Bank of America: If the Republican Party loses the Senate, U.S. stocks will drop more than 10%, and the Texas race will determine "AI policy"

Bank of America strategist Hartnett believes the market is too optimistic in its pricing of the midterm elections, and once the Democrats win the Senate, U.S. stocks face a correction risk of over 10%.

The current market pricing is based on one assumption: a moderate outcome in the midterm elections with continued deregulation and AI-friendly policies. On August 23, Hartnett pointed out in his latest research report that this assumption may be wrong.

He stated that if the Democrats win the Senate, the market’s two core narratives will take a simultaneous hit: first, financial deregulation; second, political support for AI capital expenditure. With both factors combined, he expects U.S. stocks to undergo a correction of more than 10% before year-end, accompanied by a weaker U.S. dollar and declining bond yields.

The Political Fundamentals Are Already Deteriorating

According to Bank of America data, Trump’s overall approval rating is currently 39%, economic support is just 36%, and only 30% support his handling of inflation—all significantly lower than the “pre-Iraq war level.”

This means the Republican base is not solid going into the November election. Hartnett’s hedging logic is precisely built on this political reality that the market is ignoring.

Texas Race: A “Referendum” on AI Policy

Hartnett specifically pointed out the special significance of the Texas gubernatorial election.

He believes this election is evolving into a referendum on "electricity costs and affordability" for AI infrastructure. The massive expansion of AI data centers has brought huge power consumption and infrastructure stress, which has already triggered political backlash in Texas.

If the Democrats pull off an upset win in Texas, the market will have to reassess: can AI capital expenditure still receive unconditional political support?

This is not just a political signal but directly challenges the market’s pricing logic regarding ultra-large-scale investment in AI.

Two Hedging Paths: Financials and Semiconductors

Hartnett provided specific hedging tools.

Financials (XLF): XLF is still within an upward channel, with key support around $56, where the 50-day moving average is also converging. Hartnett suggests expressing a bearish view via November-expiring bear put spread options—specifically, a Nov 56/52 put spread, with a maximum payoff ratio of about 4x.

Semiconductor ETF (SMH): The semiconductor sector lost momentum at the end of June and has continued to underperform the market. The most recent rebound was again blocked at the 50-day moving average and short-term downtrend line. Long-term support is still near the 200-day moving average, which remains quite a distance from the current price.

The key is that in recent weeks, semiconductor volatility has dropped significantly, making hedging much more attractive. Hartnett recommends hedging this risk by using downside option structures on SMH.

Fund Flows Are Already Showing Evidence

Fund flow data provides supporting evidence.

According to Bank of America data, about $6 billion has flowed out of semiconductor ETFs in the past three weeks. Hartnett notes this may not be direct hedging for the midterm elections, "but the flow of funds is moving towards the bearish scenario."

Hartnett also proposed a second logical path, with a different catalyst but the same conclusion.

He believes that quantitative easing (QE) was the starting point for the bull market of the past twenty years, and it also spawned Wall Street’s "too big to fail" narrative. After two decades of extraordinary monetary stimulus, the market generally expects current "fixed income repair" policies to be effective.

But if Treasury Secretary Bessent cannot push the 30-year Treasury yield below 5%, Hartnett believes policy failure will drive the dollar weaker, triggering a market shift: shorting risk assets, shorting leverage—especially ultra-large-scale AI operators and private credit—and shorting financials and other cyclical sectors.

"Different catalysts, same hedge: financials and AI." That’s how Hartnett summed it up.