On June 1st, 2009 Citi Group and General Motors were removed from the Dow in favor of Cisco Systems and Travelers Insurance. This study examines the impact of a company being added to the 30 stock index. Most opinions are that there should be a positive effect to the company’s stock price due to the large amount of money indexed. With over $1 Trillion dollars in tracking funds, there should be substantial funds flowing into the new companies by managers that need to hold the market portfolio.
Based on the CAPM structure of expected returns to a stock, based on the return to the market and the particular stocks sensitivity (β) plus a constant (α), we can compare the expected (CAPM) return to the actual return and analyze the reaction by investors to the news. If the news was truly unexpected, we expected to see statistical outperformance on the day of the announcement followed by a dissipation of the abnormal excess returns if markets are efficient.
This study examined a 10 day window before the announcement and a 10 day window after the announcement. Some studies try to examine the average outperformance over a year trailing the announcement of an index addition and have shown positive results. For this study though we are examining the immediate effect.
Based on a one year regression of the added stock (Travelers Insurance) versus the index (Dow Jones) there was a statistically insignificant alpha (α) of 0.002 and a market beta (β) of 1.28. It’s worth mentioning that the 5yr beta for Travelers is 0.67 (vs. S&P) and the differences between these two numbers shows the inconsistencies possible in these studies as well as the shortcomings of CAPM based expected returns.
In the 10 days prior to the announcement of the Dow addition, Travelers exhibited a consistent negative abnormal return. None of the single day underperformances were statistically significant at a 95% confidence interval. The presence of the negative performance prior to the event confirms no significant leakage of the positive event. Additionally, a director for Travelers insurance sold $3 million in company stock the Friday before the Monday announcement. If the board or executives had expected the addition, it is unlikely that they would have put in the sell order.
For the ten days following the event the results are significantly different. Travelers’ begins to accumulate abnormal returns, though none of the one day moves were significant statistically. 5.37% of abnormal returns in the seven days following the event is significant to me. This could be because the markets truly didn’t see the addition as a good thing, or could be a shortcoming of the standard error derived from the CAPM regression. In absolute terms, the outperformance for the ten days following the event was more than 5% abnormal. Those are returns over the market any investors would value regardless of their statistical significance.
Are markets efficient? Is positive news immediately priced into the security and abnormal returns erode quickly? Statistically yes in this case, though the ten day window should positive abnormal returns most days after the event.