The index effect: what happens to prices when stocks join or leave an index

Advanced10 min read · updated 11 September 2026 · 14 sources

The index effect describes how stocks react when they are added to or removed from a major index: additions rise and deletions fall, often already between announcement and implementation. Here you will learn how an index change works, using MSCI and Germany's rule-based DAX as examples, how large index fund demand is relative to trading volume, which explanations research tests and why the effect has almost disappeared for the S&P 500. What it costs index funds is covered in the article on smart indexing.

What the index effect is

When an index provider changes the composition of an index, all funds that track it have to rebalance: buy the added stocks and sell the deleted ones. Because this demand is known in advance and arrives within a short time, it moves prices. Two parts are distinguished: the announcement effect between announcement and implementation, and the effect on the effective date itself, when the index funds trade in the closing auction. Afterward, part of the move often reverses.

For research, the index effect is a rare experiment: demand changes without anything changing at the company, at least at first glance. If the price reacts permanently, this contradicts the textbook assumption that every stock has any number of equivalent substitutes and that its demand curve is therefore horizontal.

How an index change works

The major index families review their indices on fixed dates. STOXX's German DAX indices, for example, follow a clear timetable (DAX Methodology 2026):

MSCI reviews its indices quarterly in February, May, August and November (MSCI GIMI Methodology 2024); in August 2026, the results were published on August 12 and implemented as of the close on August 31 (MSCI Index Review 2026). So there are a few weeks between announcement and implementation. During this time, arbitrageurs buy the additions and sell the deletions in order to pass them on to the index funds on the effective date. This is front-running in the broader sense. Index funds, by contrast, mostly trade via market-on-close orders in the closing auction, because only the closing price on the effective date keeps their tracking error small.

A rule-based example: Germany's DAX, MDAX and SDAX

The DAX comprises the 40 largest companies on the Regulated Market of the Frankfurt Stock Exchange, the MDAX the next 50 and the SDAX a further 70; the TecDAX contains the 30 largest technology stocks. What counts is the rank by free-float market capitalization in the monthly ranking list. New DAX members also need a positive EBITDA in the last two financial years, and weights are capped at 15% each quarter (DAX Methodology 2026). Four rules decide on promotion and relegation:

Source: STOXX DAX methodology, August 2026 version (DAX Methodology 2026). Corresponding ranks apply to the MDAX, SDAX and TecDAX; STOXX may deviate in the case of takeovers or large changes in free float.

The fast entry and fast exit rules ensure that large shifts reach the index quickly; the gaps between the entry and exit ranks act as a buffer zone and prevent constant back and forth. Because the ranking lists are published every month and the rules are public, many changes can be estimated in advance. That is exactly what arbitrageurs do. Borderline cases remain open, because prices and free float change up to the cut-off date.

Example: how large is index fund demand?

Index fund demand follows from the new index weight and the assets that track the index one to one. The key question is how many days of normal trading it corresponds to:

The example explains a finding that Greenwood and Sammon (2022) came across in the working paper version of their study: more and more additions to the S&P 500 come from the S&P MidCap 400, whose funds sell on promotion; as a result, net demand turns out much smaller than the demand from S&P 500 funds alone.

Explanations: why prices react

• Price pressure: the funds buy large quantities within a short time; liquidity providers demand a premium for this, which then reverses. This interpretation was put forward by Harris and Gurel (1986), who found a short-term reversal.
• Downward-sloping demand curves: if there are no perfect substitutes, a permanently higher demand also shifts the price permanently (Shleifer 1986). Kaul et al. (2000) managed a clean test: an announced, information-free change to the free-float definition in the TSE 300 in 1996 raised the weights of 31 stocks; they earned an excess return of 2.3% in the event week, with no subsequent reversal.
• Arbitrage risk: arbitrage between stocks that are not quite identical is risky and therefore does not flatten the demand curve completely. Wurgler and Zhuravskaya (2002) show that stocks without close substitutes jump more on inclusion in the S&P 500.
• Information: an addition can reveal something new or cause it. According to Denis et al. (2003), earnings forecasts and actual earnings rise more after inclusion in the S&P 500 than for comparable companies.
• Attention: in the model of Merton (1987), investors know only some of the stocks; more attention lowers the required return and lifts the price. Chen et al. (2004) use this to explain an asymmetry: additions rise permanently, while deletions do not fall permanently, because awareness, once gained, is not lost.
• Liquidity on the effective date: the better the market can absorb the trading on the effective date, the smaller the effect. According to Greenwood and Sammon (2022), the effective date's share of trading volume around additions rose from about 15% in the 1990s to almost 30% in the 2010s.

Advanced: measuring the effect

The index effect is measured with an event study, but with two event days: the announcement A and the implementation E. Three windows are usual: the abnormal return from A to the day before E (run-up), the return on day E itself (price pressure in the closing auction) and the return in the weeks afterward (reversal). There are also a few pitfalls:

• Benchmark: after the effective date, the stock is itself in the index; adjusting with the index is then slightly biased. Benchmark portfolios of similar size and industry are cleaner.
• Clustered dates: all changes in a review fall on the same day. The abnormal returns are then correlated across the cross-section, and simple t-tests overstate the significance (Kolari and Pynnönen 2010).
• Selection: stocks that have risen strongly beforehand are the ones that get added. Without controlling for momentum, you confuse prior history with the effect.
• Net demand instead of weight: if you want to estimate the elasticity, you need the net demand of all affected funds, including the sales from the index the stock is leaving.

What the research shows

The early studies of the S&P 500 found marked price jumps but disagreed about their persistence: Shleifer (1986) found an effect that did not disappear quickly after inclusion, Harris and Gurel (1986) a largely complete reversal within two weeks. Chen et al. (2004) documented the asymmetry between the permanent gains of additions and the temporary losses of deletions. What this costs index funds was estimated by Petajisto (2011); the figures are in the article on smart indexing.

The most important recent finding: the effect has shrunk. For additions to the S&P 500, the excess return fell from an average of 7.4% in the 1990s to under 1% in the last decade, and deletions from 2010 to 2020 produced an average of only 0.1%, even though an ever larger part of the market is tied to the index (Greenwood and Sammon 2025). In the working paper version, the authors cite as the main reasons the promotions from the MidCap index with their lower net demand, and a market that can now better meet the demand on the effective date; the greater predictability of additions may also have contributed (Greenwood and Sammon 2022). They find a similar decline for other index families. The studies cited here concern US indices and the Canadian TSE 300; their figures cannot be transferred one to one to other markets, such as the German indices described above.

Limitations, misconceptions and common mistakes

• Old figures as a strategy: what amounted to several percent in the 1990s is under 1% for the S&P 500 today. If you bet on additions, you compete with specialized traders and bear the risk of wrong forecasts.
• Missing the announcement: after the announcement, most of the move is usually already in the price. The price on the effective date is of little use to outsiders.
• Gross instead of net: spreads and market impact are high for small stocks that suddenly face heavy demand.
• Not a seal of quality: an addition follows rules on size and free float, not a recommendation by the index provider. For the DAX, what counts is size, minimum requirements such as free float and liquidity, and a positive EBITDA.
• Individual case and average: research describes averages over many changes; a single case can turn out quite differently. This is not investment advice.

How Portfolio Lab calculates

Portfolio Lab does not predict index changes and has no tool for event studies. You can, however, examine a stock's price path around an addition yourself: in Settings, set the data frequency to daily, set up the stock in the Backtest as a separate portfolio and choose the index or an index ETF as the benchmark. The “Single assets” tab shows metrics and correlations, the metrics against the benchmark provide beta and alpha, and the CSV export “Period returns” contains the daily returns of the stock and the benchmark for your own calculation of abnormal returns. For index funds, the effect is already in the price, because the index and the fund bear it together. Which securities a fund holds according to its holdings list is shown by the Look-through.

This article explains a method; it is not investment or tax advice. Worked examples are simplified and rounded. All terms in the glossary.