Regression analysis of stock price effects of index additions
We now extend our analysis to an additional implication of index addition. Amihud and Mendelson (1986) argue that as stock liquidity improves, the liquidity premium in stock returns will decrease and thus lower the cost of capital. In addition, Merton (1987) presents theoretical arguments that investor recognition, or the shadow cost of incomplete information, is a priced factor in required returns. Becker-Blease and Paul (2006) test implications of the liquidity hypothesis for a sample of S&P 500 index additions and find that post-addition changes in investment opportunities are positively related to changes in stock liquidity.
They argue that this evidence is consistent with a decline in cost of capital as liquidity increases, but they do not explore whether similar effects on capital investments are present for changes in investor awareness following index addition. Similar to Becker-Blease and Paul (2006) and Lin and Smith (2007), we use capital expenditures as a proxy for investment opportunities and examine the impact of changes in both liquidity and investor recognition on change in capital investments. For each sample firm, we compute the industry-adjusted change in capital expenditures based on the median change within the same Fama-French (1997) industry during a given period. The change in industry-adjusted capital expenditures is measured from the last full fiscal year preceding addition to the first full fiscal year following addition.
As shown in Figure 1, the index fund holdings of S&P 1000 firms increased significantly during our sample period. Thus, it is possible that the effects we find are clustered in a specific period. We investigate this possibility by forming a dummy variable that equals one for firms added from 1996 to 1999 (early sample) and zero for those added from 2000 to 2003 (late sample). We then repeat the analysis in Table 4 but add interaction terms for the early sample dummy on the investor recognition, liquidity, and operating performance proxy variables. The analysis (not tabulated) shows that the effect of liquidity on permanent price effects is not constrained to either the early or late period. In contrast, changes in shadow cost appear to be impounded into announcement returns primarily in the early period. This suggests that index fund demand is not a significant determinant of whether index inclusion effects are capitalized into stock returns.
that permanent price effects are increasing in investor recognition. In contrast, analyst coverage and EPS forecasts have insignificant coefficients. Thus, we find that improved stock liquidity and better investor recognition are sources of value, but changes in analyst coverage and anticipated improvements in operating performance are not. We next investigate whether the effects of liquidity and investor recognition differ by index. Column reports estimates for the 325 SmallCap additions, column reports estimates for the 270 MidCap additions, and column reports estimates for the combined sample with interactions to test differences between the subsamples in coefficients on the proxy variables of interest.
The results are similar to column, indicating that improvement in liquidity and investor recognition (shadow cost) is an important source of value gains for both MidCap and SmallCap firms. These results are consistent with evidence in Chordia (2002) and Hegde and McDermott (2003) for the relation between stock liquidity and index returns but inconsistent with Elliott, Van Ness, Walker, and Warr (2006) who find that investor recognition is the only significant source of value for additions to the S&P 500. We do note, however, that Hegde and McDermott (2003) and Chordia (2002) do not include shareholder recognition measures in their regression models. Because S&P 500 firms are very liquid even before addition to the index (e.g., Becker-Blease and Paul, 2006), it is likely that the direct impact of improved liquidity on long-term abnormal returns is subsumed by the improvement in investor recognition. In contrast,the smaller capitalization Mid- and SmallCap firms enjoy greater direct benefit from
the improved liquidity.
Abnormal stock returns around index changes
We investigate whether the changes in investor recognition, stock liquidity, and earnings expectations explain permanent index price effects by estimating OLS regressions of returns on raw changes in the proxy variables.6 The dependent variable is the abnormal stock return from announcement day to effective day plus 60 (reported in column of Table 2). We include firm size as a control variable because it is possible that firm size varies directly with investor recognition effects that are not captured in our proxy variables. In addition, there is evidence that growth opportunities increase following index addition (Becker-Blease and Paul, 2006). Thus, we include market-to-book value of equity to control for the effect of pre-existing growth opportunities on investor perception of the economic benefit of index inclusion. The remaining control variables consist of dummy variables for trading venue, index capitalization (S&P 400 dummy), and whether a MidCap firm is new to the S&P. The trading venue dummy is included to control for any residual impact of trading venue on liquidity that is not captured by our proxy variable for liquidity, and the subindex dummies are included for tests of our hypothesis of differences in index effects between Small- and MidCap firms.
For consistency with a recent study of S&P 500 index effects by Elliott, Van Ness, Walker, and Warr (2006), we also include the measure of arbitrage risk (A1) from Wurgler and Zhuravskaya (2002) and the measure of index fund demand, PctShock. A1 is the variance of the residuals of the stock’s excess return on the market’s excess return over the T-bill rate and is computed using returns for the 250 days preceding the AD. The S&P 1000, S&P 600, or S&P 400 value-weighted index is used as a proxy for the market return, depending on the sample used in the analysis. PctShock is the percentage of the firm’s shares estimated to be held by index funds, using datam provided by S&P for yearly percentage of shares held by indexers. Results are reported in Table 6. Column (1) contains coefficient estimates for the combined samples with available data. Consistent with prior research, we find a significantly positive relation between stock liquidity and index announcement returns. We also find a significant negative coefficient on shadow cost. A lower value for shadow cost indicates better investor recognition, so this result indicates
forecasts. Consistent with this argument, they find that the post-addition revision in analysts’ earnings expectations is higher for index additions compared to benchmark firms in the IBES universe for firms added to the large capitalization S&P 500 index. They suggest that the anticipated improvement in performance stems from increased monitoring as the added firm’s visibility improves, although it is also consistent with the signaling or certification hypothesis. Following the approach in Denis, McConnell, Ovtchinnikov, and Yu (2003), which is also used in Elliott, Van Ness, Walker, and Warr (2006), we measure the change in current-year median earnings forecasts following index listing, relative to the same measure for the universe of IBES firms. The change in the forecast is standardized by actual earnings per share (EPS). Table 5 contains the pre-addition forecast, raw changes, and the percentage change in the current-year forecast. The results show that firms added to the S&P MidCap index have higher median postaddition earnings expectations relative to their benchmark. Thus, for the MidCap
The second hypothesized source of value in index addition is improvement in the trading environment of the firm’s stock. For the S&P 500, Hegde and McDermott (2003) show that improvements in stock liquidity are impounded into abnormal announcement returns, and Becker-Blease and Paul (2006) show that the improved stock liquidity is associated with better investment opportunities. Thus, for the smaller capitalization indexes examined in our study, we anticipate improvements in liquidity that will be positively related to the stock price revisions around index addition. We use CRSP data to construct three proxies for stock liquidity: the illiquidity ratio, dollar volume, and share turnover. The illiquidity ratio is first used in Amihud (2002) as a proxy for the price impact of trade and is measured as the average of the ratio of daily absolute return to the daily volume in dollars
Sources of stock price effects of index additions
In this section, we present univariate changes in stock liquidity, investor recognition, and earnings expectations following index addition. We then test whether the changes are impounded into the permanent price effects reported in column (6) of Table 2 by estimating ordinary least squares (OLS) regressions. All changes in liquidity and investor recognition are index-adjusted by subtracting the median change in the relevant S&P index during the event period. Earnings expectations are measured relative to the universe of Institutional Brokers Estimate System (IBES) firms.
We calculate abnormal returns based on market-adjusted returns, using the S&P 400 (S&P 600) value-weighted index as the market portfolio for estimates of abnormal returns for firms added to the S&P 400 (S&P 600). For estimates of abnormal returns for the combined samples (all additions), we use the value-weighted return on the combined index. We elect to employ market-adjusted returns rather than parameterbased returns from the market model or three-factor model based on evidence in Edmister, Graham, and Pirie (1994) that the estimation window of parameters surrounding index additions significantly influences estimated abnormal returns. We define the AD as the first date that S&P reports the pending index addition, regardless of the time of day that the announcement occurs. The effective date (ED) is the date
Merton (1987) contends that if investors are aware of only a subset of securities and thus are incompletely diversified, they will demand a premium (shadow cost) for the unique risk in their portfolio. Chen, Noronha, and Singal (2004) argue that if the awareness of a security increases due to its addition to a prominent index, this should reduce the shadow cost and required return for that security. This leads to our first testable hypothesis, the Investor Recognition Hypothesis:Amihud and Mendelson (1986) argue that investors impound anticipated trading costs into the price of an asset in the form of a liquidity premium. If the demand increases for securities added to prominent indexes, the result should be lower search costs and therefore lower liquidity premiums.
This will lead to higher valuations of assets in place, as argued by Amihud, and Mendelson (1986), and leads to our second hypothesis, the Liquidity Premium Hypothesis: Jain (1987) argues that addition to an S&P index is not an information-free event because the opinions of the S&P selection committee are likely formed by both private and public information. In this sense, index inclusion signals future performance improvements that were previously unknown to the investing public. Denis, McConnell, Ovtchinnikov, and Yu (2003) find that analysts’ earnings expectations increase relative to benchmark firms following index addition, lending support to the signaling or certification hypothesis. We note, however, that Denis, McConnell, Ovtchinnikov, and Yu (2003) suggest that the improved performance could also stem from increased managerial incentives as firms operate in an environment of greater implicit monitoring following index addition. Our testable hypothesis concerning
performance is as follows.
Both the Liquidity Premium and Investor Recognition Hypotheses imply a reduction in the cost of capital for a firm added to a prominent index. Although cost of capital is not directly observable, corporate actions associated with a decline in cost of capital can be directly observable. Becker-Blease and Paul (2006) argue that if cost of capital declines, it will expand the set of viable investment opportunities and should lead to increased capital expenditures. Although their paper focuses only on liquidity as a source of a lower cost of capital, the argument is also true for investor recognition. This leads to the fourth testable hypothesis, the Investment Opportunities Hypothesis:
Author:John R. Becker-Blease Donna L. Paul