How to Spot a Sector Rotation in 13F Data Without Seeing One That Isn't There

What Is a Sector Rotation?

A sector rotation is when institutional investors systematically move capital from one sector to another. For example, shifting from technology stocks to healthcare, or from growth to value. In 13F data, this shows up as reduced positions in one sector and increased positions in another across multiple filers.

The key word is “systematic.” One fund trimming tech to buy healthcare is not a rotation. Ten of the top 50 filers doing the same thing in the same quarter — that might be.

The Three Tests for a Real Rotation

Test 1: Is it broad-based?

A real rotation shows up across multiple filers with different strategies. If only one or two funds made the shift, it is an idiosyncratic decision, not a market-level rotation. Check at least 10-20 large filers before concluding that a sector rotation is underway.

Test 2: Is it delta, not level?

Many analysts look at current sector allocations and declare a rotation. But the signal is in the change, not the level. A fund with 30% technology allocation is not “rotating into tech” if it had 30% last quarter too. You need to compare quarter-over-quarter changes in sector weights.

Test 3: Is it deliberate or mechanical?

Stock price changes move sector weights automatically. If tech stocks rose 15% while healthcare was flat, every portfolio’s tech allocation increased mechanically — no buying required. A real rotation requires active changes: buying in the target sector and/or selling in the source sector, beyond what price movements would explain.

How to Check in 13F Data

1. Pick your filers: Select 10-20 large diversified managers (not sector funds) from the filer list.
2. Compare holdings QoQ: For each filer, look at the top holdings in two consecutive quarters. Identify which stocks were added, trimmed, or exited.
3. Group by sector: Classify the changes by sector. Are the adds concentrated in one sector? Are the trims concentrated in another?
4. Adjust for price: A stock that rose 20% will have a higher portfolio weight even if zero shares were added. Focus on share count changes, not value changes.
5. Count the filers: If 3 out of 20 filers shifted toward healthcare, that is noise. If 12 out of 20 did, that is signal.

Common False Positives

"Institutions are rotating into AI"

Nvidia, Broadcom, and other AI stocks have appreciated massively. Their rising portfolio weights are mostly price-driven, not purchase-driven. Check whether filers actually added shares or just rode the price up.

"Hedge funds are fleeing tech"

A few prominent hedge funds trimming tech positions does not constitute a rotation. Check the broader universe. If Vanguard, BlackRock, and State Street — who hold the index — still show the same tech allocation, the “rotation” is a few active managers rebalancing, not a market shift.

"Banks are moving to defensive sectors"

Bank 13F portfolios include client facilitation, market-making, and hedging positions. Changes in a bank’s healthcare allocation might reflect client activity, not the bank’s investment view.

When a Rotation Is Real

The clearest historical example: Q4 2022, when many large managers reduced technology exposure and increased energy and defensive allocations. This showed up across 15+ top-50 filers as active share count reductions in tech and active additions in energy names. The rotation was confirmed by both share count changes and new position openings — not just price-driven weight shifts.

Frequently Asked Questions

How many filers need to show the same pattern for a rotation to be real?

There is no magic number, but if fewer than 25% of the large diversified filers show the pattern, it is more likely noise. Above 50%, you have a strong signal.

Can I see sector rotation on 13F Insight?

You can track it manually by comparing holdings across quarters for multiple filers. The consensus holdings tool (Pro tier) can help by showing which stocks are most widely held and how that list changes over time.

Is sector rotation always a useful signal?

Not always. 13F data is 45+ days old, so any rotation you detect has already happened. The value is in understanding positioning, not in timing trades.