An index answers one question: how did the market do? It answers by tracking a weighted basket of stocks. The design choices carry all the substance: which stocks go in, how they're weighted, and when the list changes.
4.5.1 The construction math: build one by hand
Modern equity indices are free-float market-cap weighted. That phrase packs in three ideas.
- Market cap = shares outstanding × price. Weighting by size makes the index mimic the aggregate portfolio.
- Free float: exclude shares that never trade, such as promoter holdings, government stakes, strategic cross-holdings and locked-in shares. Each stock gets an Investable Weight Factor (IWF), so a float of 40% gives an IWF of 0.40. Why bother? Because an index is a shopping list for real money. Weight it by shares nobody will sell and index funds end up chasing phantom supply. India's Sensex before 2003 and Nifty before 2009 used full market cap, and both have since converted.
- The divisor:
Index level = Σ (price_i × float shares_i) / Divisor. The divisor is a scaling constant. It's chosen so the index starts at a nice base value, then adjusted whenever membership or share counts change, so the level never jumps for reasons that have nothing to do with price.
Advertisement
Worked example: a three-stock free-float index
Build it. A three-stock index of synthetic companies:
| Stock | Shares (cr) | Price (₹) | Full mcap (₹ cr) | Non-float | IWF | Float shares (cr) | FF mcap (₹ cr) |
|---|---|---|---|---|---|---|---|
| Alpha Ltd | 100 | 500 | 50,000 | Promoter 60% | 0.40 | 40 | 20,000 |
| Bravo Ltd | 200 | 150 | 30,000 | Promoter 25% | 0.75 | 150 | 22,500 |
| Chandra Ltd | 50 | 800 | 40,000 | Strategic 10% | 0.90 | 45 | 36,000 |
| Total | 120,000 | 78,500 |
Full-cap weights would be 41.7% / 25.0% / 33.3%, with Alpha biggest. Free-float weights are 25.5% / 28.7% / 45.9% (each FF mcap ÷ 78,500), so Chandra is biggest and Alpha is smallest. The ranking flipped. Float adjustment routinely reorders an index. So an Indian company that's huge but 75% promoter-owned can carry a modest index weight.
How does the index divisor work?
Now set the base. We want the index to start at 1,000, so the divisor = 78,500 / 1,000 = 78.5, in ₹ crore per index point.
Now you, Day 2. Prices move. Alpha rises 4% to ₹520, Bravo falls 2% to ₹147, and Chandra gains 1% to ₹808. Compute the new FF mcaps, the new index level, and the index return. Then, at that day's close, the committee replaces Bravo with Delta Ltd, whose FF mcap is ₹30,000 cr. Compute the new divisor.
Answers. FF mcaps: Alpha 40 × 520 = 20,800; Bravo 150 × 147 = 22,050; Chandra 45 × 808 = 36,360 → total 79,210. Index = 79,210 / 78.5 = 1,009.04, i.e. +0.90%. Cross-check with weights: 0.2548(4%) + 0.2866(−2%) + 0.4586(1%) = 1.019 − 0.573 + 0.459 = +0.905% ✓ (tiny rounding). An index return is just the float-weighted average of member returns. Replacement: new basket total = 79,210 − 22,050 + 30,000 = 87,160. Members changed but prices didn't, so the index must still read 1,009.04. That makes the new divisor 87,160 / 1,009.04 = 86.38. That divisor absorbs the discontinuity, and the same trick handles buybacks, new share issues and IWF changes. S&P's divisor and NSE's index arithmetic work this way too, only at full scale.
4.5.2 The four indices you must know
| Nifty 50 | Sensex | S&P 500 | Nasdaq-100 | |
|---|---|---|---|---|
| Owner | NSE Indices Ltd | BSE (Asia Index Pvt Ltd) | S&P Dow Jones Indices | Nasdaq |
| Members | 50 | 30 | 500 companies (~503 share lines — some have two classes) | 100 (non-financial) |
| Weighting | Free-float mcap | Free-float mcap (since 2003) | Float-adjusted mcap (since 2005) | Modified mcap with concentration caps |
| Base | 1,000 = Nov 3, 1995 | 100 = 1978–79 | 10 = 1941–43 average | 125 = Feb 1, 1985 |
| Selection | Rules-based: NSE-listed, must trade in the F&O segment, liquidity screen — impact cost ≤ 0.50% for ₹10 crore basket orders over six months; semi-annual reconstitution (effective March & September) | Rules + committee; large liquid BSE names; semi-annual review | Committee-based: US company, size floor (~$20bn+, raised periodically — check the current level in the methodology), liquidity, ≥50% public float, positive GAAP earnings in the latest quarter and summed over the trailing four; changes made as needed | Rules-based: 100 largest non-financial Nasdaq-listed by mcap; annual reconstitution each December + quarterly rebalances; capping rules trim mega-cap weights (a breach forced the July 2023 "special rebalance") |
| Flavor | India's large-cap benchmark; financials-heavy (~a third of weight) | The older, narrower cousin; tracks Nifty closely | The world's default equity benchmark | Growth/tech proxy; one share class per company; historically weights on total (not float) shares — verify in the current methodology |
Read each methodology document once. They're short, free PDFs, listed in Resources. Carry these points away from them.
- Concentration is the norm. The top 10 names are roughly 55–60% of the Nifty 50 and ~35–40% of the S&P 500. HDFC Bank alone has run around 13% of Nifty, and Nvidia/Microsoft/Apple ~6–8% each of the S&P 500 (all as of mid-2026, check in the latest factsheets). "The index rose" often means "six stocks rose."
- Entry rules shape behaviour. S&P 500 entry requires GAAP profitability, and Tesla waited years for it. Nifty entry requires F&O availability and a pass on the impact-cost screen. Index committees don't pick "the best companies". They pick the biggest investable ones by their own mechanical criteria.
- Price index vs Total Return Index (TRI). The headline Nifty/Sensex/S&P numbers ignore dividends, while the TRI reinvests them. Always benchmark fund or portfolio returns against the TRI. Compare against the price index and you flatter the manager by the dividend yield, roughly 1–1.5% per year in both markets.
Why does index inclusion move share prices?
Why indices matter mechanically, and not only as scoreboards. Trillions of dollars and lakhs of crores are contractually bound to these lists. Index funds and ETFs replicate them: SPY and its peers in the US, and in India the Nifty index funds and ETFs, plus the EPFO's equity allocation flowing in via ETFs. Futures and options settle on them too. So index inclusion is a flow event. Enter the index and passive money must buy you regardless of valuation, which Worked Example 4 puts a number on. Leave it and that money must sell. Expect entry pops and exit slumps, with volume spikes on reconstitution days. None of that is about fundamentals. It's flow.
One step is left out here: how an ETF actually keeps its price tied to the basket. That runs through the creation and redemption mechanism worked by authorised participants, and it gets its own treatment later, along with tracking error and the implementation cost of running an index portfolio.
Which other indices are worth knowing?
The family runs wider than the big four. In India you'll meet the Nifty Next 50, the Nifty Midcap 150/Smallcap 250 and Bank Nifty. In the US there's the Russell 2000 for small caps, and the Dow Jones Industrial Average. The Dow holds 30 stocks and is price-weighted, an 1896 design in which a high-priced stock outweighs a giant company. It's worth knowing as a historical artefact and as a warning. MSCI Emerging Markets matters too. India's weight in it is ~18–20%, and it pulls foreign passive flows into Indian stocks.