The timing of stock option grants has been a controversial issue in recent years.
Toronto Stock exchange rules prohibit the award of ‘in-the-money’ options. In other words, the exercise price of a stock option cannot be less than the market price of the stock when the option is granted. U.S. academics and regulators have found evidence that managers systematically gained by granting themselves and their employees ‘in-the-money’ options which were erroneously reported as ‘at-the-money’ options. They did this by reporting a false grant date to both tax authorities and the investing public. The false date was a day when the stock price was lower than on the date of the actual option grant. This practice, known as backdating, enriches managers at the expense of other shareholders. In 2008, the OSC charged Mike Lazaridis and Jim Balsillie (of RIM inc.) with option backdating. In February of 2009, Balsillie and Lazaridis agreed to pay $77 million to settle the OSC backdating charges.
An alternative method for managers to enrich themselves with stock options is spring-loading. With spring-loading, managers issue option grants on a date that precedes or follows news releases they know will have a positive or negative impact on the stock. By so doing, they use their asymmetric information to benefit themselves much in the same way as if they had purchased shares in the company prior to a good news announcement.
A number of U.S. studies have found evidence consistent with option backdating and spring- loading. Yermack (1997) examines the stock returns around option grants to U.S. CEOs between 1992 and 1994. He finds no abnormal stock returns before option grant dates but significant evidence of abnormal returns in the 50 days after grant dates. He studies earnings announcements after option grant dates and concludes that CEOs time option grant dates to occur prior to favourable earnings announcements.
Aboody and Kasznik (2000) study over 2,000 U.S. option grants over the period from 1992 to 1996. They conjecture that CEOs manipulate news announcements and not option grant dates. To test this, they analyze options that are granted on a regular annual basis (scheduled option grants). Those grant dates cannot, by definition, be manipulated by insiders. Then they analyze news announcements around grant dates and find that companies tend to rush bad news to precede option grant dates but delay good news to occur after the grant dates.
Lie (2005) and Heron and Lie (2007) analyze approximately 6,000 option grants to CEOs between 1992 and 2002. They find a pattern of declining prices prior to unscheduled option grants and rising prices afterwards. Thus, many of the companies in the Heron and Lie sample appear to have awarded their options on the day with the lowest price during the three-month span around the grant date. They observe that the timing of many unscheduled grants is too good to be lucky and conclude that many companies engage in option backdating.
The only published empirical analysis of option grant timing in Canada is Chourou and Abaoub (2008). They studied 443 unscheduled stock options to CEOs granted between January 1, 2001 and August 29, 2004. Chourou and Abaoub (2008) report a -1.72% cumulative abnormal return in the 30 days before reported option grant dates and a 1.50% cumulative abnormal return afterwards. For a subsample of 106 option grants to CEOs of cross-listed companies, they find no significant pattern following the August 29, 2002 shortening of the U.S. reporting deadline. Compton et al. (2009) explain that research on Canadian option grants is scarce because the cost of building a sample is very high. This concern has been echoed by lawyers investigating backdating in Canada. Our paper addresses the need for a comprehensive study on the timing of Canadian option grants. We analyze over 15,000 option grants involving 1,400 Toronto Stock exchange-listed companies over a period of more than ten years. We do not find evidence consistent with either backdating or spring-loading. We do find evidence of significant abnormal returns on average following unscheduled option grants.
Since the average is statistically significant, this means that the abnormal returns are experienced by a large proportion of companies. This result is consistent with a simple hypothesis: that companies award unscheduled options when they believe their shares to be undervalued. This, corporate insiders are able to profit from the information advantage they enjoy over outside investors.
Changes in Option Grant Disclosure
There were two significant changes in the regulation of option grants over our sample period: 1) a change in the disclosure timing; and 2) the introduction of SEDI (System for electronic disclosure by insiders).
The Ontario Securities Act (OSA) requires that insiders disclose option grants to the public. Before December 14, 1999, the OSA required disclosure within ten days of the beginning of the month following the option grant. After Dec. 14, 1999, the reporting window was shortened to ten days following the grant date. The second big change was the introduction of SEDI—a web-based reporting system—on June 9, 2003. McNally and Smith (2010) document how both changes improved the timeliness and accuracy of reported insider trading information. Both changes made malfeasance more visible, and should have caused a reduction in fraudulent behaviour. We divide our sample into thirds (around the two dates) to evaluate the impact of each regulatory change.
In April 2010, the insider reporting window was shortened further to five days and an alternative method of option grant reporting by issuers rather than insiders was introduced. While it would be interesting to analyze these changes, this is outside our sample period.
Data
Our option grant data is obtained from the Micromedia and SEDI databases for the period from July 1, 1995 to December 31, 2006 we include option grants if the stock traded on the Toronto Stock Exchange and if we have news data for the company. The news data are obtained from two sources: 1) for the period July 1995 to December 2000 from CANSTOCK, an electronic database of all public newswire releases for the TSX; and 2) for the period January 2000 to December 2006 from The Globe and Mail (a division of CTV Globemedia Publishing inc.). We classified news announcements as good or bad if the abnormal return on the day was positive or negative.
We identified 71,101 grant reports by insiders in 1,427 companies. In many cases, multiple insiders receive options on the same day. Our interest is in the stock price behaviour around the option grant date, so each option grant date is included only once, regardless of how many insiders receive options on that date. In our final sample, we have 15,128 option grant days in 1,427 companies over the period from July 1995 to December 2006.
Table 1 shows the sample of option grant dates by year and distinguishes between scheduled and unscheduled option grants. Following Lie (2005), scheduled grants are grants which occur within one week of the anniversary of the previous year’s option grant. Click on image to enlarge.
The other grants are classified as unscheduled. Just under a quarter (23%) of option grants occur on a regular schedule whereas three-quarters are granted at the discretion of the board of directors.
Analysis
Figure 1 presents cumulative average abnormal returns (CARS) around option grant dates. We calculate CARS separately for scheduled and unscheduled grants. We find that there are no significant negative returns over the 20-day window prior to scheduled or unscheduled grants. The finding is not consistent with backdating as companies are not selecting the date of the lowest stock price in the prior 20-day window. We do find positive returns over the 20 days following option grants, and the returns are much larger for unscheduled grants. We interpret these results as being consistent with the argument that those who award unscheduled option grants use their inside knowledge of the company to award grants when their stock price is undervalued. This is similar to the market timing ability observed by companies when they repurchase their own shares as reported in papers such as Smith and McNally (2007).
We note that this finding reflects an average across a very large sample of companies and option grants. This does not mean that there are not individual companies backdating their options. For example, RIM executives admitted to backdating. When we examined the timing of RIM option grants, we found that most RIM options are unscheduled and we observed a clear v-shaped pattern in the CARS. In general, our results reflect the behaviour of the majority of Canadian companies.
Table 2 presents another measure of abnormal stock returns around option grant dates, the buy-and- hold abnormal return (BAHR). We report average BAHRs over the 20 trading days before and after option grant dates for the three sub-periods around the regulatory changes. Click image to enlarge.
Consistent with the results from Figure 1, we observe significant abnormal positive returns following unscheduled grants in all three regulatory regimes. In the final period of the sample (after the introduction of SEDI) we find an average abnormal return of 2.87%. Keep in mind that this is for one month and is the average over many option grants. The abnormal returns following scheduled grants are positive but smaller in each regime than those following unscheduled grants. The returns following the scheduled grants are significantly positive in the first two regulatory regimes but not in the most recent (after SEDI). We observe no significant pattern of returns prior to option grant dates. This result is different than those found by Lie (2005) for the U.S. lie observed a significant decline preceding unscheduled grants. This, our results are not consistent with backdating.
Next we analyze news announcements around scheduled option grant dates. Table 3 presents the average number of good and bad news announcements in the month before and after option grants for the three sub-periods. Many firms announce good news in the month after their option grants, but there is no consistent pattern of increased frequency of good news announcements in the month after grant dates (compared to the month before). Nor do we find a pattern of increased bad news in the month preceding the option grant. We do not find evidence that companies (on average) are moving bad news before and good news after their option grant dates. While we cannot reject the hypothesis that some firms are spring-loading, it does not appear that the majority of firms are manipulating news announcements.
Conclusions
Our results indicate that most Canadian companies with unscheduled option grants award them just prior to abnormal increases in their stock prices. We do not find that the option grants follow periods of price declines, so our results do not suggest that firms backdate their grants to a date with the lowest price. We also do not find evidence that companies time their grant dates to follow bad news and precede good news, so our results are also not consistent with spring-loading.
It appears that insiders are using their information advantage to award grants on days when they know the stock is undervalued. In the subsequent month, the stock appreciates by almost 3% more than expected. This abnormal gain represents a wealth transfer from outside shareholders to insiders. The fact that this is showing up on average in our very large sample means that it is a common practice among Canadian corporations.
There were two regulatory changes during the sample period that increased the timeliness and transparency of option grants. We do not find that the regulatory changes eliminated the beneficial timing exhibited by companies granting unscheduled options. If Canadian securities regulators want to lessen the wealth transfer from uninformed outsiders to informed insiders, then they should require that all options be granted on a predetermined schedule. Meanwhile, the evidence of significant abnormal returns suggests that Canadian money managers should consider unscheduled option grants to be a signal of undervaluation.