By: Mat Lystra, Sr. Research Analyst
In a previous post, we explored how the 2007 introduction of banded methodology resulted in lower turnover for the Russell 2000 during index reconstitutions. While this is positive news for investors in index tracking funds, the broader question remains: how has this affected index performance? A closer look at the numbers reveals a significant reduction in price impact, which can be attributed to banding and other potential factors.
As a first step to measuring the impact of reconstitution on Russell 2000 performance, consider the performance of a strategy with long positions in stocks to be added to the Russell 2000 and short positions in those to be deleted. The table below reports the performance of such a strategy where both portfolios are capitalization weighted.
Negative performance of this strategy before recon day is beneficial for investors in the index tracking fund, as it is indicative of lower prices for adds to be bought and higher prices for deletes to be sold by the fund. After recon, positive performance is better for those investors as it is indicative of higher prices for stocks added and lower prices for those deleted.
Average returns of the Russell 2000 adds-minus deletes portfolios
The first table demonstrates that for the May-end to reconstitution date (June 30) time period average returns for 2007-2014 are lower than average returns for 2000-2006, which again is a positive for investors in index tracking funds. On the other hand, negative impacts after reconstitution are almost equal or smaller in magnitude for all four time periods after reconstitution, which is also favorable.
The performance of the adds-deletes strategy is not very informative regarding the effects on Russell 2000 performance for two reasons. First, it does not take into account the weight variation between the adds and deletes portfolios. Second, the adds and deletes portfolios weights are small in comparison to the Russell 2000 itself.
The next table shows the estimated impacts on Russell 2000 performance after reconstitution if adds and deletes are properly weighted. Positive returns to the adds portfolio increases index performance while positive returns to the deletes portfolio reduces index performance.
Average impact on Russell performance
The table shows estimated impacts over four time horizons after index reconstitution. In all cases the average impact is much lower in the 2007 to 2014 period than in the 2000-2006 period. Focusing on the longest study window, at the end of August, the index impact is estimated to be an average of -1.28% over 2000-2006 but only -0.03% over 2007-2014.
In addition to being much lower in magnitude, the reconstitution price effect appears to peak at 10 days after reconstitution in the 2006-2014 period, while it is at a maximum at the end of August over 2000-2006. The peak at 10 days after reconstitution is consistent with the idea that there should be a partial reversal of impact on index performance as reconstitution price pressure effects reverse, leaving only possible index-inclusion effects.
These results are further supported by two recent academic studies. Petajisto (2011) conducts a study based on price changes between May 31 and June 30 over the period 1990-2005 and estimates reconstitution price effects on the Russell 2000 as being between 0.38% and 0.77% per year, which is generally consistent with our results for the 2000-2006 period.
Chang, Hong and Liskovich (2015) study Russell 2000 reconstitution over the period from 1990 to 2012. While they do not produce impact cost estimates, their conclusions are also consistent with our findings. They report a lower cost of indexing, but cite arbitrage efficiency as another potential reason behind this trend. The study suggests that “the force of arbitrage efficiency has increased faster than the effect of indexing,” thereby putting downward pressure on indexing costs. They also find two other trends consistent with lower reconstitution price impacts: increasing volume and short interest ratios.
This mounting empirical evidence gives us good reason to believe that the average price impacts of Russell 2000 reconstitution have been significantly reduced and are currently at low levels.
For more detailed information on reconstitution see our latest research report, Russell 2000 Reconstitution Effects Revisited.
 The formula for index impact computation can be found on page 12 of Cariño and Pritamani, op. cit. Only time periods after index reconstitution are considered under the assumption that reconstitution-related decreases in the prices of stocks to be deleted that take place before index reconstitution are followed by price reversals afterward. The effects of Increases in the prices of deleted stocks are subtracted from index performance under this theory.
 See Cariño and Pritamani, op. cit., for further discussion.
 “The index premium and its hidden cost for index funds,”Antti Petajisto,Journal of Empirical Finance, Vol. 18 (2011), pp. 271-288.
 “Regression Discontinuity and the Price Effects of Stock Market Indexing,” Yen-Cheng Chang, Harrison Hong, and Inessa Liskovich, Review of Financial Studies, Vol. 28 No. 1 (2015), pp. 212-246.
© 2016 London Stock Exchange Group plc and its applicable group undertakings (the “LSE Group”). The LSE Group includes (1) FTSE International Limited (“FTSE”), (2) Frank Russell Company (“Russell”), (3) FTSE TMX Global Debt Capital Markets Inc. and FTSE TMX Global Debt Capital Markets Limited (together, “FTSE TMX”) and (4) MTSNext Limited (“MTSNext”). All rights reserved.
FTSE Russell® is a trading name of FTSE, Russell, FTSE TMX and MTS Next Limited. “FTSE®”, “Russell®”, “FTSE Russell®” “MTS®”, “FTSE TMX®”, “FTSE4Good®” and “ICB®” and all other trademarks and service marks used herein (whether registered or unregistered) are trade marks and/or service marks owned or licensed by the applicable member of the LSE Group or their respective licensors and are owned, or used under licence, by FTSE, Russell, MTSNext, or FTSE TMX.
All information is provided for information purposes only. Every effort is made to ensure that all information given in this publication is accurate, but no responsibility or liability can be accepted by any member of the LSE Group nor their respective directors, officers, employees, partners or licensors for any errors or for any loss from use of this publication or any of the information or data contained herein.
No member of the LSE Group nor their respective directors, officers, employees, partners or licensors make any claim, prediction, warranty or representation whatsoever, expressly or impliedly, either as to the results to be obtained from the use of the FTSE Russell indexes or the fitness or suitability of the indexes for any particular purpose to which they might be put.
No member of the LSE Group nor their respective directors, officers, employees, partners or licensors provide investment advice and nothing in this communication should be taken as constituting financial or investment advice. No member of the LSE Group nor their respective directors, officers, employees, partners or licensors make any representation regarding the advisability of investing in any asset. A decision to invest in any such asset should not be made in reliance on any information herein. Indexes cannot be invested in directly. Inclusion of an asset in an index is not a recommendation to buy, sell or hold that asset. The general information contained in this publication should not be acted upon without obtaining specific legal, tax, and investment advice from a licensed professional.
No part of this information may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the applicable member of the LSE Group. Use and distribution of the LSE Group index data and the use of their data to create financial products require a license from FTSE, Russell, FTSE TMX, MTSNext and/or their respective licensors.