Direct Answer
Comparing equal-weight and cap-weight versions of the same index, both indexed to 100 at a common start date, reveals market breadth: a rising equal-weight-to-cap-weight ratio means the average constituent is outperforming the largest ones, indicating broad participation; a falling ratio means gains are concentrated in the biggest names.
Key Takeaways
Most headline stock indexes are capitalization-weighted, meaning a company's index weight scales with its market value. When the largest handful of constituents account for a large share of total index weight, those few names can move the index level substantially even while most other constituents are flat or falling. Comparing a cap-weighted index to its equal-weight counterpart, where every constituent counts the same regardless of size, surfaces that gap directly, without needing a separate count of advancing versus declining issues.
- A cap-weighted index gives its largest constituents outsized influence on the index level; an equal-weight version of the same list gives every constituent the same weight.
- Both series are normalized to start at 100 on the first observation, so the comparison isolates percentage performance from each index's raw level.
- The ratio is the indexed equal-weight value divided by the indexed cap-weight value, multiplied by 100, computed for each observation.
- A rising ratio indicates broader participation; a falling ratio indicates performance concentrated in the largest names.
- This is a magnitude-weighted, complementary lens to issue-count breadth measures like the A/D line, not a substitute for them.
How Is the Equal-Weight-to-Cap-Weight Ratio Calculated?
The calculation runs in two steps. First, both the equal-weight index series and the cap-weight index series are independently normalized (indexed) to start at 100 on their first observation: each day's value is divided by the series' own first-day value and multiplied by 100. This step removes the effect of each index's absolute level, since the two indexes are typically constructed on different bases and aren't otherwise directly comparable in raw point terms.
Second, for each day, the ratio is computed as the indexed equal-weight value divided by the indexed cap-weight value, multiplied by 100. When the ratio is above 100 and rising, the equal-weight index has gained more than the cap-weight index since the starting date. When the ratio is falling, the cap-weight index is pulling ahead, meaning performance is concentrated in the largest constituents while the equal-weight "average" constituent lags behind.
Worked example
Take an equal-weight index series of 1000, 1015, 1032, 1040 across four days, alongside a cap-weight index series of 1000, 1010, 1018, 1022 over the same four days. Both start at the same raw level here for simplicity, so indexing to 100 doesn't change the first day's value, but the two series diverge from there.
| Day | Equal-weight index | Cap-weight index | Ratio (equal-weight / cap-weight × 100) |
|---|---|---|---|
| 1 | 1000 | 1000 | 100.00 |
| 2 | 1015 | 1010 | 100.50 |
| 3 | 1032 | 1018 | 101.38 |
| 4 | 1040 | 1022 | 101.76 |
The resulting ratio series, 100, 100.5, 101.38, 101.76, rises steadily across all four days. Over the full period, the equal-weight index gained 4.0% (from 1000 to 1040) while the cap-weight index gained 2.2% (from 1000 to 1022). The equal-weight index outperformed by roughly 1.8 percentage points, and the ratio's steady climb reflects that gap accumulating day by day. In this illustrative example, that pattern indicates broader participation beyond just the largest constituents driving the cap-weighted benchmark: the "average" name in the equal-weight basket did relatively better than the mega-cap-heavy version of the same list. These numbers are a constructed illustrative example, not live or historical index data.
Common mistake
The common mistake is comparing the two indexes' raw point levels directly instead of indexing them first. Two index families are usually built on different bases (a different starting divisor, a different launch date), so their raw levels aren't comparable on their own, a cap-weight index reading "5,000" and an equal-weight index reading "6,200" says nothing about relative performance until both are converted to a common starting point.
Why Does This Comparison Work as a Breadth Signal?
A capitalization-weighted index assigns each constituent a weight proportional to its market value, so the largest few companies can represent a disproportionate share of total index weight. When those largest names rally, the index level rises even if most of the remaining constituents are flat, mixed, or declining, the index-level number doesn't distinguish between "everyone participated a little" and "a few names did all the work." An equal-weight version of the identical constituent list removes that size effect entirely: every constituent contributes the same amount to the index's return regardless of its market value.
Because the two versions track the same underlying set of companies and differ only in how they're weighted, any difference in their performance over a period is attributable to the distribution of returns across constituents rather than to a different universe of stocks. That's what makes the comparison a breadth measure: it answers whether performance is broad (equal-weight keeping pace with or ahead of cap-weight) or narrow (cap-weight pulling ahead because the largest holdings are doing the heavy lifting) without requiring a separate daily count of advancing and declining issues.
How Does This Compare to Issue-Count Breadth Measures?
Issue-count measures like the advance/decline line answer a headcount question: how many stocks went up today versus down, regardless of by how much. The equal-weight-to-cap-weight ratio answers a magnitude question instead: given the actual size of each move, is performance broadly shared or concentrated. The two can and do disagree. A session where 400 of 500 index constituents advance but by tiny amounts, while the 100 largest names fall hard, could still show a positive A/D reading (more advancers than decliners) even as the equal-weight-to-cap-weight ratio falls (because the cap-weight index, dominated by those large fallers, underperforms the equal-weight index, which is dragged down proportionally less by any single name).
Because the two measures capture different aspects of participation, headcount versus magnitude, using them together gives a fuller picture than either alone. Neither is a strict subset or replacement for the other.
Common mistake
The common mistake is treating a rising equal-weight-to-cap-weight ratio and a rising A/D line as interchangeable confirmations of the same thing. They frequently move together because both tend to reflect broad participation, but they are computed from entirely different inputs (issue counts vs. weighted index levels) and can diverge from each other in specific sessions.
Misconceptions Versus Reality
| Misconception | Reality |
|---|---|
| Comparing raw index point levels directly shows relative performance | Raw levels reflect each index's own base and divisor; both series must be indexed to a common starting value (e.g. 100) before the comparison means anything |
| The equal-weight-to-cap-weight ratio is just another way of counting advancers vs. decliners | It's a magnitude-weighted comparison of two fully constructed index series, not a headcount, it captures how much constituents moved, not just how many moved which direction |
| A rising ratio means every single constituent is rising | The ratio reflects the aggregate relationship between the two index constructions; individual constituents can still be mixed even while the ratio trends in one direction |
| Equal-weight indexes are always less volatile than cap-weight indexes | Equal-weight indexes carry more relative exposure to smaller constituents, which can add volatility rather than reduce it, depending on the constituent universe |
| This ratio and the A/D line always move together | They usually move in the same general direction but are computed from different inputs and can disagree in specific periods, since one is headcount-based and the other is magnitude-based |
Risks, Limitations, and Exceptions
- The ratio only reflects the relationship between the two chosen index series over the chosen window; a different start date changes every subsequent ratio value, since both series are re-indexed from that new starting point.
- Equal-weight and cap-weight versions of an index typically use different rebalancing schedules and methodologies (S&P Dow Jones Indices, for example, rebalances its equal-weight index versions quarterly), which can introduce short-term mechanical effects around rebalance dates unrelated to genuine participation shifts.
- The comparison requires both series to cover the identical constituent list and identical date range; comparing an equal-weight version of one index to the cap-weight version of a different index would confound the result.
- Like the breadth-divergence comparison elsewhere on this site, a rising or falling ratio is a description of relative past performance, not a forecast of future performance.
- The worked example on this page uses a small, illustrative four-day dataset chosen to make the arithmetic easy to verify by hand. It is not derived from live or historical index data.
Rebalance Dates Can Move the Ratio on Their Own
Equal-weight and cap-weight versions of the same index are not maintained identically. The equal-weight version has to be rebalanced on a schedule to restore its equal weights, and that mechanical process can move the ratio around those dates for reasons that have nothing to do with how broadly the market is participating. A shift in the ratio dated close to a known rebalance deserves a second look before it is read as a change in breadth.
The start date is the other input doing quiet work. Both series are indexed to 100 at a chosen point, so moving that point changes every subsequent ratio value. The shape of the comparison is anchored to a decision, and two analysts using different start dates will produce different-looking pictures from the same underlying indexes.
What makes the comparison worth running is that it surfaces concentration without needing an issue count. If the average constituent is outperforming the largest ones, the ratio rises; if gains are concentrated in the biggest names, it falls. That is a direct read on the same question breadth counts approach from the other side.
One requirement for the comparison to mean anything: both series must cover the identical constituent list over the identical date range. Comparing an equal-weight version of one index against a cap-weight version of a different one produces a ratio that mixes the weighting question with a universe difference.
Frequently Asked Questions
How can equal-weight and capitalization-weighted indexes be compared for market breadth?
Both index series are first normalized to start at 100 on the same date, so each is measured only by its own percentage change from that starting point rather than by its raw level. The ratio of the indexed equal-weight value to the indexed cap-weight value, multiplied by 100, is then calculated for each subsequent day. A rising ratio means the equal-weight index is outperforming the cap-weight index over that period, which indicates broader participation across constituents rather than a handful of mega-cap names driving the move.
What does a rising equal-weight-to-cap-weight ratio mean?
A rising ratio means the equal-weight version of an index is gaining more (or losing less) than the capitalization-weighted version over the same period. Since every constituent counts the same in an equal-weight index, that outperformance indicates the average constituent is doing relatively better than the index's largest, most heavily weighted names, a sign of broader participation. A falling ratio indicates the opposite: performance concentrated in the largest constituents while the average stock lags.
Is this the same as counting advancing vs. declining issues?
No. Counting advancing versus declining issues, as in the advance/decline line, is a pure headcount: every issue counts once regardless of how much it moved. The equal-weight-to-cap-weight ratio instead compares two fully constructed, magnitude-weighted index series. It's a complementary lens on participation, not a replacement for issue-count breadth measures like the A/D line, and the two can occasionally disagree because they measure different things.
Do the equal-weight and cap-weight indexes hold the same constituents?
For the standard paired variants, yes, and that is what makes the ratio interpretable. Because membership is identical, the only difference between the two return streams is the weighting scheme, so the ratio isolates weighting cleanly. Pairing indexes with different membership confounds the weighting effect with a membership effect and the ratio stops answering the question it was built for.
Is the equal-weight to cap-weight ratio the same as small-cap versus large-cap?
No. Both indexes in the pair hold the same large-capitalisation companies, so neither contains small-cap exposure. What equal weighting does is give the smaller members of a large-cap universe the same weight as its largest, which is a tilt within the universe rather than a move outside it. A genuine small-versus-large comparison uses two different universes and behaves differently.
Should the ratio be built from price returns or total returns?
Whichever is used, both legs must use the same basis. Mixing a total-return series on one side with a price-return series on the other introduces a steady dividend drift into the ratio that has nothing to do with breadth. Because the two indexes have different dividend profiles, that drift can be large enough over a multi-year chart to change the apparent trend.
Why does the ratio trend for long stretches instead of oscillating?
Because it is a relative price series between two portfolios with a persistent structural difference, not a bounded statistic. Nothing in its construction pulls it back toward a central value, so it can drift in one direction for years. Treating it as mean-reverting imports an assumption the arithmetic does not supply, which is a different failure from misreading a bounded breadth percentage.
What do issue-count breadth measures usually look like when this ratio is falling?
A falling ratio means the cap-weighted version is outperforming, which typically coincides with fewer constituents keeping pace with the index. The correspondence is a tendency rather than an identity: the ratio can fall because a small number of very large names rose sharply while the median constituent was unchanged, which affects the ratio more than it affects a count-based measure.
Can this ratio be built for a single sector?
Yes, wherever an equal-weight version of that sector index exists. The read then describes participation inside the sector rather than across the market, which can be more informative for a sector with a few dominant members. The same caution applies about matching membership and return basis, and sector pairs are more likely to differ on both than the headline pairs are.
References
Equal-weight index construction and rebalancing conventions described on this page follow the general methodology published by index providers such as S&P Dow Jones Indices for products like the S&P 500 Equal Weight Index, which rebalances constituent weights back to equal on a quarterly schedule. Key reference sources include:
- S&P Dow Jones Indices, Index Mathematics Methodology: spglobal.com/spdji: the constituent-weighting formulas, including equal weighting, this comparison is built on.
- NYSE, Historical Market Data: nyse.com/market-data/historical: exchange-level advance/decline data conventions referenced for comparison.
- Nasdaq, Market Activity: nasdaq.com/market-activity: issue-level advancing/declining reporting for a comparable universe.
The worked example on this page uses a small, deterministic, illustrative four-day dataset chosen to make the indexing and ratio arithmetic easy to verify by hand. It is not live or historical index data and should not be read as a real historical episode. This content was reviewed by the Swoopr Editorial Team in August 2026 and reflects publicly available information at that time.