In Excel: use =COVARIANCE.S(array1, array2) for a sample — it measures whether two variables move together, in units that make the number hard to interpret alone.
On this page8
Syntax
Arguments
| Argument | required / optional | Description |
|---|---|---|
array1 | required | The first set of values. |
array2 | required | The second set, the same size as array1. |
Related functions
6 pairs
R² = 0.9962
Intercept = 0.1667
Need it as an auditable file?
Ships inside the linked template — formula-driven, unlocked, audit-ready.
What this does
Covariance measures joint variability: positive when two variables rise together, negative when one rises as the other falls. Its weakness is scale — the result is in the product of the two variables' units, so a covariance of 4,200 means nothing without knowing what is being measured. That is exactly why CORREL exists, dividing covariance by the two standard deviations to produce a bounded -1 to 1 figure. Covariance earns its place in portfolio mathematics, where the covariance matrix, not the correlation matrix, is what combines into portfolio variance. COVARIANCE.S is the sample form; COVARIANCE.P assumes a full population. Keep the inputs visible and clearly labelled and the whole thing stays auditable — anyone who opens the file later, including you, can see at a glance exactly what feeds the result and change one assumption without hunting through the formula. The difference between a quick fix and a sheet you can trust is the extra minute you spend validating “covariance formula excel”. Start on a copy or a tiny sample, keep the affected formula visible, and compare the result with the tool above before you touch the real workbook. When a formula is involved, keep the inputs labelled beside it, reference cells instead of typing values, and apply number formatting only after the result checks out. The point is a calculation you can defend to a CFO or an auditor, but the practical win is that someone else can open the file and understand what happened without asking you.
A worked example
Two return series in B2:B100 and C2:C100: =COVARIANCE.S(B2:B100, C2:C100) returning 0.0032 says they move together, but the size is uninterpretable on its own. Normalising it: =COVARIANCE.S(B,C)/(STDEV.S(B)*STDEV.S(C)) reproduces =CORREL(B,C) exactly, which is the figure to report. Covariance is the quantity portfolio maths runs on, and knowing it is uninterpretable alone is what keeps it out of reports. A practical tip before you scale it up: build it once on a small block of test data, confirm the number against the tool on this page, and only then point it at your real sheet. That one habit catches almost every mistake while it is still cheap to fix, long before a wrong figure reaches a report or a colleague.
In Google Sheets
Everything above works in Google Sheets too. Excel and Sheets share the formula syntax used here; only the surrounding menus are arranged differently. That portability is deliberate — learn it once and it follows you between the two tools and across Windows and Mac. The aim was to get you unstuck fast and leave you a little more capable than a copy-paste would. The answer is at the top, the tool proves it, and the detail above shows why it holds — so the next time a colleague asks, you can answer without reaching for search. The short version of “covariance formula excel”: the answer is at the top of this page, the tool proves it on your own numbers, and the sections above explain why it holds so the next variation does not stump you. Excel rewards people who reference cells instead of typing values and who keep inputs separate from formulas, because that is what makes a result you can audit months later. Build it once, deliberately, with the live tool as a check, and you convert a one-off lookup into a reusable skill — which is the whole point of learning the why and not just the what.
Common mistakes
- Reporting a raw covariance, which is uninterpretable without the units; report the correlation instead.
- Ranges of different lengths, which returns
#N/A. - Using COVARIANCE.P on sample data, which understates the figure the same way VAR.P does.
Frequently asked questions
What is the difference between covariance and correlation?
Correlation is covariance divided by both standard deviations, which bounds it between -1 and 1 and makes it comparable across datasets.
Which should I report?
Correlation, nearly always. Covariance is for the maths, not the narrative.
When does covariance matter directly?
Portfolio variance, which combines covariances rather than correlations.