Transparent method
The formula
Worked example
Example inputs
How to interpret the result
Correlation measures how two assets have moved together historically, not a fixed law about how they'll move in the future, and it's usually most reassuring right up until a crisis, when previously uncorrelated assets have a well-documented habit of moving together anyway. Two assets can show a low, comfortable correlation over years of calm markets and then converge sharply during a shared shock.
Before you rely on it
What to check
Confirm both return series use the same frequency and the same observation window. Mismatched time periods or frequencies distort a correlation figure without anyone changing the underlying assets.
The common error
Where people go wrong with asset correlation calculator
Treating a historically low correlation as a permanent diversification guarantee. Correlations shift over time and especially during market stress, exactly when diversification matters most and correlations are most likely to have already moved.
Sensitivity evidence
How return covariance changes the estimated correlation
| Return covariance | Estimated correlation | Diversification distance from +1 | Covariance entered |
|---|---|---|---|
| 0.01 | 0.556 | 0.444 | 0.01 |
| 0.011 | 0.611 | 0.389 | 0.011 |
| 0.012worked example | 0.667 | 0.333 | 0.012 |
| 0.013 | 0.722 | 0.278 | 0.013 |
| 0.014 | 0.778 | 0.222 | 0.014 |
Every input, tested
Which input moves the estimated correlation most
| Input | Tested from | To | Estimated correlation at each end | Swing |
|---|---|---|---|---|
| Asset B standard deviation | 10.0% | 14.0% | 0.8 to 0.571 | 0.229 (34%) |
| Asset A standard deviation | 13.0% | 17.0% | 0.769 to 0.588 | 0.181 (27%) |
| Return covariance | 0.011 | 0.013 | 0.611 to 0.722 | 0.111 (17%) |
Two variables at once
Estimated correlation by return covariance and asset a standard deviation
| Return covariance \ Asset A standard deviation | 11.0% | 15.0% | 19.0% |
|---|---|---|---|
| 0.01 | 0.758 | 0.556 | 0.439 |
| 0.011 | 0.833 | 0.611 | 0.482 |
| 0.012 | 0.909 | 0.667 | 0.526 |
| 0.013 | 0.985 | 0.722 | 0.57 |
| 0.014 | 1 | 0.778 | 0.614 |
Step by step
The worked example, input by input
| Input | Value used |
|---|---|
| Return covariance | 0.012 |
| Asset A standard deviation | 15.0% |
| Asset B standard deviation | 12.0% |
| Estimated correlation | 0.667 |
| Diversification distance from +1 | 0.333 |
| Covariance entered | 0.012 |
Where the inputs come from
Return covariance and the two standard deviations should come from the same return series over the same period. Enter standard deviations as percentages and covariance in matching units; the example uses decimal covariance with percent deviations. If you have only correlation from a data provider, use it directly instead. Choose at least a few years of monthly returns for a stable estimate.
When the result stops applying
Correlations change with time and rise in crises, when many assets fall together, which is when diversification is needed most. It measures only linear relationships and does not tell you about the size of the moves. A low correlation over a calm period may not hold in a downturn.
What to compare it with
Put the correlation into the two-asset portfolio risk calculator to see the effect on total volatility. The beta calculator measures relationship with the market specifically. When building a portfolio, the portfolio weighted return page shows the return side of the trade.
People also ask
Frequently asked questions
What correlation is good for diversification?
Lower is better. Values below about 0.5 give meaningful diversification, and negative values, such as between stocks and Treasury bonds in some periods, give the most.
Do correlations stay the same over time?
No. They shift with market conditions and tend to rise during crises, so diversification is weakest in exactly the worst markets.
How is correlation different from covariance?
Covariance depends on the scale of the returns and is hard to interpret. Correlation divides by both standard deviations to give a number between -1 and 1.