Principal Component Analysis
Common drivers of portfolio returns
Measure how much return variance comes from a few common components and which assets contribute to them.
What PCA helps answer
- Whether a small number of common factors dominate a portfolio.
- How much diversification is present in the return matrix.
- Which assets contribute most to each dominant component.
When to use it
Use PCA to inspect whether assets with different names share the same return patterns.
What it does not do
- It does not replace factor interpretation by itself.
- It does not prove forward-looking diversification.
- It does not eliminate the need for historical backtests.
Compare with other analyses
Factor Regression relates returns to named factors. Correlation compares asset pairs, and Portfolio Backtest replays the allocation over history.
FAQ
- Is a low number of dominant components always bad?
- Not necessarily. It is a concentration signal, not a judgment by itself. The important question is whether that concentration matches the exposure you intended to take.