Stock Repurchases Associated with Stock Options do Represent Dollars out of Shareholders Wallets

Stock Repurchases Associated with Stock Options do Represent Dollars out of Shareholders’ Wallets1 Kenneth J. Klassen2 Ranjini Sivakumar School of Accountancy University of Waterloo Waterloo, ON, Canada N2L 3G1 Abstract: This paper explores the repurchase activity and employee and executive option activity for nonfinancial firms for the period 1995-1999. We find that the value of both firm-level options outstanding and executive-level options outstanding are strong predictors of repurchase activity. When option grants are considered, only executive-level grants provide incremental explanatory power. Our results suggest that firms repurchase shares to avoid dilution from a stock option program. We also examine market reactions to repurchase authorizations. We find that the returns preceding the announcement are less negative for firms that are active executive option users, and that investors react less enthusiastically to stock repurchases when firms grant executive (but not employee) options during the repurchase year. The evidence indicates that markets recognize that particularly executive option grants represent a real cost to the firm ex post in the form of stock buybacks, possibly representing a wealth transfer from the shareholders to employees. May 2001 The authors thank the Social Sciences and Humanities Research Council of Canada for generously supporting this research, and to Albert Lo and Umang Patel for excellent research assistance. 2 Contact Author: phone: 519-888-4567 (ext. 5702); fax: 519-888-7562; email: kklassen@uwaterloo.ca 1 Stock Repurchases Associated with Stock Options do Represent Dollars out of Shareholders’ Wallets 1. Introduction Now, repurchases are all the rage, but are all too often made for an unstated and, in our view, ignoble reason: to pump or support the stock price. …Sometimes, too, companies say they are repurchasing shares to offset the shares issued when stock options granted at much lower prices are exercised. This “buy high, sell low” strategy is one many unfortunate investors have employed—but never intentionally! Managements, however, seem to follow this perverse activity very cheerfully. — Warren Buffet, Chairman, Berkshire Hathaway Inc., March 1, 2000 in his letter to the shareholders. Several recent studies have noted the phenomenal increase in open market stock repurchases in the past two decades. In 1998, for the first time, total funds paid out via repurchases have exceeded those paid out in the form of dividends. Particularly interesting has been the surge in repurchases over 1995-1999 during soaring valuations and the drop in stock buybacks in 2000 when stocks were cheap. This buyback surge has been accompanied by a rise in employee stock option awards and has been attracting an increasing amount of attention from the business press and academics. The combination of share buybacks and options has been termed lethal in Business Week (November 15, 1999) since options impose a cost on shareholders by diluting a firm’s existing equity. This article quoting Liang and Sharpe (1999) states that with executives’ compensation tied to stock market performance, firms are willing to fund buybacks and thus bolster stock prices even when stocks look overvalued. Furthermore, with companies trying to meet both buyback and capital spending goals, companies are borrowing to finance buybacks amid a market boom. Existing literature has tried to make sense of this newfound popularity of repurchases vis-à-vis dividends. For example, Grullon and Michaely (2000) show that repurchases are increasingly replacing dividends as a payout means, and that the 1982 lifting of regulatory constraints against buying back own shares was instrumental in bringing about this shift. Fama and French (1999) examine the disappearance of dividends over time and report that option-related buybacks of stock do not substitute for dividends but are associated with higher total payouts. Jolls (1998), Fenn and Liang (2000) and Weisbenner (2000) find that the probability of repurchasing stock is positively related to stock options. Fenn and Liang (2000) find that firms using more executive stock options are likely to stall dividend increases in favor of repurchases, since repurchases do not reduce stock prices and therefore preserve management held option values. Jolls (1998) reaches a similar conclusion relating management stock option ownership and use of repurchase as a payout means. When total option grants are included with executive option grants, Jolls (1998) finds that executive options increase the likelihood of a firm undertaking a repurchase, whereas total options have no incremental explanatory power. Jolls concludes that repurchases are not timed to offset the dilutive effect of options. These studies look at the repurchase decision from the perspective of the managers as agents and conclude that the increased use of repurchases has coincided with increased reliance on option based compensation. Weisbenner (2000) also examines repurchase activity in 1995. Exploring similar hypotheses, Weisbenner (1998) discovers that total options outstanding, rather than executive options outstanding, are correlated with repurchase activity. No discussion is presented why these results contradict those of Fenn and Liang (2000), and Jolls (1998). Additional tests by Weisbenner (2000) document that both total and executive options help explain total payouts, 2 and that repurchases are timed to offset the accounting effects of dilution in earnings per share, and not timed to combat the real dilution that occurs upon the exercise of the options. Liang and Sharpe (1999) observe that the sharp increase in buyback in the 1990s has been accompanied by a corresponding rise in employee stock-option awards. According to these authors, when firms repurchase shares to avoid dilution relating to option exercise, there is a wealth transfer from shareholders to employees since the cost of repurchasing is much higher than the price at which the employees exercise their options. Hence stock options do not represent dollars out of stockholders’ wallets when issued; they do so when stocks are repurchased. In this paper, we take a different tack to explore the rapid rise in repurchase activity. We argue that options impose a cost on shareholders by diluting a firm’s existing equity. Options are not set off against income and hence inflate reported profits. What about share prices? If options have a real economic cost, do investors take that into account when options are granted or exercised? If investors could assess the extent of the firms’ liabilities then share prices would reflect that information. We examine the market’s reactions to repurchase decisions, both during the period prior to the board resolution and at the decision date and investigate if markets view options as a real expense when companies repurchase shares to avoid the dilution effect. We also include more recent repurchase and options-based compensation data to improve the robustness of our tests. With these data we explore similar tests to those found in the above literature relating repurchase and option activity. Our tests find that both total options outstanding and executive options outstanding have a positive effect on repurchase behavior; however, only the influence of executives’ options outstanding is consistently statistically significant at the one percent level. However, 3 when option grants are used, both total grants and executives’ grants have incremental effects on repurchase activity. These results hold whether repurchases are measured as proportion of shares or proportion of market value, and whether options outstanding is measured as proportion of shares outstanding or proportion of value. Examining the abnormal market returns around the time of the repurchase decision reveals that, in general, the value of the firm was falling in the period preceding the announcement and the announcement has a significant positive effect on the market price. In a regression of the abnormal returns, we document that markets react to stock repurchases less enthusiastically when firms grant options. Investors recognize that option grants represent a real cost to the firm ex post in the form of stock buybacks, possibly due to resultant wealth transfer from the shareholders to employees. Overall, our evidence is consistent with markets viewing the funding of stock option programs by repurchasing shares as an expensive strategy. In the second section of the paper, we summarize the theory that has developed to explain repurchase behavior and identify the hypotheses tested in this research. Section three describes the research design, sample and summary statistics. Section four presents the results of the tests of the hypotheses. The final section summarizes and identifies conclusions. 2. Theory and Hypotheses Explaining the dramatic rise of repurchases has recently been the subject of some research (e.g., Liang and Sharpe, 1999). Prior to this increased popularity, stock repurchases were typically attributed to management’s belief that the firm’s stock was undervalued by the 4 market place (e.g., Lakonishok and Vermaelen, 1990; Healy and Palepu, 1995; and Ikenberry, Lakonishok and Vermaelen, 2000). Another explanation is that repurchases are now used to distribute temporary positive cash flow shocks (e.g., Jagannathan, Stephens and Weisbach, 1998; and Guay and Harford, 1999). Finally, the increased use of stock options has occurred concurrently with the rise in repurchases, leading several researchers to explore what links might exist between these two phenomena (e.g., Liang and Sharpe, 1999; Fenn and Liang, 2000; Jolls 1998; and Weisbenner, 2000). Liang and Sharpe (1999) present an argument that repurchases are undertaken to reduce the dilutive effects of the options’ use. Fenn and Liang (2000), and Jolls (1998), on the other hand, draw on agency theory and argue that to maintain the value of their options, managers with options will choose to undertake a repurchase, rather than distributing larger dividends, as a means of distributing cash. When Jolls (1998) includes total options as well in her empirical tests, this new variable does not provide incremental explanatory power. Weisbenner (2000) also tests both of these possible option-related explanations, and incorporates both total options outstanding, to proxy for the dilutive effects, and executive options, to proxy for the agency effects. He finds that each measure, in isolation, is related to repurchase activity. When both measures are included in the regression, the total option use is significant in explaining repurchase activity, but the options held by the top five executives does not have a significant effect. One limitation of both Jolls (2000) and Weisbenner (2000) is their use of crosssectional data. A single year can lead to unidentified spurious relations. In addition, both studies used data from the mid-1990s, before the magnitude of repurchase activity had reached the current dramatic heights. Furthermore, Jolls (1998) uses option grants in the previous year both at the firm and executive level as a proxy for options outstanding. This is 5 an imprecise measure since on average options outstanding are about four times the annual grants. In a departure from previous papers, this study uses panel data on option grants, exercises and total options outstanding at both the executive level and at the firm level for 598 firms during the period 1995-1999.3 Hence our data help us to examine whether it is firm level or executive level option grants, exercises or total options outstanding that is driving repurchase activity. Our analysis overcomes the limitation pointed out by Weisbenner (2000) that a cross-sectional analysis may give spurious results since firmspecific effects cannot be ruled out. Thus, in the tests that follow, we re-visit these two hypotheses. Since there is theoretical support for both measures of option use to be related to repurchase behavior, we test the following hypothesis in two parts (stated in alternative form): H1a The amount and value of total options outstanding will be positively related to the size of repurchases. H1b The amount and value of executive options outstanding will be positively related to the size of repurchases. Extant research has not explored the shareholders’ reactions to these explanations. Relative to the undervaluation perspective, repurchasing shares to reduce the dilutive effects of option plans or as a manager-preferred method of distributed cash to shareholders could For a sample of 595 firms during the period 1994 to 1997, Core and Guay (2001) document that while nonexecutive employees are awarded 75% of all options granted, they exercise more than 84% of all options exercised annually. 3 6 have very different market perceptions. For example, repurchasing stock because management believes the stock is undervalued conveys private information to the market place. This information will, on average, result in a positive stock price reaction. On the other hand, reducing the dilutive effects may be seen positively by the existing shareholders, but there is little new information about future firm performance that is conveyed. However, option exercises could represent a real cost to the firm representing a wealth transfer from the shareholders to employees. According to The Economist (January 27, 2001), Smithers & Co. document that if options had been accounted for at the time they were granted, the profits of large listed companies in 1998 would have been two-thirds lower. The recurring cost of options to American companies were they to be immunized would be about 20% of profits. Since few were immunized and were thus exposed to rising share prices, their true profits on a full-cost accounting measure were only 37% of those published. Thus, we expect that shareholders’ will react less strongly to repurchases that apparently have the anti-dilution objective. Similarly, if managers use repurchases rather than dividends to preserve the value of their options, existing shareholders are not expected to view this as positively. Thus, as the amount of options increase, we expect less positive market reaction. Furthermore, Core and Guay (2000) show that the existing treasury stock method of accounting for the dilutive effect of stock options systematically understates the dilutive effect and overstates earnings per share. Hence, investors cannot assess accurately the extent of the firms’ liabilities relating to exercisable options and would view repurchases associated with stock exercises less positively. In the period preceding the repurchase, repurchases made because the stock is undervalued, for example, will generally follow a downward trend in the firm’s stock price. All else being equal, management will generally prefer repurchases to dividends after a 7 period of declining stock prices. When options are present, however, the dilutive effects are largely independent of recent stock price movements. Therefore, there may be a less negative movement in stock price prior to a repurchase announcement that is driven by this motive. Again, an analogous situation would exist if the reason for the repurchase is driven by agency issues related to the use of executive stock options. In the period following the repurchase, we do not predict any abnormal market returns, independent of the reason for the repurchase. Based on these ideas, we test the second hypothesis in two parts (stated in alternative form): H2a Firms with greater total stock options, or executive options, outstanding will have a less negative return prior to the repurchase announcement . H2b Firms with greater total stock options, or executive options, outstanding will have a less positive return at the repurchase announcement. 3. Research Design, Sample and Summary Statistics 3.1 Repurchase Activity Research Design The research design to test hypothesis H1 generally follows the cross-sectional tests of Weisbenner (2000). Thus, we estimate the following regression. Repurchasesit = β0 + β1 Total Optionsit + β2 Executive Optionsit + β3 Financial Characteristicsit + β4 Year Indicatorst + εit (1) 8 Repurchases are measured using the value of repurchases, as a percentage of the firm’s market value.4 Consistent with hypothesis H1a and with the findings of Weisbenner, and contrary to the implications of the results of Jolls (2000), β1 is predicted to be positive. Consistent with H1b and with the implications of the results of Jolls, and contrary to the findings of Weisbenner, β2 is predicted to be positive as well. Options are measured in two ways. First, they are measured as the number of options held by the employees or the number of shares in the money held by the top five executives, both deflated by the total number of shares outstanding. Secondly, they are measured as the value of the options outstanding held by all employees and the value of in-the-money options held by the top five executives, deflated by the market value of common equity. While preferable, the number and value of all options held by the top five executives is not disclosed. To ensure that our results were not driven by this difference between the total option measures and the executive option measure, we also used option grants, which are highly correlated with options outstanding. Results for these specifications, not tabulated, are qualitatively similar to those reported below. The ability to distribute earnings will affect the amount of repurchases that are available. Following Fenn and Liang (2001) and Weisbenner (2000), we include capital expenditures, net operating cash flow (EBITDA – capital expenditures), each deflated by sales, to capture different aspects of cash flow availability. Repurchases are expected to be more positively influenced by high levels of net operating cash flow. Capital expenditures are clearly a competing use of cash flows generated by income, and so they are expected to negatively influence repurchase activities. In a similar manner, the other investment An alternative is to use the number of shares repurchased deflated by the number of shares outstanding. This measure has a correlation with the value measure of 0.93 and the results are very similar to those reported below. 4 9 opportunities available to the firm may influence the use of repurchases. Therefore, the market to book ratio is included in the regression as a proxy for these opportunities. An alternative view of this ratio is that relatively higher stock values are more suggestive that the firm is not undervalued, and also make the repurchase of shares more expensive. Any of these three interpretations of the market to book ratio lead to the prediction that higher levels of the ratio are negatively correlated with repurchases. Finally, leverage and the log of total assets are included to control for various other aspects of the firm that may relate systematically to repurchase activity. For example, high debt levels may make managers more reluctant to distribute current cash flows and size may relate to financing costs or asymmetric information. Finally, year indicator variables are included to control for the exogenous differences in mean repurchases that occur over the sample period. Since there is a mass at zero for the distribution of the dependent variable and the variable is non-negative, tobit estimation is used to estimate equation (1). Ideally, the equation would be estimated using a fixed effects version of tobit to ensure that the resulting test statistics are not inflated due to multiple (correlated) observations for each firm. However, we are not aware of a computerized algorithm that will perform this estimation. 3.2 Market Reactions Research Design To test hypothesis H2, we first compute the market-adjusted cumulative abnormal returns prior to, at the time of, and following the repurchase decision by the Board of Directors. Immediately following the board decision, an announcement of the repurchase plan is found in the Wall Street Journal. Due to the possible informational leaks and 10 publication delays, we consider the trading days [-20, -3] to be the pre-announcement period, [-2, +2] to be the announcement period, and [+3, +20] to the be the post announcement period. With these returns estimated, we estimate the following regressions Returnit = β0 + β1 Optionsit + β2 Year Indicatorst + εit (2) Options are measured as described in section 3.1. Separate regressions will include total and executive options. Other variables were included to capture other explanations for the repurchase decision (e.g., perceived stock undervaluation) and the market’s reaction to the repurchase (e.g., the size of the repurchase). Thus, we include the market-to-book ratio, the percentage of outstanding shares that the repurchase represents as control variables. We also included capital expenditure as a control variable to examine if the market considers repurchase as a competing use of cash flow. However, the coefficients on these variables were not statistically significant at conventional levels, so the variables were excluded from the reported results. Finally, leverage and the log of total assets are included to control for aspects of the firm that may relate systematically to repurchase activity. Equation (2) is estimated using OLS because the dependent measure is continuous and approximately normally distributed. 3.3 Sample Selection and Summary Statistics Data on repurchases is collected from the Securities Data Corporation’s (SDC) database. This database contains a comprehensive sample of authorizations of repurchases 11 and actual repurchases from 1994 onwards. To measure the actual repurchase, previous studies (Weisbenner, 2000) have used a Compustat measure of repurchase (annual data item 115) after adjustments for redemption of preferred stock. We use the SDC measure for actual repurchase of common stock that should be more accurate. From the set of firms whose common stock repurchases are recorded in these data, financial firms and utilities are excluded because they are believed to face a different incentive structure around repurchase activity. Only open-market repurchases are included consistent with prior research such as Jolls (1998) and Barth and Kaznik (1999). Option data for the top five executives is collected from Standard and Poor’s ExecuComp data base. Options data for the company as a whole is not recorded in any electronic data base and so these data were hand collected from 10-K and proxy statement filings. The Center for Research in Securities Prices (CRSP) data were used for stock return data, and Compustat was used for other financial variables found in the regressions. The intersection of these sources of data created a data set of 598 non-financial firms with data throughout most of the 1995 to 1999 period, inclusive. Summary statistics are found in Tables 1 and 2. Table 1, Panel A presents the results for repurchase and stock option activity (normalized by market value of equity). Repurchase activity has been steadily increasing from 2.65% in 1995 to 4.07% in 1999. The corresponding stock option outstanding value has increased from 4.15% to 11.37%, the option grants from 1.20% to 3.82% and the options exercised from 0.36% to 0.70% in the same period. The value of options outstanding for executives has been growing at a slower rate indicating that nonexecutive employee options have been on the rise. In comparison to options, restricted stock grants represent a much smaller component of executive compensation. In Panel B, we observe that firm-level option outstandings, grants and exercises are highly postively 12 correlated. For executives, these measures are also positively correlated but to a lesser extent. The correlation between firm level and executive level option measures while positive is even lower. The correlation between the option measures and executive restrictive stock grants is negative and fairly low. In Table 2, we present firm operating characteristics for the firms that have stock option and repurchase activity. We present descriptive statistics for the entire panel and for firm years in which firms repurchased shares. The market to book ratio and the prior year’s stock total stock return for repurchase years seems to be lower indicating firms repurchase shares when they are relatively undervalued. There does not seem to be a difference in capital expenditure and leverage in repurchase and non-repurchase years. 4. Results 4.1 Repurchase Activity Tests Results of estimating equation (1) using the 598 firms in all years with sufficient data to estimate the equation are presented in Tables 3, 4 and 5. The regressions are estimated using the tobit model because fifty percent of the dependent observations have a value of zero. Column 1 of each table presents the estimates of the regression using the firm-level option variables. Column 2 presents the estimates of the regression using executive options variables. The specification in column 3 uses both firm-level and executive options variables. 13 Table 3 estimates equation (1) using the value of outstanding options to explore the relation between repurchase activity and total option use by the firm. These estimates find support for hypothesis H1a: total outstanding options are positively related to repurchase activity. The coefficient on this variable is statistically significant at the 5% level whether the executive’s options value is included in the regression or not. The estimates are even more strongly supportive of hypothesis H1b. The estimates of the coefficients on the executive stock option use variables are statistically significant at the 1% level in both columns 2 and 3. In column 3 when both firm and executive level options are included, the results are different from the findings of Weisbenner (2000), who did not find any statistical relation between the options held by the top five executives and repurchase activity in his 1995 cross-sectional test when both option measures were included in the same regression. We find that the total value of outstanding options for both types of options are each incrementally important to the decision to repurchase shares. Of the control variables, most are statistically significant and of the predicted sign. The amount of restricted stock grants and net operating cash flow both have a positive effect of stock repurchase activity, statistically significant at the 1% level. Conversely, the marketto-book ratio, leverage and prior year’s stock return are negatively correlated with repurchase activity, these coefficients are statistically significant at the 1% level. The size of the firm is varies somewhat in importance across the three regressions. Inconsistent with prior research and expectations, the capital expenditures of the firm and shares held by insiders do not affect repurchase activity in a significant manner. In Table 4, the regressions are again estimated, but the value of option grants is used in place the of the value of grants outstanding. In column 1, when the value of all option grants for the firm is included alone, the results are very similar to those in Table 3. 14 Similarly, when only the value of option grants to the executives is included in the regression in column 2, the results are similar to those in Table 3, with somewhat larger, more statistically significant effect (coefficient (t-value) of 0.79 (3.86) in Table 4 compared to 0.26 (2.38) in Table 3). However, in a noteworthy departure from Table 3, when both variables are included in column 3, only the grants to the executives remains statistically significant at conventional levels. The coefficients on the control variables, and their statistical significance, is similar across the two tables. Thus, while the granting of stock options itself is important to repurchase behavior, the grants to executives seems to be the more important of the two types of grants. In Table 5, the exercise of options is used as the measure of option activity. The results of this table are similar to the prior tables in that when the exercise of all options, or the exercise of executives’ options are considered by themselves, the estimated coefficients are statistically significant at the 1% level. When both types of exercises are considered together in column 3, both types of options each incrementally important in explaining stock repurchases. Overall the results indicate that option activity drives repurchase activity. When trying to separate the effect of all options from those of the executives only, generally we find that both are important, except when grants are considered. In this case, only the grants to the executives was important to explain repurchase activities. The similarity of results across the three tables is not really surprising however, because the three measures of option activity are quite highly correlated as we noted on Table 2. For firm-level options, the pairwise correlation across the three measures vary from 0.75 to 0.89. For executives’ options, the pair-wise correlations across the three measures are noticeably less, and vary from 0.24 to 0.47. Taken together, the three sets of tobit regressions are consistent with both Hypotheses 15 H1a and H1b and are thus consistent with the assertion that firms avoid the stock dilution associated with option use. 4.2 Market Reactions Tests Table 6 provides the short-term market reactions to repurchase authorizations by the board. From panel A, we note that about two-fifths of the firms made only one repurchase authorization during the sample period. The total number of repurchase authorizations including multiple events was 1,171. In panel B, the results indicate that the average market reaction from –20 to –3 before the board authorization date is negative for the overall sample and for each year, each mean reaction being different from zero at a one percent level of statistical significance. The average market reaction from –2 to +2 days is positive but is not significant in the later years of our sample, in contrast to Ikenberry et al. (1995). Following the authorization, in the window +3 to +10, the abnormal return on average is not reliably different from zero. From panel C, we find that markets react more favorably to larger share repurchases because the return during the period –2 to +2 increases in size as the size of the repurchase grows. No clear pattern exists in the period prior to the announcement or following the announcement. In Panel D, we examine the market reaction according to the stated reason for repurchase. Where undervaluation was cited as a reason, markets reacted significantly positively, both in terms of average response magnitude at 3.7% and in terms of statistical significance at the 1% level. It may be noted that there was no market reaction when the reason was stated as offsetting dilution and only marginally positive (i.e., a statistically significant mean at the 10% level) when the reason was for employee benefit plans. We 16 interpret these results with caution since the reasons for repurchases were oftentimes not stated by firms and there may be multiple reasons for a particular purchase. In addition, the reason motivating the market’s reaction and management’s stated reason may also differ. In Table 7, we present the regression results relating market reactions around repurchase announcements to firm option activity. Panels A through C mirror the analysis in Tables 3 through 5 in that we employ the value of outstanding options, the value of new option grants, and the value of options exercised, respectively, in the three sets of analyses. In columns 1 to 3 we examine market reactions in the window –20 to –3 and in columns 4 to 6, market reactions in the window –2 to +2. Exploring first Panel A, from column 1, we find that in the window prior to repurchase, the evidence is consistent with management generally preferring to repurchase shares after a period of declining stock prices independent of whether it is to prevent dilution due to the total options outstanding at the firm. Hence there does not seem to be support for hypothesis 2a regarding a less negative movement in stock price prior to a repurchase announcement that is driven by overall options outstanding. For the value of outstanding executive options estimated in column 2, the coefficient on this variable is greater than zero at the 1% level of statistical significance. Thus, we find evidence consistent with hypothesis H2a—the decline in the stock price of the firm does not have to be as severe for managers to be willing to repurchase shares when there are significant executive options outstanding. At the repurchase itself (i.e., in the event window around repurchase authorizations found in columns 3 and 4), the announcement day return is independent, on average, of both the firmlevel and executives’ outstanding options. 17 Panel B repeats the analysis of Panel A, but replaces the value of options outstanding with the value of new option grants. In the period prior to the announcement, the results are similar to Panel A: total option grants have no consistent effect on the pre-announcement returns but executive option grants reduce the negative average return with a coefficient of 1.5 that is statistically significant at the 1% level. Thus, on average, an increase in executive option grants of two standard deviations, or 1.10 percent of the market value of equity, will result in a repurchase following a stock decline that is 1.65% less than otherwise would be the case. During the repurchase announcement, the shareholders are significantly less enthusiastic about the repurchase news if there are options outstanding. This statement is supported with reference to either total option grants, displayed in column 4, or executive option grants, displayed in column 5. When both types of option grants are included in the analysis, the effect of the executives’ option grants is statistically significant at the 1% level, while the effect of the total option grants is only statistically significant at the 10% level. Panel C repeats the analysis using the exercise of options as the variable of interest. In this case, there is no consistent relationship between option use, either in total or at the executive level, either prior to or at the announcement time. Taken together, there are several observations that can be made. Firstly, in the period prior to the announcement of a repurchase, the granting or holding of executive options reduces the negative stock return that is typical of that period. This is consistent with a smaller decline in the stock price being necessary to trigger repurchase activity. This behavior is consistent with a concern about the dilution that executive options pose, but also concern on the part of the executives regarding to the value of their options. The fact that 18 there is no relation between stock returns and firm-level options outstanding or grants suggests that the latter is a more likely explanation. The announcement reaction test results are consistent with the market viewing the repurchase as a less significant event if there are many options being granted, particularly to executives. In years following the grant, the market does not seem to relate consistently either the holding of the options or the exercise of the options to the announcement price movements. Overall, this provides some evidence consistent with hypotheses H2a and H2b, although in the case of H2a, alternative explanations that motivate this hypothesis may be more appropriate. 5. Conclusions This paper explores the repurchase and option activity for nonfinancial firms for the period 1995-1999. We draw upon a variety of theories that could motivate the concurrent rise in stock repurchase activity and use of employee and executive stock options. We then explore the relation between various measures of stock option activity and repurchase behavior. In a second set of tests, we explore the relation between these measure of option activity and the stock returns of the firms prior to and during the repurchase announcements. In the first set of tests, we find that both firm-level outstanding options and executivelevel outstanding options are strong predictors of repurchase activity. However, when option grants are considered together, only executive-level grants are strongly related to repurchases. The results employing exercises are similar to those for total options outstanding. Our results are consistent with firms repurchasing shares to avoid dilution from a stock option program, particularly with executive programs. 19 In the second set of tests, we find that the stock returns that preceed a repurchase announcement are less negative for firms with executive options outstanding and for firms with executive option grants. Firm-level option measures are uncorrelated with preannouncement returns. Announcement returns are only consistently related to the granting of executive-level options. In this case, investors react less enthusiastically to stock repurchases when firms are also granting executive options. The evidence indicates that markets react more consistently with the agency explanations for the concurrent rise in stock repurchases and option use. Specifically, the price reaction is consistent with option grants to executives representing a real cost to the firm ex post in the form of stock buybacks. 20 References Barth, M.E., and R. Kasznik, 1999, “Share repurchases and intangible assets,” Journal of Accounting and Economics, 28 (2): 211-241. Core, J., W. Guay, and S. P. Kothari 2000, “The economic dilution of employee stock options: Diluted EPS for valuation and financial reporting, Working paper. Core, J., and W. Guay, 2001, “Stock plans for non-executive employees,” forthcoming, Journal of Financial Economics 61, August 2001. Fenn, G.W., and J.N. Liang, 2001, “Corporate payout policy and managerial stock incentives,” Journal of Financial Economics 60, 45-72. Guay, W. and J. Harford, 2000, “The cash flow permanence and information content of dividend increases vs. repurchases,” Journal of Financial Economics 57, 385-415. Grullon, G., and R. Michaely, 2000, “Dividends, share repurchases, and the substitution hypothesis,” Cornell University Working Paper. Healy, P.M., and K.G. Palepu, 1995, “The challenges of investor communication: The case of CUC International, Inc.” Journal of Financial Economics, 38 (2): 111-140. Ikenberry, D., J. Lakonishok, and T. Vermaelen, 2000, “Stock repurchases in Canada: Performance and strategic trading,” Journal of Financial Economics, forthcoming. Jagannathan, M., C.P. Stephens, and M.S. Weisbach, 2000, “Financial flexibility and the choice between dividends and stock repurchases,” Journal of Financial Economics 57, 355-384. Jolls, C., 1998, “Stock repurchases and incentive compensation,” NBER Working Paper # 6467. Lakonishok, J., and T. Vermaelen, 1990, “Anomalous price behavior around purchase tender offers, Journal of Finance, 45: 455-477. Liang, J.N., and S.A. Sharpe, 1999, “Share repurchases and employee stock options and their implications for S&P 500 share retirements and expected returns,” Working Paper (Federal Reserve Board). Weisbenner, S.J., 2000, “Corporate share repurchases in the 1990s: What role do stock options play?” Working Paper (Federal Reserve Board). 21 Table 1: Sample characteristics Panel A: Distributional statistics Variable Statistic N Repurchases / Market value of equity Mean Std Dev Median Minimum Maximum Mean Std Dev Median Minimum Maximum Mean Std Dev Median Minimum Maximum Mean Std Dev Median Minimum Maximum Mean Std Dev Median Minimum Maximum Mean Std Dev Median Minimum Maximum 1995 134 2.65 2.54 1.75 0.11 8.51 4.15 7.10 1.72 0.27 41.82 0.71 1.02 0.27 0.00 4.01 1.20 2.02 0.40 0.01 11.15 0.19 0.30 0.06 0.00 1.19 0.36 0.70 0.09 0.00 3.43 0.14 0.28 0.01 0.00 1.07 1996 203 3.25 3.39 2.07 0.10 12.39 6.13 8.87 2.94 0.33 39.47 0.84 1.02 0.42 0.00 3.81 2.42 4.07 0.85 0.02 16.69 0.31 0.50 0.10 0.00 1.99 0.58 1.03 0.19 0.00 4.29 0.15 0.27 0.04 0.00 1.13 1997 270 2.79 2.63 1.94 0.07 9.20 7.48 10.80 3.70 0.34 42.60 1.09 1.36 0.54 0.01 4.88 3.00 4.57 1.18 0.07 17.41 0.38 0.60 0.14 0.00 2.50 0.74 1.26 0.25 0.01 5.01 0.16 0.26 0.03 0.00 1.01 1998 383 4.51 4.86 2.76 0.20 18.80 10.71 15.51 4.37 0.65 62.89 0.90 1.15 0.44 0.00 4.50 4.16 6.50 1.49 0.09 24.74 0.49 0.68 0.18 0.00 2.79 0.82 1.41 0.25 0.00 5.56 0.21 0.35 0.04 0.00 1.34 1999 285 4.07 4.21 2.63 0.12 15.78 11.37 17.19 5.10 0.72 67.09 0.95 1.23 0.46 0.00 5.14 3.82 6.91 1.47 0.13 30.85 0.41 0.53 0.20 0.00 2.02 0.70 1.40 0.21 0.00 5.95 0.12 0.20 0.02 0.00 0.78 All 1275 3.65 3.94 2.35 0.07 18.80 8.75 13.66 3.70 0.27 67.09 0.92 1.19 0.44 0.00 5.14 3.25 5.62 1.18 0.01 30.85 0.39 0.58 0.15 0.00 2.79 0.69 1.27 0.21 0.00 5.95 0.16 0.29 0.03 0.00 1.34 Firm options’ value / Market value of equity Executives options’ value / Market value of equity Firm options’ grants / Market value of equity Executives options’ grants / Market value of equity Firm options’ exercises / Market value of equity Executives options’ exercises / Mean Market value of equity Std Dev Median Minimum Maximum 22 Table 1: Sample characteristics (continued) Variable Statistic N Executives restricted stock Mean grants /Market value of equity Std Dev Median Minimum Maximum 1995 134 0.002 0.005 0.000 0.000 0.029 1996 203 0.003 0.007 0.000 0.000 0.025 1997 270 0.004 0.007 0.000 0.000 0.023 1998 383 0.004 0.008 0.000 0.000 0.028 1999 285 0.003 0.007 0.000 0.000 0.028 All 1275 0.003 0.007 0.000 0.000 0.029 Panel B: Spearman correlations EOV/MV FOG/MV ERSG/MV Firm options’ value / MV Executives options’ value / MV Firm options’ grants / MV Executives options’ grants / MV Firm options’ exercises / MV Executives options’ exercises / MV 0.11 0.89 0.14 EOG/MV FOE/MV EOE/MV 0.26 0.43 0.32 0.76 0.15 0.75 0.13 0.00 0.47 0.05 0.24 0.16 -0.07 -0.14 -0.07 -0.05 -0.04 -0.05 Notes: This table provides summary statistics for the sample of 1,275 open-market repurchase announcement by nonfinancial firms used in subsequent tests. Where multiple repurchases occurred in a single year, the first announcement only is used. Repurchases are stated at the value of the announced quantity and the price at the date of the announcement. Market value of equity is the value of outstanding common equity at the beginning of the announcement year. Firm options’ value, grants, and exercises are the value of all employee options outstanding at the end of the year, options granted during the year, and options exercised during the year, respectively, as disclosed in the firms’ 10-K documents for the year. Executives options’ value, grants, and exercises are the value of in-the-money options granted , the value of all options granted during the year, and the value of all options exercised during the year, respectively, for the firms’ top five executives as disclosed in the firms’ 10-K documents for the year. Executives restricted stock grants are grants of restricted stock during the year to the firms’ top five executives as disclosed in the firms’ 10-K documents for the year. All spearman correlations are statistically significant at the 5% level except the correlation between the firm options’ value and the executives options’ exercises. 23 Table 2: Variable summary statistics Variable Panel A: Full sample Firm options’ value / Market Value of Equity Executives options’ value / Market Value of Equity Firm options’ grants / Market Value of Equity Executives options’ grants / Market Value of Equity Firm options’ exercises / Market Value of Equity (%) (%) (%) (%) (%) 2548 2548 2548 2548 2548 2548 2548 2548 2548 2548 2548 2548 (%) (%) (%) 2548 2548 2548 2007 2016 7.94 0.93 2.90 0.36 0.62 0.16 0.003 0.08 0.12 2.26 0.19 7.04 23.93 4.08 0.00 1.47 1.42 12.33 1.19 5.01 0.55 1.14 0.28 0.007 0.13 0.12 1.83 0.16 1.53 60.60 7.01 0.01 2.04 2.00 3.62 0.44 1.10 0.13 0.21 0.03 0.000 0.05 0.10 1.73 0.17 6.90 14.68 0.92 0.00 1.28 1.24 0.27 0.00 0.01 0.00 0.00 0.00 0.000 0.00 -1.91 0.56 0.00 2.46 -81.08 0.04 0.00 -15.00 -15.00 67.09 5.14 30.85 2.79 5.95 1.34 0.029 1.78 1.89 30.55 1.61 12.91 896.15 31.59 0.06 26.08 26.00 N Mean Std Dev Median Minimum Maximum Executives options’ exercises / (%) Market Value of Equity Executives restricted stock (%) grants / Market Value of Equity Capital expenditures / Sales Net operating cash flow / Market Value of Equity Market to book ratio Leverage Log(assets) Prior year’s total stock return Executives’ shares held / Total shares outstanding Directors’ shares held / Total shares outstanding Basic earnings per share Fully-diluted earnings per share (%) (%) 24 Table 2: Variable summary statistics (continued) Variable Panel B: Firm-years with repurchases Executives restricted stock (%) grants / Market Value of Equity Capital expenditures / Sales Net operating cash flow / Market Value of Equity Market to book ratio Leverage Log(assets) Prior year’s total stock return Executives’ shares held / Market Value of Equity Directors’ shares held / Market Value of Equity Basic earnings per share Fully-diluted earnings per share Notes: This table provides summary statistics for the variables used to test hypothesis 1. Firm options’ value is the value of all outstanding employee options at the end of the firm’s fiscal year. Market value of equity is the value of outstanding common equity at the beginning of the year. Executive options’ value is the value of inthe-money options granted to the firm’s top five executives. Capital expenditures/Sales is the value of cash flows used to purchase additional capital assets, deflated by total sales for the year. Total pretax income is segregated into operating and non-operating components with operating income being as disclosed on the income statement. Market-to-book ratio is the market value divided by book value of the common equity of the firm. Leverage is the ratio of long term debt to book value of common equity. Log(assets) is the natural logarithm of total assets reported. Statistics found in Table 1 are not repeated for the subsample presented in Panel B. N Mean Std Dev Median Minimum Maximum 1275 1275 1275 1275 1275 1275 0.00 0.09 0.13 2.17 0.19 7.21 15.59 4.19 0.00 1.62 1.57 0.01 0.14 0.10 1.57 0.17 1.52 50.12 7.33 0.01 2.11 2.07 0.00 0.05 0.11 1.70 0.17 7.07 9.37 0.87 0.00 1.34 1.30 0.00 0.00 -0.34 0.56 0.00 3.64 -81.08 0.04 0.00 -8.23 -8.23 0.03 1.78 1.13 15.25 1.52 12.78 395.61 31.59 0.06 26.08 26.00 (%) (%) (%) (%) (%) 1275 1275 1275 1035 1038 25 Table 3: Repurchase activity and option value Repurchasesit = β0 + β1 Total Optionsit + β2 Executive Optionsit + β3 Financial Characteristicsit + β4 Year Indicatorst + εit ( ***, **, and * denote significance levels at 1%, 5% and 10% levels respectively) Variable Intercept Firm options’ value / Market value of equity Executives options’ value / Market value of equity Executives restricted stock grants/ Market value of equity Capital expenditures / Sales Net operating cash flow / Market value Market to book ratio Leverage log(assets) Prior year’s total stock return Executives’ shares held / Total shares outstanding Directors’ shares held / Total shares outstanding Year dummy variables Log likelihood Notes: (1) -4.026*** (-5.65) 0.020** (2.11) (2) -3.907*** (-5.65) (3) -4.400*** (-6.02) 0.020** (2.10) 0.255*** (2.38) 40.505*** (2.55) -0.107 (-0.12) 5.345*** (4.69) -0.283*** (-3.20) -2.648*** (-3.46) 0.157* (1.87) -0.015*** (-6.68) 0.014 (0.87) -5.095 (-0.44) YES*** -4581 40.206*** (2.53) 0.003 (0.00) 5.287*** (4.62) -0.337*** (-3.75) -2.780*** (-3.61) 0.159* (1.90) -0.018*** (-7.34) 0.008 (0.51) -4.019 (-0.35) YES*** -4580 0.254*** (2.37) 39.309*** (2.48) -0.033 (-0.04) 5.408*** (4.74) -0.315*** (-3.50) -2.761*** (-3.59) 0.205*** (2.37) -0.017*** (-7.05) 0.010 (0.65) -5.190 (-0.45) YES*** -4578 This table displays regression results for the 2,548 firm years for which data is available to estimate the model. Details of the variables are described in Tables 1 and 2. The model is estimated using a tobit estimation technique due tot he mass of observations at zero for the dependent variable. 26 Table 4: Repurchase activity and option grants Repurchasesit = β0 + β1 Total Option Exercisesit + β2 Executive Option Exercisesit + β3 Financial Characteristicsit + β4 Year Indicatorst + εit ( ***, **, and * denote significance levels at 1%, 5% and 10% levels respectively) Variable Intercept Firm options’ grants / Market value of equity Executives options’ grants / Market value of equity Executives restricted stock grants/ Market value of equity Capital expenditures / Sales Net operating cash flow / Market value Market to book ratio Leverage log(assets) Prior year’s total stock return Executives’ shares held / Total shares outstanding Directors’ shares held / Total shares outstanding Year dummy variables Log likelihood Notes: (1) -3.960*** (-5.67) 0.050*** (2.23) (2) -4.050*** (-5.93) (3) -4.319*** (-6.13) 0.035* (1.57) 0.793*** (3.86) 40.631*** (2.56) -0.084 (-0.09) 5.333*** (4.67) -0.292*** (-3.31) -2.652*** (-3.46) 0.154* (1.85) -0.015*** (-6.74) 0.013 (0.79) -4.078 (-0.36) YES*** -4580 37.135*** (2.34) 0.013 (0.01) 5.405*** (4.77) -0.308*** (-3.50) -2.763*** (-3.62) 0.168** (2.05) -0.015*** (-6.65) 0.005 (0.28) -4.896 (-0.43) YES*** -4576 0.734*** (3.52) 36.898*** (2.33) 0.888 (1.00) 5.470*** (4.82) -0.299*** (-3.39) -2.745*** (-3.59) 0.195*** (2.33) -0.015*** (-6.49) 0.006 (0.35) -4.939 (-0.43) YES*** -4574 This table displays regression results for the 2,550 firm years for which data is available to estimate the model. Details of the variables are described in Tables 1 and 2. The model is estimated using a tobit estimation technique due tot he mass of observations at zero for the dependent variable. 27 Table 5: Repurchase activity and option exercises Repurchasesit = β0 + β1 Total Option Grantsit + β2 Executive Option Grantsit + β3 Financial Characteristicsit + β4 Year Indicatorst + εit ( ***, **, and * denote significance levels at 1%, 5% and 10% levels respectively) Variable Intercept Firm options’ exercises / Market value of equity Executives options’ exercises / Market value of equity Capital expenditures / Sales Net operating cash flow / Market value Market to book ratio Leverage log(assets) Prior year’s total stock return Executives’ shares held / Total shares outstanding Directors’ shares held / Total shares outstanding Year dummy variables Log likelihood Notes: (1) -3.782*** (-5.48) 0.226*** (2.35) (2) -3.590*** (-5.32) (3) -3.919*** (-5.66) 0.212** (2.21) 0.976*** (2.40) -0.078 (-0.09) 5.301*** (4.63) -0.309*** (-3.49) -2.606*** (-3.39) 0.147* (1.78) -0.016*** (-6.95) 0.012 (0.76) -0.780 (-0.07) YES*** -4584 -0.097 (-0.11) 5.073*** (4.44) -0.367*** (-4.07) -2.617*** (-3.41) 0.141* (1.72) -0.017*** (-7.22) 0.008 (0.52) -0.812 (-0.07) YES*** -4583 0.921*** (2.26) -0.135 (-0.15) 5.156*** (4.51) -0.348*** (-3.85) -2.613*** (-3.41) 0.169** (2.04) -0.016*** (-7.11) 0.009 (0.56) -1.172 (-0.10) YES*** -4581 This table displays regression results for the 2,550 firm years for which data is available to estimate the model. Details of the variables are described in Tables 1 and 2. The model is estimated using a tobit estimation technique due tot he mass of observations at zero for the dependent variable. 28 Table 6: Mean Abnormal Returns on and around Board Authorizations of Open Market Share Repurchases 1995 – 1999 (***, **, and * denote significance levels at 1%, 5% and 10% levels respectively using a two-tailed test) Panel A: Number of repurchases per firm Frequency 1 239 2 206 3 104 4 37 5 12 Panel B: Repurchases by Year N All firms 1995 1996 1997 1998 1999 1171 177 238 272 311 173 -20 to -3 Mean Abnormal Returns -2 to +2 +3 to +10 0.82*** 1.31*** 1.91*** 0.26 0.27 0.71 1.35** 0.28 0.03 0.83* -0.02 1.03 Percent 39.97 34.45 17.39 6.19 2.01 Cumulative Frequency 239 445 549 586 598 -3.27*** -4.19*** -1.88*** -2.48*** -4.69*** -2.94*** Panel C: Percentage Authorized by Year N 171 392 451 113 44 Mean Abnormal Returns -20 to -3 -2 to +2 +3 to +10 -2.04** 0.50 -0.33 -3.01*** 0.72** 0.76** -4.10*** 0.79** 0.51 -2.90*** 2.14*** 0.36 -2.82** 0.01 0.41 Mean Abnormal Returns -20 to -3 -2 to +2 +3 to +10 -5.72** 6.41 -0.72 -4.90** 2.15* 0.72 -3.30*** 1.10 -0.25 -3.49*** 0.75** 0.46* 0.06 -0.41 0.28 -4.76*** 0.09 0.90 -8.76*** 3.69*** 0.85 0 to 2.5% >2.5 <= 5% >5 <= 10% >10% <=15% Not disclosed Panel D: Stated Reason for repurchase Acquisition Purposes ACQ Employee Benefit Plans Enhance Shareholder Value ESV General Corporate Purpose GEN Offset Dilution Effects DIL Stock Option Plan STP Undervaluation Notes: N 5 37 104 500 47 89 43 This table provides the mean cumulative abnormal returns prior to, at, and following the announcement of a decision by the Board of Directors to undertake open market stock repurchases. The 1,171 announcements are segregated into the number of announcements per firm in Panel A, the announcements by year in Panel B, and the percentage of total outstanding common equity authorized in the repurchase in Panel C. Panel D displays the returns by reason given in the announcement for the 825 announcements where such a reason was given. 29 Table 7: Market reaction to repurchase announcements Returnit = β0 + β1 Optionsit + β2 Year Indicatorst + εit (t-statistics in parentheses; ***, **, and * denote significance levels at 1%, 5% and 10% levels respectively) Panel A: Options outstanding Return for [-20, -3] Variable Intercept Firm options’ value / Market value of equity Executives options’ value / MV of equity Executives’ restricted stock grants/ MV Leverage log(assets) Year dummy variables Adjusted R-square Panel B: Option grants Return for [-20, -3] Variable Intercept Firm options’ grants / Market value of equity Executives options’ grants / MV of equity Executives restricted stock grants/ MV Leverage log(assets) Year dummy variables Adjusted R-square -46.25 (-0.97) -1.35 (-0.66) 1.13*** (4.72) YES 3% (1) -11.20*** (-5.77) 0.01 (0.18) 1.51*** (2.40) -56.11 (-1.17) -1.61 (-0.79) 1.26*** (5.32) YES 4% (2) -12.37*** (-6.41) (3) -12.17*** (-6.16) -0.03 (-0.48) 1.60*** (2.44) -56.85 (-1.18) -1.63 (-0.80) 1.24*** (5.11) YES 3% 38.03 (1.10) -0.46 (-0.31) 0.03 (0.19) YES 1% (4) 1.03 (0.74) -0.12*** (-2.47) -1.38*** (-3.05) 47.71 (1.38) -0.21 (-0.14) 0.01 (0.07) YES 1% Return for [-2, +2] (5) 1.22 (0.88) (6) 1.74 (1.22) -0.09* (-1.74) -1.17*** (-2.48) 45.77 (1.33) -0.26 (-0.18) -0.05 (-0.28) YES 1% -46.60 (-0.97) -1.35 (-0.66) 1.14*** (4.77) YES 3% (1) -11.31*** (-5.80) 0.01 (0.35) 0.77*** (2.54) -50.05 (-1.05) -1.44 (-0.71) 1.28*** (5.37) YES 4% (2) -12.78*** (-6.49) (3) -13.07*** (-6.34) 0.01 (0.48) 0.77*** (2.56) -50.44 (-1.06) -1.44 (-0.70) 1.31*** (5.30) YES 4% 38.83 (1.12) -0.44 (-0.30) 0.14 (0.82) YES 0% (4) 0.10 (0.07) -0.00 (-0.09) 0.20 (0.90) 37.83 (1.10) -0.46 (-0.31) 0.19 (1.08) YES 0% Return for [-2, +2] (5) -0.37 (-0.26) (6) -0.35 (-0.23) -0.00 (-0.05) 0.20 (0.90) 37.86 (1.10) -0.46 (-0.31) 0.18 (1.03) YES 0% 30 Table 7: Market reaction to repurchase announcements (continued) Panel C: Option exercises Return for [-20, -3] Variable Intercept Firm options’ grants / Market value of equity Executives options’ grants / MV of equity Leverage log(assets) Year dummy variables Adjusted R-square Notes: This table displays regression results for the 893 firm years for which data is available to estimate the model. Where more than one announcement occurred in a single year for a firm, the first announcement only was used. Panels A, B, and C estimate the model for the value of options outstanding, the value of options granted during the year, and the value of option exercised during the year, respectively. Details of the variables are described in Tables 1 and 5. The model is estimated using OLS. Return for [-2, +2] (3) (4) 0.38 (0.27) -0.18 (-0.87) 1.00 (1.15) -0.35 (-0.24) 0.12 (0.70) YES 0% -0.46 (-0.31) 0.18 (1.05) YES 0% (5) -0.22 (-0.16) (6) 0.04 (0.03) -0.19 (-0.92) 1.04 (1.20) -0.44 (-0.30) 0.15 (0.87) YES 0% (1) -11.14*** (-5.84) -0.02 (-0.08) (2) -11.13*** (-5.84) -11.10*** (-5.70) -0.02 (-0.08) -0.13 (-0.11) -1.44 (-0.70) 1.11*** (4.74) YES 3% -1.43 (-0.70) 1.11*** (4.77) YES 3% -0.13 (-0.11) -1.43 (-0.69) 1.11*** (4.68) YES 3% 31