Can sector rotation predict the stock market?

The short answer

No — not two to three days ahead, and not in any form we could measure. We tested it properly and it failed. This page is the record of that failure, because a tested no is worth more than an untested yes.

What we tested

Rotation is real. Money visibly moves between gold and equities, bonds and equities, defensives and cyclicals, energy and technology, semiconductors and the broad market. The question is not whether it happens. It is whether seeing it happen tells you anything about tomorrow.

We took ten rotation states and measured each against what the market did two and three days later — 160 tests in total. Not one of them survived.

We tested it a second way

Instead of rotation states, we used capital flow directly: eleven sectors’ inflow and outflow, market breadth, and the spread between sectors, all measured against the Nasdaq. That is 84 tests, and exactly one cleared the usual significance bar.

When you run eighty-four tests, roughly four will clear that bar by luck alone. We found one. That is fewer than chance would hand you, which is the opposite of evidence.

And a third

Finally we combined market states — risk, macro, geopolitics, energy, AI, direction and pressure — into hundreds of conditions and tested them against 1844 trades. None of the combinations survived either.

What this does not mean

It does not mean rotation is meaningless. It means rotation is description, not prediction. Knowing money left technology for energy tells you what the market is doing. Our measurements say it does not tell you what the market will do next.

That distinction matters, because a great deal of market commentary presents the first as though it were the second.

How we would know we are wrong

If a rotation state cleared a bar corrected for the whole family of tests, on a sample large enough to cut in half with the effect holding in both halves, we would reopen it. Until then it stays on the fallen list, in public.