A fund's NVDA position grew from $2B to $3B. Bullish signal, right?
Not necessarily. If the fund's total AUM also grew from $50B to $75B (due to inflows), and NVDA's stock price rose 30%, the "growth" in the position might be entirely mechanical — no new shares purchased.
Separating real conviction from AUM-driven noise is one of the most important skills in 13F analysis.
Why AUM context matters
A position's dollar value changes for three independent reasons:
- Share count change — the manager bought or sold shares (deliberate)
- Price change — the stock went up or down (market-driven)
- AUM change — the fund grew or shrank from flows (business-driven)
Only #1 reflects investment conviction. But dollar value conflates all three.
The four scenarios
| AUM trend | Position grew | What it really means |
|---|---|---|
| AUM growing | Position grew (shares up) | Could be proportional allocation to new money — check if weight changed |
| AUM growing | Position grew (shares flat) | Price appreciation only — no new buying |
| AUM shrinking | Position grew (shares up) | Strong conviction — buying while fund is losing assets |
| AUM shrinking | Position shrank (shares flat) | Price decline — no active selling |
The highest-conviction signal is buying more shares while AUM is declining. The manager is allocating a larger share of a shrinking pie to this name. That's real conviction.
The lowest-conviction signal is position growing in dollar terms while AUM is also growing and share count is flat. The position looks bigger only because the stock price went up.
How to normalize for AUM
Method 1: Track share counts, not dollar values
The simplest approach. Share count changes are unaffected by price movements or AUM flows.
- Shares up = bought more
- Shares down = sold some
- Shares flat = did nothing
Method 2: Track portfolio weight changes
Divide position value by total AUM to get weight. Compare weights across quarters.
- Weight up + shares up = active conviction increase
- Weight up + shares flat = price outperformance (passive drift)
- Weight flat + AUM up = proportional allocation (no signal)
- Weight down + shares flat = price underperformance or rebalancing
Method 3: Compare position growth to AUM growth
If AUM grew 20% and the position grew 20%, the manager maintained proportional allocation. No signal.
If AUM grew 20% and the position grew 50%, the manager actively increased exposure. Signal.
If AUM shrank 10% and the position grew 30%, the manager aggressively added. Strong signal.
Real-world example: ARK Invest
ARK's AUM went from ~$60B peak (2021) to ~$7B trough (Q1 2025) to ~$15B (Q4 2025).
During the AUM decline from $60B to $7B, many positions shrank in dollar terms. But TSLA's portfolio weight actually increased — ARK was maintaining or adding to Tesla while losing 88% of AUM. That's conviction you can measure.
Conversely, during the recovery from $7B to $15B, positions grew in dollar terms partly because of new inflows. Not all of that growth is fresh conviction — some is proportional allocation of new money.
The checklist
Before interpreting any position change as a conviction signal:
- [ ] Did the share count change? (If not, it's price-driven)
- [ ] Did AUM change significantly? (If so, normalize for flows)
- [ ] Did the weight change? (Weight change = relative conviction change)
- [ ] Is the position change proportional to AUM change? (If yes, it's mechanical)
- [ ] Is the manager adding during AUM decline? (Strongest conviction signal)
Originally published at 13F Insight
Top comments (0)