When you average a sector, you have to decide whether every company counts the same. Cap-weighted says no — bigger companies count more, in proportion to their market value. Equal-weighted says yes — every company gets one vote. The two can disagree sharply about the same sector on the same day, and the disagreement is itself the most useful thing on the page.
Take a sector containing companies 1…N. Each has some value xᵢ you want to average — on our sectors page that is either the model's expected growth for the company, or its realised return over the selected window. Each also has a market capitalisation capᵢ — the total market value of its shares.
Equal-weighted average = (x₁ + x₂ + … + x_N) ÷ N. Every company contributes exactly one Nth. This is the ordinary arithmetic mean you learned at school.
Cap-weighted average = (x₁·cap₁ + x₂·cap₂ + … + x_N·cap_N) ÷ (cap₁ + cap₂ + … + cap_N). Each company contributes in proportion to its size. A company worth ten times another moves the average ten times as much.
Our page applies the same choice to two different numbers at once: the sector's average forecast growth, and the sector's realised return over the window you have selected. Flip the toggle and both re-compute.
Example (illustrative — invented numbers chosen to show the mechanics, not a live sector). A sector with three companies. The giant is worth $800bn and rose 2% over the window. The mid-cap is worth $150bn and fell 8%. The small one is worth $50bn and fell 12%.
Equal-weighted: (2 − 8 − 12) ÷ 3 = −6.0%. The typical company in this sector had a bad window.
Cap-weighted: ((2 × 800) + (−8 × 150) + (−12 × 50)) ÷ (800 + 150 + 50) = (1600 − 1200 − 600) ÷ 1000 = −200 ÷ 1000 = −0.2%. The sector was roughly flat.
Both numbers are correct. They answer different questions. If you owned the sector through an index fund, you lived through the cap-weighted number and barely noticed anything happened. If you had picked a company in that sector at random, you probably lived through something much closer to the equal-weighted number and had a miserable time. The six-point gap between the two is the entire story: one enormous company is holding up a sector in which most members are falling.
This is not a hypothetical pathology. It is the normal condition of a modern index, where a handful of the largest companies can account for a large share of the total value. Whenever cap-weighted looks materially healthier than equal-weighted, the honest translation is: "the average is being carried by the giants, and breadth is weak."
Use cap-weighted when you want to know what happened to the money — what an index holder experienced. Almost every headline index, the S&P 500 included, is cap-weighted, so this is the number that matches the news and matches most index funds.
Use equal-weighted when you want to know what happened to the companies — whether strength is broad or narrow. This is a breadth question, and breadth is one of the more informative things you can ask about a market. See market breadth for the same idea measured across the whole market rather than one sector.
And in practice: look at both, and pay attention when they disagree. Agreement is unremarkable. A large gap is a finding.
Cap-weighted concentration is invisible in the number itself. A sector average of +3% cap-weighted might be three companies out of forty doing well. The number will not tell you that; only the comparison with equal-weighted will. Reading the cap-weighted figure alone is how people conclude a sector is "fine" while most of it is quietly deteriorating.
Equal-weighted has the mirror flaw: it hands a company worth $20bn the same vote as one worth $2tn. One small, volatile constituent having a wild week can swing the equal-weighted average around in a way that has no bearing on anything most investors own.
Two mechanical caveats specific to our page, stated so they cannot surprise you. First, a company whose market cap is missing or zero still counts in the equal-weighted mean but contributes nothing to the cap-weighted one — it simply is not in that sum. Second, if a sector has no usable market-cap data at all, the cap-weighted figure quietly falls back to the equal-weighted one rather than showing a broken number. In that case the toggle appears to do nothing, and that is why.
Finally: the compact sector heatmap on the dashboard is always equal-weighted. Only the full sectors page gives you the choice. If a heatmap tile and the sectors page seem to disagree, this is usually why.
Cap-weighted averages count each company in proportion to its market value, so the largest companies dominate. Equal-weighted averages give every company the same say regardless of size. The same sector can look healthy on one and weak on the other.
Cap-weighted to see what an index holder actually experienced; equal-weighted to see whether the move was broad or driven by a few giants. The most informative thing is the gap between them — a big gap means breadth and headline are telling different stories.
If no usable market-cap data is available for a sector, the cap-weighted average falls back to the equal-weighted one rather than displaying a broken figure. Companies with a missing market cap also drop out of the cap-weighted sum while still counting in the equal-weighted mean.
Yes. Like most headline indices, it weights its members by market value, which is why a handful of very large companies can move the index while most of its constituents go the other way.
Browse all S&P 500 tickers to see this metric applied to individual companies.
Educational research only — not investment advice.