The valuation effects of equity issues and the level of institutional ownership: evidence from analysts' earnings forecast.
This paper investigates whether the level of institutional ownership has any effect on the market reaction to announcements of a firm-level "event", namely the issuance of equity. We contribute to the literature on the relation between firm value and institutional ownership in two ways. First, previous studies that have examined the relationship between some proxy for firm value (either Tobin's q or a measure of accounting profitability) with ownership structure have a problem with the direction of causality. For instance, McConnell and Servaes |21^ document a positive relation between firm value (as proxied by Tobin's q) and institutional ownership. They interpret this result as evidence of monitoring by institutions. This result, however, can either be interpreted as indicative of the monitoring role played by institutions or that institutions tend to invest in high-q firms. If we find any relation between the announcement effects of equity issues, and therefore firm value, and institutional ownership, the direction of causality has to run from institutional ownership to firm value, not the other way around.
Second, some earlier studies have examined the relation between firm value and ownership structure of the firm by focusing primarily on the stock price reaction to announcements of specific corporate decisions. We also examine revisions in analysts' earnings forecasts around these announcements. If the announcement of the decision to finance investment projects through the issuance of equity has an effect on the stock price, it is presumably because the market perceives changes in either the expectations of future earnings or in the variability of these earnings. We, therefore, also study the relation between revisions in analysts' earnings forecasts and institutional ownership to corroborate our stock price results.
Jensen |13^ argues that the proceeds from an equity issue give more discretionary cash to managers which increases the likelihood of non-value maximizing behavior by them. Such behavior might explain the negative stock price reaction to an equity issue. We hypothesize that under the effective-monitoring hypothesis, higher institutional ownership will give institutional investors greater incentives to protect their investments in the firm's equity. They achieve this objective by carefully monitoring the use of the proceeds of the equity issue in order to ensure that the capital is used for productive purposes. The effective-monitoring hypothesis predicts a positive relation between announcement-period abnormal stock returns and the level of institutional ownership.(2) On the other hand, with higher institutional ownership, institutions may develop other profitable business relationships with the firm. They may also find it mutually advantageous to cooperate with the managers of the firm, thereby reducing the incentives to monitor the activities of the managers aggressively in order not to jeopardize these other beneficial relationships. We call this the ineffective-monitoring hypothesis. It predicts a negative relation between abnormal stock returns and institutional ownership.(3)
We first study the association between the announcement-period abnormal stock price reaction and institutional ownership. We find a significant positive relation between the announcement-period abnormal stock returns and institutional ownership. This result suggests that higher levels of institutional ownership are associated with perceptions of institutions as more effective monitors of the uses of the cash obtained from the equity issue due to their higher ownership stake in the firm. It is, however, entirely conceivable that the positive relation between the stock price reaction and institutional ownership documented in this paper may be an artifact of higher levels of institutional ownership. A higher level of institutional ownership may be associated with reduced pre-announcement asymmetry of information due to institutions either (i) systematically investing in firms that make more information available or (ii) causing more information to be generated regarding the issuing firm as a consequence of their investing. Reduced information asymmetry would result in the release of less adverse information at the time of the equity issue announcement. Alternatively, higher institutional ownership may be associated with a less negative slope of the demand curve for the stock, which would necessitate a smaller price discount to sell additional stock.
To disentangle these three possible explanations, we also examine analysts' forecasts of current-year earnings as well as five-year earnings growth around announcements of equity issues. If the only avenue by which institutional ownership affects the stock price reaction to the equity issue is through an association with the liquidity of the stock, then institutional ownership will only influence the price discount necessary to sell additional equity but will not alter analysts' forecasts of either current-year earnings or five-year earnings growth. Thus, under this liquidity explanation, we should find no relation between analysts' earnings forecast revisions and institutional ownership. If the stock price reaction is due to an association between institutional ownership and the pre-announcement asymmetry of information regarding cash flows, then institutional ownership will influence the information released by the announcement of the equity issue regarding all earnings. Specifically, if higher institutional ownership is associated with lower pre-announcement informational asymmetry, then the information asymmetry argument will predict a positive relation between abnormal forecast revisions in both current-year earnings and five-year earnings growth. Given that the cash raised through the equity issue is typically used to finance long-term projects, monitoring the uses of the proceeds of the equity issue by institutions will not impact current-year earnings but will have an effect on long-term earnings. Hence, if the documented stock price reaction is the result of effective monitoring of the use of the proceeds of the cash raised through the equity issue, there should be no relation between abnormal forecast revisions in current-year earnings and institutional ownership, but there should be a positive relation between abnormal forecast revisions in five-year earnings growth and institutional ownership.(4)
We find no relation between analysts' abnormal forecast revisions in current-year earnings with institutional ownership, but we find a significant positive relation between analysts' abnormal forecast revisions in five-year earnings growth and institutional ownership. These results are inconsistent with higher levels of institutional ownership being associated with either lower pre-announcement asymmetry of information regarding future cash flows or a less negative slope of the demand curve for the stock as plausible explanations for our stock price results by themselves. These findings, however, do support the view that institutions are effective monitors of the equity issuance decision.
Finally, we find that the significant positive relation between the announcement-period abnormal returns, as well as abnormal forecast revisions in five-year earnings growth, and institutional ownership is primarily explained by low-q (q |is less than^ 1) firms. Managers of low-q firms are more likely to misuse the proceeds from the equity offering. Hence they are the targets of closer scrutiny by institutional investors with higher ownership stakes who wish to protect their investments in the firm.(5) This result gives further credence to the effective-monitoring hypothesis.
In the popular financial press, numerous recent articles relay a common theme that institutions are starting to "attack" corporate management collectively.(6) We believe that the remarkable change in corporate ownership structure, mainly through a huge increase in the proportion of stock owned by institutional investors, and the removal of regulations that hamper the ability of institutions to monitor even more effectively will result in these investors becoming increasingly active participants in the important decisions taken by firms.(7) By studying the impact of the decision by firms' managers to issue equity on both stock prices and analysts' earnings forecasts, we find evidence supportive of this monitoring role.
The paper is organized as follows. Section I provides a description of the data. Section II describes our techniques for estimating abnormal stock return performance and abnormal earnings forecast revisions. Section III presents the results regarding the association between abnormal stock price reaction and institutional ownership. Section IV provides similar results for abnormal earnings forecast revisions. Section V replicates our earlier results for two subsamples classified by Tobin's q. Section VI provides a summary and the main conclusions of the paper.
I. Description of the Data
We identified an initial sample of seasoned equity issues from the semiannual editions of the Investment Dealer's Digest for the period January 1976 through December 1985. We obtained the announcement date for each equity offering from the Wall Street Journal Index and the weekly editions of the Investment Dealer's Digest. The announcement date used in this study is either the date the published report concerning the common stock offering appeared in the Wall Street Journal or the date of the registration of the offering as reported in the Investments Dealer's Digest, whichever comes first. Once we identified our initial sample of equity issues, we imposed the following additional data selection criteria.
First, we deleted announcements which were contaminated by other contemporaneous firm-specific events. In addition, we excluded common stock offerings that were made for a specific purpose, such as mergers, reorganizations, conversions, acquisitions, or exchange offers. For such offerings, the announcement of the specific purpose contains information about future earnings, thereby making it difficult to determine whether the change in expectations is attributable solely to the common stock offering. We also eliminated combination debt/equity offerings. Second, we restricted our sample to firms trading on either the New York or American Stock Exchanges and for which sufficient data were available on the CRSP stock returns tape.
Third, we stipulated that ownership information should be available for the firms in our sample. We are primarily interested in share ownership by institutional investors. Information concerning the percentage of shares owned by institutions was obtained from issues of the Standard and Poor's Common Stock Owners' Guide. There is a large body of theoretical and empirical literature in financial economics dealing with the effect of equity ownership by corporate insiders on the actions taken by them regarding the firm and, as a consequence, on firm value. The theoretical predictions, as well as the documented empirical relation between firm value and insider ownership, vary widely.(8) For completeness, we control for the percentage of shares owned by corporate insiders. We obtained the percentage of shares owned by insiders from issues of the Value Line Investment Survey. From our two sources, we collected information on ownership structure immediately prior to the announcement date for the equity issue.
Finally, we collected analysts' forecasts for both the current-fiscal-year's earnings and the five-year growth in earnings from Lynch, Jones, and Ryan's I/B/E/S database. The forecasts for the current fiscal year were available from January 1976 through December 1985 while forecasts for the five-year growth in earnings were only available from December 1981 onwards.
After imposing all the above restrictions, we identified a final sample (non-I/B/E/S matched sample) consisting of 379 announcements of primary common stock offerings. Exhibit I presents the distribution of the sample across years as well as information regarding the characteristics of the sample. Approximately 32% of the offerings included in the final sample occurred in 1983. The mean (median) number of shares issued is 2.62 (1.25) million shares whereas the mean (median) number of shares outstanding is 18.31 (9.91) million shares. The mean (median) market value of equity is 554.92 (250.31) million dollars. The mean (median) fraction of the shares owned by institutions is 30.40 (28.34)% and by insiders is 13.79 (7.00)%. The mean (median) number of institutional investors is 83.52 (52.00). The mean (median) stock price run-up for our sample of issuing firms is 23.13 (19.52)%.(9) Finally, the mean (median) Tobin's q is 0.99 (0.91). The characteristics of this sample are comparable with those reported in previous studies on equity offerings.(10)
This section describes the methodology we employed to measure abnormal stock price performance and to estimate abnormal earnings forecast revisions.
Exhibit 1. Sample Summary Statistics The full sample consists of 379 equity offerings by firms listed on the New York or American Stock Exchanges over the period January 1976 through December 1985. Excluded from the sample are offerings made by utilities, dual security offerings, and offerings made for a specific purpose, such as mergers, acquisitions, reorganization, conversions, or security exchanges. Run-up is the cumulative excess return over the CRSP value-weighted dally returns index over the event period -250 to -2 days prior to the equity issue announcement. PANEL A: Number of Issues Per Year Number of Offerings for Number of which Current Offerings for which Year Earnings Five-Year Earnings Full Forecasts Are Forecasts Are Year Sample Available Available 1976 31 5 -- 1977 4 2 -- 1978 18 10 -- 1979 21 12 -- 1980 42 35 -- 1981 37 20 -- 1982 43 29 25 1983 122 83 75 1984 21 13 12 1985 40 32 30 Total 379 241 142 PANEL B: Characteristics of the Full Sample of Equity Issues Description Measure Mean Median Number of Shares Issued (millions) 2.62 1.25 Number of Shares Outstanding (millions) 18.31 9.91 Market Value of Equity ($ millions) 554.92 250.31 Percentage Owned by Insiders (%) 13.79 7.00 Percentage Owned by Institutions (%) 30.40 28.34 Number of Institutional Investors 83.52 52.00 Run-Up (%) 23.13 19.52 Tobin's q 0.99 0.912
A. Estimation of Abnormal Stock Price Performance
The magnitude of stock price reactions to announcements of equity issues is measured using the residual analysis technique based upon the market model. The procedure employed in this paper is a standard "event-time" methodology. In the test, the first day on which the announcement of the equity issue appeared in the Wall Street Journal is numbered event day t = 0. Since it is not possible to determine whether public announcements come before or after the close of trading on t = -1, the impact of the announcement is measured over the two-day announcement period consisting of days t = -1 and t = 0. The security-specific parameters of the market model are estimated using 120 daily returns beginning with event day t = -180 and ending with event day t = -61.(11) The test statistics used are consistent with those reported in the literature.
B. Estimating Abnormal Earnings Forecast Revisions
I/B/E/S collects forecasts from analysts employed at over 100 brokerage firms who collectively cover more than 4,000 firms listed on the New York or American Stock Exchanges. The actual number of analysts' forecasts collected for each firm ranges from one to fifty-three analysts (the average being approximately 7.5 analysts following a firm). Specifically, I/B/E/S reports mean forecasts of earnings per share for individual firms on a monthly basis.
We used these monthly consensus forecasts to calculate the observed monthly earnings forecast revision for firm i over the announcement month t as follows:
|FR.sub.i,t^ = |F.sub.i,t^ - |F.sub.i,t-1^/|P*.sub.i^, (1)
where |F.sub.i,t^ is the mean of analysts' earnings forecasts for firm i at month t, |F.sub.i,t-1^ is the mean of analysts' earnings forecasts for firm i at month t-1, and |P*.sub.i^ is the price of a share of firm i's common stock the month before the equity issue announcement.(12)
To measure the abnormal earnings forecast revision at the announcement of an equity issue, we use a simple expectations model for expected forecast revision whose assumptions reflect the following salient time-series features of the earnings forecast data: (i) forecast revisions are subject to a bias that varies across firms and is not related in any systematic manner to the month during which a fiscal year or quarter ends, and (ii) the mean monthly percent of analysts revising their earnings forecasts is approximately 20% and again does not seem to be influenced by months during which a fiscal year or quarter ends.(13)
Since approximately 20% of the analysts revise their forecasts in any single month, there will typically be four months between individual analysts' updates. Forecast revisions in any given month, therefore, will typically be correlated with forecast revisions in the previous four months. Hence, we assume forecast revisions (|FR.sub.i,t^) follow a fourth-order moving average process and the a priori expected mean forecast revision for finn i in month t (E||FR.sub.i,t^^) is then estimated as:
E ||FR.sub.i,t^^ = |k.sub.i^ + 1/n |summation of^ ||Epsilon^.sub.i,t-s^ where s=1 to n-1. (2)
The forecastable component (|k.sub.i^), which is a measure of the bias for firm i, is estimated for each firm as the average forecast revision during an estimation period which consists of all months for which forecasts are available, excluding months -6 to +6. The unexpected component (||Epsilon^.sub.i,t-s^) is measured as the difference between the actual forecast revision in month t and |k.sub.i^. The expected forecast revision for firm i in month t, therefore, equals the forecasted component (|k.sub.i^) plus the weighted average of the four previous months' unexpected component, where the weights equal 0.20 (n=5).
The ex post abnormal, or unexpected, forecast revision for firm i in month t (|AFR.sub.i,t^) can be expressed as:
|AFR.sub.i,t^ = |FR.sub.i,t^ - E (|FR.sub.i,t-1^). (3)
Brous and Kini |3^ show that the above expectations model effectively eliminates the bias associated with raw forecast revisions. Appendix A provides details concerning the development of the expectations model.
III. Stock Price Reaction Results
The stock price reaction to announcements of primary equity issues is documented in Exhibit 2, Panel A. For the full sample of 379 issues, the two-day announcement-period abnormal return (TDAR) is -2.88%, which is significant at the one percent level. Only 20% of the issues have a positive stock price reaction, which is also significant at the one percent level. These results are consistent with those reported in the literature.(14)
Exhibit 2, Panel B presents results for the TDAR for three groups classified according to the percentage of equity held by institutions.(15) The results clearly indicate that the higher the institutional ownership, the less negative the TDAR. The TDARs for the low and high institutional ownership groups are -3.61% and -2.12%, respectively. The difference in the TDAR between these extreme groups is 1.49%, which is significant at the one percent level. These results are also borne out by looking at the difference in the percentage of positive two-day abnormal returns for these two groups.
TABULAR DATA OMITTED
Next, in order to control for other possible effects, we estimated the cross-sectional relation between TDAR and institutional ownership by estimating weighted least squares (WLS) regressions of the form:
|TDAR.sub.i^ = Control Variables + ||Gamma^.sub.1^ Institutional |Ownership.sub.i^ + ||Epsilon^.sub.i^ (4)
where control variables include the proportion of shares owned by corporate insiders, stock price run-up(16), firm size (proxied by the market value of equity)(17), and the proportion of new shares issued.(18) The results are reported in Exhibit 3. The coefficients associated with institutional ownership are positive and significant in all the equations; the coefficients associated with all the control variables, with the exception of run-up, are insignificant.(19) For example, in equation (2), the estimated coefficient for institutional ownership is 0.0245, which is significant at the five percent level. This coefficient for institutional ownership indicates that a 10% increase in institutional ownership will result, on average, in abnormal stock returns (and therefore finn value) increasing by 0.245%. The results reported in Exhibits 2 and 3 documenting a positive relation between TDAR and the level of institutional ownership are consistent with the effective-monitoring hypothesis. These results, however, can also be the artifact of higher levels of institutional ownership either being associated with lower levels of pre-announcement asymmetry of information or a more elastic demand curve for the stock. In the next section, we study analysts' forecast revisions of earnings around equity issue announcements to disentangle these plausible explanations for the significant positive relation between TDAR and institutional ownership.
IV. Analysts' Earnings Forecast Revisions and Institutional Ownership
This section provides empirical results regarding the association between abnormal earnings forecast revisions and institutional ownership.
TABULAR DATA OMITTED
A. Analysts' Forecasts of Current-Fiscal-Year-End Earnings Per Share
For the sample of 241 equity offerings for which analysts' forecasts of current-fiscal-year earnings are available, results are reported in Exhibit 4, Panel A. They indicate that the abnormal forecast revision (AFR) is -0.0026; the t-value is significant at the one percent level. If we assume an average price/earnings ratio of 15 for the issuing firms in our sample, these results tell us that announcements of equity issues are associated with an abnormal decrease in earnings per share of -0.0026x15x100, or -3.9%. We interpret our results as evidence supportive of the notion that after firms announce equity offerings analysts lower their forecasts of current-fiscal-year earnings more than expected. These results are consistent with those reported in Brous |2^, Hansen and Crutchley |8^, and Jain |12^; but are inconsistent with Healy and Palepu |9^. The significant negative abnormal forecast revision in current-year earnings is consistent with the cash flow signaling theories (see Miller and Rock |24^ and Myers and Majluf |26^), but it does not rule out the free-cash-flow hypothesis (see Jensen |13^).
When the sample is classified into groups based on institutional ownership, there seems to be no difference in the abnormal forecast revisions of earnings across groups. Cross-sectional regressions of abnormal earnings forecast revisions with the level of institutional ownership are reported in Exhibit 5. In addition to the control variables used in the previous section, we also included the announcement-period price reaction (TDAR) as an additional control variable in the event that some analysts use this information to revise their earnings forecasts.(20) We find no relation between AFR and institutional ownership. We do, however, find a significant positive relation between AFR and TDAR. A similar result was documented earlier by both Brous |2^ and Jain |12^, who interpret it to indicate that an equity issue announcement does convey information regarding the firms' earnings. The coefficient associated with the run-up variable is positive but not consistently significant.
The lack of an association between the current-year abnormal forecast revisions in earnings and institutional ownership in conjunction with the significant negative reaction for the full sample indicates that while there is unfavorable information being conveyed by the equity issue regarding earnings, it is information which is not in any way related to the level of institutional investors. This result is inconsistent with higher levels of institutional ownership being related to lower pre-announcement information asymmetry, as that would imply a positive relation between both analysts' forecast revisions in the short- and long-term with institutional ownership. The evidence presented so far is consistent with the effective-monitoring hypothesis because it predicts no relation between the abnormal forecast TABULAR DATA OMITTED TABULAR DATA OMITTED revisions in current-year earnings with institutional ownership. The reason for this prediction is that the proceeds from the equity issue are typically used for long-term projects, which will have an effect on long-term earnings. Monitoring by institutions will then influence long-term earnings only. This result is also consistent with institutional ownership being associated with a more elastic demand curve for the stock as an explanation for the previously documented stock price results. Under this explanation, we should find no relation between analysts' forecast revisions of short- and long-term earnings with institutional ownership as this argument is not earnings-related.
B. Analysts' Forecasts of Five-Year Growth in Earnings
For the sample of 142 equity issues for which five-year growth in earnings is available on the I/B/E/S tapes, the average announcement period abnormal five-year growth in earnings forecast revision (LTAFR) is -0.0406 (-4.06%), which is significant at the one percent level (see Exhibit 6, Panel A).
For groups formed for analysis according to institutional ownership, the results are reported in Exhibit 6, Panel B. The average abnormal forecast revision in five-year earnings growth is -0.0736 (-7.36%) for the low-institutional-ownership group and is -0.0172 (-1.72%) for the high-institutional-ownership group. The difference in abnormal forecast revisions in five-year earnings growth between these two extreme institutional ownership groups is 0.0564 (5.64%), which is significant at the five percent level. We obtain further confirmatory evidence by studying the TABULAR DATA OMITTED proportion of positive abnormal forecast revisions. For the low and high groups, it is 0.17 and 0.38, respectively. The difference is 0.21, which is significant at the five percent level. The results reported in Exhibit 6 indicate that with higher institutional ownership, equity issues convey less unfavorable information to analysts regarding the five-year growth in earnings and vice-versa. This finding is consistent with the notion that higher institutional ownership implies closer monitoring and, as a consequence, less non-value-maximizing behavior by managers.
Exhibit 7 reports the cross-sectional regression results for abnormal forecast revisions in five-year earnings growth regressed on institutional ownership and other control variables. There is a positive and significant relation between abnormal forecast revisions in the five-year earnings growth forecast and the level of institutional ownership in all the regressions estimated. For example, in equation (2), the coefficient associated with institutional ownership is 0.0020, which is significant at the one percent level. This coefficient tells us that a 10% increase in institutional ownership is associated, on average, with a 0.020 (or a 2.00%) abnormal increase in forecasts of the five-year growth rate in earnings per share (or an annualized rate of 0.40% per year).
In general, there is a significant and positive relation between abnormal forecast revisions in the five-year earnings growth forecast and TDAR. The positive relation between LTAFR and TDAR indicates that equity issue announcements convey information regarding long-term earnings also. Overall, our results relating the level of institutional ownership to analysts' abnormal forecast revisions provide further evidence of the effective TABULAR DATA OMITTED monitoring role played by institutions.(21) Finally, the significant positive relation between the abnormal forecast revisions in five-year earnings growth and institutional ownership is inconsistent with the liquidity-related explanation.
V. Results by Firm Type
Jensen |13^ argues that "except for firms with profitable unfunded investment projects, prices will rise with unexpected increase in payouts to shareholders (or promises to do so), and prices will fall with reduction in payments or new requests for funds." He cites the much-documented negative stock price reaction to leverage-reducing transactions (equity issues, for one) as evidence consistent with managers having more free cash and, as a result, greater incentives to invest in projects that pay below their cost of capital or waste it on organizational inefficiencies. Jensen further argues that while all firms with substantial free cash flows have a tendency to overinvest, this tendency is less severe for firms with substantial growth opportunities.
We use Tobin's q as a proxy for the growth opportunities available to a firm.(22) The q ratio is defined as the market value of the finn divided by the replacement cost of its assets. This measure is typically thought of as a profitability index for the firm's investment opportunities. Firms with profitable investment opportunities (high-growth firms) will have high q ratios, and firms with unprofitable investment opportunities (low-growth firms) will have low q ratios. Lang and Litzenberger |16^ demonstrate that an average q ratio greater than one is a necessary condition for a firm to be at the value-maximizing level of investment, and an average q less than one is the sufficient condition for a firm to be overinvesting. It is more likely that managers of firms with low q (q |is less than^ 1) ratios will misuse the proceeds from the equity issue. Hence, these firms are likely to be subject to closer monitoring by institutions, the intensity of which should increase monotonically with the size of their ownership stake in the firm.
To test this conjecture, we classify our sample into two groups based on their q ratios. The first group consists of firms with low q (q |is less than^ 1) ratios. These firms are the low-growth or overinvesting finns. The second group consists of firms with high q (q |is greater than^ 1) ratios. These firms are the high-growth firms. For each of these two groups separately, we estimated weighted least squares regressions between TDAR, AFR and LTAFR with institutional ownership. The results are reported in Exhibit 8. We find a significant positive relation between both TDAR and LTAFR with institutional ownership for the low-q group only. This result indicates that the full sample results documented earlier are explained primarily by low-q firms. These results are consistent with the need for closer monitoring of low-q firms by institutions because of the higher likelihood of overinvesting the proceeds from the equity issue. The intensity of the monitoring increases with higher ownership because the institutions have more to lose from non-value-maximizing behavior by the issuing firm's managers due to their higher ownership stake in the firm.
VI. Summary and Conclusions
This paper examines whether the level of institutional ownership results in a greater degree of monitoring of a specific corporate decision, namely the issuance of equity. To answer this question, we study the association between both the announcement-period abnormal stock price reaction and the announcement-period abnormal earnings forecast revisions (both current-year earnings and five-year earnings growth) and institutional ownership.
TABULAR DATA OMITTED
We find a significant positive relation between announcement period abnormal stock returns and institutional ownership, which lends support to the effective-monitoring hypothesis. To further corroborate our stock price results, as well as to eliminate other plausible explanations for this result, we also examined revisions in analysts' earning forecasts around equity issue announcements. We find no relation between analysts' abnormal forecast revisions in current-year earnings and institutional ownership, but we do find a significant positive relation between analysts' abnormal forecast revisions in five-year earnings growth and institutional ownership. These results are inconsistent with higher levels of institutional ownership being associated with either lower pre-announcement asymmetry of information or a more elastic demand curve for the stock. Finally, we find that the significant positive relation between the announcement-period abnormal return (as well as the five-year earnings growth forecast) is entirely explained by low-q firms. These firms are more likely to misuse the proceeds from the equity issue; hence, they are subject to closer scrutiny by institutions with higher ownership stakes in the firm. To conclude, our study supports the view that institutions are effective monitors of the equity issuance decision.
McConnell and Servaes |21^ find a positive relation between firm value (as proxied by Tobin's q) and institutional ownership. They suggest that their evidence is consistent with the efficient-monitoring hypothesis. They, however, acknowledge that they cannot be sure of the direction of causality. Our evidence suggests that the presence of higher institutional ownership influences analysts' revisions in long-term earnings forecasts that are associated with the announcement of an equity issue. This points towards a direct and active role institutional owners play in monitoring firms. It is important to stress that the equity issuance decision to finance investment projects is just one among many important decisions that managers make. Given our results, it is likely that institutional investors play a similar role regarding a whole set of other important decisions managers make. That is, however, an empirical issue which we believe will be studied in detail by other researchers.
One caveat is in order. We are not advocating that firms should unequivocally seek to attract institutional ownership because of potential monitoring benefits. They should do so only if it is an attractive option after taking into account all the potential benefits and costs of following such a course of action. Finally, the ability of institutional investors to impose further capital market discipline on managers is likely to increase as institutional ownership grows and as the public policy debate comes to recognize the substantial costs of regulations which inhibit the ability of institutions to monitor managers.
1 The negative price effect of an equity issue announcement can also represent a reaction to a perceived increase in equity risk. Healy and Palepu |9^ and Jain |12^ find an increase in systematic risk subsequent to the equity issue announcement. Lease, Masulis and Page |17^ show that the increase is due to a decrease in market return variance rather than an increase in the covariance between the returns on the firm's stock and the market. In addition, Jain |12^ finds no relation between offer announcement effects and equity risk increases.
2 We argue that institutional ownership, measured as the proportion of shares held by institutional investors, is related to the degree of monitoring by institutions. Prior research indicates that block holdings, in general, provide for more effective monitoring of managers. Hence, institutional ownership of 50% spread over 100 institutions may have different implications than the same ownership spread across only five institutions. Because current regulations inhibit large blockholdings by individual institutions, the above problem associated with our use of institutional ownership as a proxy for the degree of monitoring is alleviated. This measure has been used in a similar context by McConnell and Servaes |21^ and Pound |28^.
3 Pound |28^ studies similar issues, which are related to proxy contests.
4 We do not attempt to test directly whether institutional ownership is related to perceptions of asymmetry of information or to the liquidity of the market for the stock. We are interested in these issues only to the extent that an association between institutional ownership on the one hand and either the pre-announcement asymmetry of information or the liquidity of the stock on the other can potentially explain our documented stock price results.
5 We thank an anonymous referee for this suggestion.
6 For example, see Jacobs |11^ and Wallace |34^.
7 See Grundfest |7^ and Roe |30^ for a detailed discussion of the regulations that curtail the ability of institutional investors to take an even more active role in monitoring managers. Both authors argue that the political process is to blame for protecting managers from capital market discipline because it imposes restrictions on institutional investors. As financial institutions gain more political clout and the costs of these restrictions becomes more apparent, these regulations are likely to be relaxed.
8 The theoretical literature in this area leads to such diverse predictions as no relationship between profitability and insider ownership (Demsetz |5^), a positive relationship between firm value and insider ownership (Jensen and Meckling |14^ and Leland and Pyle |18^), or a curvilinear relationship between firm value and the percentage of shares held by insiders (Stulz |33^). The documented empirical relationship is also widely contradictory. For instance, Demsetz and Lehn |6^ document no relationship between accounting profits (a proxy for finn value) and insider ownership; McConnell and Servaes |21^ provide evidence of a curvilinear relationship between firm value (as proxied by Tobin's q) and insider ownership; Morck, Shleifer and Vishny |25^ and Hermalin and Weisbach |10^ document a nonmonotonic relationship between firm value (again proxied by Tobin's q) and the percentage of shares held by insiders.
9 Run-up is defined as the cumulative excess daily returns over the CRSP value-weighted daily index returns over the event period -250 to -2 days prior to the equity issue announcement. In all our subsequent cross-sectional regression tests, we also used run-up measured over different time periods with no effect on the main results of the paper. These further results are available from the authors upon request.
10 See, for example, studies by Asquith and Mullins |1^, Masulis and Korwar |20^, and Mikkelson and Partch |23^.
11 The abnormal stock return results reported in this paper use the CRSP value-weighted dally return index. All the tests were replicated using the CRSP equal-weighted daily stock return index with virtually identical results.
12 Christie |4^ discusses the theoretical and empirical merits of normalizing by price per share. Pound |29^ provides a simple example to explain the reason behind normalizing by price per share and not by past earnings forecasts. For a firm with a stock price of $30/share, a forecast revision from $0.01 to $0.03 per share is 200% if expressed in terms of the original forecast, and a change from $-0.01 to $0.03 is 400%. The relative importance of these revisions is better expressed as $0.02/30 and 0.04/30 respectively. The long-term forecasts available from I/B/E/S are the expected annual growth rate in earnings over the upcoming five-year period. Therefore, for long-term forecasts, we measure forecast revisions as simply the relative change in the growth rate forecasts.
13 In results not reported in this paper, we document an optimism bias for mean forecasts of both current-year earnings per share and the five-year growth of earnings over our sample period. I/B/E/S, however, does not provide information regarding the number of forecasters that update each month for mean forecasts of the five-year earnings growth. We assume a similar updating period for long-term forecasts as the short-term forecasts.
14 See Smith |32^.
15 To form the three groups, we first rank the firms in our sample in ascending order of institutional ownership. Firms within the bottom third, the middle third, and the top third are placed in the low (L), medium (M), and high (H) groups, respectively.
16 While Asquith and Mullins |1^ find a positive relation between TDAR and run-up, Masulis and Korwar |20^ find exactly the opposite relation. The sign of the coefficient associated with the run-up variable is sensitive to the time period over which the run-up is measured. See Korajczyk, Lucas and McDonald |15^ for a detailed discussion.
17 Researchers have found little relation between the valuation effects of stock offering announcements and the value of the offering, the use of the proceeds, or pre-offering leverage ratios.
18 Some evidence of a negative relationship between stock price effects and relative issue size is found in Asquith and Mullins |1^, Masulis and Korwar |20^, and Mikkelson and Partch |23^.
19 We find a negative and highly significant relationship between TDAR and run-up in all the regressions estimated. As discussed in Korajczyk, Lucas and McDonald |15^, there is no compelling theoretical reason for the relation to go in either direction.
20 We thank an anonymous referee for this suggestion.
21 It is important to emphasize that while we do find evidence supportive of the effective-monitoring hypothesis, institutional investors are not perfect monitors given the significant negative abnormal returns and abnormal forecast revisions associated with even high levels of institutional ownership.
22 Details of the computation of Tobin's q are provided in Appendix B.
1. P. Asquith and D.H. Mullins, "Equity Issues and Offering Dilution," Journal of Financial Economics (January/February 1986), pp. 61-89.
2. P.A. Brous, "Common Stock Offerings and Earnings Expectations: A Test of the Release of Unfavorable Information," Journal of Finance (September 1992), pp. 1517-1536.
3. P.A. Brous and O. Kini, "Interfirm Tender Offers and Target Firms' Future Performance: A Reexamination of Analysts' Earnings Forecasts," Journal of Financial Economics (April 1993), pp. 201-225.
4. A. Christie, "On Cross-Sectional Analysis in Accounting Research," Journal of Accounting and Economics (December 1987), pp. 231-258.
5. H. Demsetz, "The Structure of Ownership and the Theory of the Firm," Journal of Law and Economics (June 1983), pp. 375-393.
6. H. Demsetz and K. Lehn, "The Structure of Corporate Ownership: Causes and Consequences," Journal of Political Economy (December 1985), pp. 1155-1177.
7. J.A. Grundfest, "Subordination of American Capital," Journal of Financial Economics (September 1990), pp. 89-114.
8. R. Hansen and C. Crutchley, "Corporate Earnings and Financing: An Empirical Analysis," Journal of Business (July 1990), pp. 347-371
9. P. Healy and K. Palepu, "Earnings and Risk Changes Surrounding Primary Stock Offers," Journal of Accounting Research (Spring 1990), pp. 25-48.
10. B. Hermalin and M. Weisbach, "The Effects of Board Composition and Direct Incentives on Firm Performance," Financial Management (Winter 1991), pp. 101-112.
11. S.L. Jacobs, "Big Holders Resolve to Flex Their Muscle," The Wall Street Journal, February 28, 1989, p. C1.
12. P. Jain, "Equity Issues and Changes in Expectations of Earnings By Financial Analysts," Review of Financial Studies (Winter 1992), pp. 669-683.
13. M.C. Jensen, "Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers," American Economic Review (May 1986), pp. 323-329.
14. M.C. Jensen and W.H. Meckling, "Theory of the Firm: Managerial Behavior, Agency Costs, and Ownership Structure," Journal of Financial Economics (October 1976), pp. 305-360.
15. R.A. Korajczyk, D.J. Lucas, and R.L. McDonald, "Understanding Stock Price Behavior Around the Timing of Equity Issues," in Asymmetric Information, Corporate Finance, and Investment, R. Glenn Hubbard (ed.), Chicago, IL, University of Chicago Press, 1990, pp. 257-277.
16. L. Lang and R. Litzenberger, "Dividend Announcements: Cash Flow Signalling Vs. Free Cash Flow Hypothesis?," Journal of Financial Economics (September 1989), pp. 181-191.
17. R.C. Lease, R.W. Masulis, and J.R. Page, "An Investigation of Market Microstructure Impacts on Event Study Returns," Journal of Finance (September 1991), pp. 1523-1536.
18. M. Leland and D. Pyle, "Informational Asymmetries, Financial Structure, and Financial Intermediation," Journal of Finance (May 1977), pp. 371-387.
19. E. Lindenberg and S.A. Ross, "Tobin's Q Ratio and Industrial Organization," Journal of Business (January 1981), pp. 1-32.
20. R.W. Masulis and A.N. Korwar, "Seasoned Equity Offerings: An Empirical Investigation," Journal of Financial Economics (January/February 1986), pp. 91-118.
21. J.J. McConnell and H. Servaes, "Equity Ownership and Corporate Value," Journal of Financial Economics (October 1990), pp. 593-613.
22. H. McFarland, "Did Railroad Deregulation Lead to Monopoly Pricing? An Application of Q," Journal of Business (July 1987), pp. 385-400.
23. W.H. Mikkelson and M.M. Partch, "Valuation Effects of Security Offerings and the Issuance Process," Journal of Financial Economics (January/February 1986), pp. 31-60.
24. M.H. Miller and K. Rock, "Dividend Policy Under Asymmetric Information," Journal of Finance (September 1985), pp. 1031-1051.
25. R. Morck, A. Shleifer, and R.W. Vishny, "Management Ownership and Market Valuation: An Empirical Analysis," Journal of Financial Economics (March 1988), pp. 293-315.
26. S.C. Myers and N.S. Majluf, "Corporate Financing and Investment Decisions When Firms Have Information That Investors Do Not Have," Journal of Financial Economics (June 1984), pp. 187-221.
27. P. O'Brien, "Analysts' Forecasts As Earnings Expectations," Journal of Accounting and Economics (January 1988), pp. 187-221.
28. J. Pound, "Proxy Contests and the Efficiency of Shareholder Oversight," Journal of Financial Economics (January/March 1988), pp. 237-266.
29. J. Pound, "The Information Effects of Takeover Bids and Resistance," Journal of Financial Economics (March 1988), pp. 207-227.
30. M.J. Roe, "Political and Legal Restraints on Ownership and Control of Public Companies," Journal of Financial Economics (September 1990), pp. 7-41.
31. M. Smirlock, T. Gilligan, and W. Marshall, "Tobin's Q and the Structure Performance Relationship," American Economic Review (December 1984), pp. 524-532.
32. C.W. Smith, "Investment Banking and the Capital Acquisition Process," Journal of Financial Economics (January/February 1986), pp. 3-29.
33. R. Stulz, "Managerial Control of Voting Rights, Financing Policies and the Market for Corporate Control," Journal of Financial Economics (January/March 1988), pp. 25-54.
34. A. Wallace, "Pension Funds Are Acting More and More Like Big Shareholders," New York Times, May 21, 1989, Section 4, p. 6.
A. Development of Expectations Model of Forecast Revisions
In this appendix, we provide details concerning an expectations model of earnings forecast revisions which utilizes some of the salient time-series properties of earnings forecast revisions. Previous empirical studies using the I/B/E/S forecast data assume that the expected forecast revision equals zero and that the observed forecast revision is, therefore, an unbiased estimate of the unexpected or abnormal forecast revision. Both Brous |2^ and O'Brien |27^, however, present evidence that analysts' forecasts are not unbiased estimates of future earnings. The evidence from these studies suggests that the forecasts reported in the I/B/E/S database have an optimism bias in the sense that analysts tend to overestimate future earnings. Furthermore, analysts tend to lower their forecasts systematically each month up to the fiscal year-end. Therefore, the expected forecast revision in any given month is negative.
Brous |2^ provides evidence that analysts' forecasts are serially correlated. Since we are analyzing the mean forecasts of a group of analysts, of whom approximately 15 to 20 percent update their forecast each month, information released in one month will be reflected in the mean of analysts' earnings forecasts for several subsequent months. For example, if favorable information were released during January regarding a firm's earnings prospects and if 20% of the analysts following the firm revised their forecasts during January, the forecast revision for January would be positive. If an equal amount of negative information were released in February and if a different 20% of the analysts updated their forecast during February, the forecast revision for February would be zero: a result arising because the revisers would draw on all information released since their last revision, including January's favorable news. The previous months' forecast revisions are, therefore, a necessary input into a model that estimates the unexpected information released in February.
Given these statistical properties, the earnings forecast revisions process can be described by the following model:
|FR*.sub.i,t^ = |k.sub.i^ + ||Epsilon^.sub.i,t^ (A1)
where |FR*.sub.i,t^ is the unobservable "true" forecast revision, assuming that all analysts update their forecasts each month for firm i at month t; |k.sub.i^ is the forecastable component for firm i; and ||Epsilon^.sub.i,t^ is uncorrelated and, by definition, the unforecastable component for firm i at month t. If individual forecasters, however, submit a revision every n months, then the expected forecast revision for an individual analyst j who updates at month t for firm i |Mathematical Expression Omitted^ can be written as:
|Mathematical Expression Omitted^.
The first term, |nk.sub.i^, equals the expected monthly revision times the number of months since the analyst's previous update. This term represents the cumulative bias in forecast revisions over the period since the analyst's previous update. The second term, the sum of the unforecastable components, reflects all of the unexpected information since the analyst's previous forecast.
Because the mean forecast that is reported by I/B/E/S includes the forecasts of all analysts who follow the firm regardless of whether an individual forecast is changed, the mean forecast reported is a weighted average of forecasts updated |Mathematical Expression Omitted^ and forecasts unchanged. If analysts update their forecasts every n months, then 1/n forecasters update their forecasts each month. Thus, the I/B/E/S mean forecast revision at month t for firm i (|FR.sub.i,t^) is:
|Mathematical Expression Omitted^.
If the ||Epsilon^.sub.i,t^ are identically distributed with variance ||Sigma^.sub.2^, as well as independent, the covariance between |FR.sub.i,t^ and |FR.sub.i,tc^ equals:
Cov (FRi,t, FRi,t-c) = (1/n2)(n-c) s2 if n |Mathematical Expression Omitted^ c and 0 otherwise.
Given that a finite number of lagged terms are correlated, this model suggests that |FR.sub.i,t^ follows a (n-1) order moving average process. The a priori expected mean forecast revision for firm i at month t (E||FR.sub.i,t^^) is then estimated as:
E||FR.sub.i,t^^ = |k.sub.i^ + 1/n |summation of^ ||Epsilon^.sub.i,t-s^ |where^ s=1 to n-1. (A4)
The finding that 15 to 20% of the analysts revise their forecasts each month is consistent with the notion that some analysts update quarterly and others semiannually. In our estimation process, therefore, we assume n=5 and use a fourth-order moving average model to generate expected forecast revisions in which the coefficients on the lagged error terms are forced to equal 0.20. For each firm, the forecastable component (|k.sub.i^) is estimated as the average forecast revision during an estimation period, which consists of all months for which forecasts are available excluding months -6 to +6.
The expost abnormal, or unexpected, forecast revision for firm i in month t (|AFR.sub.i,t^) can be expressed as:
|AFR.sub.i,t^ = |FR.sub.i,t^ - E (|FR.sub.i,t-1^). (A5)
The announcement month abnormal forecast revision will capture most of the announcement effect of major corporate events if all those analysts who believe that the event has implications regarding future earnings adjust their forecasts in a timely fashion.
B. Estimation of Tobin's q
In computing Tobin's q, we use methods described by Lindenberg and Ross |19^; McFarland |22^; and Smirlock, Gilligan and Marshall |31^. Tobin's q is defined as the ratio of the firm's market value to the replacement cost of its assets. To estimate q, we first compute each of these two components separately.
The market value of the firm is estimated by summing the market value for its common stock, preferred stock, and debt. The market value of common stock is determined by multiplying the price per share by the number of shares outstanding. The market value of preferred stock is approximated by dividing the amount of preferred dividends by Standard and Poor's preferred stock yield index. The market value of debt includes both current liabilities and long-term debt. Current liabilities are evaluated at book value. To estimate the market value of long-term debt, we have to determine the debt's maturity structure and coupon rate. We assume that every firm in the sample issues its long-term debt with a 20-year maturity and a coupon rate equal to Moody's composite average yield on industrial bonds for the year of issue. COMPUSTAT also provides data on debt maturing in 2, 3, 4 and 5 years. These amounts were assumed to have been issued with 20-year maturities and coupons assigned as above. The remaining long-term debt is assumed to have been issued in equal annual amounts over the rest of the period. Debt was valued each year by calculating the price of each portion at the prevailing yield on industrials reported by Moody's, taking into account maturities and coupons as described above. Finally, all the miscellaneous liabilities are valued at book value.
The replacement cost of the assets of the firms is:
R = TA + (RP - BP) + (RI - BI) - DT,
where TA is total assets, including current assets; RP is the replacement cost of plant and equipment; BP is its book value; RI is the replacement cost of inventories; BI is their book value; and DT is the book value of deferred taxes. The replacement cost of plant and equipment was determined by setting up an acquisition schedule and adjusting for price changes. Each year, the value of plant and equipment was reduced by depreciation calculated at a five percent rate, then adjusted to the new price level according to the GNP implicit price deflator. New additions or sales were calculated as the change in gross plant at book value. Inventory was valued by taking into account the inventory method reported for that firm by COMPUSTAT. For LIFO, the beginning inventory was adjusted for a full year's inflation and any change in reported inventory was adjusted for one-half year's inflation. Inventories Under the Retail Cost method are adjusted by the ratio of the Wholesale Price Index to the Consumer Price Index. Inventories reported based on methods other than those listed above were valued at book. Finally, all the other assets were assumed to have replacement costs equal to their respective book values.
Peter A. Brous is an Assistant Professor at Seattle University, Seattle, Washington. Omesh Kini is an Associate Professor at Emory University, Atlanta, Georgia.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||includes appendix|
|Author:||Brous, Peter A.; Kini, Omesh|
|Date:||Mar 22, 1994|
|Previous Article:||Insider trading following material news events: evidence from earnings.|
|Next Article:||Liquidity changes associated with open market repurchases.|