Health Care has a 100% win rate in the back half of midterm election years.
Not 80%. Not 90%. ONE HUNDRED PERCENT.
I looked at every midterm year since 2006.
Here's how each sector performed from July to December:
1. Health Care $XLV: +8.46% avg, 100% win rate
2. Industrials $XLI: +7.18%, 80%
3. Materials $XLB: +6.71%, 60%
4. Financials $XLF: +6.68%, 80%
5. Cons. Discretionary $XLY: +6.35%, 60%
6. Cons. Staples $XLP: +6.29%, 80%
7. Utilities $XLU: +6.18%, 80%
8. Technology $XLK: +6.08%, 60%
9. Energy $XLE: +3.26%, 60%
10. Real Estate $XLRE: -7.94%, 0%
The S&P 500 $SPY averaged: +5.95%.
Three things nobody is talking about:
1. Health Care outperforms EVERY sector in midterm H2. Not tech. Not discretionary. Health Care. Five for five.
2. Tech drops to #8. The darling of every other year becomes middle of the pack when midterm volatility kicks in (60% win rate).
3. Real Estate has NEVER been positive Jul-Dec in a midterm year. 0 for 2. Negative every single time.
2026 is a midterm year.
The rotation is underway. It always does.
Potential attractive investment opportunities for 2H26 —
I still think they are likely to be Chinese semiconductor localization names.
A few days ago, I had meetings with Chinese sell-side analysts, and they said CXMT’s market cap immediately after IPO could be at least RMB 5 trillion.
It seems that many Western investors are underestimating the investment frenzy around CXMT within China. I was also surprised when I first heard the RMB 5 trillion number, but I don’t think it is completely unreasonable.
Most of the capital raised through CXMT’s IPO will likely flow into Chinese semiconductor localization stocks.
Names like ACMR and NAURA look promising.
I also think shorting the Apple supply chain could be an interesting idea… but there are not many Apple supply chain names with low AI exposure.
Maybe CRUS is one of the few?
But isn’t that itself a bearish point on Apple? If the Apple business were truly strong and attractive, why are so many Apple vendors trying to move away from Apple exposure and gain AI exposure instead?