Correlation and a two-fund mix
Correlation runs from -1 to +1. For two holdings with weights w1 and w2, standard deviations s1 and s2 and correlation c:
Portfolio variance = w1^2 x s1^2 + w2^2 x s2^2 + 2 x w1 x w2 x c x s1 x s2
Only when correlation is +1 does the portfolio's standard deviation equal the weighted average of the two; anything less gives a diversification benefit, provided both weights are positive.
Worked example: a client splits money equally between a stock fund (SD 16%) and a bond fund (SD 6%).
| Correlation | Portfolio standard deviation |
|---|---|
| +1.0 | 0.5 x 16 + 0.5 x 6 = 11.00% |
| +0.3 | sqrt(64 + 9 + 2 x 0.25 x 0.3 x 96) = sqrt(87.4) = 9.35% |
| 0.0 | sqrt(64 + 9) = 8.54% |
The expected return is the weighted average whatever the correlation, so lower correlation buys risk reduction at no cost in expected return.