Decoupling CEO Wealth and Firm Performance:
The Case of Acquiring CEOs
Jarrad Harford*
University of Washington Business School
Box 353200
Seattle, WA 98195-3200
206.543.4796
[email protected]
Kai Li
MIT Sloan School of Management
50 Memorial Drive, E52-457
Cambridge, MA 02142
617.452.4057
[email protected]
and
Sauder School of Business
University of British Columbia
2053 Main Mall
Vancouver BC V6T 1Z2
October 28, 2004
Abstract:
We examine CEO pay and incentives around corporate takeovers to explore whether compensation policies
in bidding firms counter or exacerbate the divergence between CEO and shareholder interests over
acquisitions. We find that CEOs are rewarded for growth through acquisition. Even in mergers where
bidding shareholders are worse off, bidding CEOs are better off three quarters of the time. In the years
following mergers, CEOs of poorly performing firms receive substantial increases in option and stock
grants that offset any effect of long-term underperformance on their wealth. As a result, the CEO’s pay and
his overall wealth become insensitive to negative stock performance, but rise in step with positive stock
performance. Corporate governance matters; bidding firms with stronger boards retain the sensitivity of
their CEOs’ compensation to poor performance following the acquisition.
Keywords: corporate governance, equity incentives, pay-performance sensitivity, wealth-performance
sensitivity, portfolio incentives
JEL Classification: G34
*
We thank an associate editor, an anonymous referee, Qiang Cheng, Wayne Guay, Rob Heinkel, Jon
Karpoff, Alan Kraus, Eric Santor, and Terry Shevlin for helpful discussions, seminar participants at
Erasmus University Rotterdam, Humboldt University, K.U. Leuven, the finance journal club at UBC,
Southern Methodist University, University of British Columbia, University of Washington, and participants
of the Northern Finance Association meetings (Quebec City) for comments. We acknowledge the financial
support from the Social Sciences and Humanities Research Council of Canada and the UBC-HSS research
grant. Li also acknowledges the financial support from the MIT Sloan School of Management and the
W.M. Young Chair in Finance from the Sauder School of Business at UBC. We are responsible for all
errors.
Decoupling CEO Wealth and Firm Performance:
The Case of Acquiring CEOs
Abstract:
We examine CEO pay and incentives around corporate takeovers to explore whether compensation policies
in bidding firms counter or exacerbate the divergence between CEO and shareholder interests over
acquisitions. We find that CEOs are rewarded for growth through acquisition. Even in mergers where
bidding shareholders are worse off, bidding CEOs are better off three quarters of the time. In the years
following mergers, CEOs of poorly performing firms receive substantial increases in option and stock
grants that offset any effect of long-term underperformance on their wealth. As a result, the CEO’s pay and
his overall wealth become insensitive to negative stock performance, but rise in step with positive stock
performance. Corporate governance matters; bidding firms with stronger boards retain the sensitivity of
their CEOs’ compensation to poor performance following the acquisition.
Keywords: corporate governance, equity incentives, pay-performance sensitivity, wealth-performance
sensitivity, portfolio incentives
JEL Classification: G34
1. Introduction
The separation of ownership and control leads to many potential conflicts of interest
between shareholders and corporate management. Since Jensen and Meckling (1976),
the literature has focused much attention on how managerial ownership and
compensation design can be used to align manager and shareholder interests. As we
discuss below, the bulk of the empirical evidence supports the hypothesis that managers
with greater ownership or managers with more equity-based incentives are less likely to
take value-destroying actions.
The 1990s witnessed explosive growth in the grants of equity-based incentives
(option and restricted stock grants) to corporate managers (Murphy (1999)). This raises
the possibility that the value of the manager’s annual flow of new grants is of the same
order as the value of the incentive effect of his existing portfolio. Consequently,
compensation policies adopted in the 1990s may have altered the relationship between a
manager’s existing ownership and the firm’s investment decisions.
This paper asks whether acquiring firm CEOs bear significant wealth losses when
they undertake acquisitions that reduce their firms’ stock performance. There is a large
literature arguing that many acquisitions destroy value for the acquirer (e.g., Agrawal and
Jaffe (2000)). Investors might assume that CEOs’ large portfolios of stock and options
provide the necessary incentives to discourage them from making bad acquisitions. Our
results show how this intuition could be wrong in the presence of dynamic compensation
changes following acquisitions. In many cases, the value of the flow of new grants after
an acquisition can swamp any incentive effect provided by the CEO’s pre-acquisition
portfolio.
We focus our study on acquisition events and on the incentives of bidding firm
CEOs in particular for several reasons. First and foremost, acquisitions may be the most
significant decisions about the allocation of corporate resources that managers make and
1
the potential wealth destruction to bidding firm shareholders is large, as documented in
Moeller, Schlingemann and Stulz (2004). Thus, it is important to understand the
managerial incentives behind corporate takeovers because of their impact on shareholder
wealth and the organization of assets in the economy. With respect to compensation, by
increasing the size of the firm and changing its scope of operations, acquisitions provide
a natural opportunity for the CEO and board to restructure his compensation. The
increase in size and complexity of integrating the two firms could lead the CEO to argue
for more pay and for pay that is less sensitive to performance for the first few years of the
acquisition, or it could result in the board arguing for more sensitivity to ensure efficient
integration. We investigate how the acquiring CEO compensation changes and whether
his wealth and compensation become more or less sensitive to the performance of the
firm following the acquisition. Second, we focus on the CEO because he is frequently
the largest stock and option holder, and thus has the largest potential gain/loss derived
from a completed merger. Although the board has to approve any merger proposals, the
CEO usually initiates the discussion of a merger attempt and, in most cases, has the
dominating influence on the final decision. Finally, by focusing on the CEO, our results
will be comparable to recent research on acquisitions and compensation, such as Bliss
and Rosen (2001), and Grinstein and Hribar (2004).
We employ a carefully constructed acquisition and compensation dataset that
overlaps with both the most recent merger wave and the massive shift in the composition
of compensation contracts for top executives. Our study provides the opportunity to
assess the applicability of extant results to the new compensation environment in the
1990s as well as to answer new questions related to current compensation structures.
These questions include whether acquisition decisions generate additional information
that is used by the board to reward or punish its CEOs. For example, in the event of a
poorly performing acquisition, do changes in the flow of new grants offset or increase the
negative effect of the acquisition on the personal wealth of acquiring CEOs? We present
2
results using various measures of CEO compensation and of the sensitivity of CEO pay
and wealth to firm performance. Evidence of the costs and benefits to bidding firm CEOs
in the takeover market will aid in understanding the forces that determine when and why
takeovers are initiated. Furthermore, our results will have general implications for
compensation policy and research on agency problems by demonstrating how new grants
of equity affect the efficacy of existing portfolio (of stock and options) incentives.
We find that bidding firm CEOs have abnormally high compensation relative to
their peers before making an acquisition. Nonetheless, CEOs are rewarded for growth
through acquisition. CEOs of bidding firms are given substantial new stock and option
grants following completed acquisitions. In fact, large grants to CEOs of poorly
performing firms offset the negative effect of poor merged-firm stock performance on
their pre-acquisition portfolio of own-firm stock and options. Consequently, CEO pay
and wealth are completely insensitive to poor post-acquisition performance, but remain
sensitive to good post-acquisition performance. Bidding firms with stronger boards
retain the sensitivity of their CEOs’ compensation to poor performance following the
acquisition. Our results brings into question the efficacy of existing equity portfolio
incentives in the face of continuous flows of large new grants, and shows that the strength
of a firm’s board affects the degree to which the new grants counter the incentives of the
CEO’s existing portfolio.
We also conduct some additional tests beyond the above main results. We carry
out robustness tests relaxing our requirement that the target be at least 10% as big as the
acquirer. We also examine the subset of CEOs who did not survive at the merged firm
following the acquisition. Finally, we check our results by controlling for whether the
merger is diversifying, the method of payment, and the announcement return. The results
are robust to all of these controls.
3
Our study bears on the self-serving management hypothesis first postulated in
Benston (1985). The hypothesis states that managers are likely to take firm-enlarging
actions that yield greater remuneration for themselves but losses for shareholders. Using
acquisition and compensation data in the 1970s and 1980s, studies by Benston (1985),
and Avery, Chevalier and Schaefer (1998)) find mixed results.
This study contributes to the recent empirical literature examining the relation
between managerial incentives and corporate acquisitions. On one hand, Datta, IskandarDatta and Raman (2001) document a strong positive relation between acquiring
managers’ equity-based compensation and merger performance. Also, Zhao and Lehn
(2003) conclude that CEOs who make value-destroying acquisitions are more likely to be
replaced subsequently. On the other hand, Bliss and Rosen (2001) show that CEO
compensation typically increases after bank mergers even if the bidder’s stock price
declines. Grinstein and Hribar (2004) find the acquiring CEOs who have more power to
influence board decisions receive significantly larger M&A bonuses. Unlike our study,
these studies examine either managerial incentives prior to acquisition, acquisitions
within a specific industry, or specific components of CEO incentives around takeovers.
Our results agree with those in Bliss and Rosen (2001) that CEOs exhibit selfserving behavior in making acquisition decisions. In addition to the generous M&A
bonuses documented by Grinstein and Hribar (2004), we find that acquiring CEOs are
richly rewarded in all firm-specific incentives. Our paper also demonstrates one specific
channel—undertaking acquisitions—through which CEOs can achieve the asymmetric
benchmarking in their pay that is shown in Garvey and Milbourn (2003).
The plan of the paper is as follows. We review the literature and present our
hypotheses in the next section. Section 3 describes the sample and variables and provides
summary statistics. Section 4 presents the empirical results and interpretation, and
4
Section 5 conducts additional investigation and some robustness check on our results.
Section 6 concludes.
2. Literature Review and Hypotheses
As discussed in the introduction, acquisitions provide bidding managers their greatest
opportunity for expressing their non-value-maximizing preferences. The empirical
findings in support of the self-serving management hypothesis have been mixed.
Starting with Benston (1985), who finds the post-merger change in equity value
of managerial holdings outweighs any increase in compensation, most early studies (e.g.,
Lewellen, Loderer and Rosenfeld (1985), and Agrawal and Mandelker (1987)) produce
evidence inconsistent with the self-serving management hypothesis.
More recently, Datta at al. (2001) document a strong positive relation between
acquiring managers’ equity-based compensation and stock price performance around and
following acquisition announcements. They conclude that their evidence indicates that
managerial incentives can be effective in shaping long-term corporate investment policies
and encourage managers to make decisions in the interests of shareholders.
On the other hand, Avery et al. (1998) find that there is no evidence that a CEO
can increase his salary or bonus by undertaking an acquisition, while finding evidence
that CEOs who undertake acquisitions obtain more outside directorships than their peers.
Thus, their findings are consistent with the hypothesis that managers can gain power,
prestige, and standing in the business community by undertaking acquisitions.
Bliss and Rosen (2001) explore the relationship between bank mergers and CEO
compensation during 1986-1995. They find that mergers have a net positive effect on
compensation, mainly via the effect of size on compensation. In addition to employing a
more general sample of mergers and focusing on a more recent sample period, our paper
differs from Bliss and Rosen (2001) in the following methodological aspects. First of all,
5
our paper is about the tension between the incentives derived from existing equity
ownership versus those from new grants. We carefully address the dual nature of grants
of options and stock on CEO incentives in the current period and in the future. Second,
we track the acquiring CEO wealth change post-merger by taking into account the effects
of stock performance, new grants, exercise of existing grants, transaction of existing
shares, and new cash compensation to provide an accurate measure of the long-run CEO
wealth effect of acquisitions.
Grinstein and Hribar (2004) examine CEO compensation and incentives around
acquisitions with a focus on M&A bonuses. They show that CEOs who have more power
to influence board decisions receive significantly larger bonuses, and that there is a
positive relation between bonus compensation and measures of CEO efforts, but not
between bonus compensation and deal performance. Their evidence is consistent with the
argument that managerial power is the primary driver of M&A bonuses.
In summary, early evidence supports the hypothesis that managers with larger
ownership stakes make better investment decisions. Subsequent research also supports
the hypothesis that managers whose compensation plans are more equity-based make
better investment decisions. Mergers enlarge firm size and scope, but the evidence is
mixed on whether this is wealth-enhancing to the bidding firm executives. Much of the
evidence that managers do not benefit from bad mergers is based on the effect of the
merger on managerial stockholdings. Prior studies have tended to examine either the
flow of compensation or portfolio (of stock and options) incentives alone and have not
examined changes in compensation policy following the merger and their possible impact
on the package of incentives facing the CEO. In this study, we will account for not only
the CEO’s pre-merger portfolio wealth but also changes in pay following the merger. We
examine how these changes affect both the level of his wealth and its sensitivity to
performance.
6
The empirical analyses in this study are designed to distinguish between the selfserving management hypothesis and the incentive alignment hypothesis. If acquisitions
are beneficial to CEOs even when they are detrimental to shareholders, then the selfserving management hypothesis is supported. If CEO compensation schemes ensure that
acquisitions improve or do not diminish the link between shareholder and CEO wealth,
then the incentive alignment hypothesis is supported.
We start by examining the flow of different components of CEO compensation
around an acquisition and whether that flow increases or decreases the CEO’s pay and
wealth sensitivity to firm performance. If acquisitions improve managerial incentives,
we would expect both the amount of pay they receive and also their total wealth to
become more sensitive to firm performance after acquisitions. Further, we would expect
sales growth through acquisitions to be rewarded less than growth of sales internally,
which is more difficult to achieve. Finally, we compute the CEO’s long-run wealth
sensitivity to the impact of the merger on acquiring firm shareholders’ wealth to provide
further evidence. If CEO wealth remains sensitive to firm performance post-merger, then
the incentive alignment hypothesis is supported. Overall, the tests are designed to
determine whether acquisitions improve or reduce CEO incentives, and exactly how
changes in compensation around the acquisition achieve that result.
3. Sample Formation and Variable Construction
We begin with all completed US mergers with announcement dates between January 1,
1993 and December 31, 2000 as identified from the Mergers and Acquisitions database of
Securities Data Company (7,076 deals).1 We first require that bidders have available
stock prices from the CRSP files, accounting information from Compustat, and executive
1
The reason for us to start our sample in 1993 is that this is the year when the coverage in the Standard
and Poor’s ExecuComp data became complete.
7
compensation data from Standard and Poor’s ExecuComp. The data availability
requirement leads to 1,508 mergers over the sample period.
If acquisitions have any impact on corporate governance and executive
compensation, the likelihood of capturing this effect will be greater when the firm makes
a large acquisition. As a result, we require the ratio of the transaction value relative to the
bidder (as measured by the market value of assets) in the fiscal year prior to the
announcement to be at least 10%. Our relative size requirement leads to 622 mergers.
To clearly delineate the effect of each acquisition on executive compensation, in
cases where a sample firm makes multiple acquisitions, only those acquisitions that do
not overlap are included. We define an overlap to occur if the gap between the
completion of one merger and the announcement of a second is no greater than two years.
The non-overlapping requirement eliminates 252 acquisitions. The final sample consists
of 370 acquisitions made by 361 firms. In sum, the largest decreases from the initial to
final sample are caused by the requirements that compensation data be available and that
the target be at least 10% of the bidding firm size.
Two event years become important in our analyses. First, year: ayr, is the fiscal
year in which the announcement of a merger bid takes place. Second, year: cyr, is the
fiscal year in which the merger is consummated. Of the 370 mergers in our sample, 244
close in the same year they are announced, 121 close in the following year, and the
remaining 5 close within 3-years of being announced. The fact that two-thirds of our
mergers close in the same fiscal year as they are announced is consistent with the
Grinstein and Hribar (2004)’s concurrent sample of mega-deals, where the median time
to completion is four months.
8
Measures of CEO Compensation
Our empirical analysis is based on three different measures of CEO compensation: total
cash compensation (cash pay), grants of stock and options for the current year (grants),
and total direct compensation (total pay).
Total cash compensation is the sum of salary and annual bonus, whereas salary
simply measures the component of compensation that is fixed at the beginning of the
year. Cash and bonus have incentive effects in the current year only, as future firm
performance does not affect CEO wealth through this dimension.
Stock and option grants are the sum of the value of the restricted stock and the
value of stock options granted during the current year. The former can be obtained
directly from ExecuComp (RSTKGRNT), while the latter is computed following Core
and Guay (2002).2 Grants of stock and options (if the underlying stock is retained) have
both a short term and a long term incentive effect. Since the level of grants usually
depends upon performance, this provides incentives in the current period. Moreover,
since the value of these grants, and thus CEO wealth will be affected by future
performance, these grants also provide incentives for firm performance in future years.
Total direct compensation is the sum of salary, annual bonus, value of restricted
stock granted, value of stock option granted, long-term incentive payouts, and any other
remuneration.
2
Core and Guay (2002) estimate the grant date value of options granted during the year using a
modified version of the Black-Scholes model. To account for the fact that executives view the life of the
option (i.e., time to maturity) as being less than their expected time to exercise, and that they frequently
exercise early, Core and Guay assume the expected time to exercise to be 70% of the option’s stated
maturity, and replace time to maturity with time to exercise in the Black-Scholes model. The expected
stock-return volatility is measured as the annualized standard deviation of daily stock returns over the 120
trading days preceding the end of the fiscal year in which the grant was made. Expected dividend yield is
estimated as cash dividends paid in the fiscal year the grant is made divided by year-end stock price, and
the treasury-bond yield corresponding to the option’s remaining time-to-maturity is used to estimate the
risk-free rate. All these option pricing parameters are updated every year in our computation. We find that
using either the Core and Guay measure or the ExecuComp’s measure (GRANTS) for the value of stock
and option grants during the year has no material effect on our conclusions (the correlation between the two
measures in our sample is .97).
9
Measures of Abnormal CEO Compensation
In this paper, we focus on examining the effect of mergers on acquiring CEO pay and
wealth. One main obstacle for us is to determine how acquiring CEOs are being paid both
pre- and post-merger. There are two possible approaches—a regression framework or the
control sample approach to provide CEO pay benchmark.
Our view is that the control sample approach requires fewer restrictions and
allows the relationship between the CEO pay and firm characteristics (i.e., match
characteristics) to be non-linear. However, once the list of match characteristics is
expanded, the control sample approach quickly reaches the possibility that there are not
enough non-event firms available for the matching. In contrast, the regression approach is
able to control for as many firm characteristics as possible as long as the regression
model does not exhaust its degree of freedom. However, the drawback of the multiple
regression setting is that the explanatory variables are usually assumed to be linearly
related to the variable of interest, which may appear to be a stringent assumption in many
corporate finance settings, thus leading to the danger of model mis-specification.
Consequently, we try to take advantage of the merits of both approaches. First, we
fit a straightforward regression model of expected pay including the explanatory
variables from the extant literature. The model is estimated over the ExecuComp CEO
population. We obtain our measure for the abnormal pay by taking the residuals for our
sample of acquiring CEOs and control firm CEOs. Second, we compare the level of
abnormal pay across the sample CEOs and control firm CEOs, to capture the effect of
acquisition on CEO pay. Given that the multivariate regression and the matching
approach take into account the respectively, linear and non-linear effects of firm size on
pay, the difference in the level of abnormal pay is more likely to be driven by the effect
of mergers. As a result, the simultaneous adoption of both approaches reduces the extent
10
of model mis-specification, but does not eliminate it as we do not know the true model
for CEO pay and we do not control for the true determinants of CEO pay perfectly in our
choice of matching firms. As a practical matter, having control firms makes some of the
comparisons, expositions and demonstration of the results more transparent in the rest of
the paper.
Our control sample of non-bidding firms is obtained using the following
procedure. We begin by identifying the set of all ExecuComp firms that were neither
targets nor bidders in acquisitions with relative size of at least 10% during our entire
sample period (1993-2000). This is the base set of potential control firms. For each
bidder in our sample, we select a control firm with the closest market value of total assets
to that of the sample firm at the fiscal yearend after the merger is consummated.
Matching is done without replacement, so a control firm is matched to only one sample
firm.3
Estimating Normal Compensation
We construct a model of normal compensation by regressing the log of CEO
compensation on firm characteristics identified in the compensation literature (for
example, Murphy (1985), Agrawal and Walkling (1994), Yermack (1995), Core,
Holthausen and Larcker (1999), and Grinstein and Hribar (2004)).
We expect that larger firms and firms with more growth opportunities will
demand higher quality managers and thus offer higher pay. We proxy for firm size with
sales. We proxy for growth opportunities in the firm’s investment opportunity set with
the firm’s yearend market-to-book ratio averaged over the previous five years. We
3
The results are robust to other matching criteria such as sales, ROA or combination of sales and
ROA. Because of the limited pool of potential matches (i.e., ExecuComp firms without acquisition
activity), the actual firms chosen for matching are relatively insensitive to which criterion we put more
weight on. Our primary tests later in the paper include controls for industry, size, M/B and other salient
characteristics.
11
include 48 Fama-French industry dummies (Fama and French (1997)) to control for
industry differences in the demand for managerial talent.
Agency theories suggest that the level of executive pay is an increasing function
of firm performance (Murphy (1985)). We employ three metrics for firm performance:
sales growth, the accounting return on assets (ROA, computed as the ratio of earnings
before interest and taxes to total assets), and the annual stock market return (computed as
the 12-month raw return of the firm’s stock in the fiscal year prior to the acquisition
announcement).
Firm risk captures both the firm’s information environment and operating
environment and is shown to be an important determinant of executive pay (see for
example, Core et al. (1999)). We have two total variance measures of firm risk: a
measure of the risk of operating performance and a measure of the risk of the stock
return. They are the standard deviation of ROA and the standard deviation of common
stock returns. The latter is obtained as the standard deviation of annual percentage stock
market return for the prior five years.
While the effectiveness of board monitoring is one factor that would affect the
CEO pay, we are primarily interested in the economic determinants of normal levels of
executive compensation. We take the relative effectiveness of corporate governance
mechanisms at a firm to be one factor that would cause its CEO pay to deviate from this
norm. Indeed, in later tests, we will be interested in the corporate governance of the
bidder as an explanation for changes in compensation policy and will use deviations from
the economically-justified pay norm as a measure of the strength of the board.
In sum, the following is our model of normal CEO compensation:
Payit = α 0 + find + ft + β1Salesit + β 2 M / Bit + β3Sales Growthit
+ β 4 ROAit + β5σ ROAit + β 6 Re turnit + β 7σ Re t it + eit ,
(1)
12
where the left-hand-side variable is the natural logarithm of cash pay, grants or total pay,
and year dummies are employed to account for economy-wide shocks.
The normal compensation model is estimated based on a panel data set that
includes all ExecuComp firms over the entire sample period. The estimation has industry
and year fixed-effects and uses standard errors that are robust to clustering at the firm
level. We obtain our measure for abnormal CEO compensation by computing the
difference between the actual compensation and the fitted value from the normal
compensation model for our sample of bidder and control firms.
Measures of CEO Wealth-Performance Sensitivities
Our first wealth-performance sensitivity measure, which Core and Guay (1999) call “new
equity incentives,” captures the change in the dollar value of the CEO’s newly granted
stock and options for a 1% change in the stock price. This measure recognizes the
argument by Hall and Liebman (1998) that large incentives can come from small
fractional holdings as long as the dollar amount is big enough. This measure of incentives
for stockholdings is straightforward as stock value increases by 1% for each 1% change
in the stock price. For options, the sensitivity of an option’s value to the stock price is the
partial derivative of option value with respect to price (the option “delta”), multiplied by
1% of the firm’s stock price. The CEO’s new equity incentives are measured by adding
the total incentives from the CEO’s current year option grants to 1% of the value of the
CEO’s holdings of newly granted stock and restricted stock.
Our two other measures of firm-specific CEO wealth-performance sensitivity
capture the immediate and long-run changes in the value of the CEO’s existing portfolio
of equity incentives to a change in firm value.4 Our immediate sensitivity measure, which
4
Unlike the previous literature, we examine the impact of acquisitions on the value of the CEO’s
existing holdings of stock and options in our performance sensitivity analysis. This is consistent with our
goal of capturing the full incentives of bidding CEOs around making takeover decisions.
13
Core and Guay (1999) call “portfolio equity incentives,” provides a snap-shot of the
CEO’s wealth sensitivity at a point in time.5 We measure a CEO’s portfolio equity
incentives by adding the total incentives from the CEO’s option portfolio to 1% of the
value of the CEO’s holdings of stock and restricted stock. This sensitivity measure
includes new equity incentives and the incentive derived from the existing portfolio
holdings of stock and options by the CEO at the beginning of either year ayr−1 or year
cyr+1. This portfolio delta is an ex-ante measure of the sensitivity of CEO wealth to a
price change at that point in time.
Our long-run wealth-performance sensitivity measure tracks the actual total value
of the CEO’s firm-specific portfolio from before to after the merger, accounting for new
grants, cash compensation, and adding back any value realized from stock and option
sales. As such, it is a measure of how closely associated CEO and shareholder wealth are
over time rather than a measure of the immediate marginal effect of a change in stock
price on CEO wealth.
To see the difference between the two wealth sensitivity measures, consider a
CEO in our underperforming (relative to the control firm) subsample with a $12 million
portfolio with a delta of 148 at the end of the fiscal year before the acquisition. This
indicates that as of that point in time, the CEO’s portfolio is such that it increases by
$148,000 for every 1% increase in shareholder wealth. From that point until the year
after the acquisition closes, the shareholders’ return is −52%, but the CEO is granted over
$10 million in new options over the same period. Due to the effects of the stock
5
In order to estimate the sensitivity of previously granted options, we follow the algorithm developed
by Core and Guay (2002) which does not require building up the CEO’s option portfolio over the years, but
rather requires only information available in the current year’s proxy statement. Specifically, the proxy
statement discloses information on unexercisable and exercisable previously granted options, which are
treated as two single grants. The strike price of each of these grants is derived from the average realizable
value of the options, assuming that unexercisable options have a time-to-maturity that is three years greater
than that of the exercisable options and out-of-the-money options have exercise prices equal to the fiscal
year end stock price. Core and Guay (2002) show that their approximation yields unbiased and accurate
estimates of pay-performance sensitivities.
14
performance, the new options, exercise of options and expiration of old out-of-the-money
options, and new cash compensation on his overall wealth, he actually ends-up 36%
richer at the end of the period. Thus, despite the fact that the portfolio equity incentive
measure suggests the CEO will suffer along with his shareholders, this static snapshot of
his wealth sensitivity misses the dynamic effect of new compensation flows to the CEO
after the acquisition.
A priori, there is no reason to believe that pay and wealth sensitivities will be
symmetric for positive and negative performance (Garvey and Milbourn (2003)). The
practice of option repricing, alone, would produce asymmetry. Therefore, most of our
tests will allow for different sensitivities to positive and negative performance.
4. Results
Table 1 presents descriptive statistics of our sample of 370 completed acquisitions. Panel
A reveals the temporal distribution of corporate acquisitions in our sample. The total
number of acquisitions has increased over most of the 1990s and experienced a drastic
decline at the turn of the century. This finding is not surprising as our sample period
coincides with the latest aggregate merger wave. Panel B gives an industry breakdown of
corporate acquisitions in our sample. The industries with the largest number of
transactions are Banking and Business Services, consistent with the finding in Bliss and
Rosen (2001) that the banking industry is experiencing massive consolidation over the
sample period.
Sample Characteristics
Financial and operating characteristics measured at the fiscal yearend prior to the
announcement of acquisitions (year ayr−1), and the first year after the acquisition (year
15
cyr+1) are reported in table 2, panels A and B, respectively. All dollar amounts are in
inflation-adjusted 2002 dollars.6 As expected of ExecuComp firms, the median firm is
quite large; the market value of equity is $1.9 billion. The book value of total assets is
$2.4 billion and total sales is $1.3 billion. The median debt ratio (book value of debt
divided by book value of total assets) is 0.21. The median annual sales growth is 8%. The
bidder firms are profitable in the year prior to the acquisition announcement with median
return on assets at 16% and median annual stock return at 25%. This equity return is also
above the market median. The value ratio of target over bidder equity (not tabulated) has
a median of 31%.
The bidder announcement return (not tabulated) has a mean of –1.3% and a
median of –1.1%. These returns are larger in magnitude than Datta et al.’s sample, but
are more comparable to Grinstein and Hribar’s sample of large deals. The method of
payment in 79% of our sample is bidder equity, which generally results in lower average
abnormal returns (Travlos (1987)). Finally, 31% of our sample deals are diversifying in
the sense that the bidder and target are not members of the same Fama-French 48
industry groups.
In the year after the completion of the acquisition (panel B), the market value of
equity is $3.1 billion. The book value of total assets is $4.0 billion and this is almost a
two-fold increase from the pre-acquisition level. Total sales increases to $2.2 billion from
the pre-acquisition level of $1.3 billion. The median debt ratio is 0.24, a moderate
increase from the pre-acquisition level at 0.21. The median annual sales growth increases
to 13% from 8%. The bidder firms remain profitable in the year after the completion of
the acquisition with median return on assets at 14% and median annual stock return at
6
In order to make the compensation comparison before and after the merger meaningful, the sample
summarized here is restricted to those CEOs who are the same from the year before the merger
announcement to the year after completion of the merger. Of the 370 original deals, this leaves 306 of the
sample CEOs and 292 of the control CEOs. We include the effect of dismissal for poor performance in our
later tests and discuss the subsample of CEOs who are replaced in the robustness section.
16
5%. However, they significantly underperform the overall equity market during the same
period.
Table 2 also presents descriptive statistics of CEO compensation before and after
the merger. In the year prior to the acquisition announcement (panel A), the median cash
pay to CEOs of bidding firms is slightly above $1 million, which is the limit for fully
deductible non-performance based cash compensation since 1993 (Hall and Liebman
(2000)). The median grants of stock and options for the current year is $0.90 million, thus
stock and options play quite an important role in compensating CEOs. The median total
pay is $2.3 million, which is about twice the median cash pay component. We also
compute the market value of the CEO’s existing holdings of stock and options in the firm
as a measure of the amount of personal wealth that the CEO has tied to the equity
performance of the firm. Table 2 shows that prior to the acquisition announcement, the
bidding firm CEOs have a median portfolio value of equity incentives of $23 million,
about ten times his annual total pay.
In the year after the completion of the acquisition (panel B), the median cash pay
to CEOs remains the same, while the median grants of stock and options for the year
increases to one and a half times of the pre-acquisition level, at $1.4 million. These
numbers indicate that from an incentive perspective, CEOs of bidding firms are rewarded
with substantially more equity-based compensation. As a result, total pay also increases
considerably with the median value equal to $3.5 million. After the merger completion,
the median total wealth of the bidding firm CEOs increases to $33 million. The $10
million change in wealth comes from a combination of new equity grants from year
ayr−1 to year cyr+1 and a median 19% return for the bidding firm over that (on average)
two-year period. A more precise breakdown and analysis of the components of the
change in CEO wealth is provided later in table 6.
17
Panel C of table 2 gives us a clear picture of how the compensation to bidding
firm CEOs changes after the completion of acquisitions. For cash pay, the median rate of
increase is 19%. For grants of stock and options, the median rate of increase is 37%. As a
result, the median change in total pay is 41%.
Table 2 also includes comparable statistics for CEOs in our control sample.
Overall, despite matching as closely as possible to the market value of total assets in year
cyr, the control firms are somewhat smaller in size than our sample firms, but are
otherwise well-matched on the metrics of leverage and performance. The data generally
show that while the pay and equity incentives of control firm CEOs are also increasing,
they are not increasing nearly at the rate of those documented for sample firm CEOs.
In summary, bidding firms are performing very well prior to the acquisition and
their CEOs are, relative to their counterparts in matched firms, highly compensated as
well. After the completion of acquisitions, the acquiring CEOs experience larger
increases in all three measures of compensation, suggesting that the acquiring CEOs
clearly have direct financial incentives in undertaking mergers. In the next subsection, we
estimate a model of normal compensation to evaluate how much of the acquiring CEO’s
compensation is abnormal or excess compensation.
Estimating Abnormal Compensation
Panel A of table 3 reports the estimation results for our normal compensation model.7
Sales, our proxy for firm size and complexity, is positively associated with CEO pay as
expected. M/B, our proxy for firm growth opportunities, and sales growth are positively
related to CEO annual grants and total pay, but not to cash pay. ROA is positive and
7
Note that in estimating the normal compensation model, we do not include firm fixed-effects. If a
firm tends to overpay its CEO in all years, then that would go into the fixed effect, leaving only over or
under payment relative to the firm’s own average pay. Since we are looking for over or under CEO
compensation relative to other firms, we do not want that to be lost to the fixed effect. Instead, to address
autocorrelation in the model, we use Huber-White robust standard errors in all panel estimations. These
standard errors are robust to both serial correlation and heteroskedasticity.
18
significant only for cash pay, presumably because of accounting-performance based
bonuses. The consistently positive coefficient on stock return for all measures of pay
demonstrates a positive pay-for-performance relation. The positive coefficient on
standard deviation of stock return shows that CEOs of more volatile firms get paid more
in terms of annual grants and total pay.
Based on our estimates from the normal compensation model in panel A, we
present evidence on whether the CEOs of the acquisition sample are over or under paid
relative to their expected normal levels of compensation. The first part of panel B gives
both the raw and abnormal levels of compensation in year ayr−1. We tabulate the pay for
the CEO of the median-performing firm, rather than the individual medians for each pay
component, so that the raw components of pay will approximately aggregate to total pay
(however, total pay to the CEO of median performing firm will be slightly larger than the
sum of cash pay and grants for the same CEO due to other payments included in total
pay). The table demonstrates that at both the raw and abnormal pay levels, the acquiring
firm CEO is overpaid relative to his counterpart in the control firm prior to the
acquisition, and his pay increases after the acquisition.8
CEO Wealth-Performance Sensitivities
Tables 2 and 3 establish that acquiring CEOs have large equity portfolios before the
merger, and that they add to them with abnormally large new grants following the
merger. As a result, they should have portfolio wealth that is highly sensitive at least to
8
The R-squared in table 3 panel A indicates our normal compensation model for grants has the poorest
fit, and panel B shows that the raw and abnormal levels of grants for our sample of bidding firm CEOs are
large. We attribute these results to several factors. First and foremost, firms do not make grants of stock and
options on an annual basis (Core and Guay (1999)), thus our annual economic determinants do a relatively
poor job in explaining grants in any one year (this is the same as saying that one year is a shorter than
optimal interval to try to predict grant behavior). As a result, the raw and abnormal levels of grants tend to
be close to each other. More importantly, the large raw level of grants for our bidding firm CEOs highlights
one point of the paper that the size of new flows of compensation in the 1990s is much larger relative to the
size of the CEO’s existing equity portfolio than twenty years ago. This has important implications for the
incentives faced by the CEO when making acquisition decisions.
19
short-term changes in the stock price. To distinguish between the incentive alignment
and self-serving management hypotheses, we directly measure the sensitivities (deltas) of
the CEOs’ portfolios before and after the acquisition and relate the changes to the
performance of the acquisition. Incentive alignment predicts that the portfolios of
bidding firm CEOs will remain sensitive to performance, even in the case of poorly
performing acquisitions.
Table 4 reports the wealth-performance sensitivity measures for the acquisition
and control samples both before and after the acquisition. These are the deltas defined in
Section 3 and, as immediate measures of the marginal (either new or total) wealth impact
of a change in the stock price, complement the long-run association measure presented
later in table 6.
Core and Guay (1999) report that the median change in the value of CEO new
grants of stock and options for 1% change in stock price is $12,251 for all ExecuComp
CEOs over 1993-1997. Our numbers in panel A show that the median new equity
incentives for our control sample are comparable to the Core and Guay number, while the
new equity incentives for our bidder sample CEOs are much stronger both before and
after the acquisition. In terms of the total portfolio equity incentives, Core and Guay
(1999) report the change in CEO median wealth for 1% change in stock price is $117,434
for all ExecuComp CEOs over 1992-1996. Our wealth-performance sensitivity numbers
in panel B show that for both control and sample CEOs, the wealth incentives are much
stronger, possibly because these are larger firms and our sample period includes the end
of the 1990s as well.
We observe that the wealth-performance sensitivities for our acquisition sample
CEOs increase substantially after the acquisition (panel B). Despite the fact that the
sensitivity of the control CEOs’ wealth also increases, the point estimate for sample CEO
wealth-performance sensitivity starts lower than that of the control firm counterpart and
20
ends higher. These sensitivity measures are economically large, as well. For example,
after the merger a 1% change in stock price leads to $424,990 change in CEO portfolio
wealth, as compared to $295,720 change in CEO portfolio wealth before the merger. A
comparison with the compensation numbers in table 2 demonstrates the economic
significance of this change in wealth-performance sensitivity. For example, after the
merger, a 3% decrease in stock price completely offsets the CEO’s entire cash
compensation (3*$424,990 = $1,274,970 compared to the median cash compensation of
$1,167,000 from table 2). Similarly, the portfolio effect of a 4% decrease in stock price
negates the value of the median CEO’s stock and option grants for the year ($1.4
million).
So far table 4 shows that, compared to their control counterparts, bidding firm
CEOs start with lower wealth sensitivities to performance and increase their sensitivities
after the merger. At first glance, this suggests that acquisitions improve the efficacy of
the compensation structure of the bidder firms by substantially strengthening the link
between CEO wealth and firm performance. However, table 4 also presents sensitivity
measures split on the performance of the merged firm. Prior to the acquisition, the
performance sensitivity of CEO portfolios is similar between the two subsamples. CEOs
who subsequently make underperforming acquisitions are receiving new grants with
performance sensitivities that are actually higher than those of CEOs who subsequently
make outperforming acquisitions. After the acquisition, a much different picture emerges.
The wealth-performance sensitivity of CEOs whose firms are underperforming drops, as
does the sensitivity of their new grants to performance. Conversely, for CEOs whose
firms outperform their control firms over the acquisition period, the sensitivities of new
grants and the overall portfolio double.
Together with the previous tables, these results add to the evidence that
compensation changes around the merger protect CEO wealth from underperformance
21
while making it very sensitive to good performance. This evidence supports the selfserving management hypothesis. To further investigate how this happens, we examine
how the amount of new flows of CEO compensation changes with respect to performance
before and after the merger in the next table.
Changes in CEO Pay-Performance Sensitivity around the Merger
The results to this point suggest that acquisitions make CEOs better off. Acquiring CEOs
usually have experienced strong firm performance and abnormally high compensation
prior to making the acquisition. Despite having performance that returns to normal on
average following the acquisition, their compensation continues to be abnormally high.
Consistent with the self-serving management hypothesis, the performance sensitivity of
their firm-specific wealth increases only when performance is good. In the next test, we
extend the normal compensation model of table 3 to capture changes in how
compensation is set after the acquisition. We attempt to control for sales growth and
ROA due solely to the merger. Based on the earlier results showing that pay-performance
sensitivity can be different for good vs. bad stock performance, we also estimate the
specifications splitting the stock returns into positive and negative returns. The results
will not only demonstrate whether internal growth of the firm, growth achieved by
acquisition, or pure acquisition effects drive the change in CEO compensation, but also
tell us how CEO pay-performance sensitivity changes after an acquisition. We run the
following pooled time series cross sectional regression over the entire sample period
using both the acquirer and control samples,
Payit = α 0 + find + ft + β1 Pr eAcqit + β 2 Acqit + β 3MultAcqit + β 4 Salesit + β 5 M / Bit
+ β 6 Sales Growthit + β 7 ISales Growthit + β8 ESales Growthit + β9 ROAit + β10 IROAit
(2)
+ β11EROAit + β12σ ROAit + β13 Re turnit + β14 Pos Re turnit + β15 Neg Re turnit
+ β16 ( Acqit * Pos Re turnit ) + β17 ( Acqit * Neg Re turnit )
+ β18 ( Strong Boardit * Acqit * Neg Re turnit ) + eit .
22
The left-hand-side variable is the logarithm of compensation in year t. PreAcq is a
dummy variable set equal to one in all years before the announcement of a merger for
acquiring firms, and zero otherwise. Acq is an acquirer dummy variable set equal to one
for acquiring firms starting in year cyr+1, and zero otherwise. The coefficient on this
variable will capture on average the pure effect of a merger on compensation change
while the PreAcq variable controls for pre-merger pay level differences for acquiring
CEOs. MultAcq is a multiple acquisition dummy variable set equal to one starting in year
cyr+1 if the firm makes more than one acquisition during the sample period, and zero
otherwise. While the number of multiple acquirers in our sample is small, this dummy
will give us some indication of the overall CEO wealth effect of an acquisition program.
ISales Growth is the change in the logarithm of sales from year-to-year assuming
there were no mergers. Hence, for the control sample, this variable will be the actual
sales growth for each year; for the acquirer sample, this variable will be the actual sales
growth for the years without the merger event, and will be the industry median sales
(excluding acquirers) for the year when the merger is consummated. ESales Growth is the
change in the logarithm of sales from year-to-year due to the merger. Hence, for the
control sample, this variable will be zero throughout and for the acquirer sample, this
variable will be zero for all non-merger years. However, for the acquirer sample during
the year when the merger is consummated, ESales Growth will be the difference between
the firm’s actual sales growth and its industry’s median sales growth. IROA and EROA
are analogous variables for the accounting return on assets from year t−1 to t.
PosReturn is holding period fiscal year return from year t−1 to t if that return is
positive, and zero otherwise. NegReturn is analogously set equal to the actual return if
that return is negative, and zero otherwise. To capture the possible differential sensitivity
of compensation to performance for sample firms after the acquisition, we interact the
return variables with the acquisition dummy. If the coefficient on the negative return
23
interaction variable is negative, then through acquisition, bidding firm CEOs are able to
detach their compensation from any poor performance by their companies. Interaction
terms are denoted by Acq* and represent the interaction of the acquisition dummy and
the identified variable. (We will discuss later the final interaction variable Strong
Board*Acq*NegReturn.) Industry dummies are employed to control for industry
compensation practices, and year dummies are employed to account for economy-wide
shocks.
The results of the estimation are presented in table 5. Because the observations
for this test are limited to the control and sample firms, the first column of the table
repeats the specification from table 3 on this sample to show that the basic compensation
model behaves similarly in this restricted sample.
Consistent with the results of table 3 showing that acquiring CEOs are abnormally
well-paid prior to the acquisition, the second column of table 5 shows a positive
coefficient on PreAcq indicating that acquiring CEOs receive abnormally high
compensation on average each year leading-up to the acquisition. The results also show
that compensation is positively related to internal sales growth and stock returns. The
coefficient on the external sales growth variable provides evidence that compensation is
increased when sales growth is increased through the effect of an acquisition. The
acquisition-related sales growth effect is equivalent to the internal sales growth effect. In
comparison, using large merger deals in the mid-1980s, Avery et al. (1998) find that
external sales growth is weighted more heavily in bidding CEO cash pay than internal
sales growth. Bliss and Rosen (2001) show that merger-related asset growth increase
CEO total pay faster than internal growth for their sample of bank mergers. We attribute
the difference in results to different samples and measures of compensation.
Splitting stock returns into positive and negative returns yields some interesting
insights. First, changes in compensation are more strongly related to negative returns
24
than to positive returns. Thus, compensation is more likely to be lowered after poor
returns. However, the negative coefficient on the interaction of negative stock returns
after acquisitions (Acq*NegReturn) counters the baseline positive link between poor
stock performance and changes in compensation. The sum of the two coefficients
(NegReturn and Acq*NegReturn) is positive, but insignificantly different from zero.
Thus, while CEOs normally are penalized for poor performance, acquiring CEOs are not
in the post-acquisition period; the acquisition and subsequent changes in compensation
sever the link between performance and compensation, but only on the downside.9
Corporate Governance and Post-acquisition Pay-for-Performance
The results thus far paint a clear picture of asymmetry in the CEO’s pay-for-performance
relationship after an acquisition. However, this downside protection afforded to the CEO
could be optimally chosen by a board seeking to provide risk-taking incentives through
convexity in pay with respect to performance. In particular, is the board simply
compensating managers for ex-ante good investments that are ex-post (perhaps due to
unforeseen factors) value-decreasing? To explore this issue, we divide sample firms by
strong versus weak board bargaining position and try to relate the degree of dampening of
the downside effect to this split. If the convexity in the payoff to making an acquisition
were optimal from a contracting perspective, then we would expect the same curvature in
compensation policy irrespective of the negotiating strength of the board vis-à-vis the
CEO.
The corporate governance literature has not settled on any one measure of the
strength of the board (see Hermalin and Weisbach (2003) for a review), and specific
9
To allow for the fact that merger decisions might be endogenous to the CEO pay determination
equation, we replace dummy variables Acq and MultAcq with instruments generated as the predicted
probabilities from a probit regression where prior year return, cash flow to total assets, leverage, M/B,
ROA, firm size and sales growth serve as explanatory variables (Harford (1999), and Grinstein and Hribar
(2004)). We find our results remain qualitatively robust.
25
measures of governance can be most relevant for different aspects of firm behavior. For
board strength, we turn to the results of Core et al. (1999), who show that a measure
based on abnormal CEO compensation is a robust proxy for the overall strength of the
board vis-à-vis the CEO. Thus we construct our strong board dummy as follows: Strong
Board is set equal to one for those (both sample and control) firms whose CEOs are in the
lowest quartile of abnormal compensation prior to the merger, and zero otherwise. In the
final column of table 5, we include an extra interaction variable Strong
Board*Acq*NegReturn to allow the post-acquisition pay-for-negative performance
sensitivity to be different for firms with strong corporate governance mechanisms. The
results cast doubt on the hypothesis that boards optimally choose to provide downside
protection following acquisitions. The coefficient on the Strong Board interaction term is
large and positive, countering the negative coefficient on the post-acquisition negative
return interaction term (Acq*NegReturn). An F-test confirms that for bidding firms with
strong boards, there remains a positive relation between pay and performance, even for
negative performance, after the acquisition. This effect is absent for firms with weak
boards, suggesting that the downside protection we observe is the result of captured
boards rather than an attempt to provide optimal risk-taking incentives to CEOs with
regard to acquisitions.10 Our results are consistent with those of Garvey and Milbourn
(2003), who employ a different metric for the strength of corporate governance
mechanisms in their sample firms and conclude that only firms with weak shareholder
protections exhibit strong asymmetry in compensation benchmarking.
In our final test, we look at the long-run relation between firm performance over
the acquisition period and the CEO total wealth over the same period. The results so far
show that, except for firms with strong boards, CEO pay is particularly insensitive to
10
One criticism of our strong board measure is that it is based on the CEO’s pay in the year prior to the
acquisition. Thus, we might mistakenly assign a firm to have weak board when in fact the abnormally low
pay to its CEO is driven by greater sensitivity to poor performance (reverse causality). However, the fact
that most of bidding firms enjoy superior stock market performance prior to the acquisition (table 2)
mitigates this concern.
26
poor performance over this period. The next set of results will reveal how these
compensation changes affect his wealth.
Long-run Effect of an Acquisition on CEO Wealth
Our analysis starts with the total value of the CEO’s portfolio of equity incentives (stock
and options) in the year prior to the bid announcement and then examines the effect of
long-run firm performance on the value of that portfolio. This represents the expected
wealth effect to a CEO contemplating an acquisition bid. In doing so, we start by
measuring the change in the total value of the CEO’s portfolio from before to after the
acquisition. This includes the effect of firm performance on the value of the equity
portfolio and the addition of any new grants to the portfolio during the period. We next
account for the CEO’s trading activities, adding back the value of any shares sold or
options exercised during the same period. For any year in which his shareholdings in the
firm decrease, we multiply the decrease in the number of shares by the midpoint of the
share price that year to estimate his income from selling shares in that year. ExecuComp
reports the value realized each year when a CEO exercises his options. We add this value
to our estimate of the value gained from selling his shares in the firm to produce an
estimate of the CEO’s income from trading activities for that year. We cumulate the
CEO’s income from trading activities and add that back to the total value of the CEO’s
remaining portfolio at the end of the measurement period. Finally, we cumulate cash
compensation to the CEO and add it to his total wealth. This approach tracks the total
impact of the firm’s performance and any changes in compensation on the CEO’s firmspecific wealth.11
11
If a CEO ceases to be employed, we set further compensation equal to zero and make a rough
estimate of how the value of his portfolio would evolve: we multiply the last known total portfolio delta by
the stock return over each succeeding year. We do this rather than dropping the CEOs who are no longer
employed in order to avoid biasing our results upward by removing CEOs who are potentially fired for
poor performance. We also perform the test without including CEOs who do not survive the merger deal
and find that the inferences are unchanged.
27
Table 6 presents the long-run association between CEO wealth and the stock
performance of the merged firm. The first three columns show the wealth change for the
CEO of the median-performing acquirer. From year ayr−1 to year cyr+1, the firm has a
raw stock return of 19%, underperforming its matched firm by almost 7%. Yet, the
acquiring CEO has a wealth increase of 39%, about half of which comes from new grants
of stock and options.
To provide some further evidence on the sensitivity of CEO wealth to
performance, in the remainder of the table, we separate acquiring CEOs based on whether
their firms outperform or underperform their matched control firms from year ayr−1 to
year cyr+1. The middle three columns in table 6 show that positive performance is very
richly rewarded, yielding a wealth increase for the CEO of the median outperforming
firm of almost $23 million at the end of the year following the merger completion (more
than doubling his starting wealth in year ayr−1). About two-thirds of these increases are
driven by the strong performance of the underlying stock, which translates into more
valuable option exercises and stock holdings. Much of the rest comes from new grants of
options and restricted stock.
For the median underperforming acquirer, we find that while a shareholder
invested in the acquiring firm would have experienced a 52% reduction in his wealth, the
CEO of the same firm experiences a net wealth increase of 36%. Offsetting the
downward pressure of the falling stock price on his wealth, are new grants of options and
stock with a value equivalent to almost 90% of his starting wealth.12
12
It is important to note that the cumulative grant date value of grants made over the period cannot be
simply aggregated with the other components of the change in CEO wealth. This value is implicitly part of
the raw change in the CEO’s portfolio value, which is driven by changes in the composition of the
portfolio, changes in the factors affecting the value of the option portion of the portfolio, and the stock
price performance. In particular, the value of the options in his portfolio at the end of year cyr+1 will differ
from their value when granted. We track not only the total change in the value of the portfolio, but the
value of the options when granted as a measure of the cost to the firm of these compensation flows.
28
Notably, as shown in panel b, the CEOs of acquiring firms that underperform
their control do not experience any wealth loss on average. Despite median stock
underperformance of −43%, offsetting grants leave the median underperforming CEO to
be 36% richer in the year following the merger.13 In fact, these large new grants are so
pervasive that 78% of the CEOs of underperforming firms are better off following the
merger. The univariate results suggest that when the merged firm does well, the CEO
enjoys exposure to the rising stock price and is additionally rewarded by new stock and
option grants. Conversely, when the merged firm performs poorly, the CEO is left
relatively unexposed to the stock performance by virtue of offsetting stock and option
grants.
In panel C of table 6, we employ a regression model to more directly examine the
association between post-merger acquiring CEO wealth and stock performance. The
results reveal a large positive intercept with an additional one-for-one increase in wealth
for increases in stock return. However, the negative coefficient on the interaction term
capturing negative stock performance completely negates any sensitivity the CEO’s
wealth has to performance on the downside. Thus, acquiring CEOs are handsomely
rewarded for positive performance, but not punished for poor performance. This
asymmetric wealth change profile makes acquisitions an attractive corporate strategy
from a CEO’s standpoint. These results are not consistent with incentive alignment, but
rather are consistent with the self-serving management hypothesis.14
13
Note that panel A of table 6 refers to the wealth effect of the CEO of the median underperforming
firm (median underperformance), while panel B contains the median wealth effect of the underperforming
subsample (the distribution presented is that of the wealth effect, conditional on performance).
14
We estimate an expanded regression model with a Strong Board interaction term similar to the final
specification in table 5. The coefficient on the Strong Board interaction variable is insignificant, indicating
that even firms with strong boards are unable to fully make their CEO wealth sensitive to negative
performance. We attribute the difference between the results on compensation in table 5 and those on
wealth in table 6 to the fact that CEO pay is completely under the control of the board while CEO firmspecific wealth is not (it is affected by CEO trading decisions, etc.). In fact, when we estimate a similar
regression, but using the scaled cumulative grants instead of total wealth change as the dependent variable,
we find that cumulative grants given in firms with strong corporate governance mechanisms are sensitive to
29
5. Robustness Checks
Much of our analysis was limited to the sample of bidding and control firm CEOs who
remained CEOs over the period of analysis. We have also examined the sample of firms
whose CEOs were different following the merger. These firms are generally somewhat
larger and their CEOs are actually paid somewhat less than CEOs who survive. For the
most part, the strength of their portfolio equity incentives is also weaker. Even the new
equity incentives granted to new CEOs following the merger are smaller than those of
surviving CEOs. We note that 78 of the control CEOs did not survive the event period
while only 64 of the sample CEOs did not survive. This suggests that if anything,
acquiring another firm actually increases the survivability of the CEO. We also note that
the analysis in table 6 treats a non-surviving CEO as one who gets no compensation at all
after leaving the firm. Thus, if poor performance leads to CEO dismissal, we are being
conservative in estimating the wealth of a dismissed CEO, and yet we still find that on
average CEOs of poorly performing firms are far better-off than their shareholders.
While our entire analysis has been based on a sample of bidder firms whose
targets were at least 10% of their size, we have repeated the analysis without imposing
that restriction, yielding 622 events. The overall inferences are unchanged in the
unrestricted sample. Naturally, because these acquisitions are less important to the
bidder, some of the post-merger changes, while still significant, are less substantial
economically than they are in the restricted sample.
Finally, we have examined whether interactions that account for the
characteristics of the bid or the abnormal stock price reaction to the announcement of the
bid have any effect on the compensation and wealth changes. We separate bids based on
negative performance. Finally, defining the negative return dummy using the cumulative raw return instead
of the cumulative abnormal return makes no material difference on our estimation results in panel C.
30
whether they are diversifying and based on the method of payment (cash or stock). We
also control for whether the bidder announcement return is positive or negative. Our
finding that compensation changes made following a merger largely insulate the
acquiring CEO from the effects of long-run performance on his wealth is robust to all of
these controls. Further, there is no significant correlation between long-run performance
and the announcement reaction. We do find some evidence that CEOs making bids that
are met with negative announcement returns have greater abnormal compensation before
the bid.
6. Conclusion
This study investigates the intersection of two areas of the literature receiving a great deal
of attention: compensation and acquisitions. Because CEO and shareholder interests can
diverge greatly in a merger attempt, the effects of CEO compensation on their incentives
to make acquisition bids are of clear interest.
In the past, CEOs suffered along with shareholders through the effect of the
merger on their stock portfolios (e.g., Benston (1985)). However, the common practice
of supplementing CEO cash compensation with large grants of new options and restricted
stock appears to have changed that. We find that current compensation practices give
CEOs incentives to undertake acquisitions as their total pay and overall wealth increase
substantially following an acquisition. More importantly, we show that, except in the
minority of the best governed firms, a CEO’s pay following a merger becomes markedly
insensitive to poor performance, while increasing in-step with good performance. Thus,
the net effect is that a CEO’s wealth actually increases even if he makes a poor
acquisition decision. It is conceivable that by stringing-together a few acquisitions, a
31
CEO could engineer a series of structural shifts in his compensation scheme that would
effectively sever the link between firm performance and his wealth over his tenure.
Our evidence suggests that current compensation schemes may actually be less
effective in controlling the agency conflict than previously thought. This is because the
flow of new incentives can actually reduce the effectiveness of existing incentives. Thus,
our results are consistently supportive of the self-serving management hypothesis rather
than the incentive alignment hypothesis. These results complement recent research by
Garvey and Milbourn (2003) who find in a broad setting that there is an asymmetry in
executive compensation benchmarking; there is significantly less pay-for-luck when luck
is down than when it is up. Together, these results suggest that compensation research
should begin to focus on how the flow of new compensation in reaction to an event or
change in performance impacts the ex-ante perceived incentive effect of pre-event
executive portfolios.
32
References
Agrawal, Anup, and Jeffrey F. Jaffe, 2000, The post-merger performance puzzle, in: Cary Cooper
and Alan Gregory eds., Advances in Mergers and Acquisitions v. 1, New York, NY:
Elsevier Science, 7-41.
Agrawal, Anup, and Gershon N. Mandelker, 1987, Managerial incentives and corporate
investment and financial decisions, Journal of Finance 42, 823-837.
Agrawal, Anup, and Ralph A. Walkling, 1994, Executive careers and compensation surrounding
takeover bids, Journal of Finance 49, 985-1014.
Avery, Christopher, Judith A. Chevalier, and Scott Schaefer, 1998, Why do managers undertake
acquisitions? An analysis of internal and external rewards for acquisitiveness, Journal of
Law, Economics, and Organization 14, 24-43.
Benston, George J., 1985, The self-serving management hypothesis: Some evidence, Journal of
Accounting and Economics 7, 67-84.
Bliss Richard T., and Richard J. Rosen, 2001, CEO compensation and bank mergers, Journal of
Financial Economics 61, 107-138.
Core, John E., Robert W. Holthausen, and David F. Larcker, 1999, Corporate governance, chief
executive officer compensation, and firm performance, Journal of Financial Economics
51, 371-406.
Core, John E., and Wayne Guay, 1999, The use of equity grants to manage optimal equity
incentive levels, Journal of Accounting and Economics 28, 151-184.
Core, John E., and Wayne Guay, 2002, Estimating the value of employee stock option portfolios
and their sensitivities to price and volatility, Journal of Accounting Research 40, 613630.
Datta, Sudip, Mai Iskandar-Datta, and Kartik Raman, 2001, Executive compensation and
corporate acquisition decisions, Journal of Finance 56, 2299-2336.
Fama, Eugene, and Kenneth French, 1997, Industry costs of capital, Journal of Financial
Economics 43, 153-193.
Garvey, Gerald T., and Todd T. Milbourn, 2003, Asymmetric benchmarking in compensation:
Executives are rewarded for good luck but not penalized for bad, Journal of Financial
Economics, forthcoming.
Grinstein, Yaniv, and Paul Hribar, 2004, CEO compensation and incentives – Evidence from
M&A bonuses, Journal of Financial Economics 73, 119-143.
Hall, Brian J., and Jeffrey B. Liebman, 2000, Taxation and the compensation of CEOs, in:
Poterba James M. ed., Tax Policy and the Economy v. 14, Cambridge, MA: MIT Press.
Hall, Brian J., and Jeffrey B. Liebman, 1998, Are CEOs really paid like bureaucrats? Quarterly
Journal of Economics 113, 653-691.
33
Harford, Jarrad, 1999, Corporate cash reserves and acquisitions. Journal of Finance 54, 1969-97.
Hermalin, Benjamin E., and Michael S. Weisbach, 2003, Boards of directors as an endogenously
determined institution: A survey of the economic literature. FRBNY Economic Policy
Review 9, 7-26.
Jensen, Michael C., and William H. Meckling, 1976, Theory of the firm: Managerial behavior,
agency costs and ownership structure, Journal of Financial Economics 3, 305-360.
Lewellen, Wilbur, Claudio Loderer, and Ahron Rosenfeld, 1985, Merger decisions and executive
stock ownership in acquiring firms, Journal of Accounting and Economics 7, 209-231.
Moeller, Sara B., Frederik P. Schlingemann, and René M. Stulz, 2004, Wealth destruction on a
massive scale: A study of acquiring firm returns in the merger wave of the late 1990's,
Journal of Finance, forthcoming.
Murphy, Kevin J., 1985, Corporate performance and managerial remuneration: An empirical
analysis, Journal of Accounting and Finance 7, 11-42.
Murphy, Kevin J., 1999, Executive compensation, in: Ashenfelter, O., Card, D. (Eds.), Handbook
of Labor Economics, Vol. 3, North-Holland, Amsterdam, 2485-2563.
Travlos, Nickolaos G., 1987, Corporate takeover bids, method of payment, and bidding firm’s
stock returns, Journal of Finance 52, 943-963.
Yermack, David, 1995, Do corporations award CEO stock options effectively? Journal of
Financial Economics 39, 237-269.
Zhao, Mengxin, and Kenneth Lehn, 2003, CEO turnover after acquisitions: Do bad bidders get
fired? J.M. Katz School of Business working paper.
34
Table 1
Distribution of Corporate Acquisitions across Time and Industries, 1993-2000
The sample consists of 370 completed acquisitions announced during the period January 1, 1993, to December 31,
2000. The bidders are listed in Securities Data Company’s Mergers and Acquisitions database and have executive
compensation data in Standard and Poor’s ExecuComp database. The industry classification follows Fama and
French (1997).
Panel A: Distribution by Year
Frequency
1993
12
1994
33
1995
57
1996
55
1997
66
1998
71
1999
55
2000
21
Total
370
Panel B: Distribution by Industry
Percent
3.24
8.92
15.41
14.86
17.84
19.19
14.86
5.68
100%
Aircraft
Apparel
Automobiles and Truck
Banking
Business Services
Business Supplies
Candy & Soda
Chemicals
Coal
Communication
Computers
Construction
Construction Material
Consumer Goods
Defense
Electrical Equipment
Electronic Equipment
Entertainment
Food Products
Healthcare
Insurance
Machinery
Measuring and Control
Medical Equipment
Miscellaneous
Mining
Personal Services
Petroleum
Pharmaceutical
Precious Metals
Printing and Publishing
Restaurants, Hotels, etc
Retail
Rubber and Plastic
Steel Works Etc
Textiles
Tobacco Products
Trading
Transportation
Utilities
Wholesale
Frequency
6
2
3
55
41
8
1
13
1
7
10
3
8
4
2
4
18
5
4
6
18
16
4
10
2
2
2
12
8
2
5
5
16
2
7
2
1
8
9
24
14
Percent
1.62
0.54
0.81
14.86
11.08
2.16
0.27
3.51
0.27
1.89
2.7
0.81
2.16
1.08
0.54
1.08
4.86
1.35
1.08
1.62
4.86
4.32
1.08
2.7
0.54
0.54
0.54
3.24
2.16
0.54
1.35
1.35
4.32
0.54
1.89
0.54
0.27
2.16
2.43
6.49
3.78
Table 2
Descriptive Statistics of Acquiring and Control Firms
The sample starts with the 370 completed acquisitions announced during the period January 1, 1993, to December
31, 2000 summarized in table 1. The control sample is obtained by matching market value of total assets at the fiscal
yearend after the merger is consummated. All variable values are obtained at the fiscal yearend either before the
merger announcement (year AYR−1) or after the merger completion (year CYR+1). All dollar values are measured
in constant 2002 dollars (millions for firm characteristics, thousands for CEO compensation). We report the median,
as well as the 5th and 95th percentile values. To make the compensation comparison before and after the merger
meaningful, we require the CEO to remain the same between year AYR−1 and year CYR+1, leaving 306 sample
firms and 292 control firms for analysis in the table. MV Equity is obtained as the product of the number of shares
outstanding and the stock price as of the fiscal yearend. MV Assets is obtained as book value of total assets minus
book value of equity plus market value of equity. Leverage is captured by the ratio of book value of long-term debt
and either book or market value of total assets. M/B is the ratio of MV Assets and book value of total assets. Sales
Growth is the difference in log sales from year t-1 to t. ROA is the accounting return on assets, obtained as the ratio
of earnings before interest and taxes to total assets. Return is the annual stock return during the fiscal year. MktAdjusted Return is the annual firm stock return adjusted for the contemporaneous annual return on the market
portfolio. Cash Pay is the sum of salary and annual bonus. Grants is the total value of all restricted stock and options
granted during the year. Total Pay is the sum of salary, bonus, other annual compensation, value of restricted stock
granted, value of new stock options granted during the year, long-term incentive payouts and all other compensation.
Portfolio Value of Equity Incentives is the market value of the CEO’s existing holdings of stock and options at the
fiscal yearend.
Table 2, continued
Firm Characteristics
MV Equity
MV Assets
Book Assets
Sales
Debt/MV Assets
Debt/Book Assets
M/B
Sales Growth
ROA
Return
Mkt-Adjusted Return
CEO Compensation
Cash Pay
Grants
Total Pay
Portfolio Value of Equity
Incentives
Firm Characteristics
MV Equity
MV Assets
Book Assets
Sales
Debt/MV Assets
Debt/Book Assets
M/B
Sales Growth
ROA
Return
Mkt-Adjusted Return
CEO Compensation
Cash Pay
Grants
Total Pay
Portfolio Value of Equity
Incentives
5th Pct
Sample
Median
Control
Median
95th Pct
266.284
408.396
184.232
161.183
0.001
0.004
1.013
-0.172
0.029
-0.277
-0.465
1949.200
4312.474
2431.040
1299.346
0.131
0.206
1.517
0.082
0.163
0.248
0.034
1848.751
3525.149
1904.955
1632.936
0.147
0.241
1.494
0.063
0.165
0.142
-0.019
11286.662
36803.060
29454.005
12206.404
0.485
0.577
5.657
0.377
0.416
1.376
0.982
410
0
558
1144
893
2449
3643
12385
16579
392
0
490
1055
489
1958
3120
7605
9959
2273
23235
342172
669
20100
508500
272.829
653.403
338.340
292.513
0.002
0.004
0.970
-0.317
0.023
-0.539
-0.661
3119.444
6690.232
3956.137
2201.771
0.167
0.244
1.425
0.131
0.135
0.052
-0.120
Panel B: CYR+1
52457.418
321.860
129456.026
591.632
75120.000
287.108
28777.000
198.567
0.458
0.000
0.515
0.000
4.391
0.989
0.618
-0.111
0.354
0.019
0.820
-0.484
0.705
-0.604
2247.029
4619.546
2304.937
1950.633
0.168
0.274
1.476
0.065
0.153
0.084
-0.065
18986.245
49361.983
35118.199
14804.777
0.511
0.612
6.712
0.322
0.358
0.953
0.791
352
0
722
1167
1425
3464
5189
17663
22087
231
0
467
1127
656
2747
3172
11305
15524
3751
33167
367097
1690
23135
864753
95th Pct
5th Pct
Panel A: AYR−1
27312.512
295.470
71465.366
443.174
46904.500
166.997
13544.000
147.617
0.362
0.000
0.491
0.000
4.535
1.012
0.465
-0.071
0.378
0.029
1.025
-0.315
0.803
-0.538
Panel C: Median CEO-Specific Change from Year AYR−1 to Year CYR+1
Sample
Control
Cash Pay
18.92%
8.45%
Grants
37.02%
27.44%
Total Pay
40.90%
23.76%
Table 3
Are CEOs of Acquiring Firms Overpaid?
This table reports the estimation results for our model of normal CEO compensation. The dependent
variable is CEO compensation measured in natural logs, and we use all three measures of CEO
compensation in our estimation. Sales is measured as its natural log. M/B is the mean market-to-book ratio
of assets from the prior five years. Sales Growth is the difference in log sales from year t-1 to t. ROA is the
accounting return on assets. Return is the annual stock return during the fiscal year. σROA and σRet are the
standard deviations of ROA and stock return, computed over the prior five years. The model is estimated
using all ExecuComp firms over the entire sample period. Corresponding p-values from Huber-White
robust standard errors are reported in brackets.
Panel A: Regression Results from Model for Normal Compensation
Cash Pay
Grants
Intercept
4.499
-1.439
[0.000]
[0.065]
Total Pay
4.058
[0.000]
Sales
0.316
[0.000]
0.682
[0.000]
0.462
[0.000]
M/B
-0.017
[0.408]
0.237
[0.002]
0.111
[0.000]
Sales Growth
-0.016
[0.744]
0.966
[0.000]
0.188
[0.019]
ROA
0.637
[0.000]
-1.129
[0.111]
0.017
[0.940]
σROA
-0.186
[0.534]
-0.290
[0.793]
0.155
[0.686]
Return
0.174
[0.000]
0.304
[0.000]
0.190
[0.000]
σRet
0.014
[0.764]
0.345
[0.082]
0.175
[0.007]
YES
0.466
6027
YES
0.154
6027
Industry and Year Dummies?
Adj R2
Number of Observations
YES
0.451
6027
Table 3, continued
Panel B: CEO Compensation Before versus After the Acquisition ($, in thousands)
The table presents the raw and abnormal CEO compensation figures for the median performing sample firm
between year AYR−1 and year CYR+1 and the control firm with the closest performance during the same
period. For comparison, we require the CEO to remain the same between year AYR−1 and year CYR+1.
Cash Pay is the sum of salary and annual bonus. Grants is the total value of all restricted stock and options
granted during the year. Total Pay is the sum of Cash Pay, Grants and other remunerations. The abnormal
compensation is computed as the difference between the actual level of compensation and the
compensation level predicted by the appropriate model from panel A (and converted to dollars from logged
values). Because there is a different model of normal compensation for each of Cash Pay, Grants and Total
Pay, the abnormal compensation figures do not need to add up.
Cash Pay
AYR−1
Control
Sample
CYR+1
Control
Sample
Grants
Total Pay
Raw
Abnormal
829.79
107.36
262.38
218.83
1095.61
-20.39
Raw
Abnormal
780.14
311.70
1096.92
1070.19
1893.05
1162.03
Raw
Abnormal
743.36
56.96
61.25
22.74
807.27
-292.65
Raw
Abnormal
767.06
213.90
4073.70
3999.14
5258.45
4123.90
Table 4
CEO Wealth-Performance Sensitivities
New equity incentives are defined as the change in the dollar value of the CEO’s newly granted stock and
options for a 1% change in the stock price. Portfolio equity incentives measure the CEO’s portfolio equity
incentives by adding the total incentives from his option portfolio to 1% of the value of his holdings of
stock and restricted stock. Note that this sensitivity measure includes new equity incentives and the
incentive derived from the existing portfolio holdings of stock and options by the CEO at the beginning of
the fiscal year. All dollar values are measured in constant 2002 dollars (in thousands). For comparison, we
require the CEO to remain the same between year AYR−1 and year CYR+1. Tests of difference between
control and sample CEOs in the same column are non-parametric Wilcoxon tests. Tests of difference
between years for a given set of CEOs are non-parametric signed-rank tests of the CEO-specific paired
differences. Corresponding p-values are reported in brackets.
Panel A: New Equity Incentives Before versus After the Acquisition
Control
Sample
Test of difference
AYR–1
13.57
20.74
[0.004]
CYR+1
13.01
23.74
[0.002]
Acquiring Firms Only, Split on Performance
Underperform Control Firm
27.07
Outperform Control Firm
14.97
20.03
29.35
Test of difference
[0.012]
[0.002]
Test of
difference
[0.015]
[0.021]
[0.072]
[0.001]
Panel B: Portfolio Equity Incentives Before versus After the Acquisition
Control
Sample
Test of difference
AYR–1
335.32
295.72
[0.222]
Acquiring Firms Only, Split on Performance
Underperform Control Firm
309.99
Outperform Control Firm
281.65
Test of difference
[0.522]
CYR+1
391.31
424.99
Test of
difference
[0.001]
[0.001]
[0.295]
290.93
563.82
[0.008]
[0.024]
[0.001]
Table 5
Changes in CEO Pay-Performance Sensitivity around the Merger
This table reports regression results examining the changes in CEO pay-performance sensitivities
subsequent to acquisitions. The dependent variable is the logarithm of CEO’s total pay for year t. PreAcq is
a dummy variable set equal to one in all years before the announcement of a merger for acquiring firms,
and zero otherwise. Acq is an acquirer dummy variable set equal to one for acquiring firms starting in year
CYR+1, and zero otherwise. MultAcq is a multiple acquisition dummy variable set equal to one starting in
year CYR+1 if the firm makes more than one acquisition during the sample period, and zero otherwise.
Sales is measured as its natural log. M/B is the firm’s mean market-to-book ratio of assets from the prior
five years. Sales Growth is the difference in log sales from year t-1 to t. ISales Growth is the change in the
logarithm of sales from year-to-year assuming there were no mergers. Hence, for the control sample, this
variable will be the actual sales growth for each year; for the acquirer sample, this variable will be the
actual sales growth for the years without the merger event, and will be the industry median sales (excluding
acquirers) for the year when the merger is consummated. ESales Growth is the change in the logarithm of
sales from year-to-year due to the merger. Hence, for the control sample, this variable will be zero
throughout and for the acquirer sample, this variable will be zero for all non-merger years. However, for
the acquirer sample during the year when the merger is consummated, ESales Growth will be the difference
between the firm’s actual sales growth and its industry’s median sales growth. ROA is the accounting
return on assets. IROA and EROA are analogous variables to ISales Growth and ESales Growth
respectively, for the accounting return from year t−1 to t. Return is the holding period fiscal year return
from year t−1 to t. σROA and σRet are the standard deviations of ROA and stock return, computed over the
prior five years. PosReturn is set equal to Return if it is positive, and zero otherwise. NegReturn is
analogously set equal to Return if it is negative, and zero otherwise. Interaction terms are denoted by Acq*
and represent the interaction of the acquisition dummy and the identified variable. Strong Board is a
dummy variable set equal to one for firms whose CEOs’ abnormal pay in year AYR−1 is in the bottom
quartile. Strong Board*Acq*NegReturn captures the post-acquisition sensitivity of pay to negative
performance for firms with strong corporate governance mechanisms. Industry dummies are employed to
control for industry compensation practices, and year dummies are employed to account for economy-wide
shocks. The sample contains all firm years for sample and control firms with sufficient data. Corresponding
p-values from Huber-White robust standard errors are reported in brackets.
Table 5, continued
Total Pay
3.677
[0.000]
Total Pay
3.702
[0.000]
PreAcq
0.133
[0.024]
0.134
[0.024]
Acq
0.021
[0.777]
0.024
[0.738]
MultAcq
0.406
[0.043]
0.409
[0.036]
Intercept
Total Pay
3.583
[0.000]
Sales
0.460
[0.000]
0.453
[0.000]
0.452
[0.000]
M/B
0.144
[0.000]
0.150
[0.000]
0.151
[0.000]
Sales Growth
0.295
[0.001]
ISales Growth
0.267
[0.007]
0.270
[0.006]
ESales Growth
0.251
[0.108]
0.261
[0.093]
IROA
-0.352
[0.297]
-0.364
[0.280]
EROA
0.104
[0.810]
0.109
[0.800]
-0.908
[0.090]
-0.930
[0.077]
PosReturn
0.171
[0.001]
0.168
[0.002]
NegReturn
0.455
[0.006]
0.462
[0.005]
Acq*PosReturn
0.069
[0.445]
0.062
[0.486]
Acq*NegReturn
-0.194
[0.451]
-0.638
[0.056]
ROA
-0.251
[0.448]
σROA
-0.697
[0.201]
Return
0.234
[0.000]
Strong Board * Acq * NegReturn
σRet
Industry and Year Dummies?
Adj R2
Number of Observations
0.746
[0.023]
0.135
[0.192]
0.174
[0.097]
0.178
[0.089]
YES
0.438
3793
YES
0.442
3766
YES
0.443
3766
Table 6
Long-run CEO Wealth Sensitivity to Firm Performance
Panel A: Decomposition of Changes in Long-run CEO Wealth
This panel presents the change in wealth for each of three representative acquiring CEOs: the CEO of the median performing acquirer, and the CEOs of the
median performing acquirers conditional on positive or negative performance. Performance is defined as the return from the fiscal year before the acquisition
announcement to the fiscal year after completion of the acquisition, adjusted for the return of the acquirer’s matching firm. The table displays the starting and
ending values of the CEO’s portfolio of equity incentives and its two components. It also shows the CEO’s total wealth change, which is the sum of the change in
the portfolio value, the effect of any stock sales or option exercises over the period, and his cash pay over the same period. The cumulative grant date value of
any option or restricted stock grants made over the period is also presented. This value is implicitly part of the raw change in the portfolio value, which is driven
by changes in the composition of the portfolio, changes in the factors affecting the value of the option portion of the portfolio, and the stock price performance.
The final two rows show what the CEO’s wealth change would have been had he invested the entire value of his equity portfolio in year AYR−1 in the stock of
his firm or the matching firm. The “% PV−1” column scales the year CYR+1 column value by the CEO’s starting portfolio value in year AYR−1.
Median Performing Acquirer
AYR−1
CYR +1
18.51%
25.37%
-6.85%
40034.0
32663.5
7370.6
38367.6
29606.2
8761.3
Return
Matching Return
Abnormal Return
Portfolio Value of Equity Incentives
Stock Value
Option Value
+ Raw Change: Portfolio Value
+ Cumulative Stock Sales
+ Cumulative Option Exercise
+ Cumulative Cash Pay
= Total Wealth Change
Cumulative Grants
If Invested in Acquiring Firm
If Invested in Control Firm
% PV−1
Acquirer with Median Positive
Performance
CYR +1 % PV−1
AYR−1
74.69%
32.53%
42.16%
20516.0
12645.1
7871.0
33589.9
14776.4
18813.5
Acquirer with Median Negative
Performance
CYR +1
AYR−1
% PV−1
-51.60%
-8.29%
-43.31%
11763.2
3060.9
8702.3
11634.8
2063.8
9571.0
-1666.4
6417.3
7312.5
3579.4
15642.8
-4%
16%
18%
9%
39%
13073.9
5571.9
1254.8
2710.6
22611.1
64%
27%
6%
13%
110%
-128.4
156.1
121.6
4074.0
4223.2
-1%
1%
1%
35%
36%
7278.0
18%
3459.5
17%
10596.4
90%
7410.3
10156.6
19%
25%
15323.4
6673.9
75%
33%
-6069.8
-975.2
-52%
-8%
Table 6, continued
Panel B: Empirical Distribution of Changes in Long-run CEO Wealth
This panel presents the distribution of the CEO’s total wealth change between year AYR−1 to year CYR+1
as a fraction of his starting portfolio value in year AYR−1 for the two subsamples of acquiring CEOs. If a
CEO departs the firm, he is assigned zero compensation after the year of departure and we multiply the last
known total portfolio delta by the stock return over each succeeding year.
Acquiring CEOs with Positive Abn. Stock Return
10th Pct
+12%
25th Pct
+45%
50th Pct
+122%
75th Pct
+228%
90th Pct
+343%
Acquiring CEOs with Negative Abn. Stock Return
-19%
+4%
+52%
+128%
+258%
Panel C: The Long-run CEO Wealth-Performance Sensitivity
This panel presents the results of a regression explaining the acquiring firm CEO’s total wealth change
between year AYR−1 to year CYR+1 as a fraction of his starting portfolio value in year AYR−1. If a CEO
departs the firm, he is assigned zero compensation after the year of departure and we multiply the last
known total portfolio delta by the stock return over each succeeding year. Negative Abnormal Return
Dummy is set equal to one if the cumulative abnormal (control firm adjusted) stock return is negative, and
zero otherwise. Cumulative Abnormal Stock Return is the cumulative abnormal return between year
AYR−1 to year CYR+1. The interaction term Negative Dummy*Cumulative Abnormal Stock Return takes
the value of Cumulative Abnormal Stock Return when it is negative, and zero otherwise. Corresponding pvalues from Huber-White robust standard errors are reported in brackets.
Intercept
Total Wealth Change
0.823
[0.000]
Negative Abnormal Return Dummy
-0.056
[0.746]
Cumulative Abnormal Stock Return
1.055
[0.000]
Negative Dummy*Cumulative Abnormal Stock Return
-1.135
[0.000]
F-Test (the sum of the interacted and non-interacted coefficients is zero)
[0.215]
Adj R2
Number of Observations
0.182
317