The Index Effect Revisited
Posted: 11 Feb 1999
Date Written: November 1998
The price reaction in stocks that are added to an index has been fairly well examined in the literature. Beneish and Whaley (BW, 1996) analyzed the effects of changes in the S&P 500 index on trading volumes, bid/ask spreads and cumulative abnormal returns. They found that there is a permanent increase in price associated with additions to the S&P, but transitory effects on volume and bid/ask spreads. Furthermore, they documented a trend among index funds to avoid this obvious distortion by trading earlier in the announcement period rather than the effective day and predicted that the "S&P Game" would disappear.
Recent evidence, however, suggests that the index effect is live and well. A recent article in the Wall Street Journal, highlighted a major disruption in the market for Safeway Inc. driven by its addition to the S&P 500 index. Moreover, it is well known in the industry that several major "risk arbitrageurs" make a regular practice of accumulating S&P addition stocks prior to the effective day of inclusion to sell to index funds. Some players have even been observed offering to sell to index funds at 1/16th below the closing price on the day prior to the effective inclusion date. Thus, the evidence suggests that not only is the index effect still alive, but many players are making a regular practice of profiting on it. In addition, the amount of funds indexed to just the S&P 500 has grown from $335 billion in 1996 to over $600 billion today. As a result, the predictions and inferences of BW are worthy of further investigation.
The purpose of this paper is to revisit the index effect, document its existence from an economic and statistical perspective and, most importantly, to examine the implementation of trading strategies that can profit from it. Using a highly detailed data set which contains the exact announcement time and date of an S&P change and estimates of the total dollars indexed on a monthly basis, we examine both intra-day and inter-day trading strategies. Furthermore, we examine the "risk arbitrage" from both sides of the coin by including deletions as well. Thus, the overall "long/short" strategy of buying additions and selling deletions is considered. We also extend the analysis to include additions and deletions to S&P Mid-Cap and Small Stock indexes. Since many times additions to one index entail deletions from another (and vice-versa) we examine the net effects as well as the incremental effects (a potentially confounding effect not previously addressed in the literature).
The results show that the index effect is indeed alive and well. In addition, we document a pronounced index effect abnormal return in mid-cap stocks although small cap stocks do not exhibit a significant index effect. The lack of an effect in small cap stocks is attributed to the smaller number of dollars indexed and to the fact that the total number of stocks in the small cap index is so large. Thus, a given index change among, for example, the Russell 2000 has a very small incremental effect on an index portfolio. It appears that the supply/demand distortion induced by an index change does not overwhelm the market making capacity in small stocks as it does in larger stocks. Our results also indicate that there are significant intra-day/inter-day trading strategies which are robust to out of sample tests.
1 The Wall Street Journal, November 19, 1998 "Spear Leeds is Criticized Over Safeway Trading", p. C1. 2 Source: Salomon Brothers Indexing Bulletin, October 1998.
JEL Classification: G1,G12,G14
Suggested Citation: Suggested Citation