Remuneration and Incentives of Managers: Executive Stock Options and Backdating in Canada

Transcription

1 Remuneration and Incentives of Managers: Executive Stock Options and Backdating in Canada Ryan A. Compton Jonathon N. Giller Lindsay M. Tedds First Draft: November 9, 2006 Please do not cite; Comments welcome Abstract Long held as a device to align managers interests with those of shareholders, the granting of executive stock options has come under close scrutiny due to concerns of backdating. This paper provides an overview of (1) the issues surrounding the practice of backdating in Canada; (2) an empirical examination of whether backdating may be occurring among companies on the Toronto Stock Exchange s S&P/TSX 60 index; and (3) suggestions for curbing the practice of backdating in Canada. Our results are consistent with a backdating explanation of executive stock option grant timing in Canada. Keywords: Backdating; Executive stock options; Compensation JEL Codes: J33; M52 Contact Author: Lindsay M. Tedds, Assistant Professor, Department of Economics, 501 Fletcher Argue Bldg, University of Manitoba, Winnipeg, MB, Canada, R3T 5V5. Phone: Fax: We would like to thank seminar participants at the University of Manitoba for their comments and suggestions. The usual disclaimer applies.

2 1. Introduction There has been a recent surge in popular interest and academic research regarding executive compensation that has been stimulated by both the escalation in the size of executive compensation, as well as the increasing array of components in executive compensation. 1 Generally, there are five components to most executive remuneration packages: annual salary, annual bonus, stock options, long-term incentive plans, and fringe benefits (e.g. pensions, chauffeur, and club memberships). The intent of each is to tie compensation to either short or long-term company performance. 2 While all of these components have been growing over time, the use of stock options has experienced the greatest increase over the last twenty years to become the single largest component of compensation among CEO s at large publicly traded companies in North America (Murphy, 1999 and The Conference Board, 2006). The increasing proportion of stock options in executive compensation packages can be tied to three key reasons. The first relates to the classic economic principal-agent problem of aligning the incentives of firm managers with shareholders. The share price of a company is generally regarded as being positively related to the financial performance of the company. Therefore, since executive stock options become more valuable the higher the stock price, options granted to executives should provide incentive for executives to maximize company performance and thus share price. Not surprisingly, shareholders have increasingly come to prefer a larger proportion of stock options in relation to base salary in executive compensation packages. Executives too have come to prefer this form of compensation likely because the overwhelming majority of stock options issued since 1980 have been exercised well in-the- 1 Since 1970, the median cash remuneration paid to S&P 500 CEO s has more than doubled and the median total remuneration has nearly quadrupled. (Murphy, 1999, p. 1) 2 It should also be noted that the size and composition of executive compensation is also intended to attract and retain key executives. 1

3 money. (Murphy, 1999, p. 23). The existence of this long-standing bull market means that executives have seen their stock prices rise regardless of their individual or collective managerial acumen. Since there has been little to no risk associated with this compensation component over the last twenty years due to a predominately bull market, many executives have amassed a substantial fortune from their stock options and have come to prefer them over annual cash awards. Third, executives also benefit from the preferential tax treatment of the income earned from executive stock options. For example, in Canada, executive stock options do not incur a tax liability at least until the options are exercised, and any proceeds are taxed similar to capital gains rather than employment income. Keep in mind, capital gains are taxed at a much lower rate than employment income. While executive stock options have gained prominence as a compensation and incentive mechanism, recent work in the U.S. has raised serious concerns about stock options as a major component of executive remuneration. These concerns raise questions about whether executive stock options align principal-agent interests or simply provide an avenue for managers to extract resources from their firm and evade personal taxes through the practice of backdating. 3 Backdating is the act of choosing a date for a stock option grant after that date has occurred, and claiming to have granted the options at that date, in order to take advantage of the historical price performance of a company s stock. In practice this would involve looking back to find a local low point for the underlying stock relative to the current day s stock price and choosing that low point for the option s grant date. Under this practice, executive stock options are claimed as being granted not-in-the-money on the date the share was trading at its lowest, however, given 3 Yermack (1997) is the first to document strange patterns surrounding the grant dates of executive stock options. Lie (2005), Narayanan and Sheyun (2005, 2006a, 2006b), Heron and Lie (2006), Heron and Lie (forthcoming), and Narayanan, Schipani and Seyhun (forthcoming) represent the new literature which examines whether firms are engaging in backdating specifically. 2

4 that the actual date the decision is made and grants are actually granted is after that date, the options are in reality likely granted in-the-money. This raises many issues such as this practice may not provide as much incentive for executives to increase company performance than it would otherwise. Further, if the options are reported as being granted not-in-the money then income taxes are being evaded. The results reported in recent empirical literature on backdating are suggestive of such a practice. Specifically, the results, which focus on the U.S., tend to show that stock returns are unusually low prior to the granting of executive stock options but then rebound rather sharply following the granting of the option. While some may argue this is coincidence or perhaps even evidence of impressive forecasting on behalf of the company, this result has raised concerns in the U.S. among regulators and shareholders concerning company practices. Although these studies are careful not to mention specific companies, a number of media outlets have identified companies with questionable granting practices and have investigated the possibility that the dating of stock options that has proved to be highly lucrative for executives has not been a mere coincidence. 4 Only recently has this issue garnered attention in Canada with the release of a September 2006 report on backdating by Sam La Bell and Chris Silvestre of Veritas Investment Research. The report examines executive stock options filed since June 2003 for S&P/TSX 60 companies, and considers for each day beginning ten days before and ending fifteen days following a grant date, the average percentage difference between the grant date price and the closing price. Their findings provide preliminary evidence which indicates that share prices were, on average, 50 basis points higher 10 days prior to the grant date and more than 100 basis points higher 15 days following the grant date. This report brought the issue of backdating in the Canadian context to 4 See for example the November 11, 2005, May 5, 2006 and May 22, 2006 editions of the Wall Street Journal. 3

5 light with stories appearing in the Financial Post and Canadian Business. 5 Shortly following this report, on September 28, Research in Motions (RIM) announced the audit committee of its board of directors would be initiating a voluntary review of its stock option grants, and a month later, the U.S. Securities and Exchange Commission began an informal inquiry into RIM s stockoption grant program. The audit committee has made a preliminary determination that a restatement of RIM s historical financial statements will be required, which has subsequently delayed RIM s filing of its current financial startements. 6 While this issue is generating attention among shareholders, regulators, and policy makers alike in Canada, academic work on backdating in Canada is nonexistent. As a result, this paper has a number of goals. First, the practice of backdating is explained in detail. Second, the tax, accounting, legal and policy issues that arise due to backdating are discussed, with special focus on Canada. Third, using data from companies listed on the S&P/TSX 60 index from 2003 to 2006, we examine empirically whether the evidence is consistent with a backdating explanation of price behaviour surrounding executive stock option grants. 7 Finally, recommendations for constraining the practice of backdating are discussed, and avenues for future research in this area are detailed. This paper contributes to the existing literature on backdating on several fronts. First, the Canadian regulatory framework is different from that which exists in the United States. The Toronto Stock Exchange requires that all stock options be granted not-in-the-money whereas no such requirement exists in the United States. Under certain circumstances, backdating of stock 5 See the September 12, 2006 edition of the Financial Post, and September 25-October 8, 2006 issue of Canadian Business. 6 See RIM press release at 7 This paper extends the work of La Bell and Silvestre (2006) by considering scheduled and unscheduled options separately, by modeling abnormal stock returns (rather than raw returns) within a 30 day window of the grant date, and considering all companies listed on the S&P/TSX 60 at any point during our coverage. 4

6 options in the U.S. is perfectly legal. Hence, when using U.S. data it is impossible to disentangle legal backdating from illegal backdating when evidence of the activity is found in the data. Whereas by using Canadian data, all backdating is illegal, meaning any evidence of backdating in Canada raises a serious regulatory issue. Second, in the U.S., as of August 2002, executives are required to disclose an option grant within two business days following the date of the grant In Canada this window is ten calendar days. It is important to know if the longer filing window results in stronger evidence of backdating. Third, examining Canadian data provides an opportunity to determine whether or not the widely documented empirical evidence of backdating holds for a different mix of firms. Canadian firms are generally much smaller than U.S. firms, and there is a substantial difference in industry structure between publicly traded companies in Canada and the U.S., with Canada have more publicly traded resource and agricultural-based companies. Hence, examining Canadian data provides an opportunity to examine backdating with a country that has different industry and firm characteristics. Finally, focusing on Canada affords the occasion to investigate the influence of a different tax regime on backdating. In particular, stock options are taxed more favourably, at least at the personal level, in Canada than in the United States with most options in Canada being taxed at capital gain rates whereas in the United States most are taxed at full rates. 8 Previewing our results, while it is not outside the realm of possibility that something else might explain the paper s findings, the evidence is suggestive of the occurrence of backdating in Canada. Given these findings, a key question for Canadian policymakers must be how to reduce this practice in Canada, and for academics and regulators alike, what further analysis is needed in the investigation of this issue? Our conclusion provides initial answers on both fronts. 8 For a detailed discussion of the tax treatment of stock options, including a comparison of the regimes in Canada and the U.S., see Sandler (2001). 5

7 2. Overview of Executive Stock Options and Backdating What are Stock Options Generally speaking, a stock option is a financial instrument which provides the holder, the right, but not the obligation, to buy or sell stock of a corporation within a stated period of time at a specified price. It is important to note however, that in the case of executive stock options, a number of differences exist relative to standard stock options, which deserve highlighting. Unlike standard stock options, executive stock options are not traded publicly on an exchange, but rather are a private contract where the board of directors or compensation committee of the firm serves as the writers of the option, while the executive acts as the holder of the option. Second, while traditional stock options can be purchased which allow the holder to buy the underlying security at a specified price or alternatively allow the holder to sell the underlying security at a specified price, executive stock options provide the executive the right only to purchase the underlying stock. As well, there is the requirement that the executive must hold the option for a pre-specified vesting period. Finally, executive stock options are usually granted at-the-money, meaning that the exercise price of the option equals the market price of the underlying stock on the day of the option grant, where as a traditional stock option is, of course, issued out-of-the-money. Since the 1990 s, stock options have become an important component of an executive s compensation package, and have long been seen as a tool for aligning management interests with those of shareholders. The underlying intuition behind executive stock options is that with an issuance of the option at-the-money combined with the required vesting period before the option can be exercised, this provides the incentive for executives to undertake actions that will bring 6

8 about an increase in the company stock price since the executive s compensation through the option is essentially tied to a rise in the stock price. In this form, stock options are not only beneficial to the executive, but they are also beneficial to the owners of the company, the shareholders. What is Backdating and is it Illegal? Backdating is essentially the practice of choosing an executive stock option grant date retroactively based on looking back and choosing a date that coincides with a low point in the company s share price. By dating the option grant with an earlier date (the one coinciding with the low point in stock price), the stock option is essentially granted in-the-money (relative to what the stock price is currently) rather than not-in-the-money, providing the executive with a higher option value. In terms of the question of the legality of backdating, as pointed out by Erik Lie, in the case of the U.S., backdating is not necessarily illegal if certain conditions are met. 9 First, documents must not have been falsified. For example, minutes of meetings and the documents surrounding the granting of the option must properly reflect the actual date the option was approved as well was granted. Second, companies must communicate to their shareholders that backdating is occurring when executive stock options are granted in-the-money based on using prior price information to set the option exercise price. Third, when backdating options to grant them in-the-money, this must be reflected in the company s statement of earnings as well as accurately stated for tax purposes. As Lie points out, these conditions are seldom met, making the act of backdating often illegal. While clearly these conditions must also be met in the case of 9 See 7

9 companies listed in Canada, one area where the U.S. and Canada differ is in the requirements of their respective stock exchanges. U.S. exchanges such as the New York Stock Exchange (NYSE), NASDAQ, and American Exchange (AMEX) allow the granting of stock options in-the-money so long as this is disclosed to shareholders. In Canada, and in particular in the case of the Toronto Stock Exchange (TSX), stock options must be granted not-in-the-money. 10 That is, granting stock options in-themoney, even if disclosed to shareholders, does not meet the exchange s requirements. Therefore, firms listed on the Toronto Stock Exchange may not, under any circumstances, offer options inthe-money. Further, the option exercise price must not be set based on undisclosed information relevant for the price of the stock, and importantly for our purposes, the option grant must be reported within ten days of the end of the month in which the grant was made to the TSX. Beyond the requirements of the TSX, the Ontario Securities Commission and Canadian securities regulators require that all insider option grants be reported on the System for Electronic Disclosure by Insiders (SEDI) within ten calendar days of the grant. 11 Should We Care About Backdating? Should Canadian companies listed on the TSX engage in backdating, a number of issues have bearing not only for the company itself, but for the executive beneficiary, shareholders, the 10 Of note however is the TSX Venture exchange (TSX-V) which funds small market capitalization companies. The TSX-V allows firms to grant options in-the-money. Options may be granted in-the-money as long as granted at a discount within a range of 15-25%. See Canadian Securities Staff Notice for more information. In the case of companies on the TSX-V, firms can issue stock options in-the-money and conditional on meeting the above requirements, not be engaging in any illegal activities when backdating. 11 Before the advent of SEDI, insider trading information was paper based and available from individual provincial securities commissions. For example, for firms listed on the TSX, the information is available in the weekly Ontario Securities Commission Bulletin. The ten calendar day filing requirement was implemented in

10 Ontario Securities Commission, the federal and provincial revenue authorities, and taxpayers in general. 12 Consider first, the Canadian tax issues surrounding backdating of stock options granted by a publicly traded company. 13 The applicable tax rules are outlined in sections 7 and 110 of the Income Tax Act, and discussed in Canada Revenue Agency Interpretation Bulletin 113R4 and Information Circular Unlike employment income (e.g. annual salary or bonus income), which is taxable in the year it is received, there is no immediate tax consequence upon the granting of stock options, regardless of whether they are granted in- or not-in-the-money. Rather, tax liability does not accrue until at least the time the option is exercised. Therefore, executives defer their tax liability by preferring options to salary and cash bonuses. In addition, options are given preferential tax treatment when compared to the taxation of other employment income such as salary and cash bonuses. In particular, employee stock options are taxed on a similar basis as capital gains. 14 An illustration of the tax treatment of shares granted not-in themoney versus in-the-money, is useful for our discussion of why executives may prefer from a tax perspective, backdating options. Consider the tax treatment of shares granted not-in-the-money (i.e. the option price is greater than or equal to the fair market value on the grant date). When an option is exercised and where the option price is less than the fair market value of the underlying stock on the exercise date, the executive must report employment income in the year the option is exercised which is equal to the difference between the option price and fair market value multiplied by the number of shares. Because the option was granted not-in-the-money, the individual can deduct 50% of 12 Some of these issues are also highlighted in Cohen and Brille (2006) and La Bell and Silvestre (2006). 13 The tax treatment of employee, and hence executive, stock options changed in The discussion that occurs in this paper therefore considers the tax treatment in effect post As of October 18, 2000, the capital gains inclusion rate is one half or 50%. 9

11 the benefit from their taxable income. 15 The inclusion of this employment income can, however, be deferred to the year in which the shares are actually sold rather than acquired, subject to an annual limit of $100, Since this portion of the tax liability can be deferred until the option has been sold, the executive can use the funds from disposition to pay for this tax liability rather than using income from other sources. When the stock is sold (assuming the stock is sold at a price above the option exercise price the stock was acquired at), the capital is calculated as the difference between the sale price and exercise price multiplied by the number of shares. The taxable capital gain is then 50% of the total capital gain. On the other hand, if options are granted in-the-money (i.e. the option exercise price is less than the fair market value of the underlying stock on the grant date), the executive must report employment income in the year the option is exercised, which is equal to the difference between the option exercise price and fair market value multiplied by the number of shares. Because the option was granted in-the-money, no deduction or deferral can be claimed. Since the stock has not been sold at this time, the executive must pay for this tax liability from other income sources. When the stock is sold (assuming the stock is sold at a price above the exercise price), the capital gain is calculated as the difference between the sale price and exercise price multiplied by the number of shares. As before, the taxable capital gain is 50% of the total capital gain. 15 Paragraph 110(1)(d) of the Act provides for an employment income deduction to partially offset the employment income triggered when an employee exercises an employee stock option. In 2000, the deduction, previously only applicable to stock options granted by Canadian-controlled private corporations, was extended to public company stock options (subsection 7(8) of the Income Tax Act) if certain prescribed conditions were met, notably that the company is listed on a prescribed stock exchange and the options are granted not-in-the money. If the individual dies or becomes a non-resident prior to disposing of the options, then the benefit must be included in that year. 16 If the fair market value of a stock option has declined below the exercise price of previously granted stock options, the company may reduce the exercise price but this prohibits the grantee from claiming the 50% benefit reduction. 10

12 To illustrate the difference in tax liability, consider a numerical example of a fictitious executive under three difference regimes: options granted in-the money, options granted not-inthe-money, and performance pay paid in cash,. We will simplify the calculation by focusing only on the federal tax liability. 17 It is assumed the executive s taxable income falls under the highest federal marginal tax rate, which is set equal to 29% in all years. In all cases, the comparison is the tax liability of a cash award or stock options granted that amount to $100,000. For the stock option components, assume the exercise price is $15 per share and the eventual sale price of the acquired stock is $19 per share. Table 1 provides a comparison of the three different compensation regimes. If the executive is provided with $100,000 in stock options that are granted not-in-the-money (or importantly for our purposes, backdated so that they appear to have been granted not-in-the-money), the tax liability is $13,050 which is payable in the disposition year. If the executive is provided with $100,000 in stock options that are granted in-the-money, the tax liability is $23,300, a portion of which is payable in the year the option is exercised and the remainder is payable in the year the stock is sold. Lastly, if the executive is provided with $100,000 cash reward, the tax liability is $29,000 which is payable in the tax year that the compensation is received. Not only is the total tax liability the lowest in the case where the options are granted not-in-the-money, but the total is also not payable until the furthest possible date. A tax situation that reduces current tax payments in favour of future tax payments is generally preferred to one weighted towards current tax payments. Hence, the most preferential compensation regime from a tax perspective is one where the options were granted not-in-themoney (or backdated as such), having a tax liability of only $13,050 as compared to the tax 17 Provinces also tax income generated from stock options and they generally follow the federal rules. However, there is some variation, notably in Ontario where, as of December 21, 2000, employees involved in research and development that exercise or sell eligible stock options are granted a $100,000 per year tax exemption. 11

13 liability of $20,300 associated with a regime where the options were granted in-the-money and $29,000 tax liability when receiving a cash bonus. Further, this compensation regime deferred the full tax liability to disposition date, as compared to the immediate tax year for the cash award and as compared to having a tax liability in both the exercise year and disposition year for the case where the options were granted in-the-money. This demonstrates a clear tax advantage of a compensation regime where stock options are granted (or at least reported as such) as being notin-the-money. It also shows that the act of reporting options that are granted in-the-money as being not-in-the-money is an act of tax evasion. The implications could be even further reaching since it has previously been found that non-compliant corporations are three time more likely to be managed by executives who have evaded personal taxes (Joulfaian 2000). Aside from the tax implications, there are important accounting implications of backdating, as any firm that has backdated stock options may be required to restate their earnings (if it is a material amount) in order to properly reflect the compensation expense resulting from the in-the-money portion of the option. This may involve restating earnings over many past years and as a result may involve a significant reduction in past earnings. For shareholders, this represents a significant concern, and likely cost, as it is reasonable to expect the stock market to price in these reductions in earnings downward. Backdating also raises issues relating to adherence to exchange regulations and Canadian securities law. As discussed earlier, backdating essentially puts grants in-the-money, and for companies listed on the Toronto Stock Exchange, this is in clear violation of the exchange s requirements. As well, with backdating occurring without the consent of shareholders, this too represents a violation of the exchange s requirements. Further, by misrepresenting the actual grant date of the option, companies breach Canadian securities laws in terms of misleading 12

14 public disclosure. By violating exchange and legal requirements, companies are subject to significant negative reaction by financial markets, and as a result, potentially a significant reduction in shareholder value. Of course it follows this also exposes the firm to potential law suits from shareholders harmed due to backdating. Lastly, there is the obvious reason to be concerned about backdating. By backdating options such that they are in-the-money, the incentive for executives to work to increase the share price may be diminished. The granting of stock options now appear more as a reward or bonus based on the past (which itself may be optimistic), rather than an incentive to create value for shareholders in the future. Clearly, backdating and the issues surrounding the discovery of backdating can be costly for the firms, executives, and shareholders involved. In the following sections, we will examine whether the empirical evidence is suggestive of backdating at the aggregate level among companies on the S&P/TSX Sample The goal of this study is to provide an analysis of stock returns around dates when senior officers and directors at S&P/TSX 60 listed companies have been awarded executive stock options. Since June 2003, the Ontario Securities Commission and Canadian securities regulators have required all senior officers report stock options on the System for Electronic Disclosure by Insiders (SEDI) within ten calendar days of their grant. The senior officers or their designated registered agents are required to report on SEDI their name, position, option grant date, filing date, number of options granted, exercise price, and currency, among other information. The information filed on SEDI is publicly accessible through SEDI s website. 18 The option grants used in this paper are based on collected information regarding all stock options filed by senior 18 See 13

15 officers and directors on SEDI from June 2003 to October 2006 for companies listed on the S&P/TSX 60 at any time during the sample period. 19 Table 2 lists the sixty-six companies used in this paper by ticker symbol, company name, and industrial sector. Based on the SEDI information for the sixty-six companies in our sample, and after removing options granted in U.S. dollars, options that were reported with a negative number, and options which do not have enough stock data either before or after the grant for the event analysis (discussed below), we have 5,644 instances of executive stock options granted. These options are then classified as scheduled, unscheduled, or uncertain. An option grant is considered scheduled if it occurs within 30 days of the one-year anniversary of the prior year s option grant date. Occasionally, for grants at the beginning of our sample, it is difficult to determine whether they are scheduled or not, and as a result, these are deemed uncertain. Otherwise, an option that does not meet the scheduled definition is considered unscheduled (ie there was no grant given within the 30 day period of the previous year or no grants we given at all in the previous year). Tables 3-7 provide more detail for those interested. In terms of daily stock prices, we use the daily closing stock price adjusted for market effects, for each of the companies included in the sample. This is obtained from the Commodity Systems Incorporated (CSI) database. 20 We also obtain from the CSI database, the daily adjusted closing price for the S&P/TSX 300 Composite Index, which is used as the market index in the construction of our abnormal returns detailed below Note that the range of grants captured using this time frame includes a small number (35) actually granted in 2002 but not recorded until after June 2003, and thus included in our SEDI sample. 20 With the exception of Agnico-Eagle Mines Ltd. (AEM) and ATI Technologies Inc. (ATY). Share prices for these companies where not available from CSI and therefore were collected from Yahoo Finance. 21 The S&P/TSX 300 Composite Index is a value-weighted index designed to measure market activity of stocks listed on the TSX. 14

16 4. Empirical Approach and Evidence on Backdating Empirical Approach A key issue of course is how does one go about identifying the occurrence of backdating? In order to do this, we must consider the behaviour of stock prices around the executive stock option grant date. This involves for each grant date that we consider the stock return behaviour around that date in terms of daily abnormal returns using an event-study methodology based on MacKinlay (1997). 22 For this study, the event of interest is an executive stock grant and the associated event window spans from fifteen trading days before to fifteen days after the granting of the option. To estimate a normal return during the event window in the absence of the event of interest, we use data for the year ending fifty days before the grant date to estimate a market model, which assumes a linear relationship between the return of the stock of the granting firm as follows: r i,t = α + β r m,t (1) where r i,t is the daily adjusted return for stock i on day t, and r m,t is the daily adjusted return based on a value weighted market index (S&P/TSX 300). Using the estimated α and β from (1), as well as the return on the market index during the event window, the predicted normal return for ^ ^ the company s stock (rp,t ) within the event window can be determined. 23 This information is then used to generate an estimate of the daily abnormal returns within the event window, calculated as simply the difference between the observed and predicted returns as follows: AR i,t = r r ) (2) ( i, t p, t 22 Yermack (1997) and Lie (2005) employ a similar methodology for their analysis. 23 An alternative approach also commonly seen in the literature to predict stock returns is based on the Fama and French (1993) three-factor model. 15

17 where r i,t is the daily adjusted return for stock i on day t, and r p,t is the predicted daily adjusted return based on a market return model. Our primary focus, the cumulative abnormal return is calculated by simply obtaining a period-by-period sum of the abnormal return as follows: T CAR i,t = = ( r i t r 1, p, t ). (3) t The above discussion is based on a single option grant for a given company. In order to draw overall inferences, it is important to note we aggregate the period-by-period cumulative abnormal returns to provide the mean cumulative abnormal returns. Evidence Before examining the cumulative abnormal returns around grant dates, Figures 1-2 detail the number of days between the option grant date and SEDI filing date for the options in our sample. Figure 1 details the filing gap for all options in our sample, and what stands out is the large number of filings which take place towards the final days of the reporting window. While roughly 25% of grants are reported by day 5, another 58% file in the latter half of the reporting window. While clearly this is not evidence of wrongdoing (it may just be insiders are busy people and this is not first priority in their list of things to do), 58% is a sizeable number, and for those interested in backdating, the later the filing, the greater the historical window to draw on in terms of historical stock prices. More interesting perhaps is the 17% that fail to file within the required 10 day window. An immediate question is what is causing this failure to meet CSA requirements? Almost 9% filed later than 50 days after their grant date. If one were interested in backdating, and willing to claim a much earlier grant date when the stock was trading at a very low historical price (say two months earlier), they could claim that the option was granted at that date two months ago and 16

18 though filing what would appear to be two months after the fact, they would certainly not be alone in the appeared tardiness of their filing. Looking at Figure 2, which considers the reporting patterns by type of grant (scheduled, unscheduled, uncertain), the results are much as in Figure 1. Figures 3-5 provide the core results of this study, detailing the average cumulative abnormal returns for a 30 day window surrounding the option grant date (15 days on either side of the grant date). 24 Recall, a story coinciding with backdating would see firms retroactively choosing a local low point and therefore retroactively dating the granting of executive options to correspond with this low point. As such this should be reflected in abnormal returns, such that the stock exhibits abnormally low returns prior to the grant date and abnormally high returns following the grant date. Figure 3 displays the results for all option grants in our sample based on the average cumulative abnormal returns around the grant date. What emerges are cumulative abnormal returns which become increasingly negative before the option grant date only to change direction with the option grant date and begin heading for positive territory immediately following the grant date. This v-shape, with the turning point corresponding with the reported option grant date, is similar to that seen in the U.S. literature (see for example Lie (2005) or Heron and Lie (forthcoming)) which first raised the possibility of backdating. This same v-shape is also seen in Figure 4 which considers those grants classified as scheduled and those classified as uncertain. Interestingly, both categories of grant classification exhibit cumulative abnormal returns which track each other very closely, suggesting the grants classified as uncertain are likely largely scheduled as well. The fact that scheduled grants show this pattern associated with possible backdating not completely unexpected given that even if 24 Yermack (1997), Aboody and Kasznik (2000), Chauvin and Shenoy (2001) and Lie (2005) all use similar graphs in their analysis. 17

19 options are scheduled to occur during the same month every year, that still leaves a significant window of time to choose a low point during the month and claim that date as the grant date retroactively while still meeting the obligation of granting the option in the particular month. Figure 5, which provides the average cumulative abnormal returns for grants deemed unscheduled (the scheduled and uncertain average cumulative abnormal returns have also been included for the sake of comparison), however provides a very different picture. Unlike Figures 3 and 4, average cumulative abnormal returns are positive throughout the 30 day event window. This is also in stark contrast with work such as Lie (2005) which generally finds a more pronounced v-pattern with unscheduled grants, the story being that unscheduled grants are more susceptible to backdating by executives as they are not tied to pre-set months or periods of option granting as is the case with scheduled grants. 25 A question then is what might be going on in Canada to suggest such a finding? One possibility is that in Canada scheduled grants may be granted for compensation purposes (congratulations for a job well done), whereas unscheduled are being used to motivate (firms are not taking advantage of low points in the share price and are expecting executives to increase the share price). This is highly speculative of course, and the finding will certainly warrant further study, however the notion that things may work differently in Canada than the U.S. is certainly a general possibility. Turning back to our general result of Figure 3, it is important to consider could something explain this v-shape? Two possibilities seen in finance literature are spring loading and bullet 25 Note that Lie s (2005) definition of scheduled differs from ours as he defines an option as scheduled if it occurs within one week of the one-year anniversary of the previous option grant, while we consider an option as scheduled if it occurs within one month of the one-year anniversary. As well, his sample is based solely on CEOs while ours includes senior executives and directors more broadly. 18

20 dodging. 26 Spring loading is the practice of issuing grants immediately before the release of good news, while bullet dodging is the practice of issuing grants immediately following the release of bad news. Perhaps what we are seeing is simply spring loading or bullet dodging? While it is not unreasonable to expect some spring loading and backdating may be occurring and partially influencing our results, these are not likely the primary drivers. In the case of spring loading, we would expect zero cumulative abnormal returns prior to the grant date and positive cumulative abnormal returns following the grant date, while with bullet dodging we would expect negative cumulative abnormal returns prior to the grant date and zero cumulative abnormal returns following the grant date. These do not appear to coincide with the cumulative abnormal returns in Figure 3. That said, these explanations are damaging as well as this involves using inside information which of course is illegal. Alternatively, one could argue that managers may release bad news and then good news in order to produce the v-effect we see. This too is unlikely as it would be rare that managers would be aware of both bad and good news at such a short time, and if they were fortunate enough to have such a combination and engage in this practice, it would more than likely raise some alarms with their respective regulators. As is often said in the literature on backdating, proving the practice of backdating is a difficult task, especially armed only with data. That said, the results are certainly suggestive of such as story and should at a minimum provoke further study of this important issue. 5. Discussion: Suggestions for Reducing Backdating Given the results of section 4 are suggestive of backdating; the next question is what should be done to reduce this practice? From our reading of the current state of Canadian 26 FIND CITATION THAT raise this general possibility of spring loading and backdating as possible explanations of the v-shape, and our analysis draws from theirs. 19

21 practice, as well as considering what goes on south of the border, a number of possible policy options become apparent for consideration. i. Reduce the reporting date to match the U.S. post Sarbanes-Oxley In the U.S., the Sarbanes-Oxley Act, enacted in 2002, saw new SEC reporting regulations which require the stock option grant holder to report grants to the SEC within two business days of receiving the grant. A study by Lie and Heron (forthcoming) show, that with the introduction of this new two-day reporting period, the return pattern associated with backdating is much weaker, while another by Lie and Heron (2006) also shows the percent of unscheduled grants backdated or manipulated fell dramatically following the introduction of the two day rule. Given these findings, a move by Canadian regulators to enforce a similar two day rule would likely also similar effects. ii. Require an immediate public press release the day of the grant. An alternative to following the U.S. lead, is to consider a practice currently in place for companies listed on the TSX Venture exchange. For TSX-V listed companies granting executive stock options, the requirement exists that a public press release be issued the day of the option grant. Through this requirement, the ability to backdate should be eliminated completely at a relatively low cost in terms of resources. iii. Remove individual responsibility for filing and make it a company responsibility. Currently, under the SEDI system, the responsibility for filing insider reports rests with the executive receiving the option grant. As a result, it is quite common to see entries occurring far after the 10 day reporting window has expired, as well as a wide range of missing information. This non-uniformity in data entry reduces transparency and potentially allows for 20

22 greater opportunity for filing misconduct. Moving this responsibility from the individual to the corporation may serve to increase uniformity and timeliness of filing. iv. Examine the cost attached to late filing of SEDI filings. A question that immediately arises when examining SEDI insider filings is what is the punishment for not filing within the 10 day window and is this enforced? It would appear that the punishment attached to late reporting is not terribly high, and/or likely not strictly enforced given the repeated occurrences of late or misfiled reports. Regulators may want to reconsider their current practice and whether increasing the costs of late filing may impact the decision to backdate. Further, imposing increased monetary costs may not be enough given the significant value backdating can create for an option grant recipient, and other more formal legal remedies may need to be considered. v. Reconsider the importance of the issue by regulators On September 8, 2006, the Canadian Securities Administration (CSA) posted a notice (CSA Staff Notice ) regarding options backdating in response to Canadian news stories on the practice of backdating. In this notice, the CSA details a number of TSX rules and securities regulations which they suggest may reduce the opportunity for Canadian companies to engage in backdating. These include things discussed earlier such as requiring options be granted not in the money, requiring options not be granted at a price which does not reflect undisclosed insider information, as well as the filing requirements in place for the TSX and SEDI. What the CSA seems to be missing is that these rules do not necessarily have bite in terms of reducing backdating. Certainly, they serve to make backdating an illegal activity, but still the opportunity to backdate remains if companies are willing to break the rules. For example, under the TSX filing requirements, companies must report option grants to the TSX 21

23 within 10 days following the month the options were granted. Clearly this provides a significant window of time to look back and engage in backdating of options. The SEDI filing requirement of 10 days is generally a shorter reporting requirement which reduces the window of opportunity to backdate, but still provides ample opportunity (not to mention the significant number of filings which take place after the 10 day deadline). The CSA staff notice should not be taken as evidence of a system which discourages backdating. It makes it clear that it is not acceptable practice, but there is only little in terms of what is actively done to investigate companies suspected of backdating, and what the CSA is doing to ensure the stated requirements are actually met and enforced. vi. Consider board-management separation and issues arising from a seemingly non-arms length relationship. Clearly a much larger task centres on the current state of affairs with how executive stock options are actually granted. There is a large literature concerning the potential for conflict of interest and opportunism with the granting of options and executive compensation due to the close relationship between compensation committees and executives. 27 Increasing the armslength relationship between the board of directors and senior management, and reconsidering how compensation committees are structured are two areas which clearly require addressing due to the fact these close relationship provide fertile ground for practices such as backdating to occur. 28 We provide no simple solutions to this problem, but clearly the problem is a glaring one. 27 See for example Yermack (1997) and Chauvin and Shenoy (2001) for more. 28 For example some companies have openly admitted that executives have a large part in structuring their own compensation with compensation committees serving only to ratify the plans, while others have their own CEOs on the compensation committee (Yermack, 1997, p. 453). 22

24 6. Conclusion A number of goals were set out at the beginning of this paper. This study has provided an overview of backdating, a practice only recently capturing the attention of academics and policy makers in the United States. The paper has also attempted to detail why policy makers and shareholders should be concerned if companies in Canada are practicing backdating. More importantly however, this paper proves to be the first academic paper to empirically investigate whether backdating may be occurring among Canadian companies. While it is impossible to determine unequivocally that companies are engaging in backdating (companies may just be extremely successful in forecasting to target their grants to coincide with these types of market returns), the aggregate evidence certainly supports such a story. These results should also serve to provoke further study in the area of backdating in Canada. For our own part, the results of this paper point for further study of the issue, and we are currently engaged in a number of areas to further our understanding of backdating in the Canadian context. Our work entails considering an expanded sample to include all companies in the S&P/TSX 300 index. With this larger sample we investigate a number of questions related to backdating such as does the evidence differ by industry, is the prevalence of backdating more pronounced the larger the gap between the option grant date and the reporting date, and does backdating differ by the gender of the recipient. As well, an interesting angle we are considering is how the 2000 Canadian tax changes may have impacted the degree of backdating due to the more favourable tax treatment received by stock options following the tax changes. Recall the Income Tax Act provides for an employment income deduction to partially offset the employment income triggered when an employee exercises an employee stock option. In 2000, the deduction, previously only applicable to stock options granted by Canadian-controlled 23

25 private corporations, was extended to public company stock options if certain prescribed conditions were met, notably for our purposes, that options are granted not-in-the money. As such, with these changes, we may see increased evidence of backdating post Given the findings of this paper, as well as the number of angles this issue of backdating can be analyzed, we expect research on testing for backdating will be an active research area for researchers for years to come and certainly deserves more study given the significant ramifications of the practice. 24