A stock index is a single number that summarizes the level (and, over time, the return) of a defined basket of securities. But the number is an artifact of a methodology, not a fact of nature: the same set of companies produces wildly different "market" readings depending on what counts as a constituent, how each one is weighted, how the running value is kept continuous through corporate actions, and how often the basket is refreshed. This section is about those construction choices and their market consequences. The core tension throughout is that an index is meant to be a passive, neutral mirror of a market, yet the trillions of dollars that track major indices turn every construction rule into a force that itself moves prices — so the measuring instrument and the thing being measured are no longer independent.

What the domain covers

There are four interlocking decisions in building any equity index, and almost every concept in this section maps to one of them:

1. Universe / eligibility. What is the target market (large-cap U.S. equity? a sector? a country?), and which securities qualify (size, liquidity, domicile, free-float, profitability screens). The S&P 500, for example, is committee-selected against eligibility rules; the Russell indexes are rules-based size cuts of the Russell 3000. 2. Weighting scheme. How much of each constituent the index holds — the single most consequential choice. The main families are price-weighting (weight ∝ share price; the Dow Jones Industrial Average is the surviving example, computed as the sum of 30 prices divided by the Dow Divisor — 0.16242563904928 as of October 30, 2025 per S&P DJI/Wikipedia), capitalization-weighting (weight ∝ market value, almost always float-adjusted so only publicly tradable shares count — the dominant modern method since MSCI's free-float transition in 2001–2002), equal-weighting (every name = 1/N), and factor / fundamental schemes. 3. Continuity maintenance — the divisor. A naive average would jump every time a stock splits, a constituent is swapped, or a member pays a special dividend. Index providers absorb these non-market events by adjusting a divisor so the index value is continuous across the change. This is the unglamorous machinery that makes an index a comparable time series; the Dow Divisor's drift to far below 1.0 over a century of adjustments is the visible scar of that process. 4. Maintenance cadence. How and when the basket is refreshed — rebalancing (re-weighting existing members) and reconstitution (adding/deleting members).

When these choices matter most: whenever you compare "the market's" return against a portfolio, choose a benchmark, build or buy a passive fund, or interpret an index move. They matter least for analyzing a single stock's fundamentals — there the index is context, not input. A recurring, decision-relevant fact is that two indices over the same stocks can behave like different asset classes (see the cap-vs-equal child).

Why the methodology is not neutral

The defining modern theme is that index construction has become price-affecting. Because passive vehicles minimize tracking error rather than price, they are price-insensitive buyers and sellers, and they all act on the same predictable rules at the same predictable times. Two empirical consequences dominate this section's children:

• Concentration as a byproduct of cap-weighting. Cap-weighting mechanically buys more of whatever has already risen, so in a narrow-leadership regime the index quietly becomes a concentrated momentum bet. The top-10 weight of the S&P 500 reached roughly 40%+ in 2025 (Visual Capitalist; RBC Wealth Management cite ~41%), above the dot-com-era peak (top-10 ~25–27% around 2000) — a construction outcome, not a deliberate allocation.
• Forced flows around basket changes. Reconstitution and inclusion events create calendar-predictable trading. The classic index effect (the abnormal return on S&P 500 addition) averaged ~3.4% in the 1980s and ~7.6% in the 1990s but fell to ~0.8% over 2010–2020 (Greenwood & Sammon 2022, NBER w30748, sample 1980–2020) — a textbook case of an anomaly arbitraged away once it became predictable.

The honest through-line: most of the tradable edges historically associated with index mechanics have decayed for large, liquid U.S. names, even as the dollar flows have grown. The mechanics remain essential for understanding prices; they are largely spent as standalone strategies.

Map of the sub-topics

• Cap-Weighted vs Equal-Weighted (001) — the two dominant weighting schemes, how each is computed and rebalanced, and the genuinely contested evidence on equal-weight's long-run "outperformance" (largely a regime-dependent size factor plus a rebalancing premium, not a free lunch). Start here; weighting is the highest-leverage construction choice.
• Index Reconstitution (002) — the rules-based add/delete process, the Russell vs S&P mechanics (annual cliff vs continuous committee), the 2026 Russell move to semi-annual recon, recon-day flow scale, and the measured replication-drag cost.
• Index Effect on Prices (003) — the addition/deletion price response specifically, its three competing explanations (downward-sloping demand, price-pressure, certification), and the well-documented disappearance of the effect for large caps.

This overview deliberately stops at the construction framework; the children carry the formulas, base rates, and source-level disputes. Not covered as their own nodes here: total-return vs price-return index variants, sector/style/factor index design, and international free-float quirks — treat those as adjacencies.

Sources

• S&P Dow Jones Indices — Dow Jones Averages Methodology and Index Mathematics methodology PDFs (divisor mechanics, price-weighting): https://www.spglobal.com/spdji/en/documents/methodologies/methodology-dj-averages.pdf
• Wikipedia / Corporate Finance Institute — Dow Divisor value (0.16242563904928 as of Oct 30 2025) and divisor adjustment for splits/spinoffs: https://en.wikipedia.org/wiki/Dow_Jones_Industrial_Average ; https://corporatefinanceinstitute.com/resources/equities/dow-divisor/
• AnalystPrep (CFA L1) & Endowus — weighting-method taxonomy (price / full-cap / float-adjusted cap / equal / factor); MSCI free-float transition 2001–2002: https://analystprep.com/cfa-level-1-exam/equity/different-weighting-methods-used-index-construction/ ; https://endowus.com/insights/equity-index-types-explained
• S&P Dow Jones Indices — "Methodology Matters" (why construction choices change outcomes): https://www.spglobal.com/spdji/en/research-insights/index-literacy/methodology-matters/
• Greenwood & Sammon, "The Disappearing Index Effect" (NBER Working Paper 30748, 2022) — addition abnormal return 3.4% (1980s) / 7.6% (1990s) / 0.8% (2010–2020): https://www.nber.org/papers/w30748
• S&P 500 concentration (top-10 ~40–41% in 2025, vs ~25–27% at the 2000 peak): Visual Capitalist https://www.visualcapitalist.com/sp/visualized-the-rising-concentration-of-the-sp-500-tema01/ ; RBC Wealth Management "The Great Narrowing"
• FTSE Russell — Russell US Indexes move to semi-annual reconstitution from 2026 (June + December): https://www.lseg.com/en/media-centre/press-releases/ftse-russell/2025/russell-us-indexes-move-to-semi-annual-reconstitution
• Child nodes (carry the detailed evidence and disputes): 001-cap-weighted-vs-equal-weighted, 002-index-reconstitution, 003-index-effect-on-prices.

Soft/qualified: concentration figures (top-10 ~40–41% in 2025) and index-effect magnitudes are summarized here at headline level and sourced precisely in the child nodes; the durability of equal-weight outperformance and the residual index effect for small caps remain genuinely contested (see children for the citations and the dispute flags).