S&P 500 Index Inclusions and Analysts’ Forecast Optimism

BACKGROUND AND HYPOTHESES

Numerous studies have documented that security analysts overreact (underreact) to extreme good (bad) news. While earlier research contends that cognitive bias exists among analysts’ earnings forecasts following extreme earnings news, more recent studies (e.g., Abarbanell and Bernard [1992], Easterwood and Nutt [1999], and Gu and Xue [2007]) attribute rational analysts’ forecast optimism to analysts’ incentive bias. The later stream of literature argues that incentives induced by extreme good or bad news may lead analysts to issue estimates that are too optimistic.

Two (non-competing) explanations have emerged as to why analysts issue more optimistic forecasts. The first explanation follows the strategic bias model (Francis and Philbrick [1993]; Das, Levine, and Sivaramakrishnan [1998]; Lim [2001]; Mest and Plummer [2003]), in which analysts have an incentive to issue favorable earnings estimates to maintain their relationship with management. Sustaining a good relationship with corporate executives is critical for analysts competing to obtain access to certain information withheld by management. In response to high earnings uncertainty, the strategic bias view suggests that analysts’ earnings estimates would be upward biased.

The second explanation follows the selection bias model (McNichols and O’Brien [1997]; Hayes and Levine [2000]; Diether, Malloy, and Scherbina [2002]), which argues that because analysts are attracted to make forecasts when a firm’s earnings reach above a certain threshold (self-censoring), forecast estimates tend to be optimistic. Furthermore, Diether et al. [2002] show that the forecast optimism resulting from self-censoring is stronger when earnings uncertainty measured by the variance of actual earnings is higher.

We study the behavior of security analysts surrounding the announcements of S&P 500 additions. Specifically, we explore the properties of analysts’ earnings forecasts around the announcements. Earlier studies have shown that S&P 500 inclusions convey favorable information about the added firms. Denis, McConnell, Ovtchinnikov, and Yu [2003] find that companies added to the S&P 500 experience significantly higher operating performance. The results of Denis et al. [2003] suggest that companies added to the S&P 500 realize economically significant increases in actual EPS during the post-inclusion period. Their finding also suggests that security analysts tend to issue upward revisions of EPS estimates for a company that is being included in the index. Dhillon and Johnson [1991] find indirect evidence of operating performance improvements for companies entering the S&P 500. Their results indicate that the value of bond increases following the firm’s addition to the S&P 500. Similarly, Cai [2007] and Jain [1987] provide indirect evidence supporting the information content viewpoint. These studies attribute higher institutional scrutiny (monitoring), lower cost of capital, and greater managerial incentives to succeed as possible factors influencing the value of the added firms.

For companies that are newly added into the S&P 500, analysts may be more likely to issue overoptimistic earnings forecasts after a firm is added into the S&P 500. This is because membership in the elite S&P 500 may be associated with an increased firm visibility (status), higher levels of investment banking relationships, and new information. Analysts’ earnings forecasts may thus be biased for incentive reasons. We examine analyst behavior surrounding announcements of Index inclusions. We test the following (null) hypotheses:

HO1: The level of analyst coverage does not change after a firm is added into the S&P 500 Index.

We consider the actual number of analysts following the added company. Index additions may attract analysts’ interest in the company. The selection bias model predicts an increase in analyst coverage for firms added into the S&P 500. We expect to find an increase in the number of analysts issuing earnings forecasts for an added firm.

HO2: The level of forecast optimism does not change when a company is added into the S&P 500 Index.

We investigate forecast optimism surrounding S&P 500 Index inclusion announcements. We compare analysts’ earnings forecasts in the pre-inclusion and post-inclusion periods to determine whether analysts’ forecasts are optimistic. We expect to find forecast optimism in the post-inclusion period.

HO3: The dispersion of analysts’ earnings forecasts remains unchanged following the announcement of index additions.

Analysts’ forecast dispersion indicates whether analysts are in agreement as to the future prospect of the companies added into the S&P 500. Analysts may agree that the future prospect of these companies would improve but disagree about the degree of operating improvement. We expect to find an increase in forecast dispersion.

HO4: Analysts’ earnings forecast accuracy does not differ after a company is included into the S&P 500 Index.

Due to forecast optimism, we predict a decrease in analysts’ forecast accuracy. This is consistent with the non-cognitive bias framework, which suggests forecast accuracy decreases as analysts are overly optimistic about future earnings.

DATA

We analyze firms that were added to the S&P 500 Index over the period 1991 through 2005. We obtain the S&P 500 addition data over the period January 1991 through December 2000 from Jeffrey Wurgler. This dataset was used in two earlier S&P index studies: Wurgler and Zhuravskaya [2002] and Barberis, Shleifer, and Wurgler [2005]. We collect the remaining index inclusion data from the Standard and Poor’s website. The initial sample consists of 386 additions over the period 1991 to 2005. Many of the S&P index additions are associated with other contemporaneous corporate events (e.g., mergers and acquisitions). We exclude inclusions related to these events from our sample. Likewise, 56 additional companies are not included in the sample due to being spin-offs from an already included parent company. We further removed 42 additions from the sample due to merger and acquisition activities.

We are left with 298 additions to the S&P 500 that are eligible for further analysis. For these companies, we collect financial statement data from CRSP/Compustat merged Industrial and Research files. Additionally, we obtain analyst forecast data from the I/B/E/S summary database. We require three-year analyst forecast data around announcements of S&P 500 Index addition be available. Finally, we dropped 104 additions due to missing or incomplete forecast data. The final sample consists of 194 index additions over the period 1991 to 2005.

METHODOLOGY

We measure bias (BIAS) consistent with prior research (Duru and Reeb [2002], Lang, Lins, and Miller [2003], Herrmann, Hope and Thomas [2008]). The variable BIAS is the difference between the consensus earnings forecast and the actual earnings, scaled by the stock price.

1

where FORECAST_t is the consensus earnings forecasts of period t. Following Herrmann, Hope, and Thomas [2008], we use the data from the four months after the end of the fiscal year t – 1 to calculate consensus earnings forecasts of fiscal year t. ACTUAL_t is the actual earnings per share before extraordinary items of fiscal year t. PRICE _t–1 is the stock price at the time of t – 1. A Positive (negative) BIAS indicates analyst optimism (pessimism).

We measure forecast accuracy based on absolute value of the difference between the firms’ forecasted and actual earnings, divided by stock price at time t:

2

where ACCURACY_t is the negative of the absolute forecast error at time t. Multiplying the absolute forecast error by (-1) gives a measure that increases with greater forecast accuracy. The ACCURACY values are negative by construction (see Lang, Lins, and Miller [2003]). The closer the accuracy value is to zero, the more accurate the forecast.

We measure forecast dispersion as standard deviation earnings forecast scaled by the stock price:

3

We also perform the following regression models to test our hypotheses:

Model A:

4

Model B:

5

where ADDITION equals 1 if the firm is added into the S&P 500 and zero otherwise. We measure firm size (SIZE) as the natural log of market value assets at the end of year t – 1. Price deflated seasonal earning change (UE) is the difference in earnings per share before extraordinary items between fiscal year t to year t – 1, scaled by the stock price at the beginning of year t. Loss equal to 1 for observations with negative earnings, zero otherwise.

Our primary variable of interest is the dummy variable, ADDITION. Our hypothesis predicts ß ₀ to be positive for Model A and to be negative for Model B, indicating analysts issue more optimistic and less accurate forecasts after the firm is added into the S&P 500.

RESULTS

Panel A of Exhibit 1 provides the summary statistics for the properties of analysts’ earnings forecasts. The time period of investigation includes the three years prior to the year the firms were added into the S&P 500 and the three years after the inclusion. The mean and median ACCURACY values are negative by construction. The closer the accuracy value is to zero, the more accurate the forecast. The mean and median signed forecast errors (BIAS) are positive, indicating that in our sample analysts were, on average, optimistic, consistent with Easterwood and Nutt [1999], Abarbanell and Lehavy [2003], Duru and Reeb [2002], and Herrmann, Hope, and Thomas [2008]. The mean (median) number of analysts is 14.83 (14.00), indicating a relatively high number of analysts following for the sample firms. Panel B of Exhibit 1 presents the correlation matrix of the forecast variables in our regression analysis (Pearson below the diagonal and Spearman above). We find a positive relationship between ADDITION and BIAS, indicating that analysts tend to issue more optimistic forecasts in the post-inclusion period.

Exhibit 2 provides the results of our comparison tests. We include seven years surrounding the announcement date of S&P 500 additions: three years before inclusion year, inclusion year, and three years after inclusion year. In our tests, we compare analyst coverage and properties of analyst forecast estimates before and after an index addition event. Consistent with prior research, we find that that once a firm enters the index, the analyst coverage increases significantly. The mean (median) number of analysts following an added stock increases from 14.20 (14.00) in the pre-inclusion period to 17.38 (17.00) in the post-inclusion period.

Furthermore, Exhibit 2 presents the properties of analysts’ earnings forecasts surrounding Index additions. We find that analysts are more optimistic following after Index inclusion announcements. The mean (median) of BIAS increases from -0.0013 (-0.0005) in the pre-inclusion period to 0.0033 (0.0000) in the post-inclusion period. The change in forecast bias is significant at the 5% level. Our results also indicate that analyst forecast dispersion is significantly higher after a firm was added into the S&P 500. The mean (median) analyst forecast dispersion increases from 0.0013 (0.0005) in the pre-inclusion period to 0.0022 (0.0010) in the post-inclusion period. The change in forecast bias is significant at the 5% level. The results show that analysts are in disagreement with respect to the future earnings prospects of a newly added firm.

Exhibit 3, Panel A, presents histograms of the proportion of forecasts according to whether forecast bias is positive (optimism), zero, or negative (pessimism) before the inclusions. Exhibit 3, Panel B, presents histograms of the same variables after the firms were added into the S&P 500. Exhibit 3 shows that the percentage of analysts issuing optimistic (pessimistic) forecasts increases (decreases) from 30% (68%) to 49% (48%). This finding supports our hypothesis that analysts are more likely to issue overoptimistic earnings forecasts for the added firms.

• Download figure
• Open in new tab
• Download powerpoint

Exhibit 3

Frequencies of Optimistic, Neutral, and Pessimistic Forecasts for Companies Added to the S&P 500 Index

Notes: Panel A presents histograms of the proportion of forecasts according to whether forecast bias is positive (optimism), zero, or negative (pessimism) before the firms were added into S&P 500. Panel B presents histograms of the proportion of forecasts according to whether forecast bias is positive (optimism), zero, or negative (pessimism) after the firms were added into S&P 500.

Panel A of Exhibit 4 provides the results based on Model A, specified in Equation (4). We find that the coefficient on ADDITION is significantly positive, indicating that analysts are more optimistic in the post-inclusion period. Panel B of Exhibit 4 presents the results based on Model B, specified in Equation (5). The coefficient ADDITION is significantly negative, indicating that analysts are relatively less accurate in the post-inclusion period.

SUMMARY AND CONCLUSIONS

Earlier studies have found that when a stock is added to the S&P 500 Index, it receives higher EPS estimates and reports significantly higher actual (realized) earnings. This article further studies how analysts respond to S&P 500 inclusion announcements. We look at the properties of analysts’ earnings estimates, surrounding the announcement, to understand the behavior of analysts. Consistent with Denis et al. [2003], we show that analysts tend to initiate coverage following the announcement. We also find that analysts overreact to S&P 500 addition announcements, because their earnings forecasts exhibit optimism in our sample of newly added companies for the post-inclusion period. We discover lower analyst forecast accuracy and higher inter-analyst forecast disagreement in the post-inclusion period. Our results suggest that analysts are not in agreement as to the future prospects of the added firm. We contend that analyst optimism in the added firms may be associated with analyst incentive bias arising from increased firm visibility (status), investment banking relationships, and new information.

This study provides novel results about the behavior of analysts following companies added to the S&P 500. Although index additions are generally associated with favorable earnings information, our investigation shows that investors should be careful when they make investment decisions based on analysts’ earnings forecasts. The dispersed earnings forecasts of these bellwether stocks tend to be overly optimistic and are less accurate. Finally, our results are consistent with previous research that S&P 500 addition is not an information-free event.