Skip to content
Menu
Finabase Learn

Funds · Indexes and index funds

Index reconstitution and rebalancing: calendars, buffers and the cost

When the big indexes change their members, how buffers limit the churn, what Tesla’s 2020 entry to the S&P 500 showed, and what the trading costs fund holders.

On this page
  1. Two kinds of change
  2. The calendars
  3. Buffers: avoiding pointless trades
  4. The index effect
  5. What it costs fund holders
  6. Three small-company funds
  7. Where this misleads
  8. How Finabase does it
  9. Frequently asked questions
  10. Summary
  11. Check your understanding

An index has to be kept up to date. Companies grow, shrink, merge, list and fail, and a list of "the largest 1,000" drawn up last year is wrong this year. Reconstitution is the scheduled review that redraws the membership. Rebalancing is the lighter adjustment that resets weights or share counts without changing who is in.

For an index these are entries in a spreadsheet. For the funds that follow the index they are real trades, made on a date announced in advance, in the same direction as every other fund on the same index. This lesson covers when the main indexes change, how their rules try to limit unnecessary changes, what happens to the prices of stocks going in and out, and what the evidence says about the cost to people who hold index funds.

Two kinds of change

Reconstitution applies the eligibility rules again from scratch: rank the universe, draw the lines, add the companies that now qualify and remove those that do not. Its size is measured by turnover, the share of the index's value that changes hands.

Rebalancing leaves the members alone and resets how much of each is held. In a market-value index that means updating share counts and float factors after buybacks and share sales. In an equal-weight or capped index it means pulling weights back to their targets (Index weighting methods: cap, price, equal, fundamental and capped).

Between scheduled dates, indexes also change for corporate events: a member acquired by another company is removed when the deal closes, whatever the calendar says.

The calendars

Index family Full membership review Takes effect Changes between reviews
S&P 500 None on a fixed date; the committee replaces members as needed Announced a few days to weeks ahead Share counts updated quarterly
Russell 1000, 2000, 3000 June, and from 2026 also December After the close on the fourth Friday of June and the second Friday of December New listings added quarterly
Nasdaq-100 December First trading day after the third Friday of December Weights reviewed in March, June and September
Vanguard's US benchmarks (CRSP, now Morningstar) Quarterly Spread over several days Companies moved between size bands in halves

Russell: a fixed date, known months ahead

The Russell reconstitution is the best-known scheduled event in US index investing because its rules are mechanical. Every eligible company is ranked by total market value on a rank day, the last business day of April for the June event. The first 1,000 form the Russell 1000 and the next 2,000 the Russell 2000. The changes take effect after the close on the fourth Friday of June.

For more than three decades that happened once a year. In January 2025 FTSE Russell announced a move to twice a year from 2026, and in September 2026 it published the schedule of the first December event.

The Russell US Indexes reconstitution of December 2026, step by step Oct. 30 Rank day: every company is sized Nov. 13 Preliminary lists of adds and deletes Nov. 16–27 Query period, weekly updates Nov. 30 Lock-down: lists are final Dec. 11 Changes take effect after the close Source: FTSE Russell schedule of September 1, 2026. Six weeks pass between ranking and trading.
The first December reconstitution of the Russell US Indexes, from FTSE Russell’s published schedule.

Companies were ranked on Friday, October 30, 2026. Preliminary lists came out on November 13, with updates during a query period that runs to November 27. The lists are locked from November 30, and the changes take effect after the close on Friday, December 11. At the preliminary stage FTSE Russell reported that the dividing line between the Russell 1000 and Russell 2000 stood at about $5.8 billion, that 21 companies were expected to join the Russell 1000 (9 of them moving up from the Russell 2000), and that 125 were expected to join the Russell 2000 while 75 left it.

The December event is lighter than June's. Size boundaries are updated at both, but the split of each index into growth and value halves is fully refreshed only in June.

Six weeks pass between rank day and the effective date, and the lists are public for four of them. Anyone can buy the additions and sell the deletions before the index funds do.

S&P 500: changes as needed

The S&P 500 does not reconstitute. Its committee replaces a member when one is acquired, no longer represents the market, or when a clearly better candidate exists, and announces each change shortly before it takes effect, typically timed with the quarterly share update. Because S&P's size and earnings criteria apply to additions and not to continued membership, a company that slips below them can remain, and turnover is low. Vanguard reported a turnover rate of 2.4% for its S&P 500 ETF and 2.6% for its Total Stock Market ETF in their latest fiscal years (fact sheets of September 30, 2026).

Nasdaq-100 and the Vanguard benchmarks

The Nasdaq-100 reconstitutes once a year. Under the 2026 methodology companies are ranked at the end of November and the new membership takes effect on the first trading day after the third Friday of December.

The US stock indexes that Vanguard's funds follow, created by CRSP and renamed Morningstar indexes in 2026, take a different approach to the same problem. They review quarterly, move a company between size bands in two halves on successive reviews (a practice called packeting), and spread each reconstitution over several trading days so that the funds do not have to trade everything at one closing price.

Buffers: avoiding pointless trades

A company ranked 999th this year and 1,003rd next year has not really changed. Moving it from a large-company index to a small-company index and back again would cost the funds on both indexes for nothing. Every provider has a rule to prevent it.

Russell's banding. A current member changes index only if it has moved beyond a band around the breakpoint. The band is 2.5 percentage points of cumulative market value on either side of the boundaries at ranks 200, 500 and 1,000, and 0.5 points at rank 2,000. A Russell 2000 member that has grown just past the 1,000th company stays in the Russell 2000 until it is clearly inside the larger index.

Nasdaq's retention rule. At the December review the 75 largest eligible companies are selected outright, and current members ranked up to 100th keep their place. A new company has to rank well inside the index to displace an existing one.

S&P's asymmetric criteria. The requirements bind on the way in and not on the way out, which works as a wide buffer.

Packeting. Moving half a position at a time means a company hovering on a boundary is only half-traded if it reverses.

Buffers reduce turnover at the price of precision. A banded Russell 2000 contains some companies larger than the smallest in the Russell 1000.

The index effect

When a stock is added to a widely followed index, funds that track the index must buy it, whatever the price. Deletions must be sold. For decades this produced a visible pattern: additions rose on the announcement and deletions fell, and part of the move reversed after the effective date.

Tesla, December 2020

Tesla was among the most valuable US companies when S&P announced after the close on November 16, 2020 that it would join the S&P 500 before the open on Monday, December 21. S&P consulted on whether to add it in one step or two and chose one. On December 11 it confirmed that Tesla would replace Apartment Investment and Management.

Tesla's split-adjusted month-end closes were $129.35 for October, $189.20 for November and $235.22 for December 2020: a rise of 46.3% in November and a further 24.3% in December. Not all of that was index buying, in a year when the stock rose many times over. But index funds had no choice about the price. They bought at the close on Friday, December 18, after the run-up, and the S&P 500 recorded Tesla as a member from that price onward.

A shrinking effect

The Tesla episode is the exception now. Greenwood and Sammon, in a study published in the Journal of Finance in 2025, measured the abnormal return of stocks added to the S&P 500 around the announcement. It averaged 7.4% in the 1990s and has been below 1% over the most recent decade. The pattern for deletions has faded in the same way.

Several things changed. Additions are now often promoted from the S&P MidCap 400, whose funds must sell as the S&P 500 funds buy. More investors trade against the event in advance. And index funds have become more willing to trade around the effective date instead of exactly at its close.

What it costs fund holders

Here is the part that is easy to miss. If a stock jumps 5% between the announcement and the effective date, an index fund buys it 5% dearer. But the index adds the stock at the same closing price. The fund matches its index perfectly. The loss is real and invisible: it is built into the index's own return, and it never shows in the fund's tracking difference.

The arithmetic is simple. Suppose 10% of an index's value is replaced in a year, the newcomers are bought 3% above the price they would have had without the event, and the leavers are sold 3% below:

cost = 10% × 3% + 10% × 3% = 0.60% of the index, in that year

Petajisto, in the Journal of Empirical Finance in 2011, estimated this hidden cost from actual additions and deletions between 1990 and 2005 at 21 to 28 basis points a year for the S&P 500 and 38 to 77 basis points a year for the Russell 2000. (A basis point is 0.01 of a percentage point.) Those figures are several times the expense ratio of a modern index fund. They come from a period when the index effect was at its largest, and the later research above suggests the cost for the S&P 500 is now much smaller.

The Russell 2000 is the most exposed of the major indexes. Its members are small and less liquid, its turnover is high because companies both graduate out of the top and fall out of the bottom, and its date is fixed and public. The summary prospectus of the largest fund on it, IWM, reported portfolio turnover of 18% for its latest fiscal year.

Three small-company funds

Growth of $1 in three small-company index funds, March 2004 to October 2026 $1 $2 $4 $8 IWM (Russell 2000) IJR (S&P SmallCap 600) VB (CRSP, now Morningstar, small cap) 2005 2010 2015 2020 2025
Monthly total returns with distributions reinvested, March 2004 to October 2026. Finabase chart data.

Three funds give a rough test. IWM follows the Russell 2000. IJR follows the S&P SmallCap 600, which is chosen by committee and applies an earnings test. VB has followed the CRSP small-cap index (renamed Morningstar US Small Cap in 2026), which uses packeting and multi-day transitions, since January 2013; before that it followed an MSCI small-cap index, so for the first nine years of this comparison it tested a different rulebook.

Fund Index Total return a year Growth of $1 Volatility
IWM Russell 2000 8.56% $6.44 19.7%
IJR S&P SmallCap 600 9.64% $8.05 19.3%
VB MSCI, then CRSP US Small Cap 9.62% $8.01 18.9%

Over the 272 months from March 2004 to October 2026, IWM returned about one percentage point a year less than either of the others. It trailed VB in 15 of the 21 full calendar years and IJR in 11.

It would be convenient to call that gap the cost of the Russell reconstitution. The data do not support saying so. If reconstitution were the cause, the shortfall should concentrate in June. It does not: IWM's average gap against IJR was −0.08% in June months and −0.08% in all other months, and against VB it was +0.46% in June and −0.12% in other months. Month-end fund prices are too coarse to isolate a few days of trading, and much of the trading against the event happens in May.

The three indexes also differ in what they hold. The S&P SmallCap 600 excludes companies without earnings at entry; the Russell 2000 takes everything in its size range. VB's index reaches into larger companies than the Russell 2000 does, and IWM's expense ratio of 0.19% is the highest of the three. Individual years show how much the mix matters: IWM beat IJR by 8.8 points in 2020 and trailed it by 12.0 in 2021. The honest reading is that index design as a whole, of which reconstitution is one part, has been worth roughly a point a year among small-company funds over this period.

Where this misleads

"Front-running" is not free money. The dates are public and many traders buy additions early. That competition is one reason the index effect shrank. Buying a stock because it is about to join an index is a bet that others have not already done so.

Low turnover is not the whole answer. A wide buffer keeps costs down and lets the index drift from its definition. A small-company index that never removed its winners would stop being a small-company index.

A fund matching its index tells you nothing here. IWM's ten-year return to September 30, 2026 was within 0.08 of a point a year of the Russell 2000's. The question this lesson raises is about the index, not the fund.

Published estimates age. The cost figures come from 1990 to 2005. Rules, liquidity and trading practice have all changed since, including the 2026 move to two Russell reconstitutions a year, whose effect cannot yet be measured.

How Finabase does it

Finabase does not build indexes or publish membership changes. Additions and deletions are announced by the index providers, and a company's own news on its symbol page will usually carry the announcement when it joins a major index.

What Finabase shows is the result in the funds. Every fund has a symbol page with quote, chart, news and a Dividends & splits tab. The comparison tab rebases each fund to its percent change from the first bar in view and tabulates price returns from one week to two years, excluding dividends; opening IWM with IJR and VB shows the three small-company indexes side by side. The Holders tab can show a fund card with issuer, expense ratio, assets, net asset value, holdings and sector weights when the data feed has it.

Frequently asked questions

What is the difference between reconstitution and rebalancing?

Reconstitution changes which companies are in the index. Rebalancing changes how much of each is held without changing the list. A capped or equal-weight index rebalances often; a market-value index mostly needs only share-count updates.

Should I sell a stock that is being removed from an index?

Removal says the company no longer meets the index's rules, usually because it has become smaller. It says nothing new about the business. Index funds must sell regardless of price, which is a reason for caution about selling alongside them, not a signal. This is general information, not advice about any holding.

Do index funds trade exactly at the close on reconstitution day?

Many do, because the index is calculated from that closing price and trading at it gives zero difference from the index. Some managers trade part of the change earlier or later to reduce the price impact, accepting a small mismatch in return.

Why did Russell move to two reconstitutions a year?

FTSE Russell's stated purpose was to keep the indexes closer to the current market between annual reviews. With a single June review, a company that grew or shrank sharply in July stayed in the wrong index for almost a year.

Does reconstitution create a tax bill for fund holders?

A fund that sells a deleted stock at a gain realizes that gain. Exchange-traded funds can usually hand such shares out through redemptions without distributing the gain to holders (ETF tax efficiency: in-kind redemptions, heartbeat trades and limits). Traditional index mutual funds have less room to do so.

Summary

Indexes are refreshed on a schedule: twice a year for the Russell indexes from 2026, once a year for the Nasdaq-100, quarterly for the Vanguard benchmarks, and as needed for the S&P 500. Buffers such as Russell's banding keep borderline companies from being traded back and forth. The funds that follow an index must buy additions and sell deletions when everyone knows they will, and the price they pay is recorded by the index itself, so it never appears as a fund falling behind. Research has found that cost to have been significant for the S&P 500 in the 1990s and much smaller since, and largest for small-company indexes with fixed, public dates.

Check your understanding

5 questions on this lesson. Each answer comes with the reason, and nothing is scored.

  1. A stock rises 6% between the announcement that it will join an index and the day it joins. An index fund buys it at the closing price on that day. How does this show in the fund’s results?

    1. The fund trails its index by the cost of the run-up
    2. It does not show against the index, because the index adds the stock at the same closing price
    3. The fund beats its index
    Show the answer

    B. It does not show against the index, because the index adds the stock at the same closing price The index records the newcomer at the effective-day close, the same price the fund pays. The cost lowers the return of the index and the fund together, so the fund’s gap to its index is unaffected.

  2. What is the purpose of banding in the Russell indexes?

    1. To cap the weight of the largest companies
    2. To keep a current member in its index unless it has clearly moved past the size boundary
    3. To delay the reconstitution in volatile markets
    Show the answer

    B. To keep a current member in its index unless it has clearly moved past the size boundary A member changes index only if it lands outside a band around the breakpoint (2.5 points of cumulative market value at rank 1,000). That avoids trading companies back and forth across the line.

  3. In a year, 8% of an index is replaced. Additions are bought 2% above and deletions sold 2% below their undisturbed prices. What is the hidden cost to the index that year?

    1. 0.16%
    2. 0.32%
    3. 4%
    Show the answer

    B. 0.32% 8% × 2% on the buying side plus 8% × 2% on the selling side is 0.16% + 0.16% = 0.32% of the index.

  4. Which index has no scheduled reconstitution date?

    1. The Russell 2000
    2. The Nasdaq-100
    3. The S&P 500
    Show the answer

    C. The S&P 500 The S&P 500’s committee replaces members as needed. The Russell indexes reconstitute in June and, from 2026, December; the Nasdaq-100 does so each December.

  5. IWM returned about one point a year less than two other small-company index funds from 2004 to 2026, but its gap was no larger in June than in other months. What can be concluded?

    1. The Russell reconstitution caused the whole gap
    2. Reconstitution has no cost
    3. The indexes differ in several ways, and monthly fund data cannot isolate the reconstitution’s share
    Show the answer

    C. The indexes differ in several ways, and monthly fund data cannot isolate the reconstitution’s share The three indexes differ in size range, earnings screens and fees as well as in how they reconstitute, and trading around the event is spread over weeks. Month-end prices cannot separate those effects.

Sources and further reading

  1. FTSE Russell — Russell US Equity Indexes, Construction and Methodology, v7.4 (November 2026): reconstitution dates, rank days and banding
  2. LSEG / FTSE Russell — press releases of January 16, 2025 (move to semi-annual reconstitution), September 1, 2026 (December 2026 schedule) and November 13, 2026 (preliminary results)
  3. S&P Dow Jones Indices — press releases of November 16, 2020 and December 11, 2020 on Tesla joining the S&P 500
  4. Nasdaq — Nasdaq-100 Index Methodology (2026 edition): annual reconstitution and quarterly rebalance
  5. Vanguard — Vanguard Small-Cap ETF (VB) fact sheet: spliced benchmark (MSCI US Small Cap 1750 Index through January 30, 2013; CRSP US Small Cap Index thereafter)
  6. Robin Greenwood and Marco Sammon, “The Disappearing Index Effect,” Journal of Finance, vol. 80, no. 2 (2025), pp. 657–698
  7. Antti Petajisto, “The index premium and its hidden cost for index funds,” Journal of Empirical Finance, vol. 18, no. 2 (2011), pp. 271–288
  8. Vanguard — fact sheets of Vanguard S&P 500 ETF and Vanguard Total Stock Market ETF, September 30, 2026: turnover rates
  9. Finabase chart data — monthly bars and dividends of IWM, IJR and VB and month-end closes of TSLA through October 2026; statistics computed by Finabase