Capital structure, equity mispricing, and stock repurchases

Alice Adams Bonaim

a
, zde ztekin
b,
, Richard S. Warr
c
a
Gatton College of Business and Economics, University of Kentucky, Lexington, KY 40506, United States
b
College of Business Administration, Florida International University, 11200 SW 8th St, Miami, FL 33199, United States
c
Poole College of Management, North Carolina State University, Box 7229, Raleigh, NC 27695, United States
a r t i c l e i n f o a b s t r a c t
Article history:
Received 4 January 2014
Received in revised form 22 March 2014
Accepted 25 March 2014
Available online 2 April 2014
We evaluate motives for share repurchases using a unified framework where a firm has a target
capital structure and has equity that can be mispriced. We document that capital structure
adjustments are a value-increasing motive for repurchases and that the extent to which adjusting
capital structure through a repurchase creates value depends on the undervaluation of the firm.
Underlevered and undervalued firms enjoy the greatest economic gains from a repurchase, as
evidenced by the stock price reaction to the repurchase announcement, and these firms are more
likely to announce a share repurchase program.
Published by Elsevier B.V.
JEL classification:
G30
G32
G35
Keywords:
Target leverage
Residual income model
Capital structure
Equity mispricing
Market timing
Share repurchase
1. Introduction
Firms repurchase stock for a variety of reasons, including signaling undervaluation, reducing the agency costs associated with excess
cash, fending off takeover attempts, and mimicking industry peers. Repurchasing stock also alters a firm's capital structure as buying
back shares decreases equity, which increases a firm's leverage ratio. Open market share repurchases, the most popular form of
repurchase, are generally interpreted as good news by the stock market and hence are greeted with abnormal returns of approximately
2% to 3% on average (e.g., Comment and Jarrell, 1991; Stephens and Weisbach, 1998). There is, however, a non-trivial variation in the
market reaction to share repurchases, implying that investors view repurchases as better news for some firms than for others.
This paper extends the previous research that investigates market reactions to share repurchases by studying the association
between stock returns to repurchase announcements and capital structure policy. Capital structure theories predict how and why
a firm could benefit from share repurchases. According to the tradeoff theory, underlevered firms can move towards their optimal
debt ratio by either issuing debt or repurchasing equity, which implies that the benefits froma share repurchase should be greater
for underlevered firms. Yet, underlevered firms must weigh the benefits of moving towards their target leverage ratio against the
cost of repurchasing stocka cost that depends on the perceived value of the stock. The market timing theory of capital structure
predicts that undervalued firms should repurchase equity to exploit mispricing opportunities while overvalued firms should
avoid repurchasing. If anything, these overvalued firms should issue stock.
Journal of Corporate Finance 26 (2014) 182200
Corresponding author.
E-mail addresses: alice.bonaime@uky.edu (A.A. Bonaim), ooztekin@u.edu (. ztekin), rswarr@ncsu.edu (R.S. Warr).
http://dx.doi.org/10.1016/j.jcorpn.2014.03.007
0929-1199/Published by Elsevier B.V.
Contents lists available at ScienceDirect
Journal of Corporate Finance
j our nal homepage: www. el sevi er . com/ l ocat e/ j cor pf i n
Previous research has attempted to differentiate among the explanations for repurchase announcements using various capital
structure theories, including the signaling/market timing and trade-off stories. In this paper, we evaluate the motive for a repurchase
in a more unified framework where a firm has a target capital structure and has equity that can be mispriced. We hypothesize that
firms that are below their optimal leverage at the time of the share repurchase announcement will benefit more from the capital
structure adjustment achieved by repurchasing stock. Further, the benefits to repurchasing will be more pronounced when a firm's
equity is also undervalued, and thus the overall cost of repurchasing is low. Correspondingly, when the firm's stock is overvalued and
repurchasing equity is relatively expensive, adjustments requiring stock repurchases will be more costly and hence less beneficial.
We study the effects of leverage and undervaluation on market reactions to open market share repurchase announcements
from 1990 to 2010. Compared to previous studies evaluating the effect of capital structure on share repurchase announcement
returns, we use more refined measures of leverage targets and equity mispricing. Our approach for estimating the target leverage
is based on Blundell and Bond (1998), who employ system generalized method of moment (GMM) estimators. Rather than
estimating a static model based on observed contemporaneous debt ratios, we estimate a dynamic panel model that produces an
estimate of the unobserved target leverage. The benefit of this partial adjustment model is that it incorporates rebalancing costs that
may slowdown the firm's rate of adjustment to its optimal leverage. We use two quite different methods to estimate mispricing: the
residual income model, as originally developed in the accounting literature (Ohlson, 1991, 1995), and the Rhodes-Kropf et al. (2005)
model, which decomposes the market-to-book ratio to separate mispricing effects from growth options.
We document significantly lower announcement returns for overlevered firms relative to underlevered firms, consistent with the
predictions of the trade-off theory. Using market debt ratios, repurchase announcements made by firms whose debt ratio is above their
target are associated with three-day cumulative abnormal returns (CARs) of 1.3%, statistically different fromthe 1.9% three-day CARs for
firms whose debt ratio is below their target. We find similar results when we use book leverage. We also find that undervalued firms
experience significantly greater announcement returns, consistent with the predictions of the market timing theory. When using the
Rhodes-Kropf et al. (2005) valuation method, the three-day announcement CARs are 1.4% for overvalued firms but 2.3% for undervalued
firms. This difference in CARs is statistically significant. We find similar results using the residual income model to measure valuation.
To examine the interaction of valuation and capital structure, we condition on firms being over or underlevered and over or
undervalued to form four groups (overlevered/overvalued, overlevered/undervalued, underlevered/overvalued, and underlevered/
undervalued). Shifting from the overlevered/overvalued group of firms to the underlevered/undervalued group causes returns to
repurchase announcements to approximately doubleregardless of how we define leverage or measure misvaluation. In sum, firms
that are above their target leverage ratio or overvalued have lower abnormal returns to repurchase announcements while firms that
are below their target leverage ratio or undervalued have higher abnormal returns to repurchase announcements.
We also study the effects of capital structure and equity mispricing on repurchase announcements in a multivariate setting,
where we control for firm characteristics and other plausible value-relevant motives for repurchasing stock. When we regress
three-day repurchase announcement CARs on indicators for over/underlevered and over/undervalued, we find evidence that the
market discounts announcements made by both overlevered and overvalued firms. However, the market only places a premium
on repurchase announcements made by undervalued firms. While the cumulative abnormal announcement returns are
significantly positive for underlevered firms, the difference is not statistically significant from our benchmark group (firms that
are close to their target with little misvaluation).
Interacting leverage and valuation reinforces our prior results: Even after controlling for other potential value-relevant
repurchase motives, we find that the market discounts repurchase announcements by firms that are both overlevered and
overvalued and places a premium on announcements by firms that are underlevered and undervalued. In the more ambiguous
cases, we document that the negative impact of being overvalued cancels out the positive impact of being underlevered. However,
market reactions are stronger (i.e., more positive) for overlevered firms that are undervalued, indicating that the benefits of
repurchasing undervalued stock outweigh the costs of being away from the target. In other words, being undervalued leads to
greater announcement returnsregardless of whether the firm is over or underlevered.
Given that the benefits to repurchasing depend on capital structure and perceived valuation, our final analysis tests whether
capital structure and misvaluation influence a firm's decision to announce a share repurchase. Specifically, we include standard
controls for alternative repurchase motives, and we model the decision to announce a share repurchase as a function of being
over/underlevered and over/undervalued. Consistent with firms understanding the economic benefits of repurchases motivated
by capital structure, we find evidence that underlevered firms are more likely to repurchase while overlevered firms are less
likely. The relationship between mispricing and the likelihood of announcing a repurchase is less clear. While undervalued firms
are unequivocally more likely to announce a repurchase, overvalued firms are not necessarily less likely.
As before, we alsointeract leverage and valuationto study howthese effects work together to influence the decisionto repurchase.
Conditional on being overvalued, the effect of being underlevered is ambiguous. However, firms that are both underlevered and
undervalued are significantly more likely to announce a share repurchase. Overlevered firms are less likely to announce a repurchase
if the firm is simultaneously overvalued, but more likely to announce a repurchase if the firm is undervalued. These results confirm
that firms recognize that the interaction between capital structure and firm valuation impacts the economic benefit to share
repurchases and incorporate this information into their decision to repurchase.
We improve uponthe existing researchby employing unique, precise measures of relative leverage andfirmvaluationwhenmodeling
a firm's choice to repurchase and the market's reaction to that choice. We contribute to repurchase and capital structure literatures by
showing that capital structure is a value-increasing motive for a share repurchase and that the extent to which adjusting capital structure
through a share repurchase creates value depends on the undervaluation of the firm. We also document that firms understand this
relationship and consider their relative leverage and undervaluation when deciding whether to announce a repurchase.
183 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
The remainder of the paper proceeds as follows. Section 2 develops our hypotheses and reviews the related literature. Section 3
describes our sample constructions. Section 4 presents our analyses. Section 5 concludes.
2. Literature review and hypothesis development
This paper contributes to the literatures on capital structure and share repurchases by reexamining the effects of capital structure
and equity mispricing on market reactions to share repurchase announcements under the assumptions that firms seek to adjust
leverage towards an optimal capital structure, conditional on the cost of adjustments as measured by equity mispricing.
2.1. Capital structure theories
The trade-off theory of capital structure states that firms have an optimal (target) capital structure. If the costs of adjustment were
zero, the firm would have no incentive to deviate from this target and adjustments would be made instantaneously. However,
because of market imperfections such as contracting costs, asymmetric information, and financing costs, firms may temporarily
deviate fromtheir target leverage. The distance of the actual leverage ratio fromthe optimal leverage ratio is the total amount that the
debt ratio must change to bring the firm back to its target debt ratio. When a firm has deviated from its optimum, it has several
options. If overlevered, it can retire debt or issue equity and if underlevered, it can repurchase shares or issue debt. To the extent that
these actions are costly, altering the leverage ratio becomes burdensome, implying slower adjustment to optimal leverage.
In this paper, we implicitly assume that firms seek to adjust leverage ratios towards a target. However, we acknowledge that
some firms could operate with a flexible target leverage ratioor with no target leverage ratio at all. While explaining how firms
determine their capital structure is beyond the scope of our paper, we note that including firms with flexible targets or no target
ratio in our sample would bias us against finding a meaningful relationship between leverage and repurchases.
The market timing theory of capital structure posits that when managers issue securities, they consider the time-varying relative
costs of issuances for debt and equity (Baker and Wurgler, 2002; Graham and Harvey, 2001; Hovakimian et al., 2001; Huang and
Ritter, 2009; Leary and Roberts, 2005; Myers, 1984). This market timing motivates the prediction that firms alter their leverage to
exploit good pricing opportunities. The market timing theory has, however, drawn criticism from Alti (2006), Flannery and Rangan
(2006), and Butler et al. (2009), among others, who question the longevity and overall economic significance of market timing.
In our study we viewmarket timing as altering the cost of adjusting to a target, and the presence of market timing behavior by firms
does not preclude the trade-off theory. Instead, we argue that the extent of equity mispricing influences a firm's incentives to adjust
toward its optimal capital structure. The trade-off theory argues that managers have discretion over their firms' leverage choices and that
they adopt capital structures with the value-maximizing level of debt. The selection of the optimal leverage target is based on a trade-off
between the relative costs and benefits of debt. Empirically, however, it is well documented that firms deviate fromtheir target leverage
ratios, and do not instantly adjust back to their target if there are significant costs associated with doing so. We recognize that the costs
and benefits of adjusting leverage can be influenced by the firm's position relative to its target and the degree of its equity mispricing.
Warr et al. (2012) document strong empirical evidence that the temporary deviationof a firm's share price fromits fundamental value
and the resulting impact on the cost of equity constitute a significant adjustment cost. They show that equity-mispricing costs have a
major impact on the rate at which firms adjust their capital structure. More specifically, overvalued firms with leverage ratios above their
target adjust back toward their target more rapidly than undervalued firms. The opposite effect is found for firms that are below their
targetovervalued firms adjust more slowly than the undervalued firms. They interpret this finding as evidence that managers exploit
equity mispricing to time the market. When the cost to issue equity is low(because stock is overvalued), managers exploit this mispricing
tothe benefit of existingshareholders andmore rapidly returntotheir leverage target. Likewise, whenthefirm's equityis undervalued, the
firm adjusts more slowly if adjustment calls for equity issuance as such an issuance would be value destroying to existing shareholders.
2.2. Capital structure and share repurchases
2.2.1. Extant literature on repurchase motives and announcement returns
Firms repurchase stock for a variety of reasons. Perhaps the most commonly cited motive for initiating a repurchase plan is
signaling undervaluation (e.g., Brav et al., 2005; Grullon and Michaely, 2004; Louis and White, 2007; Vermaelen, 1981). In addition,
firms with excess cash potentially suffer from agency problems and are more likely to repurchase stock (Jensen, 1986; Stephens and
Weisbach, 1998). Billett and Xue (2007) find a significantly positive relation between open market share repurchases and takeover
probability, and Massa et al. (2007) show that firms use repurchases to mimic their industry counterparts. Kahle (2002) identifies a
positive relationship between repurchases and outstanding options, implying that some repurchases are an attempt to undo the
effects of dilution associated with stock options.
1
Finally, Jagannathan et al. (2000) and Guay and Harford (2000) examine the choice
between dividends and repurchases and identify financial flexibility as a motive for choosing repurchases.
1
We recognize that repurchases that immediately and exactly offset stock option exercises will not result in a decrease in a rm's equity and therefore will not
alter its leverage ratio. However, the alternative is for the rm not to repurchase shares when options are exercised, which would result in an increase in equity.
Therefore, the repurchase still decreases the rm's equity relative to what it would be in the absence of the repurchase. In addition, share repurchases combatting
option-related dilution would bias us against nding that capital structure and mispricing are related to the probability of and market reaction to share
repurchases, unless being underlevered and undervalued are positively correlated with stock option exercise. Yet, if anything, large stock option exercises are
likely to be negatively correlated with undervaluation as option holders are likely to exercise when stock prices are elevated not depressed.
184 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
While earlier literature has been influential in supporting other motivations for share repurchases, credible associations between
capital structure and equity mispricingboth independently and jointlyand repurchases are not well established. We seek to fill this
gap in the literature. We also note that identifying a clear relationship between capital structure and equity mispricing and
repurchases does not invalidate previously documented repurchase motives as these motives are not mutually exclusive.
We hypothesize that capital structure and equity mispricing can influence the decision to repurchase stock. Repurchases
decrease a firm's equity and thus increase its leverage ratio. However, the benefit to adjusting capital structure must be weighed
against the costs of repurchasing stock. Using the total dollar value spent on repurchase to identify repurchasing firms, Dittmar
(2000) finds that repurchasing firms have lower distances from their target (defined as the industry median debt to assets ratio)
and lower market-to-book ratios than non-repurchasing firms and concludes that firms repurchase stock to alter their leverage
ratio and potentially to take advantage of undervaluation. Lie (2002) studies tender offers and shows that the probability of a
tender offer depends on a firm's distance from target leverage (defined using predicted values from annual OLS regressions) but
not on undervaluation (defined using four different measures based on EBITDA, assets, sales, and the residual income model).
However, tender offers, though generally large repurchase transactions that would significantly alter capital structure, are
somewhat rare. Ikenberry et al. (1995) and Grullon and Michaely (2004) note that open market share repurchasesthe focus of
our studyconstitute over 90% of the dollar value spent on repurchases.
We furthermore hypothesize that capital structure is a value-increasing motive for repurchasing and that the extent to which
adjusting capital structure through a repurchase creates value depends on the undervaluation of the firm. The literature is mixed
on the effects of leverage and valuation on repurchase announcement returns. For example, Stephens and Weisbach (1998)
identify a negative relation between three-day CARs around open market repurchase announcements and prior stock returns,
which they use as a proxy for undervaluation.
2
They do not examine leverage. Lie (2002) finds that neither the deviation from
target leverage nor undervaluation affects tender offer announcement returns. Chan et al. (2004) find that the initial market
reaction is lower for value stocks (proxied by book-to-market ratio) and insignificant for firms with low leverage, inconsistent
with both the valuation and leverage hypotheses. Defining leverage as total liabilities to total assets, i.e., not relative to a target,
and using market-to-book ratio as a proxy for undervaluation, Bonaime (2012) finds no relationship between market reactions to
share repurchase and capital structure or valuation. Bargeron et al. (2012) identify no relationship between leverage and
three-day CARs around repurchase announcements, but a negative relationship between market-to-book and announcement
CARs. In sum, the results on whether leverage and valuation impact the likelihood of a repurchase and the extent to which the
repurchase creates value are inconclusive.
Our study differs from previous studies in several important ways. First, with the exception of the Lie (2002) study of tender
offers, previous studies that examine the effect of capital structure on share repurchases have primarily focused on the actual
(observed) leverage as a proxy for target leverage and on market-to-book ratio as a proxy for equity valuation. We recognize that
observed leverage and market-to-book are poor proxies for target leverage and valuation and instead use more precise measures.
Second, we assume that firms target an optimal capital structure and that equity mispricing influences the incentives to adjust to
that capital structure. Thus, while prior studies investigate the impact of leverage and market-to-book on share repurchase
announcement returns in isolation, we recognize the importance of their interaction and evaluate their combined effect. Overall,
our paper reinforces that capital structure and equity mispricing are closely related to share repurchases and contribute to our
understanding of the motivation and economic consequences of repurchases.
2.2.2. Repurchase announcement return predictions
Table 1 graphically presents our predictions. In this table, firms are divided into four quadrants depending on whether they are
above or below their leverage targets and whether they are over or undervalued. If capital structure affects announcement
returns, then the announcement return in the top left quadrant (overlevered and overvalued) will be the lowest and the
announcement return in the bottom right quadrant (underlevered and undervalued) will be the highest. According to both
market timing and trade-off considerations, a share repurchase decision would be suboptimal for an overlevered/overvalued firm
while a share repurchase would be optimal for an underlevered/undervalued firm. Furthermore, the announcement returns in the
top right quadrant (overlevered and undervalued) and the bottom left quadrant (underlevered and overvalued) will lie between
these two extremes as in both cases the market timing and trade-off considerations would have opposite implications about the
optimality of a share repurchase decision. If relative leverage and valuation are significant sources of economic gains in share
repurchases, we expect firms to integrate these factors into their decision to announce a share repurchase plan.
3. Data and method
3.1. Sample construction
We begin with all open market share repurchase announcements available in the Securities Data Corporation (SDC) Repurchases
database between 1990 and 2010. To remain in our sample, a firm must have return data in CRSP and accounting data in Compustat.
2
We recognize that the existence of equity mispricing can imply market irrationality, but it can also result from asymmetric information between the rm and
the market. As such, we are agnostic about the source of this mispricing. In our study, however, we assume that the announcement of a repurchase does serve to
remove some of this information asymmetry, and as a result, the market's reaction to the announcement will reect a partial revelation of the true state of the
rm's valuation.
185 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
We exclude financials (SIC codes 60006999) and utilities (49004999) because of the regulated nature of their capital structure
decisions.
Table 2 presents annual summary statistics on open market repurchase announcements. The first four years of our sample
(19901993) have artificially low numbers of observations due to inconsistent reporting in the SDC Repurchases database.
3
More
reliable figures begin in 1994. We observe that the number of repurchase announcements increases during the late 1990s, and in
1998 alone there are 815 announcements. The frequency of repurchase announcements then decreases until 2003, increases until
2008, and sharply declines in 2009 during the financial crisis. Our summary statistics point to repurchases increasing during
expansionary business cycles and decreasing during recessionary periods, consistent with Dittmar and Dittmar (2008).
Next we present three-day cumulative abnormal returns (CARs) from day 1 to day +1 around repurchase announcements.
We calculate CARs with an event-study method as in Brown and Warner (1980). We estimate the market model over 255 trading
days ending 46 days prior to the announcement using the CRSP value-weighted index as a proxy for market returns. We require
at least 100 days of returns to estimate the model. Median announcement CARs are 5.7% in 1991, but this result is likely driven by
the small sample that year. Otherwise, annual median announcement CARs range from slightly below 1% to slightly above 3%. The
median three-day announcement CAR for our full sample period is 1.7%.
3.2. Measuring the distance from the target
The trade-off theory of capital structure maintains that firms select a value-maximizing leverage ratio by trading off the costs and
benefits of debt. Firms move towards this optimal leverage target gradually due to non-trivial adjustment costs.
4
The recent evidence
in the capital structure literature has convincingly demonstrated that actual leverage is a noisy approximationof target leverage in the
presence of non-trivial adjustment costs. In a frictionless world, firms would always maintain their target leverage. However, if
adjustment costs are significant, the firm's decision to adjust its capital structure depends on the trade-off between the adjustment
costs and the costs of operating with suboptimal leverage. Flannery and Rangan (2006) estimate a model that permits incomplete
adjustment toward optimal leverage eachtime period and showthat studies imposing unwarranted assumptions of homogenous and
instantaneous adjustment toward the target lead to incorrect or misleading inferences. Evaluating alternative leverage target
specifications, Flannery and Rangan (2006) and Lemmon et al. (2008) conclude that firm fixed effects are also required to capture
unobserved firm-level heterogeneity. Allowing for partial adjustment and incorporating firm fixed effects to capital structure
modeling has since become a standard practice. Accordingly, we use a partial adjustment model that incorporates the rebalancing
costs that may slow down the firm's adjustment to its optimal leverage:
LEV
j;t1
LEV
j;t
LEV

j;t1
LEV
j;t


j;t1
1
where
LEV
j,t
is the jth firm's observed leverage at the end of quarter t;
LEV*
j,t + 1
is the jth firm's target leverage for the end of quarter t; and
is the adjustment speed.
Previous researchers have estimated models that permit target leverage to vary across firms and over time:
LEV

j;t1
X
j;t
2
where X
j,t
is a vector of the jth firm's characteristics designed to capture the costs and benefits of debt. Eq. (2) thus provides a
model of the determinants of the optimal leverage, which relies only on observable variables. We use the following firm and
industry characteristics: earnings before interest and taxes as a proportion of total assets; book liabilities plus market value of
3
Results are unaffected by deleting these four years.
4
Fischer et al. (1989), Flannery and Rangan (2006), Gungoraydinoglu and ztekin (2011), Leary and Roberts (2005), Lemmon et al. (2008), Faulkender et al.
(2012), ztekin and Flannery (2012), ztekin (JFQA, forthcoming), and Warr et al. (2012).
Table 1
Predictions of the impact of equity mispricing and distance from the target leverage on repurchase announcement returns. This table presents the major
hypotheses tested. The column headings indicate whether the firm is overvalued or undervalued. The row headings indicate whether the firm is overlevered or
underlevered.
Equity overvalued
(Equity mispricing: issue equity)
Equity undervalued
(Equity mispricing: repurchase equity)
Firm overlevered
(Trade-off theory: issue equity and/or decrease debt)
Lowest announcement returns Medium announcement returns
Firm underlevered
(Trade-off theory: increase debt and/or repurchase equity)
Medium announcement returns Highest announcement returns
186 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
equity to total assets; depreciation expense as a proportion of total assets; log of total book assets (a measure of firm size); fixed
assets as a proportion of total assets; research and development expenses as a proportion of total assets, where missing values are
set equal to zero and the median leverage ratio for the firm's industry, based on the Fama and French (1997) industry categories.
To control for unobservable factors that could affect leverage, we include firm(Flannery and Rangan, 2006; Lemmon et al., 2008)
and year fixed effects. Because optimal leverage LEV
ij,t

is unobservable, substituting Eq. (2) into Eq. (1) yields the following:
LEV
j;t1
X
j;t

1 LEV
j;t

j;t1
3
Eq. (3) requires instruments for the endogenous transformed lagged-dependent variable and a correction for the short panel
bias (Blundell and Bond, 1998; Huang and Ritter, 2009). Blundell and Bond's (1998) system generalized method of moment
(GMM) estimation method provides adequate estimates in the presence of these estimation issues (Flannery and Hankins, 2013).
We therefore use a two-step systemGMMto estimate Eq. (3) using 1990 to 2010 data for our sample firms, and we control for the
potential endogeneity of the right-hand-side variables by using lags of the same variables as instruments.
The estimated coefficients fromEq. (3) indicate each firm's target ratio (Eq. (2)) and the distance fromits target debt ratio:
Distance
ij;t1
L
^
EV

ij;t1
LEV
ij;t
: 4
Throughout the empirical analysis, we use measures of both book leverage and market leverage. Book leverage is defined
as:
BLEV
Long Term Debt Short Term Debt
Total Assets
: 5
Market leverage is defined as:
MLEV
Long Term Debt Short Term Debt
Long Term Debt Short Term Debt Market Value of Equity
6
where Market Value of Equity is calculated as the stock price times common shares outstanding.
Table 2
Summary statistics. This table presents summary statistics on repurchase announcements reported in the SDC repurchase database between 1990 and 2010. N
represents the number of announcements in each year and period. Other statistics represent the median value each year. 3-day CAR represents 3-day cumulative
abnormal return (in %), calculated with an event-study method. We estimate the market model over 255 trading days ending 46 days prior to the announcement
using the CRSP value-weighted index as a proxy for market returns and require at least 100 days of returns. Dist BB_BL (ML) is the total amount that the book
(market) debt ratio must change to bring the firm back to its target debt ratio [TL
t + 1
DR
t
] using Blundell and Bond targets; VP_RIM is the value to price ratio
measured as the valuation calculated from the Residual Income Valuation Model divided by the stock price; and VP_RKRV is the value to price ratio measured as
the valuation calculated from the market-to-book decomposition divided by the stock price. The first four years of our sample (denoted by

) have particularly
low numbers of observations due to inconsistent reporting in the SDC Repurchases database.
Year N 3-day CAR (%) Dist BB_BL Dist BB_ML VP_RIM VP_RKRV
1990

10 2.156 0.035 0.209 0.647 0.562

1991

5 5.693 0.001 0.000 0.525 1.492

1992

17 1.810 0.004 0.016 0.693 0.721

1993

40 1.112 0.000 0.018 0.937 0.800

1994 364 2.309 0.005 0.036 0.629 0.948
1995 384 1.266 0.019 0.013 0.698 0.954
1996 461 1.839 0.020 0.011 0.587 0.880
1997 497 1.405 0.034 0.000 0.552 0.984
1998 815 2.085 0.075 0.150 0.526 0.939
1999 551 2.738 0.050 0.113 0.619 0.907
2000 548 3.079 0.016 0.128 0.517 0.988
2001 458 2.221 0.011 0.003 0.443 1.003
2002 380 2.399 0.001 0.074 0.609 0.943
2003 294 1.162 0.000 0.014 0.852 0.809
2004 380 0.952 0.004 0.012 0.745 0.876
2005 460 1.360 0.000 0.000 0.622 0.859
2006 477 0.932 0.009 0.000 0.520 0.902
2007 564 1.725 0.052 0.073 0.441 0.892
2008 576 2.113 0.080 0.332 0.572 0.966
2009 232 2.588 0.018 0.071 1.026 0.895
2010 367 0.888 0.003 0.033 0.850 0.938
19901999 3144 1.867 0.042 0.064 0.596 0.926
20002010 4736 1.623 0.012 0.025 0.571 0.912
All 7880 1.722 0.023 0.041 0.581 0.918
187 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
We present our target leverage models for both book and market leverage in Table A1 of the Appendix and the median
distances from the target leverage in Columns 4 and 5 of Table 2.
3.3. Measuring equity misvaluation
We estimate misvaluation by two methods: the residual income model (RIM) (Ohlson, 1991, 1995) and the Rhodes-Kropf et al.
(2005), (RKRV) model. While some studies have successfully used market-to-book as a proxy for valuation (notably Baker and
Wurgler, 2002), Warr et al. (2012) argue that market-to-book is a poor proxy for valuation because it captures other effects,
including growth options and debt overhang problems, and the actual performance of market-to-book as a valuation measure is
weak.
5
Furthermore, they argue that the relationship of market-to-book with other variables is not consistent over time. For
example, the premise that high market-to-book firms underperform low market-to-book firms (Frankel and Lee, 1998; La Porta,
1996) is time dependent (Kothari and Shanken, 1997) and has substantially lower predictive power in the second half of the 20th
century.
The residual income model was originally developed in the accounting literature (see Ohlson, 1991, 1995).
6
The residual
income model is estimated by adding the discounted expected earnings in excess of the expected return on book value (this is
similar to economic value added [EVA]) to the book value of equity. Eq. (7) is a formal representation of the model.
V
RIM
B
0

X
n
t1
E
t
r B
t1
1 r
t

E
t
r B
t1
E
t1
r B
t
2 r 1 r
n
7
V
RIM
is the value of the firm's equity at time zero, B
0
is the book value at time zero, r is the cost of equity, and E
t
is the expected
future earnings for year t at time zero. Time zero is the beginning of the fiscal year, and n equals two years. We use the perfect
foresight version of the residual income model that uses realized earnings.
7
B
0
is the book value of equity. E
t
is the income before
extraordinary items. The perfect foresight model does suffer the fact that it uses information that is unknown at the time of the
capital structure decision and implicitly assumes that managers possess an unbiased expectation of future earnings. The cost of
capital, r, is estimated using Fama and French's (1997) three factor model (with monthly returns) to calculate with the short-term
T-bill as a proxy for the risk-free rate of interest. The final term, the terminal value, is the average of the last two years of the finite
series and is restricted to be nonnegative, as a negative terminal value implies that the firm would continue to invest in negative
NPV projects in perpetuity. The estimated intrinsic value of the stock V
RIM
is compared to the market value of the stock to
determine the valuation error. Estimated mispricing is measured as:
VP
RIM

V
RIM
P
0
8
where VP
RIM
is the mispricing at time zero, P
0
is the market price of the stock at time zero, and V
RIM
is the intrinsic value of the
stock at time zero, which is the beginning of the firm's fiscal year. We present our residual income model estimation in Table A2 of
the Appendix.
The second approach used is that of Rhodes-Kropf et al. (2005) who use the model developed by Rhodes-Kropf and Viswanathan
(2004) to decompose the market-to-book ratio into a valuation and growth component, using the following identity:
M
B

M
V

V
B
9
where M is the market value of equity, B is the book equity and V is the fundamental value of equity. Taking logs of the terms in
Eq. (9) (represented by lower case letters) we get:
mb mv vb : 10
RKRV then decompose the component m
it
v
it
(where i and t represent the ith firm at year t) into a sector component and a
firm specific component. Thus, there is the potential for market-wide misvaluation and firm specific misvaluation. RKRV defines
the valuation measure, v as a function of the firm specific accounting information at t,
it
and a vector of accounting multiples, .
So v(
it
; ) is the predicted value as function of accounting multiples, , at time t. To account for industry or sector specific
valuation multiples, they add the subscript j.
5
Lee et al. (1999) show that the market-to-book ratio only predicts about 0.33% of the variation in real stock returns. Flannery and Rangan (2006) and Liu
(2009) nd little evidence of market timing when using market-to-book.
6
This valuation approach is widely use in the nance literature. See for example D'Mello and Shroff (2000), Lee et al. (1999), Frankel and Lee (1998), Penman
and Sougiannis (1998) and Elliott et al. (2007, 2008).
7
The perfect foresight approach is also used in D'Mello and Shroff (2000), Lee et al. (1999), Dong et al. (2006), and Elliott et al. (2007).
188 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
The resulting Eq. (11) presents the market to book ratio as a function of firm, sector and long-run components.
m
it
b
it
m
it
v
it
;
jt

v
it
;
jt

v
it
;
j

v
it
;
j

b
it
11
The first part, m
it
v(
it
;
jt
) represents firm valuation where the valuation is conditioned on the sector j multiples at
time t. This component will also include market wide valuation effects as these will affect the sector as well. The second component
v(
it
;
jt
) v(
it
;
j
) is the difference between the full time series valuation and the sector specific valuation (that is not time
dependent). This component captures valuation effects that are specific to the sector. The final component v(
it
;
j
) b
it
captures the
long run difference between the sector and the firm's book value.
To implement the RKRV valuation approach, we use their model that includes book value, net income and leverage as the
it
vector (this approach is the same as in Hertzel and Li (2010), and we follow those authors closely). Market value is estimated as a
simple linear function of the accounting variables (b, NI and LEV
RKRV
) with their accounting multiples (
1jt
,
2jt
,
3jt
,
4jt
). For the
purpose of this paper, we are estimating the first component of Eq. (11), m
it
v(
it
;
jt
), the firm specific misvaluation. This
component is the residual of the following equation:
m
it

0jt

1jt
b
it

2jt
ln NI

it

3jt
I
b0
ln NI

it

4jt
LEV
RKRV it

it
: 12
Net income (NI)
+
is the absolute value of net income, and is also interacted with a negative net income dummy variable, I
b0
.
Leverage (LEV
RKRV
) is computed as:
LEV
RKRV
1
Market value
Market value AssetsDeferred taxesEquity
: 13
We estimate Eq. (12) annually for each of the 12 Fama French Industries (j = 1, 12) for the years 1970 to 2011 and present
these results in Table A3 of the Appendix. Firm value, V
RKRV
is the predicted value from Eq. (12). The final valuation measure is
computed as:
VP
RKRV

V
RKRV
P
0
: 14
We present the average valuation ratios for both the RIM and RKRV methods in Table 2, columns 6 and 7. It is worth noting
that while the mean of the RKRV method is close to 1 (because it is based on the residual of the accounting multiples regression),
the mean of the RIM model is significantly less than 1. This is not an error in the model, but merely the result of using a rolling
historic risk premium. We could have used a risk premium that on average results in a valuation ratio of 1, but there would be no
benefit in doing so. What is important is that the valuation ratio tells us the relative misvaluation of one stock compared to
another. The median valuation ratio could provide a rough estimate of fair value, assuming that on average, the market was fairly
valued over the sample period. It is important to note that we remain agnostic as to the cause of the mispricing. For example, the
mispricing captured by our models could be due to asymmetric information between managers and shareholders or irrationality
on the part of shareholders. We recognize that the mispricing may be only perceived mispricing. In other words, the firm is
trading at its fundamental value, but management and/or shareholders believe that the firm is mispriced. In our study, the source
of mispricing is unimportant as long as managers are aware of the (true or perceived) mispricing and use it to the firm's
advantage when making capital structure adjustments.
The RIM method imposes several quite restrictive assumptions. In particular the model (as implemented) assumes that
managers possess perfect foresight, that growth options can be capitalized in three periods, and that the firm's forward-looking
cost of capital is based on the historical cost of capital. We could increase the complexity of the RIM model, but complexity comes
at a costparticularly as it excludes firms with insufficient data. Instead we opt to also present the RKRV model as an entirely
different method of estimating firm value. In addition to not suffering the limitations above, the RKRV model also includes market
evidence as to the value of growth and discount rates. Furthermore, the RKRV approach allows for greater cross-industry
variations in discount rates. Ultimately, however, the performance of these valuation measures is an empirical question.
4. Analysis and results
4.1. Share repurchase announcement returns
4.1.1. Univariate results
We explore how leverage and undervaluation relate to market reactions to open market share repurchase announcements.
Given that repurchasing stock increases a firm's leverage ratio, we expect stronger market reactions, i.e., greater positive
abnormal returns, around repurchase announcements made by firms whose leverage ratio is below their target. We also expect
announcement returns to be related to perceived valuation. Specifically, if the firm appears undervalued (overvalued), then we
anticipate that investors will react more (less) favorably to a repurchase.
189 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
Table 3 presents median three-day CARs around open market repurchase announcements, segmented on our two measures of
relative debt ratio and our two measures of undervaluation.
8
Using either market or book debt ratios, we observe significantly
lower announcement returns for overlevered firms relative to underlevered firms, consistent with the predictions of the trade-off
theory. Using market debt ratios, repurchase announcements made by firms whose debt ratio is above their target are associated
with three-day CARs of 1.3% relative to 1.9% for firms whose debt ratio is below their target. Results are similar using book debt
ratios: Three-day CARs are 1.4% for overlevered firms, statistically different from1.8% for underlevered firms. In addition, a greater
number of underlevered than overlevered firms are buying back stock: Approximately two-thirds of repurchasing firms are
underlevered at the time of the announcement. Importantly, for a significant portion of underlevered firms (60% and 50%, using
book and market leverage, respectively), the size of the repurchase program is large enough to fully close the deviation from their
optimal leverage. These findings are consistent with capital structure rebalancing being an important and value-relevant motive
for repurchasing.
Undervaluation also affects repurchase announcement returns. Using either the residual income model or the Rhodes-Kropf et al.
method, we find that undervalued firms experience significantly greater announcement returns, consistent with the predictions of
the market timing theory. Three-day announcement CARs are 1.5% for overvalued firms but 1.7% for undervalued firms when using
the residual income model. The return differential is greater when conditioning on the Rhodes-Kropf et al. method: Overvalued firms
experience three-day CARs of 1.4%, relative to 2.3% for undervalued firms. All in all, market reactions to repurchase announcements
are dampened if firms are overlevered or overvalued. We also note that the numbers of undervalued and overvalued firms that are
buying back stock are similar by construction as we use the sample median as the cutoff.
Table 4 presents univariate pairwise comparisons of median three-day CARs around open market repurchase announcements.
Using both leverage definitions and both undervaluation measures, we categorize firms into one of four groups: (1) overlevered/
overvalued, (2) overlevered/undervalued, (3) underlevered/overvalued, and (4) underlevered/undervalued. The number of
observations in each group provides insight into the potential correlation between leverage and valuation. In Panel A, we observe
that 2302 firm-year observations out of 5322, or 43.25% of the sample are undervalued. Similarly, 45.17% (1621 firm-year
observations out of 3589) of underlevered firm-year observations are undervalued. Further, 3589 or 67.44% of all firm-year
observations are underlevered, and 70.41% (1621 out of 2302) of undervalued firm-year observations are underlevered. Given
that the probabilities of being undervalued conditional on being underlevered and vice versa are similar to the unconditional
probabilities, we are confident that our leverage and valuation measures are capturing distinct firm characteristics. Consistent
with this observation, the correlations between distance and valuation measures are weak, ranging from 10% to 16%, depending
on the methodology.
Consistent with our prior findings, overvalued firms that are also above their target debt ratio experience the lowest CARs around
open market share repurchase announcements, while undervaluedfirms belowtheir target debt ratio experience the greatest CARs of
any category. Differences in CARs fromthe underlevered/undervalued group and the overlevered/overvalued range from0.6% to 1.4%
depending on the leverage and valuations measures. All differences in CARs are statistically significant at the 1% level.
When we examine conditional differences in CARs, we observe that, within overlevered firms, repurchase announcement CARs are
between 7 and 89 basis points greater for overvalued than undervalued firms. This return differential is similar within underlevered
firms: Repurchase announcement CARs are between 25 and 81 basis points greater for overvalued than undervalued firms.
Table 3
Univariate comparisons of median 3-day announcement CARs. This table presents median 3-day CARs to repurchase announcements reported in the SDC
repurchase database between 1990 and 2010, segmented on measures of relative leverage and valuation. Panel A and Panel B report results from the Blundell and
Bond targets using the market debt ratio and book debt ratio, respectively. Panel C and Panel D report results from the residual income model (VP_RIM) and
Rhodes-Kropf et al. method (VP_RKRV), respectively. Significance at the 1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk, respectively.
N Median 3-day CAR Difference (P-value)
Panel A: leverage: market debt ratios
Overlevered (distance ML b 0) 2483 1.337 0.575
Underlevered (distance ML N 0) 4593 1.912 b0.001***
Panel B: leverage: book debt ratios
Overlevered (distance ML b 0) 2642 1.402 0.415
Underlevered (distance ML N 0) 4552 1.817 0.002***
Panel C: valuation: RIM
Overvalued (VP_RIM b VP_RIM
Median
) 2892 1.540 0.188
Undervalued (VP_RIM N VP_RIM
Median
) 2891 1.728 0.100*
Panel D: valuation: RKRV
Overvalued (VP_RKRV b VP_RKRV
Median
) 3616 1.373 0.878
Undervalued (VP_RKRV N VP_RKRV
Median
) 3606 2.251 b0.001***
8
We focus on medians to mitigate the effect of outliers and to conserve the sample size, but results are similar when we focus on means and/or tercile
groupings. Furthermore, focusing on ve-day CARs around open market repurchase announcements to take into account any possible information leakage yields
same conclusions.
190 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
When we instead condition on valuation and study return differentials based on relative leverage ratios, we find that, within
overvalued firms, CARs surrounding repurchases announced by underlevered firms are between 30 and 55 basis points greater
than those announced by overlevered firms. Finally, conditioning on a firm being undervalued, announcement CARs are between
16 and 74 basis points greater for underlevered than overlevered firms. These results suggest that both componentsrelative
leverage ratio and firm valuationcontribute to the market's reaction to open market share repurchase announcements.
9
4.1.2. Multivariate results
To further explore how the interaction between capital structure and firm valuation influences open market repurchase
announcement returns, we move to a regression setting. Table 5 reports regressions of three-day CARs around repurchase
announcements on indicator variables that proxy for relative leverage ratio and valuations. The models presented are fully
specified: 1) leverage targets are a function of observed firm and industry characteristics as well as unobserved firm and year
fixed effects, 2) equity mispricing is estimated either via the residual income (the RIM approach) or as a function of firm-specific
accounting variables and accounting multiples, (the RKRV approach), and 3) the announcement returns are market adjusted.
Consequently, we do not expect significant differences in our results when we introduce additional measures to control for
potential motives of share announcement returns other than the capital structure policy. Nevertheless, we show all models with
and without the following additional variables controlling for commonly cited repurchase motives and determinants of open
market repurchase announcement returns.
Repurchase plan size is the announced size of the repurchase plan, expressed as a percentage of shares outstanding. Log size is the
natural log of market capitalization(number of shares outstanding times price per share). Following Acharya et al. (2007), cash flowis
operating income before depreciation less the sum of depreciation and amortization, total income taxes, total interest and related
expense, common dividends, and preferred dividends divided by total assets. Non-operating cash flow is non-operating cash flow,
9
The results in Tables 3 and 4 are robust to using the alternative trade-off model presented in Ogden and Wu (2013). We present these additional results in
Tables A4 and A5 of the Appendix.
Table 4
Univariate pairwise comparisons of median 3-day announcement CAR. This table presents median 3-day CARs to repurchase announcements, segmented on the
interaction of measures of relative leverage and valuation. Panel A and Panel B report results from the residual income model (VP_RIM) using the Blundell and
Bond targets with the market debt ratio and book debt ratio, respectively. Panel C and Panel D report results from the Rhodes-Kropf et al. method (VP_RKRV)
using the Blundell and Bond targets with the market debt ratio and book debt ratio, respectively. We present differences in medians across rows (undervalued
overvalued) and columns (underleveredoverlevered), as well as along the diagonal (undervalued/underleveredovervalued/overlevered). Significance at the
1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk, respectively.
Overvalued
(VP b VP
Median
)
Undervalued
(VP N VP
Median
)
Difference
(P-value)
Panel A: Market debt ratios and VP_RIM
Overlevered (distance ML b 0) 1.200 1.317 0.118
N = 1052 N = 681 (0.336)
Underlevered (distance ML N 0) 1.702 2.058 0.357
N = 1968 N = 1621 (0.042**)
Difference (P-value) 0.502 0.741 0.859
(b0.023**) (b0.001***) (0.001***)
Panel B: book debt ratios and VP_RIM
Overlevered (distance BL b 0) 1.306 1.376 0.070
N = 839 N = 1076 (0.347)
Underlevered (distance BL N 0) 1.651 1.898 0.249
N = 1864 N = 1632 (0.068*)
Difference (P-value) 0.345 0.522 0.592
(0.032**) (0.006***) (0.005***)
Panel C: market debt ratios and VP_RKRV
Overlevered (distance ML b 0) 1.049 1.942 0.893
N = 1340 N = 921 (b0.001***)
Underlevered (distance ML N 0) 1.597 2.409 0.812
N = 2050 N = 2268 (b0.001***)
Difference (P-value) 0.549 0.467 1.360
(b0.001***) (0.048**) (b0.001***)
Panel D: book debt ratios and VP_RKRV
Overlevered (distance BL b 0) 1.228 2.096 0.868
N = 1505 N = 910 (b0.001***)
Underlevered (distance BL N 0) 1.529 2.258 0.730
N = 1917 N = 2358 (b0.001***)
Difference (P-value) 0.301 0.162 1.030
(0.090*) (0.138) (b0.001***)
191 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
scaled by total assets. Standard deviation of cash flowis calculated over a 5-year period ending prior fiscal year. We require a minimum
of 3 years of available cash flow data. Cash is cash and short-terminvestments, divided by assets. Property, plant and equipment is the
property, plant, and equipment, divided by assets. Sales growth is the percentage change in sales over the past year. Standard deviation
of returns is the standard deviation of daily stock returns from 255 to 46 days before. We require 100 days of available returns.
Illiquidity is defined as the ratio of the daily absolute return to the dollar trading volume on that day. This ratio gives the absolute
percentage price change per dollar of daily trading volume, or the daily price impact of the order flow (Amihud (2002)). We average
daily illiquidity for each firmover the period starting 255 trading days prior to the repurchase announcement and ending 46 trading
days prior to the repurchase announcement. All control variables are demeaned to allowfor a more straightforward interpretation of
the intercepts and are measured at the end of the prior fiscal year to alleviate endogeneity concerns. Standard errors are clustered at
the firm level and are heteroskedasticity-robust.
10
10
Our inferences are unchanged when standard errors are clustered along two dimensions, both at the rm and year level, allowing observations which share a
rm or share a year to be correlated (Peterson, 2009). We choose to report one-way clustered standard errors as it is computationally less intensive.
Table 5
Multivariate comparisons of 3-day announcement CARs. This table presents regressions of 3-day CARs to repurchase announcements on measures of relative
leverage and valuation. Columns 1 and 2 (3 and 4) report results from the residual income model (VP_RIM) using the Blundell and Bond targets with the market
(book) debt ratio. Columns 5 and 6 (7 and 8) report results from the Rhodes-Kropf et al. method (VP_RKRV) using the Blundell and Bond targets with the market
(book) debt ratio. Underlevered (overlevered) and undervalued (overvalued) are binary variables taking a value of 1 for firms that are in the highest (lowest)
distance and valuation terciles. Repurchase plan size is the announced size of the repurchase plan, expressed as a percentage of shares outstanding. Log size is the
natural log of market capitalization (number of shares outstanding times price per share). Cash flow is operating income before depreciation less the sum of
depreciation and amortization, total income taxes, total interest and related expense, common dividends, and preferred dividends divided by total assets.
Non-operating cash flow is non-operating cash flow, scaled by total assets. The standard deviation of cash flow is calculated over a 5-year period ending prior
fiscal year. We require a minimum of 3 years of available cash flow data. Cash is cash and short-term investments, divided by assets. Property, plant and
equipment is the property, plant, and equipment, divided by assets. Sales growth is the percentage change in sales over the past year. Standard deviation of
returns is the standard deviation of daily stock returns from 255 to 46 days before. We require 100 days of available returns. Illiquidity is defined as the ratio of
the daily absolute return to the dollar trading volume on that day. This ratio gives the absolute percentage price change per dollar of daily trading volume, or the
daily price impact of the order flow Amihud (2002). We average daily illiquidity for each firm over the period starting 255 trading days prior to the repurchase
announcement and ending 46 trading days prior to the repurchase announcement. All variables are demeaned and all control variables are measured at the end of
the prior fiscal year. P-values are reported beneath the coefficient estimates in parentheses. Standard errors are clustered at the firm level and are
heteroskedasticity-robust. Significance at the 1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk, respectively.
Valuation model VP_RIM VP_RKRV
Leverage measure Market Book Market Book
(1) (2) (3) (4) (5) (6) (7) (8)
Constant 2.770*** 2.879*** 2.659*** 2.520*** 2.169*** 2.534*** 1.967*** 2.184***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Underlevered 0.315 0.121 0.338 0.358 0.217 0.156 0.332 0.317
(0.238) (0.664) (0.173) (0.173) (0.414) (0.575) (0.180) (0.228)
Overlevered 0.492** 0.565** 0.279 0.068 0.326 0.467* 0.147 0.040
(0.037) (0.021) (0.248) (0.790) (0.164) (0.055) (0.548) (0.878)
Undervalued 0.552** 0.489** 0.539** 0.479* 1.858*** 0.837*** 1.874*** 0.840***
(0.023) (0.049) (0.028) (0.054) (0.000) (0.003) (0.000) (0.003)
Overvalued 0.766*** 0.435 0.725*** 0.403 0.165 0.263 0.195 0.260
(0.003) (0.113) (0.004) (0.138) (0.466) (0.267) (0.388) (0.272)
Repurchase plan size 0.006 0.006 0.006 0.006
(0.263) (0.269) (0.280) (0.284)
Log size 0.411*** 0.413*** 0.395*** 0.394***
(0.000) (0.000) (0.000) (0.000)
Cash flow 3.972** 3.911** 3.973** 3.907**
(0.029) (0.031) (0.028) (0.030)
Non-operating cash flow 4.613 4.612 4.165 4.206
(0.304) (0.309) (0.366) (0.364)
Standard deviation of cash flow 0.453 0.492 0.474 0.505
(0.699) (0.674) (0.685) (0.664)
Cash 0.819 0.889 1.295* 1.297*
(0.264) (0.233) (0.086) (0.089)
Property, plant and equipment 0.481* 0.501* 0.462* 0.477*
(0.075) (0.063) (0.086) (0.077)
Sales growth 0.000*** 0.000*** 0.000*** 0.000**
(0.003) (0.005) (0.009) (0.011)
Standard deviation of returns 54.942*** 55.346*** 53.743*** 54.019***
(0.000) (0.000) (0.000) (0.000)
Illiquidity 8696.237 7870.375 8758.307 8103.855
(0.311) (0.357) (0.309) (0.344)
Observations 7880 6606 7880 6606 7880 6606 7880 6606
192 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
Columns 1 and 2 (3 and 4) of Table 5 report results from the residual income model using the Blundell and Bond targets with
the market (book) debt ratio. Columns 5 and 6 (7 and 8) report results from the Rhodes-Kropf et al. method using the Blundell
and Bond targets with the market (book) debt ratio. Underlevered (Overlevered) is a binary variable taking a value of 1 for firms
that are in the highest (lowest) distance tercile. Undervalued (Overvalued) is a binary variable taking value of 1 for firms that are in
the highest (lowest) valuation tercile.
11
Therefore, our indicator variables must be interpreted relative to firms in the middle
tercile with respect to distance from target leverage and valuation (firms close to their target with little misvaluation).
Consequently, the announcement returns for each group of firms would on average be equal to the sum of the coefficient
estimates on the constant term and the relevant indicator variable. This approach allows us to establish the relative importance of
leverage and valuation considerations in contributing to the observed announcement returns.
Consistent with our prior results, undervaluation appears to be a strong determinant of repurchase announcement CARs: All
coefficients on our undervaluation indicator variable are positive and statistically significant. The economic magnitude of these
variables is also impressive: Undervalued firms experience CARs that are between 0.5% and 1.9% greater than firms in the middle
relative leverage and valuation terciles. Our overvaluation indicator variable is negative in the majority of cases, and statistically
significant in two specifications, indicating that the market reaction to repurchases announced by overvalued firms is significantly
reduced. We also find that the most overlevered firms generally experience lower CARs around repurchase announcements,
though the coefficient on our overlevered indicator variable only achieves statistical significance in our specifications using
market leverage. While the announcement returns are significantly positive for underlevered firms, the difference is not
statistically significant from our benchmark group (firms close to their target with little misvaluation).
Coefficients on control variables generally take on the expected sign. Firms that are larger, have greater cash flows, have more
property, plant, and equipment and have greater sales growth have lower announcement returns while firms with more volatile
stock prices tend to have higher announcement returns. Repurchase announcement CARs are statistically unrelated to the
announced plan size, non-operating cash flow, the standard deviation of cash flow, cash and liquidity. These results are generally
consistent with prior studies of short-run returns around repurchase announcement such as Stephens and Weisbach (1998),
Bonaime (2012) and Bargeron et al. (2012).
12
Next, we extend our pairwise comparisons to a multivariate setting in Table 6. Similar to Table 5, Columns 1 and 2 (3 and 4)
report results from the residual income model using market (book) debt ratio while Columns 5 and 6 (7 and 8) report results
from the Rhodes-Kropf et al. method using market (book) debt ratio. We interact our indicator variables, which implies that our
comparison group (reflected in our constant term) is the union of firms in the middle leverage and valuation terciles.
Consistent with our expectations, overlevered/overvalued firms generally experience lower CARs than our comparison group
and underlevered/undervalued firms generally experience higher CARs. The coefficients on the interaction of overlevered and
overvalued indicators are negative in seven of eight cases and statistically significant in five cases. All coefficients on the
interaction of underlevered and undervalued indicators are positive, and seven are statistically significant. In terms of magnitude,
the significant positive impact on repurchase CARs of being underlevered/undervalued appears to be greater than the negative
effect of being overlevered/overvalued.
The remaining interaction terms help to determine if one effectthat of distance from target leverage or that of valuation
dominates. We expect underlevered (undervalued) firms to have higher returns, but overvalued (overlevered) firms to have
lower returns. Therefore, the impact of the underlevered/overvalued (overlevered/undervalued) interaction term is theoretically
ambiguous. Empirically, we find six coefficients on the overlevered/undervalued interaction are positive and two are statistically
significant, which provides weak evidence that the valuation effect dominates. However, if valuation is the dominant effect then
we would expect the coefficient on the interaction terms of underlevered and overvalued to be negative. Coefficients are negative
for five of the eight coefficients, but none are statistically significant. Therefore, the effect of being simultaneously underlevered
and overvalued is unclear and neither effect dominates.
While the addition of the control variables for potential motives behind share announcement returns weakens the economic
and statistical significance of some of our leverage and valuation proxies (underlevered/overvalued and overlevered/
undervalued), our main conclusions continue to hold. The cross-sectional pattern for the announcement returns is in line with
our hypotheses outlined in Table 1, even after controlling for other potential motives of share repurchases. We observe the
greatest announcement returns for underlevered/undervalued firms and the smallest announcement returns for overlevered/
overvalued firms. The differences across these two groups are always economically and statistically significant regardless of how
we define leverage and measure misvaluation. The announcement returns for overlevered/undervalued and underlevered/
overvalued lie between these two extremes. For example, in column 2 where we control for potential drivers of share
repurchases, using market debt ratios and the residual income model, announcement returns are smallest for the overlevered/
overvalued firms (1.8%), followed by underlevered/overvalued firms (2.4%) and overlevered/undervalued firms (2.7%), and then
followed by underlevered/undervalued firms (3.3%).
11
Recall that the underlevered rms have a Distance N0, while overlevered rms have a Distance b0. Also, undervalued rms have high V/P ratios, while
overvalued rms tend to have low V/P ratios, relative to the median V/P ratio.
12
Results are robust to alternative denitions of free cash ow and including market-to-book, dened as market capitalization divided by book value of equity.
RIM model could yield downward biased estimates of the true valuation for growth rms (with low or negative earnings in the three years of the valuation),
making them appear overvalued. The advantage of including market-to-book is to distinguish the effect of growth opportunities on rm valuation. However,
market-to-book also proxies for rm value and therefore may cause multicollinearity issues. In general, the introduction of control variables does not inuence
the economic and statistical relevance of our leverage and valuation proxies.
193 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
4.2. Decision to announce a share repurchase plan
Given that relative leverage and valuation contribute to the economic impact of share repurchases, firms may consider their
current position relative to target leverage and their misvaluationindividually or concurrentlywhen deciding to announce a
share repurchase in the first place. In Table 7 we use a logistic model to examine the decision to announce a share repurchase as a
function of being overlevered/underlevered and overvalued/undervalued. As opposed to our prior tables, which focus only on
firms that announce a share repurchase during our sample period, we include the full population of industrial firms (excluding
financials and utilities) for which we can calculate our leverage and valuation measures and our control variables. Our dependent
variable is an indicator variable equal to one if the firm announced a share repurchase during the year. Columns 1 and 2 report
results from the residual income model using market and book debt ratios, respectively, while Columns 3 and 4 report results
from the Rhodes-Kropf et al. method using market and book debt ratios, respectively.
We find that underlevered firms are more likely to announce a share repurchase, though the coefficients only achieve
statistical significance when using market leverage. On the other hand, overlevered firms are less likely to announce a share
repurchase, though in this case the coefficients only achieve statistical significance when using book leverage. The effect of being
Table 6
Multivariate pairwise comparisons of 3-day announcement returns (3-AR). This table presents regressions of 3-day CARs to repurchase announcements on
interacted measures of relative leverage and valuation. Columns 1 and 2 (3 and 4) report results from the residual income model (VP_RIM) using the Blundell and
Bond targets with the market (book) debt ratio. Columns 5 and 6 (7 and 8) report results from the Rhodes-Kropf et al. method (VP_RKRV) using the Blundell and
Bond targets with the market (book) debt ratio. Underlevered (overlevered) and undervalued (overvalued) are binary variables taking a value of 1 for firms that
are in the highest (lowest) distance and valuation terciles. Repurchase plan size is the announced size of the repurchase plan, expressed as a percentage of shares
outstanding. Log size is the natural log of market capitalization (number of shares outstanding times price per share). Cash flow is operating income before
depreciation less than the sum of depreciation and amortization, total income taxes, total interest and related expense, common dividends, and preferred
dividends divided by total assets. Non-operating cash flow is non-operating cash flow, scaled by total assets. The standard deviation of cash flow is calculated over
a 5-year period ending prior fiscal year. We require a minimum of 3 years of available cash flow data. Cash is cash and short-term investments, divided by assets.
Property, plant and equipment is the property, plant, and equipment, divided by assets. Sales growth is the percentage change in sales over the past year.
Standard deviation of returns is the standard deviation of daily stock returns from 255 to 46 days before. We require 100 days of available returns. Illiquidity is
defined as the ratio of the daily absolute return to the dollar trading volume on that day. This ratio gives the absolute percentage price change per dollar of daily
trading volume, or the daily price impact of the order flow Amihud (2002). We average daily illiquidity for each firm over the period starting 255 trading days
prior to the repurchase announcement and ending 46 trading days prior to the repurchase announcement. All variables are demeaned and all control variables are
measured at the end of the prior fiscal year. P-values are reported beneath the coefficient estimates in parentheses. Standard errors are clustered at the firm level
and are heteroskedasticity-robust. Significance at the 1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk, respectively.
Valuation model VP_RIM VP_RKRV
Leverage measure Market Book Market Book
(1) (2) (3) (4) (5) (6) (7) (8)
Constant 2.670*** 2.684*** 2.657*** 2.617*** 2.350*** 2.528*** 2.420*** 2.459***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Overlevered & overvalued 1.211*** 0.877** 1.102*** 0.728* 0.613** 0.183 0.426 0.121
(0.002) (0.038) (0.006) (0.079) (0.024) (0.517) (0.127) (0.678)
Underlevered & overvalued 0.379 0.292 0.271 0.154 0.011 0.109 0.347 0.227
(0.320) (0.456) (0.466) (0.691) (0.977) (0.774) (0.306) (0.520)
Overlevered & undervalued 0.137 0.008 0.100 0.014 1.010*** 0.118 1.381*** 0.318
(0.650) (0.979) (0.768) (0.967) (0.004) (0.751) (0.003) (0.508)
Underlevered & undervalued 1.014** 0.621 0.902** 0.741* 2.366*** 1.264*** 1.895*** 1.052***
(0.018) (0.163) (0.020) (0.065) (0.000) (0.003) (0.000) (0.006)
Repurchase plan size 0.006 0.006 0.006 0.006
(0.276) (0.274) (0.266) (0.269)
Log size 0.427*** 0.433*** 0.410*** 0.418***
(0.000) (0.000) (0.000) (0.000)
Cash flow 3.839** 3.918** 4.007** 3.920**
(0.033) (0.030) (0.027) (0.030)
Non-operating cash flow 4.701 4.531 4.069 4.385
(0.297) (0.314) (0.368) (0.338)
Standard deviation of cash flow 0.506 0.456 0.472 0.450
(0.664) (0.695) (0.683) (0.698)
Cash 0.882 1.047 1.114 1.062
(0.229) (0.152) (0.135) (0.151)
Property, plant and equipment 0.484* 0.502* 0.472* 0.472*
(0.073) (0.064) (0.079) (0.080)
Sales growth 0.000*** 0.000*** 0.000*** 0.000***
(0.003) (0.003) (0.004) (0.004)
Standard deviation of returns 54.093*** 54.062*** 53.040*** 53.776***
(0.000) (0.000) (0.000) (0.000)
Illiquidity 8961.466 8453.674 9423.685 8588.412
(0.297) (0.323) (0.279) (0.311)
Observations 7880 6606 7880 6606 7880 6606 7880 6606
194 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
overvalued is unclear: Overvalued firms are less likely to repurchase shares according to the Rhodes-Kropf et al. method, but more
likely to repurchase shares when valuation is measured with the residual income model. However, being undervalued clearly
positively affects the likelihood of announcing a share repurchase.
To gage the economic magnitude of these effects, we calculate the odds ratios that capture the effect of a change from zero to
one in our indicator variables on the implied likelihood of announcing a repurchase program. The odds of undertaking share
repurchases are 19 percentage points higher for underlevered firms and 812 percentage points lower for overlevered firms.
Undervaluation increases the likelihood of repurchase by 2531 percentage points and overvaluation decreases it by 22
percentage points using Rhodes-Kropf et al. method. Using the residual income model, both undervaluation and overvaluation
seem to increase the probability to repurchase shares, however the economic impact is much stronger with undervaluation (31
vs. 11 percentage points).
With the exception of volatility of cash flow and sales growth, our control variables are statistically significant and carry the
expected signs. While larger firms, firms with high cash and high cash flow are more likely to announce share repurchases; firms
with more property, plant, and equipment, less stock market liquidity, and more variable stock prices are less likely to announce
repurchase plans.
Next, we interact our leverage and valuation indicators in Table 8, which implies that our comparison group is the union of
firms in the middle leverage and valuation terciles. Similar to Table 7, Columns 1 and 2 report results from the residual income
model using market and book debt ratios, respectively, while Columns 3 and 4 report results from the Rhodes-Kropf et al. method
using market and book debt ratios, respectively.
Table 7
Determinants of the share repurchase decision: leverage and misvaluation effects. This table presents logistic regressions modeling the decision to announce a
share repurchase. Column 1 (2) reports results from the residual income model (VP_RIM) using the Blundell and Bond targets with the market (book) debt ratio.
Column 3 (4) reports results from the Rhodes-Kropf et al. method (VP_RKRV) using the Blundell and Bond targets with the market (book) debt ratio.
Underlevered (overlevered) and undervalued (overvalued) are binary variables taking a value of 1 for firms that are in the highest (lowest) distance and
valuation terciles. Log size is the natural log of market capitalization (number of shares outstanding times price per share). Cash flow is operating income before
depreciation less the sum of depreciation and amortization, total income taxes, total interest and related expense, common dividends, and preferred dividends
divided by total assets. Non-operating cash flow is non-operating cash flow, scaled by total assets. The standard deviation of cash flow is calculated over a 5-year
period ending prior fiscal year. We require a minimum of 3 years of available cash flow data. Cash is cash and short-term investments, divided by assets. Property,
plant and equipment is the property, plant, and equipment, divided by assets. Sales growth is the percentage change in sales over the past year. Standard
deviation of returns is the standard deviation of daily stock returns over the prior fiscal year. We require 100 days of available returns. Illiquidity is defined as the
ratio of the daily absolute return to the dollar trading volume on that day. This ratio gives the absolute percentage price change per dollar of daily trading volume,
or the daily price impact of the order flow Amihud (2002). We average daily illiquidity for each firm over the prior fiscal year. All variables are demeaned and all
control variables are measured at the end of the prior fiscal year. P-values are reported beneath the coefficient estimates in parentheses. Standard errors are
clustered at the firm level and are heteroskedasticity-robust. Significance at the 1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk, respectively.
Valuation model VP_RIM VP_RKRV
Leverage measure Market Book Market Book
(1) (2) (3) (4)
Constant 3.411*** 3.350*** 3.311*** 3.250***
(0.000) (0.000) (0.000) (0.000)
Underlevered 0.176*** 0.037 0.172*** 0.034
(0.000) (0.311) (0.000) (0.343)
Overlevered 0.056 0.124*** 0.024 0.087**
(0.131) (0.001) (0.507) (0.017)
Undervalued 0.273*** 0.269*** 0.224*** 0.221***
(0.000) (0.000) (0.000) (0.000)
Overvalued 0.097*** 0.105*** 0.254*** 0.254***
(0.008) (0.004) (0.000) (0.000)
Log size 0.179*** 0.180*** 0.209*** 0.210***
(0.000) (0.000) (0.000) (0.000)
Cash flow 3.254*** 3.265*** 3.109*** 3.117***
(0.000) (0.000) (0.000) (0.000)
Non-operating cash flow 2.858*** 2.950*** 2.675*** 2.761***
(0.001) (0.000) (0.001) (0.001)
Standard deviation of cash flow 0.089 0.089 0.071 0.072
(0.425) (0.429) (0.503) (0.501)
Cash 0.633*** 0.553*** 0.460*** 0.386***
(0.000) (0.000) (0.000) (0.000)
Property, plant and equipment 0.128** 0.133*** 0.141*** 0.145***
(0.013) (0.010) (0.006) (0.005)
Sales growth 0.000 0.000 0.002 0.002
(0.940) (0.956) (0.659) (0.672)
Standard deviation of returns 8.082*** 7.112*** 7.802*** 6.883***
(0.000) (0.000) (0.000) (0.000)
Illiquidity 5980.988** 6817.304*** 5423.331** 6178.741***
(0.011) (0.005) (0.016) (0.008)
Observations 66,188 66,188 66,188 66,188
195 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
The effect of being overlevered and overvalued on announcing a repurchase is generally negative (three out of four cases)
and statistically significant in two of four cases. Overlevered and overvalued firms are 1921 percentage points less likely to
repurchase their shares. On the other hand, the effect of being underlevered and undervalued is unambiguous: Using either
measure of leverage or valuation, underlevered/undervalued firms are significantly more likely to announce a repurchase.
The economic magnitude of this effect ranges from 26 to 48 percentage points using Rhodes-Kropf et al. method and 5277
percentage points using the residual income model.
The effect of being underlevered and overvalued or overlevered and undervalued is theoretically unclear, and the empirical
results appear to depend on the measure of valuation. The positive effect of being underlevered dominates when using the
residual income model while the negative effect of being overvalued dominates when using the Rhodes-Kropf et al. method.
When they are underlevered and overvalued, firms are 1527 percentage points more likely to repurchase their shares with
the residual income model but 1018 percentage points less likely to undertake share repurchases with the Rhodes-Kropf et al.
method. The positive effect of being undervalued overshadows the negative effect of being overlevered, but only when
valuation is evaluated with the residual income model: the overlevered and undervalued firms are 1522 percentage points
more likely to repurchase their shares. The effects cancel each other out leading to a significant impact on the decision
to repurchase with the Rhodes-Kropf et al. valuation measure. To summarize, relative leverage and misvaluation both play
an important role in the decision to undertake a share repurchase plan as well as in determining the economic gains from
repurchase announcements.
Table 8
Determinants of the share repurchase decision: combined effects of leverage and misvaluation. This table presents logistic regressions modeling the decision to
announce a share repurchase. Column 1 (2) reports results from the residual income model (VP_RIM) using the Blundell and Bond targets with the market (book)
debt ratio. Column 3 (4) reports results from the Rhodes-Kropf et al. method (VP_RKRV) using the Blundell and Bond targets with the market (book) debt ratio.
Underlevered (overlevered) and undervalued (overvalued) are binary variables taking a value of 1 for firms that are in the highest (lowest) distance and
valuation terciles. Log size is the natural log of market capitalization (number of shares outstanding times price per share). Cash flow is operating income before
depreciation less the sum of depreciation and amortization, total income taxes, total interest and related expense, common dividends, and preferred dividends
divided by total assets. Non-operating cash flow is non-operating cash flow, scaled by total assets. The standard deviation of cash flow is calculated over a 5-year
period ending prior fiscal year. We require a minimum of 3 years of available cash flow data. Cash is cash and short-term investments, divided by assets. Property,
plant and equipment is the property, plant, and equipment, divided by assets. Sales growth is the percentage change in sales over the past year. Standard
deviation of returns is the standard deviation of daily stock returns over the prior fiscal year. We require 100 days of available returns. Illiquidity is defined as the
ratio of the daily absolute return to the dollar trading volume on that day. This ratio gives the absolute percentage price change per dollar of daily trading volume,
or the daily price impact of the order flow Amihud (2002). We average daily illiquidity for each firm over the prior fiscal year. All variables are demeaned and all
control variables are measured at the end of the prior fiscal year. P-values are reported beneath the coefficient estimates in parentheses. Standard errors are
clustered at the firm level and are heteroskedasticity-robust. Significance at the 1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk, respectively.
Valuation model VP_RIM VP_RKRV
Leverage measure Market Book Market Book
(1) (2) (3) (4)
Constant 3.366*** 3.354*** 3.300*** 3.278***
(0.000) (0.000) (0.000) (0.000)
Overlevered & overvalued 0.022 0.037 0.205*** 0.231***
(0.734) (0.559) (0.000) (0.000)
Underlevered & overvalued 0.238*** 0.137** 0.110** 0.202***
(0.000) (0.015) (0.037) (0.000)
Overlevered & undervalued 0.137** 0.197*** 0.087 0.099
(0.017) (0.000) (0.185) (0.135)
Underlevered & undervalued 0.571*** 0.420*** 0.389*** 0.233***
(0.000) (0.000) (0.000) (0.000)
Log size 0.176*** 0.177*** 0.190*** 0.189***
(0.000) (0.000) (0.000) (0.000)
Cash flow 3.302*** 3.319*** 3.203*** 3.229***
(0.000) (0.000) (0.000) (0.000)
Non-operating cash flow 3.075*** 3.181*** 2.971*** 3.112***
(0.000) (0.000) (0.000) (0.000)
Standard deviation of cash flow 0.092 0.095 0.093 0.092
(0.415) (0.409) (0.422) (0.430)
Cash 0.647*** 0.615*** 0.517*** 0.506***
(0.000) (0.000) (0.000) (0.000)
Property, plant and equipment 0.131** 0.135*** 0.135*** 0.139***
(0.011) (0.009) (0.008) (0.007)
Sales growth 0.001 0.001 0.001 0.001
(0.834) (0.840) (0.788) (0.767)
Standard deviation of returns 7.857*** 7.372*** 7.860*** 7.309***
(0.000) (0.000) (0.000) (0.000)
Illiquidity 6203.818*** 6788.505*** 5944.767** 6466.443***
(0.009) (0.005) (0.010) (0.006)
Observations 66,188 66,188 66,188 66,188
196 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
5. Conclusion
This paper extends previous research that investigates market reactions to share repurchases by studying the association
between stock returns to repurchase announcements and capital structure policy. The tradeoff theory predicts that the benefits
froma share repurchase should accrue to underlevered firms because they move towards their optimal debt ratio by repurchasing
equity. But underlevered firms must weigh the benefits of adjusting to their target capital structure against the cost of
repurchasing equity. The market timing theory predicts that undervalued firms should repurchase equity to exploit mispricing
opportunities while overvalued firms should not. We evaluate these theories in a unified framework within the context of a firm
having a target capital structure. We hypothesize that firms will benefit more from the capital structure adjustments achieved by
repurchasing stock when a firm's equity is undervalued. On the other hand, capital structure adjustments requiring repurchases
of overvalued stock will be more costly and hence less beneficial.
Consistent with our predictions, we find that market reactions to open market share repurchase are magnified (dampened) if
the firmis underlevered and undervalued (overlevered and overvalued). Our results are robust to different methods of measuring
equity mispricing, alternative definitions of leverage, and adding controls for some well-known determinants of share
repurchases. For example, using market debt ratios and the residual income model, the announcement returns are 1.5% higher for
underlevered and undervalued (3.3%) firms compared to overlevered and overvalued (1.8%) firms, even after taking into account
other potential motives for share repurchases. Furthermore, firms recognize the relationship between capital structure and
valuation as the source of economic benefits to share repurchases and are most likely to announce a share repurchase when they
are underlevered and undervalued.
To conclude, we contribute to both the repurchase and the capital structure literatures by showing that capital structure
adjustments are a value-increasing motive for a share repurchase and that the extent to which adjusting capital structure through
a share repurchase creates value depends on the undervaluation of the firm. Our findings suggest directions for further research.
With similar reasoning, the likelihood of equity issuance should be higher and the market reaction should be more favorable (less
value destroying) for overlevered and overvalued firms. Put differently, the likelihood of issuance should decrease and the
negative market reaction should be magnified if the firm is overlevered and overvalued in the context of seasoned equity
offerings.
Table A1
Target capital structure estimations. The presents target capital structure estimations using Blundell and Bond's (1998) during the sample period (19902010) for
book and market leverage, respectively. Profit is earnings before interest and taxes as a proportion of total assets; Market-to-book is book liabilities plus market
value of equity to total assets; Depreciation is depreciation expenses as a proportion of total assets; Size is the log of total book assets (a measure of firm size);
Tangibility is fixed assets as a proportion of total assets; R&D dummy is a binary variable equal to one if research and development expenses are positive and zero
otherwise; R&D expenses are research and development expenses as a proportion of total assets, where missing values are set equal to zero; and Median industry
leverage is the median leverage ratio for the firm's industry, based on the Fama and French (1997) industry categories. All variables used in the estimations are
lagged one period to avoid reverse causality. The targets are the predicted values from these regressions. The top row reports the coefficient estimates. Robust
standard errors are reported beneath the coefficient estimates in parentheses. Significance at the 1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk,
respectively.
Book debt ratio Market debt ratio
Leverage 0.7694*** 0.7592***
(0.012) (0.010)
Profit 0.0033 0.0030
(0.007) (0.007)
Market-to-book 0.0014** 0.0027***
(0.001) (0.001)
Depreciation 0.1382** 0.3888***
(0.061) (0.062)
Size 0.0047*** 0.0064***
(0.002) (0.002)
Tangibility 0.1007*** 0.1402***
(0.019) (0.020)
R&D dummy 0.0208 0.0260
(0.029) (0.024)
R&D expenses 0.0089 0.0674
(0.085) (0.086)
Median industry leverage 0.7694*** 0.7592***
(0.012) (0.010)
Firm fixed effects Yes Yes
Year fixed effects Yes Yes
Observations 46,162 46,120
Appendix A
197 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
Table A2
Parameter estimates for the residual income model.
This table presents means and medians of the parameter estimates used in the residual income model (Eq. (7))
V
RIM
B
0

X
n
t1
E
t
r B
t1
1 r
t

E
t
r B
t1
E
t1
r B
t
2 r 1 r
n
:
B is the book value of equity, E is expected earnings before extraordinary items, and r is the industry cost of capital using the Fama and French 48 industry
classifications. In our implementation, n = 2, which results in the following model:
V
RIM
B
0

E
1
r B
0
1 r
1

E
2
r B
1
1 r
1

E
2
r B
1
E
3
r B
2
2 r 1 r
2
:
Mean Std. dev. Median
r 0.129 0.049 0.125
B
0
852.826 4193.451 76.307
B
1
923.587 4437.286 85.056
B
2
1007.064 4822.014 93.077
E
1
107.748 859.062 5.421
E
2
115.932 872.759 5.418
E
3
125.870 925.763 5.526
Table A3
RKRV valuation model parameter estimates.
This table presents the average Fama Macbeth regression coefficients and Fama Macbeth standard errors (in parenthesis) for annual cross-sectional regressions
for firms in each of the 12 Fama and French industry classificationsspecifically Eq. (12):
m
it

0jt

1jt
b
it

2jt
ln NI

it

3jt
I
b0
ln NI

it

4jt
LEV
RKRV it

it
:
where the variable E
t
c
0

represents the time series average constant for each regression. E
t
c
k

is the time series average slope for the kth accounting variable.
The accounting variables are as follows: m is the log of the market value of equity, b is log of the book value of equity, NI
+
is the absolute value of net income (also
interacted with an indicator variable I
b0
for negative net income). Finally, LEV
(RKRV)
is the firm leverage estimated using the method of Rhodes-Kropf et al. (2005).
The time series average R
2
is reported.
Fama and French industry classification
1 2 3 4 5 6 7 8 9 10 11 12
E
t

0

2.20 2.44 2.06 2.18 2.28 2.29 2.71 2.41 2.29 2.45 2.14 2.23
(0.07) (0.09) (0.07) (0.08) (0.07) (0.05) (0.13) (0.13) (0.06) (0.04) (0.07) (0.05)
E
t

1

0.65 0.60 0.68 0.70 0.64 0.62 0.61 0.81 0.65 0.62 0.58 0.64
(0.01) (0.02) (0.01) (0.01) (0.02) (0.01) (0.02) (0.03) (0.01) (0.01) (0.01) (0.01)
E
t

2

0.31 0.30 0.28 0.22 0.35 0.33 0.29 0.16 0.30 0.33 0.38 0.29
(0.01) (0.02) (0.01) (0.01) (0.02) (0.01) (0.02) (0.04) (0.01) (0.01) (0.01) (0.01)
E
t

3

0.02 0.04 0.03 0.01 0.02 0.08 0.01 0.01 0.08 0.10 0.20 0.07
(0.01) (0.02) (0.01) (0.02) (0.03) (0.01) (0.06) (0.10) (0.01) (0.02) (0.02) (0.01)
E
t

4

2.71 2.62 2.43 2.39 2.77 2.65 2.75 2.82 2.38 2.62 1.38 2.10
(0.07) (0.10) (0.09) (0.12) (0.16) (0.09) (0.19) (0.21) (0.06) (0.08) (0.05) (0.06)
R
2
0.86 0.83 0.86 0.86 0.87 0.86 0.85 0.92 0.87 0.87 0.84 0.84
198 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
Table A4
Univariate comparisons of median 3-day announcement CARs. This table presents median 3-day CARs to repurchase announcements reported in the SDC
repurchase database between 1990 and 2010, segmented on measures of relative leverage and valuation, where relative leverage is calculated using the modified
Blundell and Bond targets, following Ogden and Wu (2013). More specifically, we re-run our leverage regressions documented in Table A1 replacing the
market-to-book variable (MB) with its inverse exponential function (i.e., exp
MBi;t1
). Panel A and Panel B report results from the modified Blundell and Bond
targets using the market debt ratio and book debt ratio, respectively. Panel C and Panel D report results from the residual income model (VP_RIM) and
Rhodes-Kropf et al. method (VP_RKRV), respectively. Significance at the 1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk, respectively.
N Median 3-day CAR Difference (P-value)
Panel A: leverage: market debt ratios
Overlevered (distance ML b 0) 2483 1.337 0.561
Underlevered (distance ML N 0) 4593 1.898 b0.001***
Panel B: leverage: book debt ratios
Overlevered (distance ML b 0) 2642 1.387 0.462
Underlevered (distance ML N 0) 4552 1.849 0.002***
Panel C: valuation: RIM
Overvalued (VP_RIM b VP_RIM
Median
) 2892 1.540 0.277
Undervalued (VP_RIM N VP_RIM
Median
) 2891 1.817 0.100**
Panel D: valuation: RKRV
Overvalued (VP_RKRV b VP_RKRV
Median
) 3616 1.373 0.761
Undervalued (VP_RKRV N VP_RKRV
Median
) 3606 2.135 b0.001***
Table A5
Univariate pairwise comparisons of median 3-day announcement CAR. This table presents median 3-day CARs to repurchase announcements, segmented on the
interaction of measures of relative leverage and valuation, where relative leverage is calculated using the modified Blundell and Bond targets, following Ogden
and Wu (2013). More specifically, we re-run our leverage regressions documented in Table A1 replacing the market-to-book variable (MB) with its inverse
exponential function (i.e., exp
MBi;t1
). Panel A and Panel B report results from the residual income model (VP_RIM) using the modified Blundell and Bond targets
with the market debt ratio and book debt ratio, respectively. Panel C and Panel D report results from the Rhodes-Kropf et al. method (VP_RKRV) using the
modified Blundell and Bond targets with the market debt ratio and book debt ratio, respectively. We present differences in medians across rows (undervalued
overvalued) and columns (underleveredoverlevered), as well as along the diagonal (undervalued/underleveredovervalued/overlevered). Significance at the
1%, 5% or 10% levels are shown with 3, 2, or 1 asterisk, respectively.
Overvalued
(VP b VP
Median
)
Undervalued
(VP N VP
Median
)
Difference
(P-value)
Panel A: market debt ratios and VP_RIM
Overlevered (distance ML b 0) 1.126 1.387 0.261
N = 1052 N = 681 (0.336)
Underlevered (distance ML N 0) 1.731 2.066 0.335
N = 1968 N = 1621 (0.042**)
Difference (P-value) 0.606 0.680 0.941
(b0.023***) (b0.001***) (0.001***)
Panel B: book debt ratios and VP_RIM
Overlevered (distance BL b 0) 1.246 1.450 0.204
N = 839 N = 1076 (0.204)
Underlevered (distance BL N 0) 1.716 1.930 0.214
N = 1864 N = 1632 (0.073*)
Difference (P-value) 0.470 0.480 0.685
(0.015**) (0.004***) (0.002***)
Panel C: market debt ratios and VP_RKRV
Overlevered (distance ML b 0) 1.061 1.712 0.651
N = 1340 N = 921 (b0.001***)
Underlevered (distance ML N 0) 1.594 2.277 0.683
N = 2050 N = 2268 (b0.001***)
Difference (P-value) 0.533 0.565 1.216
(b0.001***) (0.013**) (b0.001***)
Panel D: book debt ratios and VP_RKRV
Overlevered (distance BL b 0) 1.133 1.779 0.646
N = 1505 N = 910 (0.004***)
Underlevered (distance BL N 0) 1.545 2.210 0.665
N = 1917 N = 2358 (b0.001***)
Difference (P-value) 0.412 0.431 1.077
(0.011**) (0.046**) (b0.001***)
199 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200
References
Acharya, Viral V., Almeida, Heitor, Campello, Murillo, 2007. Is cash negative debt? A hedging perspective on corporate financial policies. J. Financ. Intermed. 16,
515554.
Alti, Aydogan, 2006. How persistent is the impact of market timing on capital structure? J. Financ. 61, 16811710.
Amihud, Yakov, 2002. Illiquidity and stock returns: cross-section and time-series effects. J. Financ. Mark. 5, 3156.
Baker, Malcolm, Wurgler, Jeffrey, 2002. Market timing and capital structure. J. Financ. 57, 132.
Bargeron, Leonce, Bonaim, Alice, Thomas, Shawn, 2012. Returns over the life-cycles of open market repurchase programs. Working Paper. University of Pittsburghand
University of Kentucky.
Billett, Matthew, Xue, Hui, 2007. The takeover deterrent effect of open market share repurchases. J. Financ. 62, 18271850.
Blundell, Richard, Bond, Stephen, 1998. Initial conditions and moment restrictions in dynamic panel data models. J. Econ. 87, 115143.
Bonaime, Alice, 2012. Repurchases, reputation, and returns. J. Financ. Quant. Anal. 47, 469491.
Brav, Alon, Graham, John, Harvey, Campbell R., Michaely, Roni, 2005. Payout policy in the 21st century. J. Financ. Econ. 77, 483527.
Brown, Stephen J., Warner, Jerold B., 1980. Measuring security price performance. J. Financ. Econ. 8, 205258.
Butler, Alexander W., Cornaggia, Jess, Grullon, Gustavo, Weston, James, 2009. Corporate financing decisions andmanagerial market timing. WorkingPaper. Rice University.
Chan, Konan, Ikenberry, David, Lee, Inmoo, 2004. Economic sources of gain in stock repurchases. J. Financ. Quant. Anal. 39, 461479.
Comment, Robert, Jarrell, Gregg A., 1991. The relative signaling power of Dutch-auction and fixed-price self-tender offers and open-market share repurchases. J.
Financ. 46, 12431271.
D'Mello, Ranjan, Shroff, Pervin K., 2000. Equity undervaluation and decisions related to repurchase tender offers: an empirical investigation. J. Financ. 55, 23992425.
Dittmar, Amy K., 2000. Why do firms repurchase stock? J. Bus. 73, 331355.
Dittmar, Amy K., Dittmar, Robert F., 2008. The timing of financing decisions: an examination of the correlation in financing waves. J. Financ. Econ. 90, 5983.
Dong, Ming, Hirshleifer, David, Richardson, Scott, Teoh, Siew H., 2006. Does investor misvaluation drive the takeover market? J. Financ. 61, 725762.
Elliott, William B., Koter-Kant, Johanna, Warr, Richard S., 2007. A valuation-based test of market timing. J. Corp. Finance 13, 122128.
Elliott, William B., Koter-Kant, Johanna, Warr, Richard S., 2008. Market timing and the debt-equity choice. J. Financ. Intermed. 17, 175197.
Fama, F. Eugene, French, Kenneth R., 1997. Industry costs of equity. J. Financ. Econ. 43, 153193.
Faulkender, Michael W., Flannery, Mark J., Hankins, Kristine W., Smith, Jason M., 2012. Cash flows and leverage adjustments. J. Financ. Econ. 103, 632646.
Fischer, Edwin, Heinkel, Robert, Zechner, Josef, 1989. Dynamic capital structure choice: theory and tests. J. Financ. 44, 1940.
Flannery, Mark J., Hankins, Kristine W., 2013. Estimating dynamic panel models in corporate finance. J. Corp. Finance 19, 119.
Flannery, Mark J., Rangan, Kasturi, 2006. Partial adjustment and target capital structures. J. Financ. Econ. 79, 469506.
Frankel, Richard, Lee, Charles M.C., 1998. Accounting valuation, market expectation and cross-sectional stock returns. J. Account. Econ. 25, 283319.
Graham, John R., Harvey, Campbell R., 2001. The theory and practice of corporate finance: evidence from the field. J. Financ. Econ. 60, 187243.
Grullon, Gustavo, Michaely, Roni, 2004. The information content of share repurchase programs. J. Financ. 59, 651680.
Guay, Wayne, Harford, Jarrad, 2000. The cash-flow permanence and information content of dividend increases versus repurchases. J. Financ. Econ. 57, 385415.
Gungoraydinoglu, Ali, ztekin, zde, 2011. Firm- and country-level determinants of corporate leverage: some new international evidence. J. Corp. Finance 17,
14571474.
Hertzel, Michael G., Li, Zhi, 2010. Behavioral and rational explanations of stock price performance around SEOs: evidence from a decomposition of market-to-book
ratios. J. Financ. Quant. Anal. 45, 935958.
Hovakimian, Armen, Opler, Tim, Titman, Sheridan, 2001. The debt-equity choice. J. Financ. Quant. Anal. 36, 124.
Huang, Rongbing, Ritter, Jay, 2009. Testing theories of capital structure and estimating the speed of adjustment. J. Financ. Quant. Anal. 44, 237271.
Ikenberry, David, Lakonishok, Josef, Vermaelen, Theo, 1995. Market underreaction to open market share repurchases. J. Financ. Econ. 39, 181208.
Jagannathan, Murali, Stephens, Clifford P., Weisbach, Michael S., 2000. Financial flexibility and the choice between dividends and stock repurchases. J. Financ.
Econ. 57, 355384.
Jensen, Michael, 1986. Agency costs of free cash flow. Am. Econ. Rev. 76, 323329.
Kahle, Kathleen, 2002. When a buyback isn't a buyback: open market repurchases and employee options. J. Financ. Econ. 63, 235261.
Kothari, S.P., Shanken, Jay, 1997. Book-to-market, dividend yield, and expected market returns: a time-series analysis. J. Financ. Econ. 44, 177210.
La Porta, Rafael, 1996. Expectations and the cross-section of stock returns. J. Financ. 51, 17151742.
Leary, Mark, Roberts, Michael, 2005. Do firms rebalance their capital structure? J. Financ. 60, 25752619.
Lee, Charles, Myers, James, Swaminathan, Bhaskaran, 1999. What is the intrinsic value of the Dow? J. Financ. 54, 16931741.
Lemmon, Michael, Roberts, Michael R., Zender, Jaime F., 2008. Back to the beginning: persistence and the cross-section of corporate capital structure. J. Financ. 63,
15751608.
Lie, Erik, 2002. Do firms undertake self-tender offers to optimize capital structure? J. Bus. 75, 609639.
Liu, Laura X., 2009. Historical market-to-book in a partial adjustment model of leverage. J. Corp. Finance 15, 602612.
Louis, Henock, White, Hal, 2007. Do managers intentionally use repurchase tender offers to signal private information? Evidence from firm financial reporting
behavior. J. Financ. Econ. 85, 205233.
Massa, Massimo, Rehman, Zahid, Vermaelen, Theo, 2007. Mimicking repurchases. J. Financ. Econ. 84, 624666.
Myers, Stewart C., 1984. The capital structure puzzle. J. Financ. 39, 575592.
Ogden, Joseph P., Wu, Shanhong, 2013. Reassessing the effect of growth options on leverage. J. Corp. Finance 23, 182195.
Ohlson, James A., 1991. The theory of value and earnings, and an introduction to the Ball-Brown analysis. Contemp. Account. Res. 7, 119.
Ohlson, James A., 1995. Earnings, book values, and dividends in security valuation. Contemp. Account. Res. 11, 661687.
ztekin, zde, 2014. Capital structure decisions around the world: which factors are reliably important? J. Financ. Quant. Anal. (forthcoming).
ztekin, zde, Flannery, Mark J., 2012. Institutional determinants of capital structure adjustment speeds. J. Financ. Econ. 103, 88112.
Penman, Stephen H., Sougiannis, Theodore A., 1998. A comparison of dividend, cash flow, and earnings approaches to equity valuation. Contemp. Account. Res. 15,
343383.
Peterson, Mitchell A., 2009. Estimating standard errors in finance panel data sets: comparing Approaches. Rev. Financ. Stud. 22, 435480.
Rhodes-Kropf, Matthew, Viswanathan, S., 2004. Market valuation and merger waves. J. Financ. 59, 26852718.
Rhodes-Kropf, Matthew, Robinson, David T., Viswanathan, S., 2005. Valuation waves and merger activity: the empirical evidence. J. Financ. Econ. 77, 561603.
Stephens, Clifford, Weisbach, Michael, 1998. Actual share reacquisitions in open-market repurchase programs. J. Financ. 53, 313333.
Vermaelen, Theo, 1981. Common stock repurchases and market signalling. J. Financ. Econ. 9, 139183.
Warr, Richard S., Elliott, W.B., Koter-Kant, Johanna, ztekin, zde, 2012. Equity mispricing and leverage adjustment costs. J. Financ. Quant. Anal. 47, 589616.
200 A.A. Bonaim et al. / Journal of Corporate Finance 26 (2014) 182200