Correlation
Correlation measures how closely two variables move together, on a scale from -1 (perfectly opposite) to +1 (perfectly together), with 0 meaning no linear relationship. In investing, the correlation between two assets' returns determines how much diversification actually reduces portfolio risk — combining assets with low or negative correlation smooths swings more than combining assets that move in lockstep.
Why it matters
Correlation is the mechanism behind diversification (Diversifikation, spreading money across several assets to reduce risk). A portfolio split across German, European, and US equities, or a broad ETF-Sparplan (fund savings plan), only reduces risk if the pieces do not all fall together — during sharp downturns, asset classes that normally look independent can shift toward correlation near +1 at the worst possible moment. Reading correlation critically helps a saver evaluate a fund factsheet or robo-advisor allocation instead of assuming that a long list of different ticker names automatically means lower risk.
Worked examples
1. Reading a correlation number. A German equities fund and a US technology fund show these annual returns:
| Year | Fund A (German equities) | Fund B (US tech) |
|---|---|---|
| 2022 | -12% | -18% |
| 2023 | +9% | +14% |
| 2024 | +4% | +2% |
| 2025 | -3% | +6% |
Funds A and B mostly move in the same direction — both down in 2022, both up in 2023 — but not in exact proportion each year. That pattern is illustrative of a moderate positive correlation, well short of +1. A correlation of exactly +1 would mean the two funds' returns move in perfect proportion every single year; -1 would mean they move in exactly opposite proportion.
2. Why correlation changes portfolio risk. Anna splits €10,000 evenly between two funds. If the funds have a correlation near +1, a year that costs Fund A 15% is likely to cost Fund B a similar amount, and the combined portfolio drops close to 15% too. If the funds have a correlation near 0 or negative, a bad year for Fund A can coincide with a flat or positive year for Fund B, so the combined portfolio's drop is smaller than either fund's individual drop. That dampening effect — not stock-picking skill — is what diversification actually delivers, and it only works to the extent the underlying assets are not highly correlated.
Check yourself
A correlation of -1 between two assets' returns means:
What is the maximum possible value of a correlation coefficient?
Which of these statements about correlation are correct? (Select all that apply.)