How Powerful CFOs Camouflage and Exploit Equity-Based Incentive Compensation

“Pay Duration,” discussed more thoroughly in “Methods” section, is a measure of how quickly CFOs have unrestricted access to their total compensation earned in a year. While CFOs generally have full rights to cash compensation (salary and bonus) in the year granted, they must often wait a period of time before they have unrestricted rights to stock options or restricted stock grants. Hence, our measure takes into account the vesting schedules as well as the type (cliff versus graded vesting) of equity-based compensation. To the best of our knowledge, our paper is the first to investigate the impact of CFOs’ power on their pay duration, and the corresponding impact of pay duration on a potentially unethical practice: earnings management.

Schipper (1989, p. 92) defines earnings management as “a purposeful intervention in the external financial reporting process, with the intent of obtaining some private gain (as opposed to, say, merely facilitating the neutral operation of the process).”

CFO compensation is typically comprised of a combination of salary, bonuses, restricted stock grants, and stock option grants. While cash compensation (salary and bonus) is available in the contemporaneous year, CFOs must wait, usually between 1 and 5 years, before their grants of restricted stock and stock options vest and they have full rights to the shares. Two CFOs may have the same level of total compensation but one may place a higher value on his/her compensation because he/she has unrestricted access to it sooner. For example, consider two CFOs who both receive total compensation of one million dollars. CFO A receives a $750,000 salary and a $250,000 bonus, so CFO A's pay duration is zero years because he/she receives all of his/her compensation by the end of the fiscal year. CFO B receives a $500,000 salary and $500,000 in stock options that vest in four years, so CFO B's pay duration is two years (we present details on this calculation in “Methods” section below).

As an example, see the 2006 disclosure for Michael Herbert, CFO of Simpson Manufacturing here: https://​www.​sec.​gov/​Archives/​edgar/​data/​920371/​0001104659070196​62/​a07-6189_​1def14a.​htm. Page 11 shows Mr. Herbert’s cash compensation for the year, but the option awards number includes the value of all options vesting that year (and thus expensed for GAAP purposes), not what he was awarded during the current year. To find that information, the reader must look on page 17 which shows the stock options Mr. Herbert received and their vesting information. Using page 18, the reader can determine what stock options have not yet vested.

Our study employs Jones’s (1991, p. 367) definition of an unethical decision to include that which is either illegal or morally unacceptable to the larger community. An ethical decision is one that is both legal and morally acceptable to the larger community. We contend in our study that the incentive compensation strategies that powerful CFOs employ constitute unethical behavior because, although they are not illegal, the strategies would likely be morally unacceptable to the larger community if the actions were publicly known.

Graham et al. (2005, p. 35) identify some of the discretionary transactions managers can choose to undertake or forgo in order to achieve financial reporting objectives. These include (1) decreasing/increasing research and development, advertising, and/or maintenance expenditures, (2) increasing or decreasing inventories, (3) liberalizing credit terms or discounts to encourage customers to purchase more product, and (4) selling investments or assets to recognize gains.

“Rents” refer to the excess compensation or the more favorable compensation arrangements that an executive receives over what he/she would have received had his/her compensation contract been obtained in a true arm’s length negotiation (Bebchuk and Fried 2003).

Hodge et al. (2009) provide survey evidence that executives value stock option grants more when the vesting period is shorter. In many instances, a large portion of an executive’s annual compensation is tied up in grants that they are not guaranteed to receive unless the grants fully vest (Jiang et al. 2010). Fu and Ligon (2010) show that a relatively high proportion of stock options (e.g., 36%) are exercised in the first year after they vest, and that the decision to exercise early is highly motivated by an executive’s need to manage her/his wealth and to diversify equity portfolios. Longer vesting periods expose the executive to increased liquidity risks and limit her or his ability to diversify. Besides increasing liquidity risk, longer vesting periods also increase forfeiture risk. Dahiya and Yermack (2008) find that 96% of firms require executives that voluntarily resign to forfeit any unvested stock options, which leads to a reduction in voluntary executive turnover (Balsam and Miharjo 2007).

Under efficient contracting, the duration of the executive’s pay package would be determined by firm characteristics and economic factors (Core and Guay 1999). Further, an assumption of efficient contracting is that the board of directors and the executive are negotiating at “arms-length.” While we cannot conclusively state that the board and the executive are not negotiating efficiently, the managerial power theory would suggest a departure from efficient contracting toward a negotiation where the executive can wield influence and power to improve the outcome of the negotiation in his/her favor.

Gopalan et al. (2014) find that the pay duration of the CEO is inversely related to the level of accruals manipulation. We extend this by looking at earnings management and CFO pay duration.

All CFOs with short durations may desire to manage earnings upward, but not all may have the influence to accomplish it. Because of the influence they wield in the firm, powerful CFOs with short duration may be more likely to manage earnings to maximize their compensation.

Our hypothesis, based on theory as discussed above, implies managerial opportunism. We acknowledge, however, there are cases where it is very difficult to disentangle the opportunistic vs. informational factors contained in earnings management. In some cases, managers may exercise discretion in the financial reporting process to reveal forward-looking private information to the market (Hodder et al. 2006). Moreover, in some cases, managers may engage in earnings management as the “lesser of two evils.” For example, Graham et al. (2005) find that CEOs and CFOs admit to engaging in real earnings management (i.e., they pass up positive NPV projects that require an initial large expense) to meet short-term Wall Street analysts’ EPS benchmarks. Managers state that meeting a benchmark limits volatility and stock price declines, which benefits shareholders, at least in the short-term. Since a first-order effect resulting from income-increasing earnings management is to boost or maintain stock prices, whether it is for opportunistic reasons or to reveal private information, the executive would benefit from higher stock prices, if her or his pay duration is short. To sum, even if the CFO engages in earnings management for informational reasons, the impact would be the same: a boost to stock prices, which in turn benefits the manager.

While not using tenure directly, Geiger and North (2006) find that a firm’s use of discretionary accruals decreases significantly after the appointment of a new CFO. Presumably, this may be because the new CFO lacks the influence over personnel and the specific firm knowledge that allows CFOs the ability to manage earnings.

The calculation as presented assumes restricted stock grants and stock option grants vest on a cliff schedule (all at the end of the vesting period). Similar to Gopalan et al. (2014), if grants vest on a graded schedule (i.e., a portion of the grant vests each year until fully vested), we use (t _i  + 1)/2 in place of t _i and (t _j  + 1)/2 in place of t _j .

Consistent with Core and Guay (1999, 2002) and Gopalan et al. (2014), we calculate PPS as the change in the grant’s value corresponding to a 1% change in the firm’s stock price.

Gopalan et al. (2014) provide a detailed description of how they estimate the vesting periods for pre-2006 grants in Section 2.1.3 of their paper. We briefly describe their method (which we follow) here. Using the vesting detail in the Equilar database, we calculate the number of new option grants that remain unvested at the end of each year by exercise price and expiration date (also rolling forward previous years’ unvested options and updating the unvested amount at the end of future years). Using ExecuComp, we aggregate unvested option grants at the end of each year by exercise price and expiration date. We merge ExecuComp and Equilar data together by executive id, year, exercise price, and expiration date. We then subtract the Equilar unvested options each year from the ExecuComp totals to get the number of unvested pre-2006 grants that remain at the end of each year. We are able to track the changes from 2006 through 2009 to estimate the vesting schedules of those pre-2006 grants. For any pre-2006 grants remaining at the end of 2009, we assume these grants vest at the end of the next year. We follow a similar procedure for estimating the vesting schedules of pre-2006 stock grants using totals by executive and year only as these grants do not have an exercise price or expiration date. If we are unable to estimate the vesting period for pre-2006 grants or ExecuComp and Equilar disagreed on the number of new grants in any particular year, we drop these observations from our sample.

As in Eq. (1), we adjust our calculation for grants that vest on a cliff schedule. See footnote 14.

Equilar Inc. is an executive compensation research firm. They collect executive compensation data from firm proxy statements.

The ExecuComp database includes data item CFOANN to indicate what executive holds the title of CFO. The Equilar database includes a similar data item. In cases where they differed (often in years of turnover), we read CFO profiles from the Corporate Library database to determine who was CFO at the end of the year.

There are a few instances when this field is blank and we code the number of titles as one when in fact, the CFO may hold any number of titles. If instead we exclude these observations from our analysis, our reported results are the same.

The number of observations for 2006 is the smallest because firms were not required to disclose the detailed vesting data Equilar collects until fiscal years ending on 12/31/06 and later.

Specifically, the CEO power index ranges from zero to three. First, we assign each CEO a one if she or he serves as chairman of the board of directors. Second, we assign each CEO a one if her or his tenure as CEO of the firm is greater than the median CEO tenure. Finally, we assign each CEO a one if he or she holds more titles than the median CEO titles held.

We scale the ranks as follows: quartile rank of zero = zero, quartile rank of one = 1/3, quartile rank of two = 2/3, and quartile rank of 3 = 1. We follow a similar procedure for the CFO power index.

Kothari et al. (2005) show that researchers can improve the reliability of results in studies involving earnings management by using performance matching to calculate abnormal accruals.

We require twenty observations within a two-digit SIC code to perform our analysis.

We calculate total accruals (TACC) using information from the balance sheet (indirect approach) as well as using information from the statement of cash flows (direct approach). Hribar and Collins (2002) argue that using the statement of cash flows provides better estimation of accruals. However, in many current studies, the balance sheet method is still used. For completeness and to show the robustness of our results, we report the results of both in “Results” section.

Two other recent studies differ on how they calculate total RTM. Cohen and Zarowin (2010) conduct their analysis using two different measures of total RTM. Their first measure combines abnormal production costs and abnormal discretionary expenses. Their second measure combines abnormal discretionary expenses and abnormal cash flows from operations. Zang (2012) uses one measure of RTM that combines abnormal production costs and abnormal discretionary expenses. Following either paper, we obtain similar results to the Badertscher's (2011) measure for total RTM (results not tabulated).

Consistent with other papers that study earnings management, we eliminate financial institutions and utility firms from our analysis. If we include financial institutions and utility firms and rerun the regressions reported in Table 5, our results remain quantitatively similar (results not tabulated).