding.

Core views

The report begins by using the failure to meet the U.S. government's “three arrows” to show that policy credibility is declining: GDP growth has been below 2% over the past six quarters, missing the 3% target; the budget deficit is approximately 6% of GDP, twice the 3% target; and daily oil production has increased by only 0.3 million barrels since 2024, far below the target increase of 3 million barrels. The report treats bonds and foreign exchange as thermometers of credibility, arguing that rising yields and a weakening dollar signal an erosion of trust. The government is therefore attempting to defend three “Maginot Lines”: gasoline at $4 per gallon, USD/JPY at 160, and long-term U.S. Treasury yields at 5%. Yet gasoline has risen from approximately $3 before the war to above $4, while the U.S.-Iran “economic war,” together with U.S. crude inventories and the Strategic Petroleum Reserve standing at 40- to 50-year lows, makes suppressing energy prices more difficult. Intervention in USD/JPY, meanwhile, would require the Bank of Japan to deliver a meaningful rate hike on September 18 to stabilize Japan's long-term bond market. Policy attention has therefore shifted toward preventing the 30-year Treasury yield from breaking above 5%. The report views the Treasury's doubling of long-term Treasury buybacks as a new “Bessent put” and a form of quasi-quantitative easing. The backdrop is that total U.S. government debt has just exceeded $40 trillion, with net Treasury issuance of approximately $2 trillion in each of 2026 and 2027. At the same time, total investment-grade bond issuance in 2026 is approximately $2 trillion; AI-related bond issuance has already reached $200 billion to $300 billion year-to-date, and hyperscaler bonds account for approximately 9% of total investment-grade bond supply in 2026. Treasury financing will crowd out corporate financing capacity, while the U.S.-China AI race is regarded by the U.S. government as a paramount national-security and macroeconomic priority. The fact that the 30-year yield still rose to a 20-year high despite no growth in nonfarm payrolls and zero inflation readings further pressures the authorities to restore credibility quickly. On this basis, the report identifies a sequence of policy actions, including U.S. dollar swap lines with Asia and the Gulf region, yen intervention, and the doubling of long-term Treasury buybacks. It judges that these measures can constrain further yield increases but cannot genuinely lower yields; materially reducing the cost of capital for sectors such as AI would still require deflation or recession. The report argues that the past 20 years of unconventional monetary stimulus were a major source of the asset bull market and Wall Street's belief that it is “too big to fail.” The market should therefore begin by assuming that the current policies aimed at “fixing fixed income” will succeed. The report even argues that it would represent an exceptionally severe policy failure if a new round of quantitative easing still could not bring the 30-year yield back below 5%. In a successful scenario, upside in yields would be constrained and neglected long-duration assets should receive support. In a failure scenario, the U.S. dollar could fall sharply, prompting investors to short risk, leveraged, and cyclical assets, particularly AI hyperscalers, private credit, and financials. The report also argues that fiscal easing requires support from a “just hawkish enough” Warsh at Jackson Hole to avoid further damage to the dollar. The relevant speech is scheduled for 10:00 a.m. on August 28, coinciding with the release of the U.S. Bureau of Labor Statistics' annual nonfarm payroll revision. In terms of specific contrarian trades, the report counters the “Anything But Bonds” consensus by favoring long-duration REITs, XBI, KRE-represented regional banks, and small caps. The rationale is that another step higher in yields would damage both Wall Street and the real economy, forcing an escalation in policy intervention. Against the “Anywhere But China” consensus, the report is bullish on Hong Kong real estate, arguing that the two major bear markets of the 2020s—China and real estate—may be nearing an end. Regarding the “Anything But the Dollar” consensus, the report agrees that the dollar remains under pressure and regards gold as the most direct dollar-depreciation trade, simultaneously capturing a bond rout, asset inflation, political populism, and sanctions demand stemming from the U.S.-Iran “economic war.” On AI, the report does not reject the structural theme in AI stocks, but recommends shorting AI bonds. The government can cap yields but cannot materially reduce financing costs without deflation or recession, while the supply of AI-related bonds is expanding rapidly. For investors who remain long AI stocks, the report proposes a barbell with commodities and natural resources, because these resources are inputs for the power and infrastructure required by the U.S.-China AI race. Semiconductor ETF flows have also weakened, with outflows of $0.7 billion during the week and cumulative outflows of approximately $6.3 billion over the past three weeks, showing that enthusiasm across AI-related markets is not moving uniformly. The U.S. midterm elections represent another important risk channel. The market consensus does not expect a Democratic sweep in November, but the report recommends hedging by shorting financials. If Democrats take the Senate and Republicans lose the Texas governorship, the report expects equities could fall by more than 10% before year-end, while the U.S. dollar and bond yields would also decline. Its evidence includes Trump's overall approval rating of only 39%, with approval on the economy and inflation at 36% and 30%, respectively, all below prewar levels of 46%, 42%, and 38%. This leaves financials exposed simultaneously to policy, cyclical, and electoral surprise risks. Fund-flow data show that risk appetite remains strong: equities recorded inflows of $40.1 billion during the week, comprising $45.9 billion of ETF inflows and $5.8 billion of mutual-fund outflows; bonds recorded inflows of $21.4 billion, marking 69 consecutive weeks of inflows; gold received $3.8 billion, crypto assets $1.0 billion, and cash saw outflows of $1.2 billion. U.S. equities attracted $28.9 billion, the largest inflow in three weeks; U.S. Treasuries received $7.4 billion, the largest in six weeks; investment-grade bonds attracted $7.5 billion, extending their inflow streak to 20 weeks; and emerging-market bonds received $3.3 billion, the largest in 11 weeks. Relatively weak areas included South Korean equities, with outflows of $0.8 billion for their first outflow in eight weeks; Japanese equities, with outflows of $1.7 billion for their first outflow in 11 weeks; and financials, with outflows of $2.0 billion, the largest in 11 weeks. Positioning indicators, however, are issuing a clear warning. The BofA Bull & Bear Indicator rose from 9.3 to 9.5, driven by improved global equity-market breadth and long positioning in S&P 500 and gold futures. Among its components, Fund Manager Survey positioning is in the 99th percentile, equity fund flows in the 93rd percentile, and hedge-fund positioning in the 83rd percentile. BofA characterizes overall positioning as “extremely bullish,” with the Bull & Bear Indicator maintaining its sell signal. Since the signal was triggered on May 26, the S&P 500 has risen 1.6%, although it experienced a 4% drawdown between June 2 and July 29; over the same period, the MSCI All Country World Index rose 1.3%, while HYG fell 0.8%. This indicates that policy support and strong inflows have not ended, but crowding has increased significantly. BofA private clients have $4.7 trillion in assets under management, allocated 66.5% to equities, 16.9% to bonds, and 9.4% to cash. Weekly bond inflows were the largest since May 2026. Over the past four weeks, private clients bought Japanese equities, municipal bonds, and TIPS ETFs, while selling emerging-market bonds, utilities, and financial-sector ETFs. Of their cumulative inflows into U.S. Treasuries since 2020, $52 billion went into intermediate-term Treasuries and $28 billion into Treasury bills, also reflecting a tendency to add interest-rate assets against a backdrop of high equity allocations. Year-to-date cross-asset returns further demonstrate market divergence: oil rose 53.0%, global equities 13.9%, U.S. equities 11.6%, gold 4.0%, high-yield bonds 2.6%, cash 2.3%, the U.S. dollar 0.6%, and investment-grade bonds 0.1%, while government bonds fell 1.1% and Bitcoin fell 16.9%. The report therefore does not provide a single, broad directional judgment on risk appetite. Instead, it combines policy support, U.S. dollar depreciation, extremely bullish positioning, and election tail risk into a strategy that favors gold, selectively goes long duration assets, and avoids highly leveraged and cyclical risks.

Analysis framework

The report first measures U.S. policy credibility against three objectives—economic growth, the fiscal deficit, and oil production—and then uses gasoline prices, USD/JPY, and the 30-year Treasury yield to identify policy defense lines. It subsequently combines Treasury supply with investment-grade and AI bond issuance to assess whether quasi-QE measures such as long-term Treasury buybacks can ease financing pressure; on that basis, it constructs two asset-response paths for policy success and failure. Finally, it uses global fund flows, BofA private-client allocations, the Fund Manager Survey, futures positioning, and the Bull & Bear Indicator to test market consensus and crowding, and derives contrarian cross-asset expressions accordingly.

Methodology notes

• Event-driven strategy and behavioral financeFund flow/positioning analysis

  Analysis of global fund flows, private-client allocations, and futures positioning

  The report uses weekly subscriptions and redemptions across assets and regions, private-client holdings, and positioning in S&P 500 and gold futures to identify where capital is moving toward or away from and to judge whether trades are crowded.

• Event-driven strategy and behavioral financeExpectation gaps/expectation management

  ABB, ABC, ABD, and AI contrarian trades

  The report first defines the market's long-standing consensus, then searches for areas where policy pressure, neglected valuations, or financing constraints could reverse that consensus, forming contrary expressions through gold, Hong Kong real estate, duration assets, and AI bonds.

• Event-driven strategy and behavioral financeEvent-driven analysis

  Treasury buybacks, Jackson Hole, the Bank of Japan meeting, and U.S. midterm-election scenarios

  The report treats specific policy and political events as catalysts and analyzes how they transmit to long-term yields, the U.S. dollar, equities, financials, and AI financing costs.

• Quantitative/factor/portfolio theory

  BofA Bull & Bear Indicator

  This proprietary indicator combines hedge-fund positioning, equity and bond fund flows, credit technicals, global equity-market breadth, and Fund Manager Survey positioning. The current reading rose from 9.3 to 9.5, which the report interprets as “extremely bullish” and a sell signal.

• Quantitative/factor/portfolio theory

  Deviation from the 200-day moving average

  The report compares assets' deviations from their 200-day moving averages in U.S. dollar terms to identify overbought and oversold conditions across asset markets.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Gold

  The report identifies it as the preferred long and the most direct U.S. dollar depreciation trade.

  Strengths

  It can simultaneously express U.S. dollar depreciation, bond stress, asset inflation, political populism, and sanctions demand; it recorded inflows of $3.8 billion during the week.

  Weaknesses

  Long positioning in S&P 500 and gold futures has already pushed the Bull & Bear Indicator into extremely bullish territory.

  Comparison

  Compared with other U.S. dollar depreciation trades, the report regards gold as the most direct expression.

  Risks

  A restoration of policy credibility or reversal of crowded positioning could weaken the trade.

• Long-duration REITs, XBI, KRE, and U.S. small caps

  The report treats these neglected duration assets as contrarian longs against the “Anything But Bonds” consensus.

  Strengths

  If the economic damage from a further rise in yields forces an escalation in policy intervention, these rate-sensitive assets could benefit.

  Weaknesses

  The current cost of capital remains high, and the report expects policy only to cap yields rather than lower them materially.

  Comparison

  Relative to high-momentum assets pursued by the market, the report defines these areas as out-of-favor contrarian trades.

  Risks

  If the 30-year Treasury yield cannot be held below 5%, duration- and financing-sensitive assets may remain under pressure.

• Hong Kong real estate

  The report uses it to express a contrarian trade against the “Anywhere But China” consensus.

  Strengths

  The report believes that the two major bear markets of the 2020s—China and real estate—may be nearing an end.

  Weaknesses

  The asset is simultaneously exposed to two themes that have faced prolonged pressure: China and real estate.

  Comparison

  Against the consensus of avoiding Chinese assets, the report selects Hong Kong real estate as the clearest contrarian expression.

  Risks

  The report does not provide a specific timeline or quantitative confirmation criteria for the end of the bear market.

• AI bonds and hyperscalers

  The report advocates shorting AI bonds and, in a policy-failure scenario, shorting highly leveraged risks such as AI hyperscalers.

  Strengths

  The AI race is regarded by the U.S. government as a national-security and macroeconomic priority and may therefore receive policy financing support.

  Weaknesses

  AI bond issuance has already reached $200 billion to $300 billion year-to-date, while the cost of capital may decline materially only under deflation or recession.

  Comparison

  The report distinguishes between AI stocks and AI bonds: investors may retain long exposure to the former but should remain cautious on the latter.

  Risks

  Treasury supply crowding out corporate financing, persistently high long-term yields, and failure of quasi-quantitative easing.

• Commodities and natural resources

  The report recommends using them to construct a barbell portfolio with long positions in AI stocks.

  Strengths

  These resources are important inputs for the power and infrastructure required by the U.S.-China AI race.

  Weaknesses

  The report does not provide a uniform direction or target level for specific commodities.

  Comparison

  Compared with holding AI stocks alone, commodities and natural resources are used to diversify financing-cost and valuation risks.

• U.S. financials

  The report uses short positions in financials as a hedge against U.S. midterm-election and cyclical risks.

  Weaknesses

  The sector recorded outflows of $2.0 billion during the week, the largest in 11 weeks; a policy-failure scenario would also be unfavorable for cyclicals.

  Comparison

  Compared with duration assets supported by policy, financials are used to express election and cyclical downside risks.

  Risks

  If Democrats sweep the midterm elections, the report expects equities could fall by more than 10%, with financials potentially suffering a greater impact.

• Private credit

  The report identifies it as a leveraged asset to short in a scenario where quasi-quantitative easing fails.

  Strengths

  A new round of liquidity support could theoretically support risk assets.

  Weaknesses

  High leverage and high financing costs make it relatively sensitive to policy failure.

  Comparison

  Together with AI hyperscalers, the report identifies it as a primary leveraged risk in the event of policy failure.

  Risks

  If a new round of quantitative easing cannot control long-term yields, private credit could face simultaneous pressure from asset prices and financing.

Key data

• Progress on the U.S. “three arrows”GDP growth below 2% over the past 6 quarters; fiscal deficit approximately 6% of GDP; daily oil production up 0.3 million barrels since 2024Respectively below the policy targets of 3% GDP growth, a 3% deficit ratio, and an increase of 3 million barrels per day.
• Policy “Maginot Lines”Gasoline at $4/gallon, USD/JPY at 160, long-term U.S. Treasury yields at 5%The report believes breaches of these levels would threaten economic, AI-financing, and asset-market policy objectives.
• Size of U.S. government debtMore than $40 trillionThe report cites this as one reason for the Treasury's expansion of long-term Treasury buybacks.
• Net U.S. Treasury issuanceApproximately $2 trillion in 2026 and approximately $2 trillion in 2027The report believes large-scale Treasury supply will crowd out corporate bond issuance.
• 2026 investment-grade bond supplyApproximately $2 trillionAI-related bond issuance is increasing rapidly within this total.
• AI bond issuance$200 billion to $300 billion year-to-dateHyperscaler bonds account for approximately 9% of total investment-grade bond supply in 2026.
• BofA Bull & Bear Indicator9.5Up from 9.3; positioning is characterized as “extremely bullish,” and the indicator signal is sell.
• Performance since the sell signal was triggeredSPX +1.6%, ACWI +1.3%, HYG -0.8%Since the signal was triggered on May 26; SPX experienced a 4% drawdown between June 2 and July 29.
• Weekly equity fund flowsInflows of $40.1 billionETF inflows of $45.9 billion and mutual-fund outflows of $5.8 billion.
• Weekly bond fund flowsInflows of $21.4 billionBonds have recorded inflows for 69 consecutive weeks.
• Weekly gold, crypto-asset, and cash flowsGold inflows of $3.8 billion, crypto-asset inflows of $1.0 billion, and cash outflows of $1.2 billionThe asset-class summary separately shows precious-metal inflows of $3.9 billion, marking 7 consecutive weeks of inflows.
• Major fixed-income fund flowsU.S. Treasuries +$7.4 billion, investment-grade bonds +$7.5 billion, emerging-market bonds +$3.3 billionRespectively the largest in six weeks, the 20th consecutive week of inflows, and the largest in 11 weeks.
• Major regional equity fund flowsUnited States +$28.9 billion, Japan -$1.7 billion, South Korea -$0.8 billionThe United States recorded its largest inflow in three weeks; Japan and South Korea recorded their first outflows in 11 and 8 weeks, respectively.
• Semiconductor ETF fund flowsOutflows of $0.7 billion during the week and cumulative outflows of $6.3 billion over the past 3 weeksIndicates that some capital is beginning to exit AI-related trades.
• Financial-sector fund flowsOutflows of $2.0 billionThe largest outflow in 11 weeks.
• BofA private-client asset allocation$4.7 trillion in assets under management; equities 66.5%, bonds 16.9%, cash 9.4%Weekly bond inflows were the largest since May 2026.
• Trump approval ratingsOverall 39%, economy 36%, inflation 30%Below prewar levels of 46%, 42%, and 38%.
• 2026 year-to-date cross-asset returnsOil +53.0%, global equities +13.9%, U.S. equities +11.6%, gold +4.0%, high-yield bonds +2.6%, cash +2.3%, U.S. dollar +0.6%, investment-grade bonds +0.1%, government bonds -1.1%, Bitcoin -16.9%All are year-to-date returns in U.S. dollar terms.

Impact & implications

The report argues that long-term Treasury buybacks and Federal Reserve cooperation can prevent Treasury yields from spiraling out of control, but cannot materially reduce the cost of capital without deflation or recession. This combination benefits gold and previously out-of-favor duration assets that are sensitive to rising yields, but AI bond supply, highly leveraged private credit, and financials still face financing and policy risks. Meanwhile, equity and bond flows are strong even as the Bull & Bear Indicator reaches 9.5, showing that policy support and crowded positioning coexist. If policy or election outcomes deviate from expectations, the U.S. dollar, equities, and cyclical assets could correct rapidly.

Risks

• If quasi-quantitative easing cannot bring the 30-year Treasury yield back below 5%, the report expects the U.S. dollar could fall sharply while risk, leveraged, and cyclical assets come under pressure.
• With total U.S. government debt above $40 trillion and net issuance of approximately $2 trillion in each of 2026 and 2027, corporate and AI bond financing could be crowded out.
• The U.S.-Iran “economic war,” gasoline prices above $4 per gallon, and crude-oil and strategic reserves at 40- to 50-year lows make energy inflation difficult to alleviate.
• If the Bank of Japan does not raise rates sufficiently, yen intervention may lack support and Japan's long-term bond market may remain unstable.
• The BofA Bull & Bear Indicator has reached 9.5 and signals extremely bullish positioning, increasing the risk of a market drawdown.
• If Democrats sweep the U.S. midterm elections, the report expects equities could fall by more than 10% before year-end, while the U.S. dollar and bond yields would also decline.
• AI bond supply is increasing rapidly, while yield policy can only cap yields rather than materially reduce the cost of capital, potentially intensifying pressure on AI financing.

What to watch

• Monitor whether the 30-year U.S. Treasury yield can be held below the 5% policy defense line.
• Monitor the Jackson Hole speech at 10:00 a.m. on August 28 and the simultaneous release of the U.S. Bureau of Labor Statistics' annual nonfarm payroll revision.
• Monitor whether the Bank of Japan raises rates sufficiently on September 18 to stabilize Japan's long-term bond market.
• Track whether U.S. gasoline prices can fall below $4 per gallon, as well as changes in crude-oil inventories and the Strategic Petroleum Reserve.
• Track the policy defense line near USD/JPY 160 and the effectiveness of foreign-exchange intervention.
• Monitor the U.S. midterm elections in November, particularly whether Democrats can win the Senate and whether Republicans can retain the Texas governorship.
• Track whether the BofA Bull & Bear Indicator remains above 9.5 and whether equity breadth, futures positioning, and Fund Manager Survey positioning become more crowded.
• Monitor bond issuance by AI companies and hyperscalers, semiconductor ETF flows, and changes in corporate financing costs.