The price effects of FTSE 100 index revision: what drives the long-term abnormal return reversal?

Abstract

We examine short- and the long-term price effect associated with the FTSE 100 index revisions. We control for both heteroskedastic nature of the residual and the change, between the estimation and the test period, in the beta coefficient of the standard market model. Our findings reveal no relationship between the long-term price reversals and the change in the discount rate, as approximated by the beta coefficient of the market model. Overall, we provide strong evidence in favour of the price pressure hypothesis, where the price increase (decrease) gradually starting before the announcement an inclusion (exclusion) and reverses completely in less than two weeks after the index revision date.

Acknowledgement

The authors are very grateful to Shuxing Yin for her helpful research assistance. The comments provided by colleagues at Aston University in particular, Patricia Chelley-Steeley, Jim Steeley, Nathan Joseph and Samuel Agyei-Ampomah on earlier versions of the article are gratefully appreciated. The article has also benefited from the comments of an anonymous referee and the editor, Mark Taylor.

Notes

¹ A detailed explanation of the volatility hypothesis is provided in the next section.

² The change in the beta coefficient of the market model associated with the FTSE 100 index revision is reported by Coakely and Kougoulis (2004).

³ Separation between pre- and post-event betas is recommended by Jain (1987) and Copeland and Mayers (1982).

⁴ This model is discussed in Karafiath (1988) and Kryzanowski and Zhang (1993). It is also used in the context of corporate layoff announcements by Hahn and Reyes (2004).

⁵ The release ending day is defined by Lynch and Mendenhall (1977) as the day when the index funds have completed their trades. Opong and Hamill (2004) suggest that the release ending day as the day with mean abnormal trading volume (AV_t ) that is lower than the average AV_t for all later days through to +21 day. We initially use day CD + 3 as the end release day to compare our results with that of Opong and Hamill (2004). However, an alternative definition is used to test the sensitivity of the results to the window selection. More details are provided in Section ‘Price volatility effect’.

⁶ The decision to include (exclude) a particular stock to (from) FTSE 100 index is based solely on the market capitalization which rules out the information hypothesis as satisfactory explanation to the observed price behaviour.

⁷ However, it is important to point out that the long horizon tests have low power due to the greater sampling variability, see among others, Brown and Warner (1985), Chapter 4 of Campbell et al . (1997) and Hedge and McDermott (2003).

⁸ The fact that our end-release day is different from that reported by Opong and Hamill (2004) is likely to be due to the use of different methodologies in examining the volume effect associated with the index revision.