Correlation
Correlation measures how closely two things move together, on a scale from −1 to +1:
- +1: they always move in the same direction, in step.
- 0: no link; one's move tells you nothing about the other's.
- −1: they always move in opposite directions.
It is computed from returns (daily or monthly percentage changes), not from prices. In a spreadsheet: =CORREL(B2:B500, C2:C500) on two columns of returns.
Diversification
Diversification means spreading money across assets that don't move in step, so one bad outcome doesn't sink everything.
The useful surprise: it lowers the bumps without lowering the average. A mix's expected return is the weighted average of the parts; its volatility is lower than that, unless the correlation is +1.
Made-up example: two assets, each with 20% annualised volatility and the same expected return, held 50/50:
| Correlation | Volatility of the 50/50 mix |
|---|---|
| +1 | 20.0% |
| +0.95 | 19.7% |
| +0.5 | 17.3% |
| 0 | 14.1% |
| −0.5 | 10.0% |
Same expected return in every row, less wiggle as correlation falls. +0.95 barely helps.
(For the zero row: √(0.5² × 20%² + 0.5² × 20%²) = √0.02 = 14.1%. The full two-asset formula is in the OpenStax reading below.)