Electronic copy available at: http://ssrn.com/abstract=1972680

Accelerated Share Repurchases, Bonus Compensation, and CEO Horizons

Carol Marquardt

Associate Professor

Baruch College – CUNY

Zicklin School of Business

Stan Ross Department of Accountancy

55 Lexington Ave

New York, NY 10010 USA

Carol.Marquardt@baruch.cuny.edu

646.312.3241

Christine Tan

Assistant Professor

Fordham University and Financial Accounting Standards Bard

1790 Broadway

New York, NY 10019

ctan@fordham.edu

Ph: 646 312 8213

and

Susan M.Young

Associate Professor

Fordham University

1790 Broadway

New York, NY 10019

Syoung16@fordham.edu

Ph: 646 312 8245

November 2011

We thank Bo Zhang for research assistance and John Graham for providing marginal tax

rate data. We also thank Monica Banyi, Mary Ellen Carter, Victoria Dickinson, Valentin

Dimitrov, Peter Easton, Nicole Thorne Jenkins, Joshua Livnat, Patricia O‘Brien, Joshua

Ronen, Richard Sloan, Tom Stober, Christine Wiedman, and workshop participants at

Notre Dame, Indiana, USC, Waterloo, Edinburgh, Fordham, the Columbia-NYU-Baruch-

Rutgers Joint Accounting Conference, the Seventh Annual University of Utah Winter

Accounting Conference, and the 2007 AAA Annual Meeting for helpful comments and

suggestions. An earlier version of this paper was entitled, ―Managing EPS through

Accelerated Share Repurchases: Compensation versus Capital Market Incentives.‖

Electronic copy available at: http://ssrn.com/abstract=1972680

2

Accelerated Share Repurchases, Bonus Compensation, and CEO Horizons

Abstract

We examine whether short-term financial reporting objectives related to executive compensation

and employment horizons affect managers‘ decisions to undertake accelerated share repurchases

(ASRs) versus open market repurchases (OMRs). In an ASR, the firm repurchases borrowed

shares and simultaneously enters into a forward contract with an investment bank. This structure

provides potential financial reporting advantages over OMRs in that earnings per share (EPS)

benefits are recorded immediately (i.e., the reporting effects are ―accelerated‖) while the actual

share repurchases and potential costs associated with the forward contract are deferred to a future

date. Consistent with this short-term focus, we find that firms are more likely to choose ASRs

over OMRs when the repurchase is accretive to EPS, when annual bonus compensation is

explicitly tied to EPS performance, when CEO horizons are short, and when CEOs are more

entrenched. These results are robust to controlling for endogeneity in the decision to repurchase

shares. In addition, we find no evidence that compensation committees adjust executive pay for

the effects of the ASR. Overall, our results suggest that short-term financial reporting benefits are

a significant determinant of decisions to undertake ASRs, consistent with theories of managerial

myopia.

Key words: Accelerated share repurchase, compensation, voluntary turnover, horizon problem

Electronic copy available at: http://ssrn.com/abstract=1972680

Accelerated Share Repurchases, Bonus Compensation, and CEO Horizons

1. Introduction

While stock repurchases have recently soared in popularity and now represent the

dominant form of cash payout (see Skinner, 2008), they have also come under increasing

criticism in the financial press (e.g., Morgenson, 2006; Shaw, 2006; MacDonald, 2007).

The main thrust of the criticism relates to their use in boosting short-term earnings per

share (EPS). More pointedly, a report by Audit Integrity (2007) argues:

―The most consistent benefit is that they tend to be good for management

by producing the appearance of earnings growth where none may exist,

while providing management with additional bonuses and compensation at

those companies which reward executives for increased EPS.‖

In this paper, we empirically examine whether short-term financial reporting

incentives related to executives‘ compensation contracts and employment horizons affect

firms‘ decisions to undertake an innovative new form of stock buybacks – accelerated

share repurchases (ASRs). In an ASR, the firm repurchases borrowed shares and

simultaneously enters into a forward contract with an investment bank. This structure

provides potential financial reporting advantages over typical open market repurchases

(OMRs) in that earnings per share (EPS) benefits are recorded immediately (i.e., the

reporting effects are ―accelerated‖) while the actual share repurchases and potential costs

associated with the forward contract component of the agreement are deferred to a future

date. These transactions thus provide a powerful setting in which to examine whether

assertions linking managerial myopia to repurchase decisions are valid.1

1 For example, Maremont and Ng (2006) have criticized the increasing use of ASRs because they can be

used to obtain short-term EPS increases but potentially damage shareholder value in the long run due to the

2

Using a sample of actual stock repurchases from 2004 to 2006, we empirically

examine the determinants of ASRs using a two-stage Heckman (1979) procedure that

corrects for any self-selection bias related to the repurchase decision. In the first stage,

we follow Dittmar (2000) and model the repurchases as a function of firm cash levels and

cash flows, market-to-book ratio, leverage, dividend payout, firm size, stock price

performance, and option usage. In the second stage, which is our primary focus, we

model the decision to execute the repurchase as either an ASR or an OMR and include

the Inverse Mills Ratio (IMR) from the first stage analysis to control for potential

endogeneity problems.

Consistent with the ―accelerated‖ reporting effects on EPS associated with ASRs,

we find that short-term financial reporting incentives related to executives‘ compensation

contracts and employment horizons are significant determinants of the decision to

undertake ASRs. As expected, we find that firms are more likely to execute an ASR

when the reporting effects are accretive to EPS and when firms‘ bonus compensation

contracts explicitly reward executives on EPS performance. We also report evidence that

short CEO horizons are a determinant of repurchase decisions -- voluntary CEO turnover

in the year following the repurchase is significantly higher for the firms that choose ASRs

over OMRs. We also find that the ASR firms have CEOs that are more entrenched than

the OMR firms, as measured by the frequency with which CEOs also hold a dual position

as chairman of the board. These results are robust to controlling for alternative

motivations for repurchases, as well as other determinants of compensation structure and

CEO turnover. Our findings suggest that the basic premise of the criticisms offered by

guaranteed nature of the repurchase agreement. Some analysts hold similar views (e.g., McConnell, Pegg,

Senyek, Mott, and Calingasan, 2006).

3

the financial press – that repurchases are often motivated by short-term reporting effects

– has validity. More particularly, our results show that managerial myopia is more

clearly linked to ASRs than to OMRs, consistent with our expectations.

We also present additional analyses. First, we examine whether compensation

committees adjust reported EPS for the effects of the ASRs when determining cash

compensation levels. Using a model similar to Healy et al. (1987), we find no evidence

that compensation committees adjust reported EPS in setting executive pay, which is

consistent with our finding of greater CEO entrenchment for firms that choose ASRs over

OMRs.

Second, to address the issue of whether ASRs might impair firms in the long run,

we examine quarterly operating performance over a six-quarter window, beginning with

the quarter before the repurchase and ending four quarters after the repurchase. Using

unadjusted data, we find that ASRs significantly underperform OMRs in every quarter;

however, this finding does not hold when we adjust for industry and size. We thus

cannot conclude that the use of ASRs damages the firm in the long run. This finding is

consistent with those of Bowen, et al. (2008), who link managerial opportunism to poor

corporate governance, but are unable to document a deterioration in future firm

performance.

Finally, we examine whether firms with performance-contingent supplemental

executive retirement plans and that also use EPS as a performance metric for bonus

determination are more likely to choose an ASR over an OMR. We do not find evidence

to suggest that these firms are more likely to choose one form of share repurchase over

another.

4

This paper contributes to the accounting literature in several ways. First, we

extend the literature on the use of stock repurchases as a potential earnings management

tool. While prior work by Bens, et al. (2003) and Hribar, Jenkins, and Johnson (2006)

have shown that benchmark-beating is a significant determinant in the decision to

undertake OMRs, we show that different incentives related to EPS reporting are at play in

the decision to undertake ASRs. This finding is important in that it deepens our insight

into managers‘ motivations behind basic financing decisions that affect the firm.

Our results also contribute more broadly to the literature that links executive

compensation to corporate financing decisions. Compensation policies have been shown

to affect firms‘ leverage ratios (Smith and Watts, 1992; Berger, et al., 1997), convertible

debt choices (Marquardt and Wiedman, 2005), and the agency costs of debt (Duru, Mansi

and Reeb, 2005). While prior literature has also related compensation policies to stock

repurchases, this stream of research has tended to focus on the use of repurchases to

offset potential dilution from stock option grants (Bens et al., 2003; Dittmar, 2000). In

contrast, we document that the reliance on short-term EPS-based bonus contracts is a

significant determinant of repurchase decisions.

In addition, our research contributes to the literature on managerial myopia. A

number of studies have investigated the associations between short-term horizon

incentives and suboptimal investment decisions (see, e.g., Narayanan, 1985; Dechow and

Sloan, 1991; Bebchuk and Stole, 1993; Bushee, 1998), as well as accrual management

(see Pourciau, 1993; Murphy and Zimmerman, 1993; Reitenga and Tearney, 2003). Our

evidence complements and extends these findings by documenting that stock repurchase

decisions are associated with CEO employment horizons.

5

The remainder of this study is organized as follows. The next section describes

ASR agreements in greater detail. We develop our hypotheses in Section 3 and discuss

our research design in Section 4. Section 5 describes our sample selection process. We

present results in Section 6 and conclude with Section 7.

2. Background on accelerated share repurchases

An ASR is an arrangement in which a company purchases a block of borrowed

shares and simultaneously enters into a forward contract with an investment bank. The

investment bank buys the company shares back in the open market over time, generally

six months to one year, to replace the borrowed shares (see Figure 1). The forward

contract is typically settled at the volume-weighted average price (VWAP) over the

contract period.

For financial reporting purposes, an ASR is treated as two separate transactions.

First, when the agreement is initiated, stockholders‘ equity is decreased by the number of

shares repurchased times the current share price, and cash is decreased or a liability is

increased by an equal amount; this affects EPS calculations (on a weighted-average basis)

for the current period. In addition, any fees charged by the investment bank are

expensed. Second, because the firm typically has the choice to settle the forward contract

in either cash or shares, they are not required under current reporting standards to mark

the forward contract to market.2 Thus, there are no reporting effects associated with the

2 The implicit assumption behind the lack of mark-to-market accounting for the forward contract is that the

company intends to settle the forward contract in shares and therefore need not consider the change in the

fair value of the forward contract in the calculation of net income. In reality, most ASR forward contracts

are settled in cash.

6

forward until the settlement date, at which point the firm will deliver to (or receive from)

the investment bank the difference between the VWAP and the stock price at the

initiation date of the ASR. At settlement, the accounting treatment is to decrease cash (or

increase liabilities) and to decrease equity, assuming the price of the company‘s stock has

increased. The repurchased shares may be kept in treasury or retired (see the Appendix

for a numerical example of the accounting treatment for ASRs).3

A key difference in accounting treatments between ASRs and OMRs is the timing

of the recognition of the decrease in shares outstanding. In an OMR, this decrease is

recorded when shares are actually repurchased, while in an ASR, the financial reporting

effects are based on the repurchase of borrowed shares, preceding actual share

repurchases. Therefore, one advantage to ASRs is the accelerated financial reporting

benefits. The disadvantage is that cash must be provided up front, and the firm must pay

the average share value over the life of the contract regardless of the increase in share

price. Firms do not have an option to discontinue repurchasing shares once the ASR has

been entered into as they would with an OMR program.4

One likely reason for the increased frequency of ASR agreements since 2004 is

related to the issuance of Statement of Financial Accounting Standard (SFAS) No. 150,

―Accounting for Certain Financial Instruments with Characteristics of both Liabilities and

Equity.‖ SFAS 150 became effective for interim periods after June 15, 2003. Prior to

SFAS 150, firms commonly wrote put options on their own shares to hedge against price

3 Dickinson, et al. (2008) examine whether the current accounting for ASRs is representationally faithful

and document significant differences in the value relevance of the assets and liabilities of ASR firms versus

an industry-matched sample. 4 In fact, Lie (2005) shows that almost 25 percent of firms that announce an OMR plan do not repurchase

any shares at all in the announcement quarter.

7

increases. SFAS 150 requires that firm use mark-to-market accounting on puts and

forward options, reducing the benefit to the firm by requiring changes in value to be

recorded as increases or decreases to net income. However, as noted above, the forward

contracts associated with ASRs are not required to be marked-to-market when the firm

has the option of settling the contract in cash or shares. An additional discussion of

accounting for ASRs can be found in EITF 99-7, ―Accounting for an Accelerated Share

Repurchase Program.‖

3. Hypothesis development

The unusual structure of ASRs, as outlined in the previous section, has piqued the

interest of academic researchers. Chemmanur, et al. (2008) explore the determinants of

the decision between executing an ASR versus an OMR, with a focus on testing various

corporate financing theories of repurchases, and Akyol, et al. (2009) ask whether ASRs

are used to deter takeovers, finding no evidence consistent with this motivation.

This paper differs substantially in focus from the concurrent work on ASRs in that

we concentrate on the role that financial reporting objectives play in the choice between

an ASR and an OMR. In particular, we rely on two aspects of the ASR agreement in

developing our hypotheses: 1) the financial reporting benefits are ―accelerated‖ relative

to OMRs in that the effect on EPS is recorded prior to the actual repurchase of shares in

the open market over time; and 2) any cost associated with settlement of the forward

contract is shifted to a future period and is recorded directly in equity, bypassing the

income statement.5 We predict that these features will make ASRs relatively more

5In an untabulated analysis, we examine final settlement costs for 45 ASR firms in our sample that

disclosed this information. On average these firms make an additional cash payment to the investment bank

8

attractive to ―myopic‖ managers, i.e., those with strong incentives to increase EPS in the

short run. Our hypotheses below all relate to this basic tenet.

3.1 Accretive EPS Effect

We predict that managers are more likely to choose ASRs over OMRs when the

impact of the repurchase is accretive to EPS. One main advantage to an ASR is that the

reporting effects are ―accelerated‖ relative to the actual repurchase of shares. If the net

effect of the repurchase does not increase reported EPS, the manager arguably has less to

gain in choosing the ASR structure. While, in theory, OMRs could be sized such that

they would have reporting effects equal in magnitude to ASRs, in practice, the vast

majority of OMRs are not accretive to EPS. For example, Hribar, Jenkins, and Johnson

(2006) find that only 9% of OMRs in their sample increase reported EPS by one cent or

more.6 Our first hypothesis is thus as follows:

H1: Accretive EPS effects are positively associated with the decision to

undertake an ASR versus an OMR.

at the end of the contract period that represents 5.7% of the original ASR cost. Further, the pattern of settlement strongly suggests that ASR firms are concerned with the financial reporting effects of the

forward contract. We discuss this further in the results section. 6 The relatively small effects of OMRs on reported EPS suggest that OMRs are well-suited as an EPS

management tool when firms fall just short of an earnings benchmark – indeed, Hribar et al. (2006)

empirically show that firms exploit OMRs‘ inherent flexibility by strategically timing OMRs when they

―need a penny‖ to make an analyst forecast. While it is possible that firms also use ASRs to meet or exceed

analyst forecasts, it is a relatively straightforward exercise to adjust expected EPS for the effects of the

repurchase because the number of shares repurchased is known in advance with certainty; further,

anecdotal evidence shows that managers themselves explicitly disclose the reporting effects of the ASR on

future EPS, and analysts adjust their forecasts accordingly. For example, Rockwell Collins issued a press

release on September 29, 2006, announcing an ASR of 4.7 million shares at an initial cost of $257 million.

The press release also included the following statement: ―With the execution of this agreement, the

company now expects fiscal year 2007 earnings per share in the range of $3.10 to $3.20, a 5 cent increase

over the previously announced guidance range of $3.05 to $3.15.‖

9

3.2 EPS-Based Bonus Compensation

Prior research shows that the use of EPS-based bonuses affects firms‘ financing

decisions. For example, Marquardt and Wiedman (2005) find that firms are more likely

to structure convertible bond transactions to increase EPS when manager bonuses are

based on reported EPS figures.7 In addition, there is practitioner evidence relating stock

repurchases to executive compensation. In their recent report from The Center for

Financial Research and Analysis and The Corporate Library, Lehman and Hodgson

(2006) examine firms in the S&P 500 with negative cash flows prior to or during share

repurchase programs and found that a greater percentage of the CEOs for these firms

were rewarded on ―per share‖ performance metric (43.11 percent) compared to S&P 500

firms as a whole (27.85 percent). They also found that bonuses were more likely paid out

to the CEOs of these firms—88 percent of these CEOs received annual bonuses versus 78

percent of the S&P 500 firms and 37.5 percent received cash bonuses versus 22.2 percent

of CEOs in the S&P 500. The authors conclude that share repurchase programs may be

used to generate higher levels of EPS and EPS growth in order to increase payout of

incentives.

We predict that firms with bonus compensation contracts that are explicitly based

on EPS performance will be more likely to choose ASRs over OMRs. The immediate

boost to EPS and the deferral of contract settlement costs to a future period makes ASRs

a more attractive than OMR when executives are explicitly rewarded on short-term EPS

performance. In addition, prior research has linked OMRs to firms‘ reliance on stock

7 On a related note, Altamuro (2006) finds that firms are more likely to use synthetic leases, which increase

reported earnings, when bonuses are tied to net income.

10

option compensation (see Bens et al., 2003; Kahle, 2002; Dittmar, 2000), which tends to

be negatively associated with cash-based compensation (see Bryan, et al., 2000).

Our second hypothesis is therefore as follows:

H2: The use of bonus compensation contracts that explicitly reward EPS

performance is positively associated with the decision to undertake an ASR

versus an OMR.

3.3 Voluntary Turnover and the Horizon Problem

Prior research finds that executives make opportunistically-motivated decisions as

their employment horizon decreases. For example, Dechow and Sloan (1991) find that

CEOs invest less in R&D expenditures in their final years in office in an attempt to

increase their earnings-based compensation, and Reitenga and Tearney (2003) report

evidence of accruals management in the last two years of CEOs‘ employment.

Because part of the cost associated with an ASR is shifted into future fiscal

reporting periods (typically the following year) through the forward contract component

of the agreement, this repurchase structure is likely more attractive to executives with

very short employment horizons. Specifically, we hypothesize that CEOs would most

profit by timing an ASR for the year prior to their departure. We focus on CEO turnover

in the year following the repurchase because Yermack (2006) documents that separation

payments (also known as ―golden handshakes‖) are common when the CEO leaves the

firm, either voluntarily or through dismissal; further, the large majority of these payments

are awarded on a discretionary basis by the board of directors and not according to the

terms of an employment agreement. CEOs could benefit from the reporting effects in the

year of the repurchase, and the additional costs of the ASRs that would be recognized in

11

the year of departure would likely have little impact on the departing CEO‘s separation

package, particularly if the CEO is entrenched.8 Our third hypothesis is stated as follows:

H3: Voluntary CEO turnover in the year following the repurchase is positively

associated with the decision to undertake an ASR versus an OMR.

3.4 Managerial Entrenchment

We hypothesize that CEOs will be best able to benefit from the accelerated

financial reporting effects of an ASR if they have sufficient managerial power (Bebchuk,

et al., 2002) to influence board decisions and compensation arrangements. Bebchuk et al.

(2002) argue that managers can influence all facets of the pay-setting process, including

the appointment or reappointment of directors who are less likely to challenge his/her

compensation. Consistent with this theory of managerial power, Davila and Penalva

(2006) find that firms with more entrenched executives are more likely to have

compensation contracts that put more weight on accounting-based measures of

performance, which are arguably easier to manipulate, than on market-based performance

measures.

We also observe that compensation committees shield CEO pay from a variety of

factors, including accounting choices related to inventory and depreciation methods

(Healy, et al., 1987); restructuring charges (Dechow, et al., 1994); accounting losses

(Gaver and Gaver, 1998); strategic expenditures (Duru, et al., 2002); merger and

acquisition activity (Grinstein and Hribar, 2004); and recurring operating expenses

8 There is also evidence that compensation committees pay special attention to earnings management that

occurs in a CEO‘s final year of employment, after the executive‘s decision to leave the firm has been

revealed (see, e.g., Cheng, 2004, or Huson, et al., 2006).

12

(Comprix and Muller, 2006).9 It is therefore not unreasonable to expect that boards might

similarly ignore the effects of stock repurchases in determining executive compensation,

particularly if managers are more entrenched. Our fourth and final hypothesis is thus

stated as follows:

H4: Managerial entrenchment is positively associated with the decision to

undertake an ASR versus an OMR.

4. Research design

We examine whether financial reporting objectives related to short-term executive

compensation contracts and employment horizons affect stock repurchase decisions using

a two-stage Heckman self-selection model. In the first stage, we model the manager‘s

decision to repurchase shares; in the second stage, which is our main focus, we model the

decision to execute an ASR versus an OMR after controlling for the effects of self-

selection bias.

4.1 First stage analysis: Repurchase decision

We first estimate a tobit model of the decision to repurchase shares similar to

Dittmar (2000):

REPt = αt + β1CFOt-1 + β2CASHt-1 + β3MKBKt-1 + β4PAYOUTt-1 + β5SIZEt-1

+ β6RETURNt-1+β7LEVt-1 + β8DILUTIONt-1 + εt (1)

where REPt is the dollar volume of repurchases, defined as Compustat data item

Purchase of Stock minus any change in preferred stock from t-1 to t, divided by the prior

9 In contrast to these findings, Jackson et al. (2008) find that when a firm has losses or decreases in

earnings, accounting fundamentals can explain changes in CEO bonuses. They conclude that compensation

committees do rely on accounting fundamentals in setting CEO bonuses when earnings are negative or

decreasing.

13

year-end market value of equity and truncated at 1%; CFO is operating cash flows,

divided by assets, at t-1; CASH is cash and cash equivalents, divided by assets, at t-1;

MKBK is the ratio of the market value of equity plus debt to the book value of assets at t-

1; PAYOUT equals cash dividends divided by net income before extraordinary items at t-

1; SIZE equals the log of total assets at t-1; RETURN is the one-year stock return from

year-end t-2 to year-end t-1; LEV is defined as the difference between a firm‘s net debt-

to-asset ratio (where net debt equals debt minus cash) and the median net debt-to- asset

ratio of all firms with the same two-digit SIC code at t-1; and DILUTION equals the

difference between the shares outstanding used to calculate diluted earnings per share and

basic earnings per share, divided by the latter at t-1.10

Consistent with findings in prior

research, we expect repurchase activity to be positively associated with CFO, CASH,

SIZE, and DILUTION and negatively associated with MKBK, PAYOUT, RETURN, and

LEV. Following Heckman (1979), we include the Inverse Mills Ratio (IMR) from

equation 1 in our second-stage analysis to control for endogeneity in the decision to

repurchase stock in our main tests.

4.2 Second stage analysis: ASR vs. OMR

Our main focus is on the manager‘s decision between executing an ASR versus an

OMR. We use a probit regression model to test H1-H4, where the dependent variable,

10

Dittmar‘s (2000) model includes an OPTIONS variable, defined as the percentage of shares outstanding

held in reserve to cover stock options, as reported by Compustat. This variable is not available after 1998;

we therefore include our DILUTION variable as an alternative measure of stock option reliance. One

advantage of using this variable to measure dilution is that it includes the effects of all potentially dilutive

securities on stock price, not only stock options. Dittmar (2000) also includes a takeover variable for a

portion of her sample period. While we do not explicitly include takeover attempts in our model, we note

that firm size, growth opportunities, and leverage, all of which are represented in equation 1, have been

found to be negatively associated with the probability of takeover (see Palepu, 1986). In addition, Akyol, et

al. (2009) find no evidence that ASRs decrease a firm‘s attractiveness as a takeover target.

14

ASR, equals one if the firm chooses to undertake an ASR and zero if the firm chooses an

OMR.

To test H1, where we predict that accretive EPS effects are positively associated

with the ASR structure, we estimate the impact of the repurchase on diluted EPS for the

fiscal quarter in which the repurchase occurs. We follow Hribar et al. (2006) and compare

reported diluted EPS with an ―adjusted‖ EPS that removes the effects of the repurchase.

Specifically, we add back (on a weighted-average basis) the shares repurchased to the

denominator and after-tax interest charges to the numerator of diluted EPS. If this

―adjusted‖ EPS measure is less than reported EPS, the repurchase is accretive to EPS.

We create an indicator variable, ACCRETIVE, that equals one if the repurchase is

accretive to EPS and zero otherwise. We use this general measure of reporting impact

because managers of the repurchasing firm do not know actual diluted EPS for the

quarter at the time repurchasing decisions are made, but are likely able to determine

whether a repurchase will be accretive or dilutive to reported EPS. We allow for this

uncertainty by using an indicator rather than a continuous variable to estimate reporting

effects. We expect a positive estimated coefficient on ACCRETIVE.

To proxy for compensation incentives related to EPS performance (H2), we

follow Marquardt and Wiedman (2005) and create an indicator variable, BONUS, which

equals one if EPS is explicitly mentioned as a determinant of annual bonuses in the firms‘

proxy statement, and zero otherwise. We believe that this is the most direct measure of

whether a manager is compensated based on reported EPS.11

If managers of ASR firms

11

We hand-collect this information from the ―Report of the Compensation Committee‖ within firms‘ proxy

statements. This report typically outlines the general compensation philosophy and provides details about

the CEO‘s compensation more specifically.

15

are motivated by short-term compensation contracts, we expect a positive estimated

coefficient on BONUS.

To test H3, we create an indicator variable, TURNOVER, which equals one if the

CEO of the repurchasing firm voluntarily leaves the firm in the fiscal year following the

repurchase, and zero otherwise. Using a procedure similar to Huson, et al., (2001), we

classify turnover as ―forced‖ if a search of Factiva reveals that the CEO was fired, forced

from the position, or departs due to policy differences; we classify all other turnover as

voluntary. We expect a positive estimated coefficient on TURNOVER.

To test H4, we create an indicator variable, COB, which equals one if the CEO is

also the chairman of the board (COB), and zero otherwise. We choose this measure as it

is well-documented in prior research that duality in CEO and COB leadership rules

reflects managerial entrenchment and power. For example, Brickley, et al. (1997) find

that dual roles are associated with longer tenure and higher compensation levels. Ryan

and Wiggins (2004) find that CEOs who are also COB use their position to influence

directors‘ compensation, potentially interfering with effective monitoring. Grinstein and

Hribar (2004) find that M&A bonuses are positively related to shared CEO and COB

roles – i.e., CEOs who also serve as COB receive significantly higher bonuses and

engage in larger M&A deals. In addition, they find that compensation committees do not

appear to adjust bonuses for the performance of the deal. To the extent that the duality in

CEO and COB roles reflect managerial entrenchment, we expect a positive estimated

coefficient on COB.12

12

We use COB as our primary measure of managerial entrenchment because it is the most commonly used

in the prior literature. However, in sensitivity tests, we also use CEO age and CEO tenure as additional

measures of managerial entrenchment (see e.g. Berger et al. 1997; Davila and Penalva 2006). Results are

qualitatively similar.

16

We also control for other firm characteristics that might affect the choice between

an ASR and an OMR. For example, it is possible that the same factors that motivate

compensation policies also affect repurchase structure. One such variable is suggested by

the ‗informativeness principle,‘ which states that when a performance measure reflects

the influence of factors other than the executive‘s actions, i.e., the performance measure

is ‗noisy,‘ firms will reduce the weight placed on that measure in CEO compensation

packages (see Holmstrom, 1979; Milgrom and Roberts, 1992). Yermack (1995) and

Bryan, et al., (2000) find support for this idea in that they find that firms with noisy

accounting earnings relative to stock returns tend to rely more on stock option awards for

compensation purposes. We thus expect that firms that rely on EPS-based compensation

contracts will have earnings realizations that are less noisy relative to stock returns. We

control for this effect by including in our model the variable NOISE, which is defined as

the ratio of the standard deviation in return on assets to the standard deviation of stock

returns, measured over the five-year period preceding the repurchase. We expect NOISE

to be negatively associated with the likelihood of an ASR.

We also include the firm‘s marginal tax rate (MTR) as a control. Section 162(m)

of the Internal Revenue Code eliminates the tax deductibility of non-performance based

executive compensation over $1 million, and Balsam and Ryan (1996) show that

compliance with 162(m) is positively associated with firms‘ marginal tax rates.13

To

ensure that our BONUS variable, which would qualify as performance-based

compensation measure under the Code, is not confounded with a tax effect, we include

13

On a related note, Perry and Zenner (2001) document increased pay-for-performance sensitivity for the

firms affected by 162(m). More recently, Carter et al. (2008) find increased reliance on earnings

information in bonus contracts in the post-SOX era.

17

MTR in our model. We use simulated marginal tax rates as calculated by Graham and

Mills (2008) using financial statement data, which are highly correlated with marginal

rates based on actual tax returns.14

We expect MTR to be positively associated with the

likelihood of an ASR.

We also control for alternative financial reporting objectives associated with stock

repurchases. Barth, et al. (1999) and Myers, et al. (2007) find that the market rewards

patterns of increasing earnings, and Bens, et al. (2003) find that firms increase the level

of their firms‘ stock repurchases when earnings are below the level required to achieve

the desired rate of EPS growth. We therefore control for firms‘ incentives to use

repurchases to maintain strings of consecutive earnings increases by including in our

model the variable STRING, which equals the number of consecutive quarters prior to the

announcement date of the repurchase that the firm has met or exceeded the benchmark of

the prior year‘s EPS for the same fiscal quarter, up to a maximum of 20 quarters.

Because OMRs are arguably better suited to managing EPS over a period of several

quarters – managers can easily vary the timing and number of shares repurchased with an

OMR plan – we expect that STRING will be negatively (positively) associated with the

likelihood of an ASR (OMR).

In addition, because Skinner and Sloan (2002) find that market penalties for

negative earnings surprises are stronger for high growth than for low growth firms,

incentives to maintain earnings strings should be stronger for high growth firms. We

therefore also include SALESGROWTH, defined as annual sales growth over the fiscal

14

We thank John Graham for providing data on marginal tax rates. When these data are not available, we

use Graham and Mill‘s (2008) ―PseudoStatutory‖ variable, which they show is a second-best alternative to

their simulated rates.

18

year prior to the repurchase announcement, as an additional control variable.15

As with

the STRING variable, we expect a negative association between SALESGROWTH and the

likelihood of an ASR.

Our second-stage model is as follows:

Pr(ASRi) = αt + β1ACCRETIVEi + β2BONUSi + β3TURNOVERi + β4COBi

+ β5NOISEi + β6MTRi + β7STRINGi+β8SALESGROWTHi + β9IMRi + εi (2)

where i denotes firm i. In summary, we predict positive estimated coefficients on

ACCRETIVE, BONUS, TURNOVER, COB, and MTR, and negative estimated coefficients

on NOISE, STRING, and SALESGROWTH. A significant estimated coefficient on IMR

would indicate the existence of endogeneity in the repurchase decisions.16

5. Sample selection

We identify our sample of ASR firms by conducting key word searches on

Factiva and the SEC‘s EDGAR database for the terms ―accelerated‖ and ―repurchase‖

and ―stock or share.‖ The following is an example of a typical announcement of an ASR:

‗Laboratory Corp. of America Holdings (NYSE: LH) has entered into an

accelerated share repurchase agreement with an affiliate of Lehman

Brothers Inc. (NYSE: LEH) to repurchase around $250 million of

LabCorp stock. Under the agreement, LabCorp purchased around 3.4

million shares for subsequent delivery for a prepayment of $250 million.

The purchase price for these shares is subject to an adjustment based on

the volume weighted average price of LabCorp's stock during a period

15

We also use market-to-book as an alternative measure to proxy for growth and our results are

qualitatively similar. 16

It could be argued that a number of the variables in equation 1 might also affect the decision to

undertake an ASR versus an OMR. For example, Chemmanur, et al. (2008) hypothesize that differences in

firm size, cash levels, payout ratios, leverage, takeover measures, and valuation ratios could play a role in

choosing between an ASR and an OMR. We effectively include these variables as controls by including

the IMR from the first stage analysis in the second stage probit model. However, in sensitivity tests, we

examine the robustness of our main analysis to inclusion of these variables as individual regressors in the

second-stage model.

19

following execution of the agreement. The purchase price adjustment

should be settled in the first quarter of 2007.‘

When a company‘s announcement of share repurchase type was unclear, the company

was excluded from the sample. For example, we exclude Seagate Technology who made

the following announcement on August 6, 2008:

‗Stock repurchases under this program may be made through a variety of

methods, which may include open market purchases, privately negotiated

transactions, block trades, accelerated share repurchase transactions or

otherwise, or by any combination of such methods.‘

Our initial search yielded 109 firms that had engaged in one or more ASR during

2001-2006.17

Consistent with reports in the financial press stating that the prevalence of

ASRs has recently increased dramatically, we note that we could identify only six ASRs

prior to 2004; we therefore limit our focus to the 2004-2006 period.

For the first stage tobit analysis, the sample consists of all firms listed on

Compustat and CRSP over 2003-2006 that have all the data necessary to estimate

equation (1).18

This selection criterion results in 21,499 firm-years, of which 3,531 are

identified as repurchasers and 17,968 as non-repurchasers. The proportion of

repurchasers (16.4%) is comparable to Dittmar (2000), where repurchases occur in 16.7%

of sample firm-years.

For the second stage analysis, our probit model requires a sample of firms that

executed OMRs over 2004-2006. Because ASR transactions are typically announced

publicly through a press release at the time that they are executed, we limit our sample of

17

There were 18 firms that engaged in an ASR as well as an OMR during the 2004-2006 period. We

included these firms in our final sample. We also re-ran the tests excluding these firms and the results are

qualitatively the same. 18

Estimation of equation (1) requires one year of lagged data; hence we require data from 2003 to examine

repurchase activity over 2004-2006.

20

OMRs to those where firms announce the initiation of an OMR program and

subsequently follow through with an actual repurchase of shares during the same fiscal

quarter as the OMR program announcement. We believe these selection criteria best

enhance the comparability of the firms undertaking ASRs versus OMRs. Following Lie

(2005), we use the SDC Platinum database to identify OMR program announcements and

identify actual repurchases as occurring when quarterly Compustat data item #93

(Purchases of Common and Preferred Stock) exceeds 1% of the firm‘s market value.

After eliminating observations with missing data, our final sample consists of 70 ASR

firm-years and 201 OMR firm-years.

6. Results

6.1 First stage analysis: Tobit model of repurchases

Table 1, Panel A provides univariate comparisons of the independent variables

used to estimate equation (1). We compare both means and median across repurchasing

and non-repurchasing firm-years. Consistent with Dittmar (2000), repurchasing firms

generally have higher cash flows (CFO) and cash levels (CASH), have lower market-to-

book ratios (MKBK), are larger in firm SIZE, have lower stock RETURNS, and higher

dividend PAYOUT, are less levered (LEV), and have a greater proportion of potentially

dilutive securities (DILUTION) than non-repurchasing firms.

Table 1, Panel B provides the results of the first-stage tobit regression. The

estimated coefficients on CFO, CASH, SIZE, and DILUTION are significantly positive,

consistent with our expectations. The estimated coefficients on MKBK, PAYOUT and

RETURN are significantly negative (p = 0.0275, 0.0012, and 0.0001, respectively). The

21

pseudo-R2 is 9.1%. Overall, our results from the first-stage regression suggest that the

firm characteristics of repurchasers and non-repurchasers are significantly different in the

predicted fashion over our sample period. We use the results from this analysis to

calculate the inverse mills ratio (IMR), which we require as a control in our second-stage

model.

6.2 Second stage analysis: ASRs versus OMRs

In Table 2, we present univariate results for our second-stage variables. As

expected, the accretive effects to EPS are significantly greater for ASR firms, with the

repurchase increasing reported EPS by one cent or more for 44.3% of the ASR firms

versus 13.9% of the OMR firms (p=0.0001). In addition, a significantly higher

proportion of ASR firms explicitly reward their executives on EPS performance than do

OMR firms (57.5% vs. 30.0%, p=0.0001). We also document striking differences in

CEO turnover. For the ASR group, 23.8% of CEOs voluntarily leave the firm in the year

following the repurchase versus 7.6% of the CEOs in the OMR group (p = 0.0022). The

CEOs of the ASR firms also yield significantly more power, as measured by the

frequency with which they hold dual CEO and COB positions: 81% versus 51% for OMR

firms (p=0.0001). The univariate results are thus consistent with H1 – H4.

The control variables are also generally consistent with our expectations. ASR

firms have significantly lower median levels of relative noise in earnings (NOISE) than

OMR firms (0.033 versus 0.052, p=0.0019), consistent with the informativeness

principle. ASR firms also have significantly higher marginal tax rates (MTR), shorter

strings of consecutive earnings increases (STRING), and lower sales growth

22

(SALESGROWTH) than OMR firms. The inverse mills ratio (IMR) from the first-stage

regression is significant, highlighting the importance of controlling for the repurchase

decision.

We present Pearson and Spearman correlation coefficients in Table 3. The

strongest pair-wise correlation is between STRING and SALESGROWTH (Spearman ρ =

0.228), which are both proxies for financial reporting incentives related to benchmark

beating, followed by the correlation between ACCRETIVE and COB (ρ = 0.188), which

suggests that entrenched managers may be more likely to undertake repurchases in order

to secure financial reporting benefits. In general, however, the low correlation levels

suggest that multicollinearity is not likely to be a major concern.

Table 4 provides the second-stage probit regression of the ASR versus OMR

repurchase decision on our variables of interest. For completeness, we report results with

and without controlling for selection bias (IMR); as a sensitivity test, we also include

specifications with SIZE and RETURNS as additional regressors. Consistent with our

expectations, ACCRETIVE is significantly positively associated with the likelihood of

undertaking an ASR versus an OMR across all five specifications, as is BONUS and

COB, while TURNOVER is at least marginally significantly positive in Models 1 through

4. The estimated coefficients for the control variables NOISE, MTR, and STRING are

generally not significantly different from zero, but the estimated coefficient on

SALESGROWTH is at least marginally significantly negative across all five models. The

significance of the IMR variable, which proxies for self-selection bias, is sensitive to

model specification -- it is not significantly different from zero in Models 2 and 4, but if

RETURNS is added to the analysis, as in Model 5, it becomes significantly negative.

23

Lastly, if we include firm SIZE as an additional regressor, as in Models 3 and 4, the

significance levels of the other variables increase slight, but the basic tenor of the results

does not change.19,20

Overall, our findings are consistent with H1 – H4 and do not appear

to be driven by self-selection bias: short-term financial reporting incentives related to

executive compensation contracts and employment horizons are significant determinants

of the ASR versus OMR choice.

6.3 Additional analysis: Effect on executive cash compensation

In our main tests, we implicitly assume that compensation committees do not

adjust reported EPS for the effects of the ASR when setting executive pay. In this section,

we directly examine the validity of this assumption. We estimate a cross-sectional model

in which CEO cash compensation for the fiscal year of the repurchase is regressed on the

change in reported EPS and the market value of equity. Following Healy et al. (1987), we

transform both cash compensation and firm size by taking natural logs of both variables

so that their distributions more closely approach linearity. Our models are as follows:

log(Cash Compensationt) = β1+β2 ∆EPSt + β3 LogSIZEt + ε (3)

log(Cash Compensationt) = β1+β2 ∆EPSASRt + β3 LogSIZEt + ε (4)

Cash compensation is defined as the sum of annual salary and bonus; ΔEPS is

defined as the difference between current and prior year diluted EPS, divided by the

absolute value of prior year diluted EPS; ΔEPSASR is defined as the difference between

19

As a sensitivity test, we include the remaining independent variables from equation (1) as additional

regressors in equation (2). Only SIZE and RETURNS are significant determinants of the ASR versus OMR

choice after controlling for selection bias, and inclusion of the other variables from equation (1) does not

qualitatively change the results in Table 4. 20

In untabulated analyses, we also include CEO age as an additional regressor, as Desai, Hogan, and

Wilkins (2006) find that CEO age is a significant determinant of executive turnover. CEO age is not a

significant determinant of repurchase structure, and our inferences drawn from Table 4 remain unchanged.

24

diluted EPS for the current fiscal year, with the numerator adjusted for after-tax interest

charges related to the repurchase and the denominator adjusted for the time-weighted

number of shares repurchased, and prior year diluted EPS, divided by prior year diluted

EPS; and Size is the (logged) market value of equity as of the fiscal year-end. We

compare adjusted R2‘s across the two specifications using a Vuong Z-test. If

compensation committees adjust reported EPS for the effects of the ASR, we should

observe a higher adjusted R2 in the second model (equation 4). We obtain cash

compensation data from Execucomp when available; otherwise, we hand-collect the

information from firms‘ proxy statements filed with the SEC. We examine both the

CEO‘s cash compensation, as well as the sum of cash compensation for the top five

executives of the firm.

Results are presented in Table 5. In Panel A, we present descriptive statistics on

cash compensation and EPS. Mean (median) CEO cash compensation in the fiscal year of

the ASR is $1.886.0 ($1.261.3) million, which is slightly higher than the mean (median)

CEO cash compensation of $1.707 ($1.120) million for all Execucomp firms during our

sample period. The average CEO bonus comprises almost one-third of total cash

compensation, which is comparable to the average for Execucomp firms. Average cash

compensation for the top five executives is slightly higher than the Execucomp average

of $4.823 million. The average bonus for the top five executives in ASR firms comprises

31 percent of total cash compensation compared to 36 percent for all Execucomp firms.

The average cash compensation of the mean (median) diluted EPS is $2.646 ($2.380),

which indicates that ASR firms are fairly profitable. When diluted EPS is adjusted for the

effects of the ASR, the mean (median) figure is $2.589 ($2.372). The mean (median)

25

percentage effect of the ASR on diluted EPS is 2.2% (1.1%), which is relatively large. In

addition, mean (median) annual growth in diluted EPS is 20.1 percent (10.4 percent), but

would be 16.4 percent (8.5 percent) if the ASR had not taken place.

Panel B presents the results from the OLS regressions. The adjusted R2‘s are

similar across models and Vuong Z-tests comparing adjusted R2‘s from Models 1 and 2

and from Models 3 and 4 are not significant at conventional levels. Based on these

findings, it does not appear that compensation committees adjust reported EPS for the

effects of ASRs, consistent with the managerial entrenchment exhibited by firms that

undertake these transactions. We caution, however, that our lack of results could also be

due to low statistical power.

6.4 Additional analysis: Post-repurchase operating performance

To assess whether there are negative long-run consequences associated with

ASRs, we compare the post-repurchase operating performance of ASRs versus OMRs.

We eliminate firms that are missing necessary accounting data on quarterly Compustat to

compute performance-adjusted ROA. Similar to Lie (2005), we develop matched samples

based on: 1) announcement year and quarter, 2-digit SIC code, and firm size as measured

by total assets; and 2) announcement year and quarter, 2-digit SIC code, and performance

as measured by return on assets (ROA) and market to book ratio (M/B) in the

announcement quarter. We define performance as operating income scaled by cash-

deflated assets at the beginning and end of the quarter.

We present results for the firms matched on industry and size in Table 6

(untabulated results for firms matched on performance measures are qualitatively

similar). The sample size is limited due to both to the availability of an appropriate

26

matching control firm and by the recency of our sample firms. In particular, firms that

completed ASRs in 2006 would have little data available on quarterly Compustat with

which to carry out the analysis.

As shown in Panel A, mean and median unadjusted ROA are significantly smaller

for ASR firms in quarters -1 through +2. However, when operating performance is

adjusted for industry and size, we no longer observe any significant differences between

ASRs and OMRs. Similarly, in Panel B, where changes in operating performance are

presented, no significant differences between the two groups are apparent. These findings

suggest that either the choice between an ASR and OMR has no significant effects on

future performance or, alternatively, that our small sample size does not allow us to

detect underlying performance differences. Hence, we do not find conclusive evidence

that the use of ASRs damages the firm in the long run. This finding is consistent with

those of Bowen, et al. (2008), who link managerial opportunism to poor corporate

governance, but are also unable to document a deterioration in future firm performance.21

6.5 Additional analysis: Executive pensions

Finally, we examine whether certain types of pension plans may also increase the

likelihood of choosing an ASR. According to recent articles in the business press (see e.g.

21

We also examined final settlement costs for 45 ASR firms in our sample that disclosed this information.

Eight of these firms chose to settle the forward contract using shares, and 37 chose to settle in cash. Of the

cash settlement firms, 33 had an increase in share price over the contract period and therefore had a further

obligation to the investment bank, while none of the stock settlement firms had an increase and therefore

were not required to issue additional stock to the investment bank. A chi-squared test reveals that this

difference in settlement patterns is highly significant (χ2 = 26.76, p <= 0.0001) and is consistent with

companies choosing to settle in cash to avoid issuing additional shares that further dilute EPS. This pattern

of settlement strongly suggests that ASR firms are concerned with the financial reporting effects of the

forward contract.

27

Schultz and McGinty, 2009), another area of executive compensation that has been

overlooked as a means of excessive pay is supplemental executive retirement plans

(SERPs). These plans, which are non-qualified and non-contributory, are designed for

key employees and provide benefits above and beyond those covered in qualified

retirement plans. Sometimes termed ‗camouflage‘ or ‗stealth‘ compensation, they are a

potential area where compensation can be hidden from investors and analysts (Bebchuk

and Fried, 2004). In a period where many employers have frozen pension plans for

regular employees, the Wall Street Journal found that pensions for top executives rose

almost twenty percent in 2008. The value of these contracts can be difficult to determine

from information in corporate footnotes yet can result in tens of millions of dollars in

post-retirement pay (Bebchuk and Fried, 2004).

Kalyta (2009) suggests that these supplementary retirement plans present an

excellent opportunity to examine managers‘ discretionary accounting choices. He finds

that managers make income increasing discretionary accounting choices in the years prior

to retirement when CEO pension is based on firm performance. Kalyta and Magnan

(2008) provide evidence that the vague disclosure of these SERPs also prevents effective

monitoring of this compensation mechanism. They find that proxies for CEO power are

significantly related to the occurrence and size of CEO SERP benefits.

To examine this relationship, we create an indicator variable which equals one if

the firm has a performance-contingent SERP in the year of the repurchase and zero

otherwise. We collect this information from the pension and deferred compensation

benefits section of firm proxy statements. Our results provide some evidence that the

existence of a SERP is related to the probability of choosing an ASR over an OMR.

28

Results show this pension variable is significant in the multivariate analysis at the p<.05

level (untabulated) when size is not included in the regression. However, when size is

added to the full model, the results are not significant at traditional levels.

We also examine the interaction of SERP with BONUS and include this in the

probit regression. If the firm has a performance-contingent supplemental executive

retirement plan and EPS is used as a performance benchmark, then we would expect

these firms to more likely choose an ASR over an OMR. We correct for the interaction

effect in the probit regression using the approach of Ai and Norton (2003) and Norton,

Wang and Ai (2004). While the multivariate results are not significant, we do find in

untabulated results, that approximately 19% of OMR firms have SERPs and one-third of

these firms have SERPs that are linked to EPS as a performance indicator (Pearson Χ2 =

0.217). More than half of the ASR firms have SERPs (53%) and nearly two-thirds of

these firms have SERPs linked to EPS (Pearson Χ2 = 1.984). Overall, our results provide

only limited evidence that firms with performance-contingent SERPs are more likely to

choose an ASR over an OMR.

7. Conclusions

In this paper, we examine whether short-term financial reporting incentives

related to executives‘ compensation contracts and employment horizons affect firms‘

decisions to undertake ASRs versus OMRs. We find that firms are more likely to choose

ASRs over OMRs when the repurchase is accretive to EPS, when annual bonus

compensation is explicitly tied to EPS performance, when CEO horizons are short, and

when CEOs are more entrenched. Importantly, our findings are robust to controlling for

endogeneity in the decision to repurchase shares. Overall, these results suggest that

29

short-term financial reporting benefits are a significant determinant of decisions to

undertake ASRs, consistent with assertions in the financial press. These findings have

particular relevance in light of the recent intensified focus on executive compensation and

managerial myopia in the investment and regulatory communities.

In additional analyses, we examine whether compensation committees adjust for

the reporting effects of the ASR when setting CEO pay; we find no evidence of

adjustment, consistent with the managerial entrenchment exhibited by these firms. We

also examine whether ASRs impair the long-run performance of firms; our findings are

inconclusive, consistent with Bowen et al. (2006), who similarly document a link

between managerial opportunism and poor governance yet are unable to document

deleterious subsequent firm performance. Finally, we do not find evidence that firms

with performance-contingent supplemental executive retirement plans and that treat EPS

as a performance benchmark are more likely to choose and ASR over an OMR. However,

we caution that the lack of results in our additional analyses may be due to low statistical

power stemming from small sample size.

30

Appendix

The following is an example of the accounting treatment for an ASR. Suppose Company

X wants to buy back 1 million shares of stock. Currently, the company has 10 million

shares outstanding and the stock price is $10 per share. After the decision, the company

has net earnings of $2 million for the quarter ended March 31.

Scenario 1: The stock price stays the same over the quarter.

Company X enters into an ASR agreement on January 1 and agrees to repurchase

1 million shares of stock. The ASR has an end contract date of March 31. The

current stock price is $10 per share.

Jan. 1: Treasury Stock $10,000,000

Cash or Liability $10,000,000

Forward agreement: no entry made as the forward contract has no significant

value at the contract‘s initiation date. However, any fees charged by the

investment bank would be recorded as an expense.

March 31: No entries required

Effect on EPS at 3/31:

With ASR: Without ASR:

$2,000,000/9,000,000 = $.22 $2,000,000/10,000,000 = $.20

Scenario 2: The stock price increases to $15/share on January 31 and remains there for

the rest of the quarter.

Jan. 1: Treasury Stock $10,000,000

Cash or Liability $10,000,000

March 31: The VWAP over the three months is $13.33 per share.

If settled in cash:

Treasury Stock $3,333,333

Cash or Liability $3,333,333

EPS on 3/31 is same as above.

If settled in stock:

An adjustment would be made to the shares outstanding. The company would

now show that approximately 888,888 shares have been repurchased, versus

1,000,000. However, as EPS is calculated on a weighted-average basis, there is no

change in reported EPS for the first quarter. There is no impact on the balance

sheet.

31

References

Audit Integrity. 2007. Insights. Share repurchases revisited. June 18.

Ai, C. and E. C. Norton. 2003. Interaction terms in logit and probit models. Economics

Letters 80: 123-129.

Norton, E. A., H. Wang and C. Ai. 2004. Computing interaction effects and standard

errors in logit and probit models. The Stata Journal 4: 154-167.

Akyol, A.C, Kim, J.S., Shekhar, C. 2009. Do accelerated stock repurchases deter

takeovers? An empirical analysis. Working paper, The University of Melbourne.

Balsam, S., Ryan, D. 1996. Response to tax law changes involving the deductibility of

executive compensation: A model explaining corporate behavior. Journal of the

American Taxation Association 18 (Supplement), 1-12.

Barth, M., Elliott, J., Finn, M. 1999. Market rewards associated with patterns of

increasing earnings. Journal of Accounting Research 37 (2), 387-414.

Bebchuk, L., Stole, L.A. 1993. Do short-term objectives lead to under- or over-

investment in long term projects? Journal of Finance 48 (2), 719-729.

Bebchuk, L., Fried, J., Walker, D. 2002. Managerial power and rent extraction in the

design of compensation. The University of Chicago Law Review 69 (3), 751-846.

Bebchuk, L., Fried, J., Walker, D. 2004. Pay without performance. First Harvard

University Press.

Bens, D., Nagar, V., Skinner, D., Wong, F. 2003. Employee stock options, EPS dilution,

and stock repurchases. Journal of Accounting and Economics 36 (1-3), 51-90.

Berger, P.G., Ofek, E., Yermack, D.L. 1997. Managerial entrenchment and capital

structure decisions. Journal of Finance 52 (4), 1411-1438.

Bowen, R., Rajgopal, S., Venkachatalam, M. 2008. Accounting discretion, corporate

governance, and firm performance. Contemporary Accounting Research 25(2),

351-405.

Brickley, J., Coles, A.L., Jarrell, G. 1997. Leadership structure: Separating the CEO and

chairman of the board. Journal of Corporate Finance 3 (3), 189-220.

Bryan, S., Hwang, L., Lilien, S. 2000. CEO stock-based compensation: An empirical

analysis of incentive-intensity, relative mix and economic determinants. Journal

of Business 73 (4), 661-693.

Bushee, B. 1998. The influence of institutional investors on myopic R&D investment

behavior. The Accounting Review 73(3), 305-333.

32

Chemmanur, T.J., Cheng, Y., Zhang, T. 2008. Why do firms undertake accelerated share

repurchase programs? Working paper, Boston College.

Cheng, S. 2004. R&D expenditures and CEO compensation. The Accounting Review 79,

305-328.

Comprix, J., Muller, K. 2006. Asymmetric treatment of reported pension expense and

income amounts in CEO cash compensation calculations. Journal of Accounting

and Economics 42 (3), 385-416.

Davila, A., Penalva, F. 2006. Governance structure and the weighting of performance

measures in CEO compensation. Review of Accounting Studies 11 (4), 463-493.

Dechow, P., Huson, M., Sloan, R.G. 1994. The effect of restructuring charges on

executives‘ cash compensation. The Accounting Review 69 (1), 138-156.

Dechow, P.M., Sloan, R.G. 1991. Executive incentives and the horizon problem: An

empirical investigation. Journal of Accounting and Economics 14 (1), 51-89.

Desai, H., Hogan, C., Wilkins, M. 2006. The reputational penalty for aggressive

accounting: Earnings restatements and management turnover. The Accounting

Review 81 (1), 83-112.

Dittmar, A. 2000. Why do firms repurchase stock? Journal of Business 73 (3), 331-356.

Duru, A., Iyengar, R., Thevaranjan, A. 2002. The shielding of CEO compensation from

the effects of strategic expenditures. Contemporary Accounting Research 19 (2),

175-193.

Duru, A., Mansi, S.A., Reeb, D.M. 2005. Earnings-based bonus plans and the agency cost

of debt. Journal of Accounting and Public Policy 24, 431-447.

Gaver, J., K. Gaver, K. 1998. The relation between nonrecurring accounting charges and

CEO cash compensation. The Accounting Review, 73 (2), 235–253.

Graham, J., Mills, L. 2008. Using tax return data to simulate corporate marginal tax rates.

Journal of Accounting and Economics 46 (2-3), 366-388.

Grinstein, Y., Hribar, P. 2004. CEO compensation and incentives: Evidence from M&A

bonuses. Journal of Financial Economics 73 (1), 119-143.

Healy, P., Kang, S., Palepu, K. 1987. The effect of accounting procedure changes on

CEOs‘ cash salary and bonus compensation. Journal of Accounting and

Economics 9 (1), 7-34.

33

Heckman, J. 1979. Sample selection bias as a specification error. Econometrica 47 (1),

153-161.

Holmstrom, B. 1979. Moral hazard and observability. Bell Journal of Economics 10, 74-

91.

Hribar, P., Jenkins, N., Johnson, W.B. 2006. Stock repurchases as an earnings

management device. Journal of Accounting and Economics 41 (1,2), 3-27.

Huson, M., Parrino, R., Starks, L.T. 2001. Internal monitoring mechanisms and CEO

turnover: A long-term perspective. Journal of Finance 56 (6), 2265-2297.

Huson, M.R., Wiedman, C., Wier, H. 2006. Cash compensation and earnings components

around CEO turnover: Does the compensation committee see over the horizon?

Working paper, University of Alberta and University of Waterloo.

Jackson, S.B., Lopez, T.J., Reitenga, A.L. 2008. Accounting fundamentals and CEO

bonus compensation. Journal of Accounting and Public Policy 27, 374-393.

Kahle, K. 2002. When a buyback isn‘t a buyback: Open market repurchases and

employee options. Journal of Financial Economics 63 (1), 236-261.

Kalyta, P. 2009. Accounting discretion, horizon problem, and CEO retirement benefits.

The Accounting Review 84 (5), 1553-1573.

Kalyta, P., Magnan, M. 2008. Executive pensions, disclosure quality and rent extraction.

Journal of Accounting and Public Policy 27, 133-166.

Lehman, J., Hodgson, P. 2006. Fueling EPS through share repurchases—Who

benefits? Report of The Center for Financial Research and Analysis and The

Corporate Library.

Lie, E. 2005. Operating performance following open market share repurchase

announcements. Journal of Accounting and Economics 39 (3), 411-436.

MacDonald, E. 2007. Buyback binge. Forbes.com, May 21.

Maremont, M., Ng, S. 2006. Moving the market – Tracking the numbers outside audit:

Buybacks via loophole can have hidden cost. The Wall Street Journal, January 31,

C1.

Marquardt, C., Wiedman, C. 2005. Earnings management through transaction structuring:

Contingent convertible debt and diluted EPS. Journal of Accounting Research 43

(2), 205-243.

34

McConnell, P., Pegg, J. Senyek, C., Mott, D., Calingasan, A. 2006. Accounting Issues:

Accelerated Share Repurchase – Share Repurchase Using Derivatives. Bear

Stearns Equity Research.

Milgrom, P., Roberts, J. 1992. Economics, Organization and Management. London:

Prentice-Hall.

Morgenson, G. 2006. Why buybacks aren‘t always good news. The New York Times,

Nov. 12.

Murphy, K.J., Zimmerman, J. 2003. Financial performance surrounding CEO turnover.

Journal of Accounting and Economics 16 (1-3), 273-315.

Myers, J., Myers, L., Skinner, D. 2007. Earnings momentum and earnings management.

Journal of Accounting Auditing and Finance 22 (2), 249-284.

Narayanan, M.P. 1985. Managerial incentives for short-term results. Journal of Finance

40 (5), 1469-1484.

Palepu, K. 1986. Predicting takeover targets: A methodological and empirical analysis.

Journal of Accounting and Economics 8 (1), 3-35.

Perry, T., Zenner, M. 2001. Pay for performance? Government regulation and the

structure of compensation contracts. Journal of Financial Economics 62 (3), 453-

488.

Pourciau, S. 1993. Earnings management and non-routine executive changes. Journal of

Accounting and Economics 16 (1-3), 317-226.

Reitenga, A., Tearney, M.. 2003. Mandatory CEO retirements, discretionary accruals,

and corporate governance changes. Journal of Accounting Auditing and Finance

18 (3), 255-280.

Ryan, H.E., Wiggins, R.A. 2004. Who is in whose pocket? Director compensation, board

independence, and barriers to effective monitoring. Journal of Financial

Economics 73 (3), 497-524.

Schultz, E.E., McGinty, T. 2009. Pensions for executives on rise: Arcane techniques,

generous formulas boost payouts as share prices fall. Wall Street Journal, Nov. 3.

Shaw, H. 2006. Are buybacks financial cover-ups? CFO.com, Nov. 16.

Skinner, D., Sloan, R. 2002. Earnings surprises, growth expectations and stock returns, or

don‘t let an earnings torpedo sink your portfolio. Review of Accounting Studies 7

(2-3), 289-311.

35

Skinner, D. 2008. The evolving relation between dividends, earnings, and stock

repurchases. Journal of Financial Economics 87 (3), 582-609.

Smith, C.W., Watts, R.L. 1992. The investment opportunity set and corporate financing,

dividend, and compensation policies. Journal of Financial Economics 32 (3), 293-

332.

Yermack D. 1995. Do corporations award CEO stock options effectively? Journal of

Financial Economics 32 (2,3), 237-270.

Yermack, D. 2006. Golden handshakes: Separation pay for retired and dismissed CEOs.

Journal of Accounting and Economics 41 (3), 237-256.

36

FIGURE 1

Overview of ASR transaction*

Pays cash and enters

E into forward contract Loan shares to bank

Borrows shares and enters

into forward contract

Purchases shares

and returns to investors

*See McConnell et al. (2006).

Investment

Bank

Open

Market

Investors

ASR

Firm

37

TABLE 1 First stage tobit regression of stock repurchase model Panel A: Univariate comparisons of first stage independent variables

Non-repurchasers Repurchasers p-value

CFO

Mean

Median

0.0221

0.0625

0.1155

0.1145

0.0001

0.0001

CASH

Mean

Median

0.1955

0.1258

0.2216

0.1368

0.0001

0.7859

MKBK

Mean

Median

2.2313

1.5779

2.1708

1.7862

0.1331

0.0001

SIZE

Mean

Median

5.7731

5.6417

7.3521

7.3700

0.0001

0.0001

RETURNS

Mean

Median

0.1866

0.1352

0.1361

0.1290

0.0001

0.0001

PAYOUT

Mean

Median

0.1653

0.0000

0.1765

0.0361

0.2576

0.0001

LEV

Mean

Median

0.0343

-0.0037

0.0237

-0.0178

0.0388

0.0872

DILUTION

Mean

Median

0.0261

0.0059

0.0349

0.0209

0.0001

0.0001

Panel B: Tobit estimation results

REPt = αt +β1CFOt-1+β2 CASHt-1+β3MKBKt-1+β4PAYOUTt-1+β5SIZEt-1

+β6RETURNSt-1+β7LEVt-1+β8DILUTIONt-1+εt

Estimated coefficient p-value

Intercept -0.158 0.0001

CFO 0.248 0.0001

CASH 0.056 0.0001

MKBK -0.003 0.0275

PAYOUT -0.017 0.0012

SIZE 0.018 0.0001

RETURNS -0.620 0.0001

LEV -0.015 0.1687

DILUTION 0.139 0.0001

Log likelihood -760.22

Pseudo R2 0.091

N=21,499. Number of non-repurchasers = 17,968; number of repurchasers = 3,531. CFO equals operating cash flows,

divided by assets, at t-1; CASH equals cash and cash equivalents, divided by assets, at t-1; MKBK equals the ratio of the

market value of equity plus debt to the book value of assets at t-1; PAYOUT equals cash dividends divided by net

income before extraordinary items at t-1; SIZE equals the log of total assets at t-1; RETURN equals the one-year stock

return from year-end t-2 to year-end t-1; LEVER equals the difference between a firm‘s net debt-to-asset ratio (where

net debt equals debt minus cash) and the median net debt-to-asset ratio of all firms with the same two-digit SIC code at

t-1; DILUTION equals the difference between the shares outstanding used to calculate diluted earnings per share and

basic earnings per share, divided by the latter at t-1. All variables are winsorized at 1% and 99%. P-values in Panel A

are based on two-sided t-tests for means and Wilcoxon tests for medians. REP is the dollar volume of repurchases

divided by the prior year-end market value of equity, truncated at 1%.

38

TABLE 2

Univariate comparisons of second stage variables

Variable

ASRs

(n=70)

OMRs

(n=201)

p-value

ACCRETIVE

Proportion

BONUS

Proportion

0.443

0.575

0.139

0.300

0.0001

0.0001

TURNOVER

Proportion

0.238

0.076

0.0022

COB

Proportion

0.810

0.510

0.0001

NOISE

Mean

Median

0.084

0.033

0.108

0.052

0.3901

0.0019

MTR

Mean

Median

0.258

0.350

0.191

0.350

0.0020

0.0032

STRING

Mean

Median

3.038

2

4.857

3

0.0013

0.0057

SALESGROWTH

Mean

Median

0.097

0.069

0.169

0.123

0.0031

0.0015

IMR

Mean

Median

1.386

1.373

1.550

1.563

0.0001

0.0001 ACCRETIVE equals 1 if the repurchase is accretive to diluted EPS by one cent or more for the fiscal quarter

of the repurchase and 0 otherwise. BONUS equals 1 if EPS if explicitly mentioned in the firm‘s proxy

statement as one of the determinants of annual cash bonuses and 0 otherwise. TURNOVER equals 1 if the

CEO voluntarily leaves the firm in the year following the repurchase and 0 otherwise. COB equals 1 if the

CEO is also Chairman of the Board and 0 otherwise. NOISE is the ratio of the standard deviation in return

on assets to the standard deviation of stock returns, measured over the five-year period preceding the

repurchase. MTR is the firm‘s marginal tax rate. STRING is the number of consecutive quarters prior to the

announcement date of the share repurchase that the firm has met or exceeded the benchmark of the prior

year‘s EPS for the same fiscal quarter, up to a maximum of 20 quarters. SALESGROWTH is measured

annually. All variables are measured in the fiscal year prior to the repurchase announcement, except where

indicated, and are winsorized at 1% and 99%. P-values for BONUS, TURNOVER, and COB are based on

two-tailed test of differences in population proportions; p-values for differences in means are based on two-

tailed t-tests; p-values for differences in medians are based on two-tailed Wilcoxon tests.

39

TABLE 3

Pearson/Spearman correlation coefficients

Variable

ACCRETIVE

BONUS

TURNOVER

COB

NOISE

MTR

STRING

SALES

GROWTH

IMR

ACCRETIVE

1.000

-

0.007

(0.916)

0.094

(0.137)

0.188

(0.003)

-0.087

(0.167)

0.091

(0.149)

0.046

(0.462)

-0.125

(0.048)

-0.159

(0.011)

BONUS 0.007

(0.916)

1.000

-

0.040

(0.494)

0.104

(0.082)

-0.117

(0.045)

0.140

(0.017)

-0.140

(0.017)

-0.071

(0.227)

-0.151

(0.012)

TURNOVER 0.094

(0.137)

0.040

(0.494)

1.000

-

0.092

(0.125)

-0.190

(0.001)

0.129

(0.001)

-0.002

(0.971)

-0.037

(0.523)

-0.123

(0.041)

COB 0.188

(0.003)

0.104

(0.082)

0.092

(0.125)

1.000

-

-0.188

(0.002)

0.179

(0.002)

-0.080

(0.181)

-0.037

(0.537)

-0.147

(0.017)

NOISE -0.074

(0.242)

-0.123

(0.036)

-0.093

(0.112)

-0.013

(0.826)

1.000

-

-0.194

(0.001)

-0.090

(0.124)

-0.019

(0.742)

0.178

(0.003)

MTR 0.091

(0.149)

0.139

(0.017)

0.129

(0.027)

0.180

(0.003)

-0.197

(0.001)

1.000

-

-0.104

(0.075)

0.092

(0.115)

-0.012

(0.837)

STRING 0.063

(0.316)

-0.115

(0.051)

-0.003

(0.958)

-0.098

(0.103)

-0.003

(0.957)

-0.096

(0.103)

1.000

-

0.228

(0.001)

0.062

(0.302)

SALESGROWTH -0.015

(0.813)

-0.008

(0.892)

0.010

(0.871)

-0.018

(0.759)

-0.015

(0.806)

0.112

(0.058)

0.223

(0.001)

1.000

-

0.054

(0.367)

IMR -0.182

(0.004)

-0.147

(0.014)

-0.120

(0.045)

-0.141

(0.023)

0.082

(0.175)

-0.022

(0.722)

0.056

(0.353)

0.052

(0.389)

1.000

- N=271. Pearson (Spearman) correlation coefficients are presented below (above) the diagonal. Variable definitions are provided in Table 2. P-values (in

parentheses) are based on two-tailed significance levels.

40

TABLE 4

Probit regressions of repurchase type on firm characteristics

iiiiiii

iiii

RETURNSSIZEIMRHSALESGROWTSTRINGMTR

NOISECOBTURNOVERBONUSACCRETIVEASR

)(

)Pr(

11109876

543210

Variable Predicted

Sign

Model 1

Model 2

Model 3

Model 4

Model 5

Intercept ? -1.4087

(0.0001)

0.3507

(0.6375)

-3.4488

(0.0001)

-4.8360

(0.0008)

1.1784

(0.1627)

ACCRETIVE + 0.8682

(0.0001)

0.8191

(0.0004)

0.8219

(0.0004)

0.8558

(0.0003)

0.8163

(0.0006)

BONUS + 0.4860

(0.0208)

0.4720

(0.0258)

0.3902

(0.0378)

0.3971

(0.0386)

0.3837

(0.0313)

TURNOVER + 0.5670

(0.0530)

0.4991

(0.0478)

0.5297

(0.0815)

0.5478

(0.0732)

0.3248

(0.1525)

COB + 0.7013

(0.0023)

0.6819

(0.0035)

0.6681

(0.0113)

0.6016

(0.0124)

0.6035

(0.0123)

NOISE - 0.2131

(0.6262)

0.2948

(0.4998)

0.5397

(0.2341)

0.5397

(0.2375)

0.4303

(0.3670)

MTR + 0.1784

(0.7747)

0.1765

(0.7792)

0.1373

(0.8343)

0.2258

(0.7341)

0.0158

(0.9810)

STRING - -0.0401

(0.0919)

-0.0387

(0.1047)

-0.0334

(0.1768)

-0.0330

(0.1849)

-0.0349

(0.1637)

SALES GROWTH - -1.2119

(0.0517)

-1.2223

(0.0588)

-1.1031

(0.0958)

-1.1031

(0.0911)

-1.4899

(0.0217)

IMR ?

-0.7040

(0.1308)

0.6323

(0.2800) -2.3087

(0.0003)

SIZE +

0.2644

(0.0002)

0.3217

(0.0003)

RETURNS ?

8.6336

(0.0001)

Log likelihood -101.02 -99.34 -93.47 -92.89 -111.45 N=271. The dependent variable, ASR, is an indicator variable that equals 1 if the stock repurchase is

accelerated and 0 if open market. All other variables are defined in Table 2. P-values are presented in

parentheses and are based on two-tailed significance levels.

41

TABLE 5

Regression of logged executive cash compensation on EPS and size

Panel A: Descriptive Statistics

Variable

Mean

Median

Std. Dev.

Total CEO Cash Compensation ($ 000‘s) $1,886.0 $1,261.3 2,115.7

CEO Bonus / Total CEO Cash Compensation 0.314 0.409 0.309

Total Top5 Cash Compensation ($ 000‘s) $5,409.3 $4,005.5 3,718.4

Top5 Bonus / Total Top5 Cash Compensation 0.310 0.367 0.264

EPS (as reported) $2.646 $2.380 1.694

EPSASR (as if ASR had not occurred) $2.587 $2.372 1.667

% Difference (EPS-EPSASR)/EPS 0.022 0.011 0.061

∆EPS 0.201 0.104 0.741

∆EPSASR

0.164 0.085 0.723

Panel B: OLS Regressions of Cash Compensation on EPS

Model 1: log(CEO Cash Compensationt) = β1+β2 ∆EPS + β3 SIZE + ε

Model 2: log(CEO Cash Compensationt) = β1+β2 ∆EPSASR + β3 SIZE + ε

Model 3: log(Top5 Cash Compensationt) = β1+β2 ∆EPS + β3 SIZE + ε

Model 4: log(Top5 Cash Compensationt) = β1+β2 ∆EPSASR + β3 SIZE + ε

Model

β1

β2

β3

Adj. R2

1 4.692

(0.0001)

0.128

(0.0112)

0.287

(0.0001)

25.89%

2 4.673

(0.0001)

0.128

(0.0124)

0.288

(0.0001)

25.71%

3 6.311

(0.0001)

0.098

(0.0141)

0.237

(0.0001)

26.87%

4 6.299

(0.0001)

0.100

(0.0138)

0.238

(0.0001)

26.65%

N=70. Cash compensation figures for the CEO and Top 5 executives are obtained from Execucomp where

possible or hand-collected from proxy statements filed with the SEC otherwise. All variables are measured

as of the end of the fiscal year of the repurchase. EPS is diluted EPS (Compustat data item #57). EPSASR

is diluted EPS with the numerator adjusted for the time-weighted interest earned on the cash used in the

repurchase and the denominator adjusted by the time-weighted number of shares repurchased. ΔEPS is

current year‘s diluted EPS minus lagged diluted EPS, divided by lagged diluted EPS. ΔEPSASR is

EPSASR minus lagged diluted EPS, divided by lagged diluted EPS. SIZE is the log of the market value of

equity. P-values are presented in parentheses.

42

TABLE 6

Quarterly operating performance for firms with share repurchases during the announcement quarter

Unadjusted Industry and Size Adjusted

ASRs OMRs ASRs OMRs ASRs OMRs ASRs OMRs

Quarter n n Mean Median Mean Median n n Mean Median Mean Median

Panel A: Levels of operating performance

-1 54 174 .0353 a**

.0289 a***

.0504 a .0462

a 46 156 .0078 -.0006 -.0024 .0101

0 54 175 .0359 a**

.0290 a**

.0489 a .0410

a 48 161 .0071 .0016 -.0090 .0082

1 45 161 .0343 a*

.0307 a***

.0493 a .0426

a 37 142 .0054 .0008 -.0031 .0085

2 38 117 .0339 a**

.0294 a***

.0554 a .0458

a 31 97 .0079 .0038 .0071 .0111

3 34 73 .0336 a .0294

a .0489

a .0357

a 30 52 .0078 .0001 .0001 .0064

4 24 36 .0372 a .0291

a .0412

a .0388

a 22 27 .0121 .0033 .0054 .0078

Panel B: Changes in operating performance

0 to +1 45 159 -.0005 -.0004 .0015 .0005 36 136 .0035 .0004 -.0068 -.0021

0 to +2 38 117 -.0008 -.0017 -.0049 b .0002 31 95 -.0010 -.0004 -.0109

b -.0005

0 to +3 34 73 .0011 -.0007 .0035 .0023 30 51 -.0011 -.0008 -.0154 b .0028

0 to +4 24 35 .0037 .0014 .0067 b .0013 21 26 .0004 .0018 .0020 .0008

Operating performance is measured as operating income scaled by cash-adjusted assets at the beginning and end of the fiscal quarter. Quarter 0 is the

announcement quarter. Industry adjusted performance is the difference between the operating performance of the sample firms and the operating performance of

their respective date, industry and asset matched firm. OMRs used in this analysis are those that repurchase shares totaling at least one percent of total assets in

the same quarter as the repurchase announcement. a and

b denote that the mean or median is different than zero at the .01 and .10 level, respectively.

***, **, * indicates that ASRs are significantly different from OMRs at the .01, .05 and .10 level, respectively.