The Index Effect Revisited

January 01, 2005

Illustration by Garrian Manning

Summary of Research Findings:

  • There has been an intertemporal shift in the index effect. Excess returns associated with index addition announcements have diminished sharply.
  • The median excess returns between announcement date and effective date for S&P 500 adds was 8.9% between mid-1998 and mid-2000, 4.5% between mid-2000 and mid-2002 and only 3.6% between mid-2002 and mid- 2004. This declining pattern is also observable for S&P MidCap 400 and S&P SmallCap 600 additions.
  • Excess volumes remain robust, suggesting the decline has not been caused by reduced trading interest in index changes.
  • Index fund demand for added shares has not shrunk. The portion of shares outstanding of a company demanded by S&P 500 index funds has remained at 10% to 11% over
    the past six years. It has more than doubled for the S&P MidCap 400 and S&P SmallCap 600, to 6% and 5% respectively.
  • Neither does an increase in interindex moves explain the decline. The diminution of the index effect across time persists in the subsample of companies added from outside the S&P Composite 1500.
  • There are three structural issues behind the decline:
  • First, the growing popularity of the index effect has resulted in a surge of proprietary trading activity that seeks to exploit this arbitrage opportunity. As the competition among arbitrageurs has increased, the profit opportunity has diminished.
  • Second, many index funds increasingly have begun to trade around the effective date to minimize the cost of index changes.
  • Third, the increase in indexed assets for the midand small-cap indexes provides better offset for moves, thus reducing net demand.

Index Effect Redux
From tests of demand curves of stocks to more recent industry research on "wealth loss" associated with changes to popular stock indexes, the index effect has been a muchresearched topic. The fact that stocks added to a popular index experience excess returns and volumes between the announcement date (AD) and effective date of change (ED), followed by a small post-ED correction, has been well documented over the last two decades. Several theories exist to
explain this phenomenon.1

Our aim here is not to revisit these old issues. Rather, we wish to highlight an intertemporal shift in the index effect, which we believe has important ramifications. Specifically, the index effect has begun to diminish thanks to changes in the trading environment and the way index funds handle
these events. This diminution of the index effect in recent years needs to be factored into design and evaluation of trading strategies. It also means that future research on the subject, most of which take into account five years of data or more, would need to account for intertemporal shifts in the
research design.

The Shrinking Of The Post-Addition Price Pop
Exhibits 1, 2 and 3 show the median excess returns of additions to S&P 500, S&P MidCap 400 and S&P SmallCap 600 over three time periods: mid-1998 to mid-2000, mid-2000 to
mid-2002 and mid-2002 to mid-2004. These time periods encompass different market environments. They also differ in the number of index changes because of differences in level
of corporate activity such as mergers, acquisitions or spinoffs. (Appendix 1 has more details on the index changes during these time periods.)

Across all three indexes, the median excess return between the announcement date and effective date of change has declined from the first time period to the second, and from the second to the third. In all the three exhibits, the excess returns line for the latest time period (7/2002 to 6/2004) falls sharply below the other two periods.


The Shrinking Index Effect Is Not Explained By
Reduced Trading Interest

Has the decline in the post-addition price pops been caused by reduced trading interest in index changes? The answer is a clear no. Exhibits 4, 5 and 6 chart the median abnormal volumes for the day after announcement (AD+1), ED, and for five days after ED. We measure abnormal volume
by dividing the daily volume by the average volume of the prior 20 days. To account for seasonal changes in trading activity, we then divide the result by a similar measure for

NYSE aggregate volume.
While there is no clear pattern emerging across time (except perhaps the increase in trading activity with S&P SmallCap 600 and S&P MidCap 400 changes post-2000), it is clear that trading interest remains strong and excess volumes remain robust. Clearly, index additions continue to be traded
heavily by speculators and index fund managers.

Demand For Shares By Index Funds
Has Not Diminished

The shrinking excess returns are also not explained by reduced demand from index funds. The portion of shares outstanding of a company demanded by index funds upon its admission into an index is shown in Exhibits 7, 8 and 9. While this percent has remained at 10% to 11% over the past six years for the S&P 500, it has more than doubled for the S&P MidCap 400 and S&P SmallCap 600 following an increase in assets linked to those indexes.
 

Addition candidates to Standard & Poor's indexes have been subjected to minimum liquidity and public float requirements through all three periods considered. Therefore, stock
 

level liquidity constraints cannot be a reason either.

Neither Does An Increase Interindex Moves Explain
The Decline

Additions caused by interindex moves (deletion from one component of the S&P Composite 1500 and addition to another) have a less pronounced index effect because of offsetting buy and sell requirements of different index fund groups. For example, if an S&P MidCap 400 company is moved to the S&P 500, buy trades from S&P 500 index funds will be offset quite a bit by sell trades from S&P MidCap 400 index funds.
Exhibit 10 shows the median excess returns between AD and ED for the sample excluding the interindex moves. The diminution of the index effect across time persists in the subsample of companies added from outside the S&P Composite 1500.

Exhibits 11, 12 and 13 show the percentage split of additions coming from outside the 1500 and those that come from interindex moves. The percentage of moves has actually declined for the S&P 500. For the S&P MidCap 400 and S&P SmallCap 600, there has been an increase in interindex moves.
 

SoWhat Explains The Decline In The Index Effect?

There are several structural issues behind the decline of the index effect that suggest the declining trend will persist in the longer term. Arbitrage profits decline as competition among arbitrageurs increases. The index effect generates an arbitrage opportunity caused by sudden demand for shares of a particular stock.
As with arbitrage opportunities such as the January effect, the index effect will diminish as the number of arbitrageurs increase. Over the past six years, the growing popularity of the index effect has resulted in a surge of proprietary trading activity that seeks to exploit this arbitrage opportunity. As the competition among arbitrageurs increases, the profit opportunity dissipates quickly.
Index funds trading around the effective date. In theory, an index fund buys shares of the company being added to the index at close of trading on the effective date. That is increasingly not the case. Discussions with index fund managers suggest that over recent years, many have increasingly begun to trade around the effective date to minimize the cost of index changes. Further, they also may not buy their required shares in a single block.
Increase in indexed assets for mid- and small-cap indexes provides better offset for moves. The increase in S&P Mid-Cap 400 and S&P SmallCap 600 assets to 6% and 5% of shares of their constituents means that interindex moves between those two indexes are well matched by index fund drive trades. The differential between S&P MidCap 400 and S&P 500 in terms of the fraction of shares held by index funds has shrunk from 8% in 1998 to 5% in 2003. Therefore, moves between those two indexes are also better matched than before.

And What Are The Implications?
 

  • Index arbitrage profits from additions will become more difficult as competition increases and index fund trading becomes more unpredictable.
  • The index effect will probably never vanish. Rather it will, on average, settle at a level much below those seen in the past decade.
  • The "wealth loss" associated with changes to popular size indexes will decline. This will make index-based tactical asset allocation cost competitive versus broad-market indexing. Also, this means index funds preferring to trade exactly at the close of the effective date will benefit from the trading of those index funds that prefer to trade around it.
  • Future research on the index effect needs to factor in this structural shift in the market. A more fertile area for research is price changes associated with share changes in indexes. Similarly, most academic research focuses on price impact on a daily basis. Intraday pricing data or the price impact in different markets (i.e., specialist or dealer based) might be a more interesting way to look at the index impact.

* "Others" includes:

  • Spin-off by an index constituent in which the spun-off company is added to the index.
  • Acquisition of one index constituent by another, and the combined company being added to the index.
  • A company acquiring an index constituent and the merged company being added to the index. In each of these cases, the index fund owns fully or in part the required shares of the added company. Therefore, these are not pure additions from an index effect perspective. These are removed from our sample in the data presented in this report.

1. For a summary of these theories, see "Price Changes Associated with S&P 500 Deletions: Time Variation and Effect of Size And Share Prices," July 9, 2002 at www.indices.standardandpoors.com

 

Top    
 

Find your next ETF

Reset All