BofA's Hartnett Warns of Autumn Stagflation Risk: Diesel, Bonds, and AI Triple Pressure Converging
nashnova research
BofA chief strategist Michael Hartnett warns that record diesel prices, 30-year Treasury yields at their highest since 2007, and $1.5 trillion in AI spending with no productivity payoff are converging into a fall stagflation risk.
Why is he watching diesel, not crude oil?
Crude oil back at $100 a barrel grabs headlines, but Hartnett points the alarm at diesel: the diesel crack spread — how much a refinery earns turning crude into diesel — has hit a record $102 per barrel, and retail diesel has reached $6 per gallon, also a record.
This means → crude is still well below its Ukraine-crisis peak, yet diesel has already broken records. The pressure sits in the refining segment, not upstream.
Shipping, trucking, farming, construction, and mining all run on diesel. It is the real economy's single largest input cost, and at this level it "simply cannot be ignored."
His verification trigger: the Transportation ETF (IYT) is testing its 200-day moving average at 80. A decisive break below confirms the fall stagflation scenario.
30-year yields at 2007 levels — can bonds still diversify?
The 30-year Treasury yield has climbed to its highest since June 2007, yet markets show no panic. Hartnett cites history: the day after QE3 was announced in 2012, the 30-year yield jumped 16 basis points; after QE4 landed in 2020, it surged 51 basis points within days — "markets usually test policy resolve."
In plain terms = every time the central bank floods the system, the bond market doesn't buy in immediately — it pushes yields higher first, forcing policymakers to show their hand.
The deeper problem: during the 2000–2019 "secular stagnation" era, bonds and equities were negatively correlated, and 2%–3% yields were almost a free bonus. Now the two asset classes are positively correlated again, and bonds' diversification benefit is fading fast.
This means → allocators may need higher yields before shifting into bonds at scale — bonds have to get "cheaper" before buyers step in.
$1.5 trillion spent — where is AI's productivity?
AI-related investment over the past three years has topped $1.5 trillion, yet economy-wide evidence of a productivity lift remains scarce. Total factor productivity — TFP, the gauge of how much technology adds to growth — is falling below its long-term trend line.
This reflects a sharp contradiction: Wall Street is pricing AI euphoria, but the real economy's efficiency data has not followed. Hartnett's line: "Sometimes Main Street knows what Wall Street doesn't."
The AI boom's side effect shows up in valuations: the S&P 500's free-cash-flow yield has been compressed to a historic low.
Hedging data backs the call: a high free-cash-flow stock basket is up 37% year-to-date, outperforming the broad market. In 2022 — a textbook stagflation year — the same basket rose 9% while the S&P 500 ETF fell 18%.
Is capital voting with its feet?
Over the past three weeks, average weekly net inflows into U.S. equities were just $7 billion, down sharply from July's peak of $52 billion. U.S. stocks recorded their largest three-week net outflow since January 2026, totaling $14.2 billion.
Latest weekly global flows: bonds drew $17.5 billion, cash $12.9 billion, equities $9.8 billion, crypto $1.3 billion, gold $0.6 billion. Global bonds have averaged $18 billion in weekly net inflows for four straight weeks.
Hartnett's proprietary sell indicator eased from 9.6 to 9.5, driven by slowing equity inflows and outflows from healthcare. Since the sell signal triggered on May 2, the S&P 500 is up just 1.0% and the ACWI 1.5% — no meaningful correction yet.
This means → he sees "a calm market plus hawkish policy" as "a breeding ground for volatility." No correction does not mean no risk — it means pressure is building.
What is the verification logic for fall stagflation?
Hartnett braids the three threads together: diesel pushing up real-economy costs + bonds losing diversification power + AI valuations detached from productivity — all three resonating equals stagflation.
He is explicit: "It's not too late to hedge AI-bubble equity indices." On the defensive side, high free-cash-flow stocks have outperformed in past stagflation episodes — that is his recommended direction.
The single verification node: whether IYT holds the 200-day moving average at 80. If it holds, the pressure pauses; if it breaks, the fall stagflation scenario is formally in play.
市场有风险，内容仅供研究参考，不构成投资建议。