Min correlation is an allocation method that sets weights so the instruments in a strategy move together as little as possible: instruments closely tied to the rest get less weight, those that behave differently get more. Where Inverse volatility looks at each instrument on its own, Min correlation looks at the relationships between them.
Also seen as: minimum-correlation algorithm, MCA, minimum correlation weighting
How does Fincanva handle it?
Included from Starter upwards. See what each plan includes.
- It is a single-strategy method: a Combined does not offer it when it splits capital across its strategies.
- Its one setting, Correlation matrix, chooses how the correlations are estimated; the weights fall out of them. It starts on Sample, for a newly chosen method and for a strategy saved before the setting existed alike, and a chip under the method names any other choice — see Choose how a method estimates risk.
- Marchenko-Pastur is included from the Ultimate plan, like the advanced covariance matrix estimators; the other five choices come with the method. On a lower plan it stays in the list with the plan mark, and choosing it opens the plan dialog. See what each plan includes.
- It is not the Combined's correlation matrix: the correlation matrix on a Combined's Correlations page is a separate analysis, unaffected by it.
- The calculation window (In-sample), over which correlations are measured, defaults to 12 months and is a choice of 1, 2, 3, 6, 12, 18 or 24 months, because the method also reads annualised volatility over it — see which window lengths you can choose.
- Correlations are re-measured at every rebalance, so the weighting can shift sharply over a backtest — a rolling correlation shows that movement. With a single instrument there is nothing to correlate.
Why does low correlation matter?
A portfolio's volatility depends on how its holdings move relative to each other, not only on how volatile each one is: two instruments with identical volatility give a less volatile combination the less correlated they are. That is the mechanism behind diversification, and the correlation term in the mean-variance volatility formula. For two instruments held at equal weight with the same volatility:
- the volatility of the equal-weight pair
- each instrument's own volatility
- their correlation
At a correlation of 1 the pair is no calmer than one instrument alone; the lower it falls, the more the pair's volatility drops below each instrument's.
Which correlation estimate does Min correlation use?
Sample, unless you choose another under Correlation matrix — "How the method measures how closely the instruments move together." Change opens six choices, none with a setting of its own:
| Choice | What it does |
|---|---|
| Sample (default) | "The historical estimate, computed on prices." |
| Pearson on returns | "The classic correlation, computed on daily returns instead of prices." |
| Spearman (rank-based) | "Compares the order of returns, not their size: extreme days weigh less." |
| Kendall (rank-based) | "Counts how often two instruments rise and fall together: the most robust to extreme days." |
| Exponentially weighted moving average (EWMA) | "Gives more weight to recent days: reacts sooner when correlations change." |
| Marchenko-Pastur | "Separates the signal from the noise and keeps only the signal." |
Sample correlates prices, which picks up shared long-run trends as well as day-to-day co-movement; the other five work on daily returns, which measure the day-to-day co-movement alone. Pearson is the textbook coefficient; Spearman and Kendall compare the order of returns, so a few extreme days move them less; EWMA follows a change in correlations sooner, with no decay setting here; Marchenko-Pastur, in the list's Advanced group, keeps only the structure distinguishable from noise — the idea behind the same option under Covariance matrix.
What does it look like in practice?
A strategy holds three instruments, each with 15% volatility. Their pairwise correlations over the window are: A with B, 0.85; A with C, 0.20; B with C, 0.35. Held at equal weight, a pair's volatility is 15% × √((1 + ρ) ÷ 2):
| Pair | Correlation | Volatility of the equal-weight pair |
|---|---|---|
| A + B | 0.85 | 15% × √0.925 = 14.4% |
| B + C | 0.35 | 15% × √0.675 = 12.3% |
| A + C | 0.20 | 15% × √0.600 = 11.6% |
Every pair is built from instruments with the same 15% volatility, yet the calmest pair is nearly three percentage points calmer than the most correlated one. Only the correlations explain it, and picking on that difference is what Min correlation is for. Fincanva reports the same pair-by-pair reading for a Combined's strategies in its correlation matrix.
How is Min correlation different from Risk parity?
Both take correlations into account, but aim at different things. Risk parity targets the split of risk — every instrument supplies an equal share of the total. Min correlation targets the co-movement itself, favouring instruments that behave least like the rest. A highly correlated instrument can still receive a substantial risk-parity weight; under Min correlation it is the very thing being weighted down.
Fincanva describes how these methods work; it does not recommend one. See Is this financial advice?.