Financing Payouts *

Joan Farre-Mensa
Harvard Business School

Roni Michaely
Cornell University
and IDC

Martin Schmalz
University of Michigan

December 8, 2014

We would like to thank Malcolm Baker, Alexander Ljungqvist, Uday Rajan, Matt Rhodes-Kropf, Jeffrey Zwiebel, and seminar
audiences at Harvard Business School. Schmalz is grateful for generous financial support through an NTT Fellowship from the
Mitsui Life Financial Center.

Address for correspondence: Harvard Business School, Rock 218, Soldiers Field, Boston, MA, 02163; phone 617-495-6963.
E-mails: jfarremensa@hbs.edu (Farre-Mensa); rm34@cornell.edu (Michaely); schmalz@umich.edu (Schmalz).

Financing Payouts

Abstract
Despite the obvious interest in payout policy, no paper to date has systematically analyzed
how payouts are funded, perhaps because the answer might have appeared just too
obvious: payouts are funded with free cash flowat least over long enough time periods.
In stark contrast to this commonly held view, we find that firms rely on the capital markets
to finance a third of aggregate payouts, mainly with debt but also with equity. Such
financed payouts are widespread, persistent, prevalent both among dividend-paying and
repurchasing firms, and large in magnitude. Standard interpretations of agency or signaling
theories are unable to explain this behavior. We argue, however, that our findings are
consistent with a reinterpretation of ideas related to agency conflicts and a holistic view of
corporate financial strategy that examines payout and capital structure decisions jointly.

Key words: Payout policy, financing decisions, debt issues, equity issues, capital structure.
JEL classification: G35; G32.

The established conventional wisdom is that firms rely on internally generated funds to fund
their payouts, whether these payouts are motivated by agency, signaling, or other considerations.
For example, Grullon, Michaely, and Swaminathan (2002) and DeAngelo, DeAngelo, and Stulz
(2006) present a lifecycle view of payouts where mature, cash-rich firms distribute excess free
cash flow to their investors while young, growing firms raise but do not pay out capital. Ross,
Westfield, and Jaffe (2013) conclude that a firm should begin making distributions when it
generates sufficient internal cash flow to fund its investment needs now and into the foreseeable
future. Accordingly, they recommend managers to set their level of payouts low enough to
avoid expensive future external financing (p. 607). While it is a theoretical possibility that firms
could also raise outside funds to finance their payouts, such behavior is costly and thus
considered, at least by some authors, uneconomic as well as pointless (Miller and Rock, 1985).
The results in this paper counter this conventional thinking: We find that 42% of firms with
positive payout initiate an equity or net debt issue during the same year; the vast majority of
them, 36% of all payers, could not have funded their payout without the proceeds of these issues.
In addition, we show that 35% of the cash firms raise is paid out by the same firms during the
same year. Our findings thus indicate that a substantial fraction of aggregate payouts is not
funded by internally generated funds but rather is actively financed in the capital markets.
In addition to being widespread, actively financed payouts are also important because of their
dollar magnitude: 32% of the aggregate capital paid out by public U.S. firms is simultaneously
raised by the same payers via firm-initiated equity or net debt issues.1 Most of the capital that
firms simultaneously pay out and actively raise26% of aggregate payoutsis needed to fill the
gap between the firms payouts and the sum of their internal funds and the proceeds of employee
stock option exercises. If we follow a less conservative approach and include employee-initiated
1

Much of the proceeds of gross debt issues are simply used to roll over prior debt. Our conservative focus on net
debt issues (i.e., debt issues net of debt repurchases) allows us to capture those proceeds that firms can actually use
to fund investment, cash flow shortfalls, oras it turns outspayouts.

equity issues as a source of payout financing instead of treating them as internal funds, the
percentage of financed payouts increases by nine percentage points: 41% of the aggregate capital
paid out by public firms is simultaneously raised by the same payers via firm- or employeeinitiated security issues.
Financed payouts are markedly pro-cyclical: In 2007, right before the onset of the Great
Recession, industrial public firms in the U.S. actively financed $200 billion of payouts by
simultaneously initiating debt or equity issues. Two years later, in the midst of the recession, the
amount of financed payouts was cut by a factor of four, barely reaching $50 billion. Notably,
though, payout financing is not a transitory phenomenon: All our results remain virtually
unchanged if we measure firms sources and uses of funds over four-year intervals. This finding
indicates that the use of external capital to finance payouts is persistent and is not the result of
payout smoothing or, more generally, of time mismatches between free cash flow and payouts.
The frequency, magnitude, and persistence of financed payouts is striking, particularly in
light of the obvious costs associated with this behavior. In addition to direct issuance costs, these
costs include asymmetric information discounts on newly issued securities, due to which
financed payouts are inconsistent with pecking order models (Myers and Majluf, 1984). The
possibility that firms may have to pass up profitable investment opportunities when resorting to
the capital markets to finance payouts instead of investment is another potential cost of financing
payouts (Asker, Farre-Mensa, and Ljungqvist, 2014).
The fact that approximately 40% of all payers finance their payouts implies that there must
be benefits that offset these well-known costs. In order to explore what these benefits are, we
first examine the form of payouts that firms finance. We find that firms are as likely to finance
share repurchases as dividends. Thus, actively financed payouts cannot be simply explained by
firms desire to avoid the well-known costs associated with cutting dividends (Linter, 1956;
Brav, Graham, Harvey, and Michaely, 2005). We do observe, however, that in recessionary
2

years, a larger fraction of financed payouts take the form of dividends than repurchases. This
finding is consistent with the notion that the desire to avoid dividend cuts is an important driver
of payout financing behavior in times when access to capital is most limited.
Next, we examine the extent to which firms choose debt or equity issues to actively finance
their payouts. This choice has direct capital structure implications, and thus it points to different
motives for why firms raise and pay out capital at the same time. Our analysis shows that net
debt is by far the most important source of payout financing: up to 30% of aggregate payouts are
financed via simultaneous net debt issues made by the same payers. Conversely, 39% of the
proceeds of net debt issues $135 billion of the $350 billion of net debt issued by public U.S.
firms in our average sample yearare paid out during the same year by the same issuing firms.
These findings are consistent with firms using financed payouts as a way to adjust their capital
structure without depleting their cash reserves.2
Given that secondary equity offerings (SEOs) and private placements are relatively rare, it is
not surprising that only 3% of aggregate payouts are financed via firm-initiated equity issues.
Yet, when firms do initiate equity issues, they pay out as much as 19% of the proceeds during the
same year. This unexpected finding of equity recycling indicates that capital structure
adjustments are an important but not the only reason why firms choose to set payout levels that
they need to finance by actively raising capital.
On the face of it, equity recycling may appear hard to justify due to the obvious costs
associated with equity-financed payouts. It can however be the result of a monitoring strategy in
which boards make firms pay out all their operating cash flow, with the only exception of those
funds required for maintenance-type investments. As a result, managers are forced to raise
2

Such capital structure adjustments can be motivated by agency (Jensen, 1986), taxes (e.g., Modigliani and Miller,
1963), or risk-rebalancing considerations. Debt-financed payouts also allow firms to avoid paying repatriation taxes
when their cash is held overseas (Foley, Hartzell, Titman, and Twite, 2007). In addition, in the case of debt-financed
repurchases, they can increase earnings-per-share and thus help firms meet or beat analysts forecast (Bens, Nagar,
Skinner, and Wong, 2003).

capital to fund any new investment projects, even those they could have funded internally had
they cut their payouts (Easterbrook, 1984). This monitoring strategy mitigates the agency costs
of free cash flow by allowing shareholders to decide which new projects should be funded; to the
extent that debtholders can exercise a similar monitoring role, it can also motivate debt-financed
payouts. Importantly, though, this monitoring strategy, which results in payouts well in excess of
free cash flow, is very different from the commonly held notion that cash rich firms simply pay
out their free cash flow to mitigate agency problems (Jensen, 1986).
The empirical evidence is consistent with the notion that the desire to prevent wasteful
investment plays a significant role in explaining both equity- and debt-financed payouts: we find
that 30% to 40% of actively financed payouts correspond to firms with unusually high
investment levels. These firms internal funds would have been enough to fund their payouts in a
counterfactual scenario in which the firms invested at routine levels. Conversely, these firms
would have had to raise less or no capital had their cut their payouts when investment needs
surged. Yet by forcing the firms to maintain their payouts, their boards ensured that managers
had to raise capital and thus had to subject themselves to the scrutiny of the capital markets.
By contrast, it appears unlikely that financed payouts are motivated by signaling
considerations. Financed payouts are inconsistent with well-known payout signaling models such
as Miller and Rock (1985). In other models, such as Bhattacharya (1979), financed dividends can
arise when a firm signals its quality by committing to a dividend level that, ex ante, it expects to
be able to fund internally; yet, ex post, a low-probability cash flow shock forces the firm to raise
funds to finance its dividend. We find that close to 80% of firms that need to finance their
dividends do not have abnormally low profitability levels. In addition, dividend signaling models
such as Bhattacharyas cannot explain financed repurchases, which are just as prevalent as
financed dividends: Repurchase announcements are typically seen as significantly less binding

than dividends (Jagannathan, Stephens, and Weisbach, 2000), and so they can be cancelled
following a negative shock with little signaling cost.
Our paper makes two contributions. First, ours is the first paper to systematically analyze
how firms fund their payouts. The vast literature on payout policy has investigated in detail the
determinants of the form of payouts (dividends or repurchases), their motivations, and the effect
that payout decisions have on equity returns. (See Allen and Michaely (2003), DeAngelo,
DeAngelo, and Skinner (2008), and Farre-Mensa, Michaely, and Schmalz (2014) for recent
reviews.) Yet despite the obvious interest in payouts, no paper to date has analyzed how payouts
are actually funded, perhaps because the answer might have appeared just too obvious: payouts
are funded with free cash flowat least over long enough time periods.
In fact, we find that firms rely on the capital markets to finance a third of aggregate payouts,
with another nine percentage points of paid-out capital provided by employees through stock
option exercises. While it has long been known that, occasionally, some firms rely on the capital
markets to finance large payouts (Denis and Denis, 1993; Wruck, 1994; Andrade and Kaplan,
1998), the pervasiveness, economic magnitude, and persistence of financed payouts that we
document is unexpected.
Second, our paper enhances our understanding of what firms do with the proceeds of security
issues. As it turns out, in addition to using these proceeds to fund investment, cover cash flow
shortfalls, or build up their cash reserves (e.g., Kim and Weisbach, 2008; Denis and McKeon,
2012), a substantial fraction of the capital firms raise is used to fund payouts. By simultaneously
raising and paying out capital, firms can accomplish a number of objectives, from rebalancing
their capital structure, mitigating the agency costs of free cash flow, or increasing earnings-pershare. Regardless of the motives at play, a key insight of our paper is that payout and issuance
decisions are intrinsically related, and thus much can be gained by studying them jointly as
interdependent elements of the financial ecosystem.
5

The paper proceeds as follows. Section 1 provides aggregate statistics on payout and capital
raising activity. Section 2 examines simultaneous payouts and security issues. Section 3
investigates whether firms that simultaneously raise and pay out capital could have funded their
payouts without raising capital. Section 4 discusses the costs and benefits of financing payouts
with the goal of shedding light on the motives behind this behavior. Section 5 concludes.
1. Aggregate payout and capital raising activity
1.1 Sample selection
Our baseline sample consists of all public U.S. firms that appear in the Compustat-CRSP
merged files between 1989 and 2012.3 We exclude firms in the year of their IPO because we are
interested in payout and issuance decisions after firms initial public offering. We also exclude
financial firms (SIC 6) and utilities (SIC 49). The final sample consists of 10,591 unique firms
and 90,791 firm-year observations.
1.2 Description of aggregate payout and capital raising activity
We start by describing the aggregate payout and capital raising activities of industrial public
U.S. firms during our sample period. Table 1 shows the percentage of firms that pay out or raise
capital for each year in our sample period (Panel A) and the aggregate dollar amount paid out or
raised (Panel B).4 We report annual figures averaged over four-year intervals to conserve space.
To better visualize time trends, Figures 1 and 2 show the underlying annual figures that we have
averaged in Table 1. Importantly, in Table 1 and Figures 1 and 2 there is no link at the firm level
between payouts and security issues, as capital may be paid out and raised by different firms.
Several significant trends emerge from this analysis. Both the percentage of firms with
positive total payout (the sum of dividends and share repurchases) and the dollar amount paid out
have experienced a substantial increase over our sample period. This increase in payouts has
3

We start our sample in 1989 because this is the first full year for which data from the Statement of Cash Flow were
standardized following the adoption of Statement of Financial Accounting Standard 95.
4
All dollar figures reported in the paper are in real dollars of year 2012 purchasing power.

largely been driven by share repurchases, which have been the most important payout vehicle
since 1997 (with the only exception of 2009). In particular, Figure 1 shows that repurchases
exploded from 2002 to 2007, going up by a factor of almost four, from $143 billion in 2002 to
$562 billion in 2007. They fell back to $150 billion in 2009 and have been recovering since then.
Dividends, on the other hand, have experienced a different dynamic: As noted by Fama and
French (2001), the number of dividend payers declined through the 1990s and reached a lowpoint in 2001, when less than a quarter of all public U.S. firms paid a dividend. But dividends
have made a remarkable comeback since then, with the percentage of dividend payers peaking at
39% in 2012, the last year of our sample period.5 At the same time, the aggregate dollar amount
paid out through dividends stayed relatively flat during the 1990sreflecting the fact that the
firms that stopped paying dividends accounted for a relative small fraction of aggregate
dividends payments (DeAngelo, DeAngelo, and Skinner, 2004)and has been slowly but
steadily increasing since 2001. Figure 1 also shows that repurchases have been much more
volatile and pro-cyclical than dividends throughout our sample; this is consistent with the wellknown tendency of firms to pay smooth dividends (e.g., Jagannathan et al., 2000; Leary and
Michaely, 2011).
On the securities issuance side, net debt issues, defined as the absolute value of the difference
between the amount of debt issued and the amount repurchased or retired, have been by far the
most important source of external funds for U.S. firms. In the average year in our sample period,
U.S. firms raised $255 billion in net debt, representing over 70% of the capital they raised that
year. Figure 2 shows that the aggregate dollar amount raised via net debt issues has been mostly
pro-cyclical, though it was slow to recover after the 2001 recession, when it continued declining
through 2005. The figure also shows that between 20% and 45% of all firms issued net debt in

Farre-Mensa, Michaely, and Schmalz (2014) review the recent payout literature and, in particular, papers that have
examined the forces behind this disappearance and reappearance of dividends.

any given year, a percentage that was also highly pro-cyclical.

In addition to issuing debt, firms can raise capital by issuing equity. We break down equity
issues in firm-initiated issues (SEOs and private placements) and employee-initiated issuesthe
latter being mostly the result of employee stock options exercises (McKeon, 2013). Employeeinitiated issues have become increasingly prevalent over the last three decades: In any given year
since 1997, over 60% of all public firms received a cash inflow from employee stock option
exercises. These exercises have become an importantthough perhaps unintendedsource of
capital for public U.S. firms, with the aggregate proceeds of employee-initiated equity issues
peaking at $100 billion in 2007.
Firm-initiated equity issues, on the other hand, are less frequent: During our sample period,
only between 10% and 20% of all public firms initiated an equity issue in any given year. The
annual proceeds of firm-initiated equity issues averaged $45 billion during our sample, reaching
a peak of $123 billion in 2000 but not surpassing $50 billion since 2002. As a result, employeeinitiated equity issues have been a quantitatively more important source of funds than firminitiated issues in every year since 2002, except in 2009 (see Figure 2).
There is a key conceptual difference between a firm that chooses to raise capital by initiating
a debt or an equity issue, and a firm that raises capital as a by-product of its employees option
exercises, over which the firm has little direct control.6 Throughout the paper, we therefore
distinguish between i) instances in which firms actively raise capital, either by initiating the
issuance of net debt or equity, and ii) instances in which the capital-raising events are triggered
by actions of the firms employees.

Firms do control if and when they issue options to their employees, but arguably this decision is motivated by the
need to provide incentives and typically takes place several years before the options are exercised. Also, as argued
by McKeon (2013), the vast majority of stock option grants are to employees below the top five executives. Even in
the case of options exercised by top managers, these exercises are most naturally seen as personal investment
decisions rather than as firm-level capital-raising events.

To sum up, Table 1 and Figures 1 and 2 show that the aggregate payout and capital raising
activities have both trended upwards over our sample period. The aggregate annual amounts paid
out and raised are positively correlated over time, with a correlation coefficient of 0.58, and their
average values are almost identical ($353.7 billion and $350.4 billion, respectively).
It is important to emphasize that the aggregate statistics presented so far do not imply that
payouts and issuances are related at the firm level: For example, it could be that firms paying out
and raising capital are different firms that find themselves at different stages of their life cycles,
as predicted by standard lifecycle theories (e.g., Grullon et al., 2002; DeAngelo et al., 2006). The
goal of this paper, however, is to understand the extent to which firms rely on the capital they
raise to finance their payouts. To that end, we provide evidence that payouts and issuances are
related at the firm level in the next sections.
2. An upper bound for financed payouts: Simultaneously raising and paying out capital
The following example illustrates the conceptual difficulties of measuring firms reliance on
the proceeds of security issues to finance their payouts: Consider the case of a firm that, in a
given year, pays out $50, has free cash flow (operating cash flow net of investment) of $50, and
issues $100 of net debt. Can we say that this firm is using debt to finance its payout? On the one
hand, the firm is simultaneously issuing debt and paying out capital, and so one could argue that
it is; on the other hand, the firms free cash flow is enough to fund its payout, and so the firm did
not in fact need to issue debt to finance its payout. These difficulties are, of course, a direct
consequence of the fact that money is fungible.
In this section, we start our analysis of financed payouts by investigating how common it is
for firms to pay out and raise capital simultaneously, i.e., during the same fiscal year. In doing
so, we examine joint payout and issuance decisions at the firm level, in contrast to the aggregate
statistics presented in Section 1. As the previous example illustrates, this will give us an upper
bound of the extent to which firms rely on the proceeds of security issues to finance their
9

payouts. The underlying assumption is that issuance proceeds are the source of capital of first
resort when funding payouts.7 In Section 3, we will provide a lower bound of how prevalent
financed payouts are by assuming that the proceeds of security issues are the source of capital of
last resort when it comes to funding payouts.
We report both firm counts (e.g., the percentage of firms that simultaneously raise and pay
out capital) and dollar magnitudes (e.g., the dollar amount of capital that firms simultaneously
raise and pay out). Similarly as Table 1, the tables in this section report annual figures averaged
over four year intervals to conserve space. We supplement the tables with graphs plotting annual
figures to better illustrate time trends.
In Section 2.1., we first examine the prevalence of firms that pay out capital and issue any
type of security during the same year. We then analyze how our findings change if we focus only
on actively issued securities, i.e., instances in which firms initiate a debt or an equity issue, thus
excluding equity issues that are the result of employee stock options exercises. In Section 2.2, we
further disaggregate the securities that are issued simultaneously with payouts into net debt
issues, firm-initiated equity issues, and employee-initiated equity issues.
2.1 All security issues vs. firm-initiated security issues
Table 2, Panel A, provides an upper bound of the extent to which firms finance their payouts
by examining firms that pay out capital and issue any type of security (net debt, firm-initiated
equity, or employee-initiated equity) during the same year. Columns 1, 2, and 3 report the
number of firms that, in any given year, both raise and pay out capital, presented as a fraction of
the whole population of public firms (Column 1), the population of firms that pay out capital
(Column 2), and the population of firms that raise capital (Column 3).

The only exception are the proceeds of debt issues: By using net debt in our analysis, we are assuming that debt
proceeds are only used to finance payouts after funding debt retirements.

10

The takeaway from these three columns is clear: A substantial fraction of firms raise and pay
out capital during the same year.8 Specifically, Column 1 shows that, in the average year over
our sample period, 40% of all public U.S. firms paid out and raised capital during the same fiscal
year. This percentage has been growing over time, peaking in 2012 at 52%. Relatedly, Column 2
shows that a striking 82% of payout payers raise capital in the same year; Column 3 conditions
the sample on firms that raise capital, showing that over 50% of firms that issue securities also
simultaneously pay out capital.
Taken at face value, these findings may be viewed as surprising: Insofar as information
asymmetries between firms and investors cause a wedge between the costs of external and
internal capital, leading firms financial policies to follow a pecking order (Myers and Majluf,
1984), firms should cut their payouts before issuing any securities. Similarly, many standard
payout models do not predict the frequent occurrence of simultaneous payouts and security
issues (see, e.g., Miller and Rock, 1985). In what follows, our goal is to better understand why so
many firms appear to find it optimal to simultaneously raise and pay out capital, despite the costs
of such a policy of financing payouts.
We start by analyzing the degree to which our findings are driven by employee-initiated
equity issues. Table 1 shows that, in the later years of our sample, over two-thirds of all public
firms experienced such an issue. In light of this, it is all but unavoidable that many payout payers
will simultaneously raise capital. In fact, to the extent that the proceeds from these issues can be
seen as additional free cash flows that firms do not actively seek to raise and may not need, one
might expect firms to pay these proceeds back to their investors.
In order to investigate to which extent our results are driven by such a possibility, Table 2,
Panel B, excludes employee-initiated equity issues from our analysis and focuses only on payout
8

All firm counts we report throughout the paper require variables to be greater than $100,000 to be considered
positive. In this section, this ensures that we identify firms that pay out and raise economically meaningful amounts
of capital and thus do not pick up rounding errors.

11

payers that actively initiate security issues in the same year. This analysis provides an upper
bound for the degree to which payouts are actively financedi.e., the degree to which firms
initiate securities issues in order to finance their payouts. Column 1 in Panel B shows that, in the
average sample year, 20% of all public firms simultaneously pay out and actively raise capital by
initiating a net debt or an equity issue. This means that an average of 42% of all payers actively
raise capital during the same year (Column 2); conversely, 46% of all firms that actively raise
capital also pay out capital during the same year (Column 3). Thus, a comparison of Panels A
and B indicates that employee-initiated equity issues make up just under a half of the instances of
firms that simultaneously pay out capital and issue securities.
Figure 3 (top graph) shows that the percentage of public firms that simultaneously pay out
and actively raise capital is highly pro-cyclical:9 It fell from a peak of 26% in 1998 to 13% in
2002 (in the aftermath of the 2001 recession), and then again during the Great Recession, from
22% in 2007 to 12% in 2009, before bouncing back to 25% in 2012. Therefore, it is in years
when capital is arguably easiest to raise that firms choose to actively issue securities and
simultaneously pay out (at least part of) the proceeds. The dotted and dashed lines show similar
time trends when plotting the percentage of firms with simultaneous payouts and firm-initiated
security issues relative to the total number of firms paying out capital and initiating security
issues, respectively.
A natural question follows: Are simultaneous payouts and security issues economically
important, or are the above results driven by small firms, and therefore the dollar magnitudes are
negligible? Columns 4 through 9 of Table 2 investigate this question. Panel A examines the
dollar amount that firms simultaneously raise and pay out during the same fiscal year, measured
as follows: For each firm-year, we calculate the minimum of the proceeds of its security issues

The underlying data are the same as in Column 1 of Table 2, Panel B, but in the figure we can plot each annual
percentage instead of averaged over four-year intervals, as we do in Table 2 to conserve space.

12

(net debt and equity, denoted SIit) and its total payout (TPit): min{SIit, TPit}. For example, for the
firm in our previous example raising $100 of net debt, paying out $50, and with $50 of free cash
flow, we define the amount that the firm simultaneously raises and pays out to be min{100, 50}
= $50. (Note that this firm increases its cash holdings by $100.)
As explained above, by measuring the capital that firms simultaneously raise and pay out, we
obtain an upper bound of the extent to which firms rely on the proceeds of security issues to
finance their payouts. The firm in our example could in fact have paid out $50 without raising
any capital, and so one can argue that the debt issuance decision was not related to the payout
decision. That said, had the firm not simultaneously issued debt and paid out $50, it could have
reduced the debt it raised by this amount while still achieving the same cash increase.
To get a sense of how large simultaneous payouts and security issues are relative to the total
amount of capital firms pay out and the amount they raise, we construct the following two ratios
for each year in our sample period: RtTP
RtSI

min SI

iPubt

it

, TPit /

iPubt

iPubt

min SI it , TPit /

TP

iPubt

it

and

SI it , where Pubt denotes the set of all public firms in our sample

in year t. RtTP captures the fraction of aggregate total payouts that is simultaneously raised
through securities issued by the same payers during the same year, while RtSI captures the
fraction of aggregate proceeds of security issues that is paid out by the same issuers during the
same year.10
Column 4 in Table 2, Panel A, shows how the annual ratio RtTP (averaged over four-year
intervals to conserve space, as before) has evolved over our sample period: On average, 41% of
the aggregate capital paid out by public U.S. firms was raised during the same year by the same
10

Alternatively, we could construct the ratio min{SIit , TPit}/ TPit for each payout payer, and then average this ratio
across all firms paying out capital in any given year (and analogously with the ratio min{SIit , TPit}/ SIit for security
issuers). Doing this yields similar patterns to those we report using RtTP and RtSI , respectively.

13

firms that made the payouts. This ratio has decreased somewhat in recent years, indicating that
simultaneous payouts and security issues have not kept pace with the payout explosion captured
in Figure 1. This is despite the fact that Column 5 shows that the fraction of aggregate proceeds
of security issues that are simultaneously paid out, RtSI , has been growing over time, averaging
39% over our sample and reaching 57% in 2011.
Columns 4 and 5 in Panel B show that the large dollar magnitude of simultaneous payouts
and security issues is not primarily the result of firms paying out the proceeds of employee stock
options exercises. Columns 4 and 5 in Panel B show the analogous versions of the RtTP and RtSI
ratios with SI substituted by AI, where AI includes only capital actively raised through either net
debt issues or firm-initiated equity issues. Remarkably, we still find that, on average over our
sample period, close to a third (32%) of the aggregate capital paid out by public U.S. firms is
financed through firm-initiated security issues during the same year. The results in Column 5 are
equally if not more surprising: A staggering 34% of the aggregate proceeds of active issuances
are paid out by the same firms during the same fiscal year, a ratio that has increased markedly
since 2003. The bottom graph in Figure 3 shows that simultaneous payouts and firm-initiated
security issues are strongly pro-cyclical, with the aggregate annual sum of simultaneous payouts
and active issues,

min AI

iPubt

it

, TPit , peaking in 2007 at $227 billion.

Taken together, our findings paint a very different picture from the commonly held view that
firms that (actively) raise capital and those that pay out capital are different firms that are at
different stages of their lifecycles and face different growth opportunities (e.g., Grullon,
Michaely, and Swaminathan, 2002; DeAngelo, DeAngelo, and Stulz, 2004). Rather, we have
shown that, on average over our sample period, over 82% of firms that pay out capital also issue
securities in the same year they pay out. Even if we focus only on firm-initiated security issues,
this ratio remains a very substantial 42%. Perhaps even more remarkable, the amount of capital
14

that is actively raised and paid out by these firms represents 32% of the aggregate total payout;
conversely, 34% of the proceeds of security issues initiated by public U.S. firms are paid out to
shareholders during the same year.
We now investigate whether the results are primarily driven by one particular form of
payout. In particular, one might expect that because it is costlier to cut dividends than
repurchases (e.g., Brav et al., 2005), firms might be particularly keen on maintaining dividend
payments, even if doing so requires raising capital. Under this hypothesis, dividends that are
financed through security issues should be more prevalent than financed repurchases. To see
whether this is the case, Columns 6 and 7 in Table 2 perform the same analysis as Columns 4 and
5 but substituting total payouts (TP) with dividend payout (Div); similarly, Columns 8 and 9
substitute total payouts with capital paid out via share repurchases (Rep).
The results in Panel A indicate that firms pay out the proceeds of security issues via
dividends and share repurchases to a similar extent (23% and 24%, respectively).11 When we
restrict the sample to firm-initiated security issues (Panel B), we find very similar results: 21% of
the proceeds of firm-initiated security issues are paid out via dividends and also 21% via
repurchases. If anything, a comparison of Columns 7 and 9 indicates that in recent years, a larger
share of the proceeds of firm-initiated issues has been paid out via share repurchases than via
dividends. The only exception is 2009 (not captured in the table), when 30% of the proceeds of
firm-initiated security issues were paid out via dividends and only 17% via repurchases. (Figure
1 shows that repurchase volume collapsed during the Great Recession.)
Therefore, it appears that a desire to maintain dividends can be the primary motive for
actively financed payouts only in recessionary years (when raising capital is likely costliest). On
the other hand, in expansionary years, particularly in the later years of our sample, over a third of
11

The sum of Columns 7 and 9 does not equal Column 5. To illustrate why, consider the case of a firm that raises
$80 of debt, pays out $50 in dividends and another $50 via share repurchases. For this firm, min{SI, TP} = $80 <
min{SI, Div} + min{SI, Rep} = $50 + $50 = $100.

15

the capital that firms actively raise is simultaneously paid out via share repurchases. To the
extent that repurchases are typically seen as less sticky than dividends, maintaining repurchase
levels is unlikely to be the motivation for such financed repurchases.
Our findings show that simultaneous payouts and firm-initiated security issues represent a
large fraction of both payout and capital raising activity. The implications of these findings may
critically depend on the type of security issues: For example, as we discuss in Section 4, whether
firms finance their payouts via equity or debt issues has very different capital structure
implications and thus would point to different motives for why firms raise and pay out capital at
the same time. The next section breaks down the role that debt and equity issues play in
explaining our findings.
2.2 Breaking down the role of debt and equity issues
Table 3 examines the extent to which firms simultaneously pay out capital and issue net debt
(Panel A), firm-initiated equity (Panel B), and employee-initiated equity (Panel C), following the
same structure as Table 2. Three results stand out.
First, debt appears to be the dominant form of payout financing: Column 4 indicates that up
to 30% of aggregate payouts are financed via net debt issues (Panel A), while firm- and
employee-initiated equity issues finance at most 3% (Panel B) and 11% (Panel C) of payouts,
respectively. Columns 6 and 8 show that debt dominates the financing of both dividends and
share repurchases.
Second, when examining firm counts, a somewhat different picture emerges: Column 1
shows that the percentage of public firms with a simultaneous payout and net debt issue is a
substantial 18% (Panel A). That said, this fraction is smaller than the 34% of firms with a
simultaneous payout and employee-initiated equity issue (Panel C), reflecting the large
prevalence of firms with capital inflows from employee stock option exercises. On the other

16

hand, simultaneous payouts and firm-initiated equity issues are rare: less than 4% of all public
firms initiate equity issues and pay out capital during the same year (Panel B).12
Third, relative to the total amount of capital firms raise, Column 5 shows that a remarkable
39% of net debt is paid out during the same year (Panel A); this represents $135 billion of the
$350 billion of net debt issued by public U.S. firms in our average sample year. Panel B shows
that an equally remarkable 19% of firm-initiated equity issuance proceeds are also paid out.
Therefore, while the fact that SEOs and private placements are relatively infrequent implies that
they finance only a small fraction of payouts, when firms do actively raise equity, they pay out
close to a fifth of the proceeds during the same year.
As for employee-initiated equity issues, as much as 79% of these proceeds are paid out
(Panel C). This finding is consistent with the notion that for most firms, the cash inflows from
option exercises represent unneeded capital that is a by-product of the firms compensation
policies, and so they pay the cash back to their shareholders. At the same time, our results are
also consistent with firms using employees as a source of capital for making payouts to
shareholders.
3. A lower bound for financed payouts: The payout-funding gap
Our analysis of simultaneous payouts and issuances in Section 2 provides an upper bound for
the extent to which firms rely on the capital markets to finance their payouts. In this section, we
provide a lower bound for the prevalence and dollar magnitude of financed payouts by focusing
only on the subset of simultaneous payers and issuers that could not have funded their payouts
without the proceeds of security issues.

12

The fact that some firms simultaneously raise and pay out equity has been noted by Weld (2008) and Grullon,
Paye, Underwood, and Weston (2011).

17

3.1 Prevalence and magnitude of payout-funding gaps

In order to identify firms that, ceteris paribus, would have been unable to fund their payouts
without simultaneously raising capital, we need to measure the gap between a firms payout and
its free cash flow. To define this gap, it is helpful to consider the following cash flow identity,
which expresses a firms total payout in terms of the firms potential sources and uses of cash:
Total payout (TPit) = Free cash flow (FCFit) Change cash (CCit) + Security issues (SIit)

(1)

Free cash flow (FCFit) is the sum of operating cash flow (OCFit) and investment cash flow
(ICFit).13 As in Section 2, total payout (TPit) is the sum of dividends and share repurchases, while
security issues (SIit) is the sum of the proceeds of a firms debt and equity issues. In addition, we
define a firms cash reduction as minus its change in cash holdings if the change in cash is
negative, and zero otherwise (i.e., CRit = min{CCit, 0} 0).
It then immediately follows from equation (1) that whenever a firms total payout exceeds
the sum of its free cash flow and cash reduction (i.e., TPit > FCFit + CRit), the firm needs to
issue securities to finance (part of) its payout (i.e., SIit > 0). We label such a firm as having a
payout-funding gap, and define its gap as follows:
Payout gap (PGit) min{max{TPit (FCFit + CRit), 0}, TPit}

(2)

To illustrate our definition of payout gap, consider the following three situations. First, a firm
that pays out $50, has free cash flow of $25 and no changes in cash, has a $25 payout gap.
Second, a firm that pays out $50, has negative free cash flow of $100 and no changes in cash,
has a $50 payout gap. That is, our definition ensures that a firms payout gap is never larger than
the payout itself, even if free cash flow is negative.
Lastly, consider a firm like the one in our example in Section 2 that pays out $50, has free
cash flow of $50, and issues $100 worth of net debt, which it uses to build up its cash reserves

13

The two main components of investment cash flow are capital expenditures and acquisitions, both of which enter
the definition of investment cash flow with a negative sign as they represent capital outlays.

18

(i.e., SIit = CCit = $100). This firm pays out and raises capital during the same year, and hence
was captured as such in Section 2 (min{TPit, SIit} = $50). By contrast, according to the payout
gap definition introduced in this section, this firm does not have a gap because its free cash flow
is sufficient to fund its payout, i.e., PGit = 0.
This last example illustrates the complementary nature of our analysis in Sections 2 and 3:
Section 2 gave us an upper bound of the extent to which payouts are financed because by
measuring financed payouts as min{TP, SI}, we effectively assumed that the proceeds of security
issues were the source of capital of first resort when funding payouts. On the other hand, our
definition of payout gap in equation (2) provides a lower bound for financed payouts by
assuming that security issues are the source of capital of last resort for payouts, only being used
after firms exhaust their free cash flow and cash reductions. This allows us to address the
conceptual problem posed by the fungibility of money without making further assumptions.
The first four columns of Table 4, Panel A, examine the prevalence of firms with a payout
gap and the dollar magnitude of their gaps. Columns 1 and 2 show that, in our average sample
year, 22% of all public firmsrepresenting 46% of all firms that pay out capitalhave a payoutfunding gap.14 Column 3 shows that, across firms with a payout gap, the ratio of the payout gap
to total payout, PGit / TPit, averages 73%; therefore, conditional on having a gap, payout gaps
are large. Importantly, the magnitude of payout gaps is also substantial at the aggregate level:
Column 4 shows that the annual ratio of the aggregate sum of payout gaps over the aggregate
capital paid out by all public firms,

iPubt

PGit /

TP

iPubt

it

, averages 30%. This means that, in our

average sample year, $98 billion or 30% of the total capital paid out by public U.S. firms could
not have been funded without the proceeds of security issues. Therefore, our lower bound for

14

Analogously as in the previous tables, we require PGit > $100,000 to classify a firm as having a payout gap.

19

financed payouts indicates that no less than 46% of all payers finance their payouts, with the
dollar amount of financed payouts representing at least 30% of aggregate total payouts.
How do our findings change if we examine only payouts that, ceteris paribus, could not have
been funded without the proceeds of firm-initiated security issues? To investigate this, Columns
5 through 8 in Table 4, Panel A, perform an analysis analogous to that in Columns 1 through 4
but focusing on what we call active payout gaps, defined as follows:
Active payout gap (APGit) min{max{TPit (FCFit + CRit + EEit), 0}, TPit}

(3)

where EEit captures the proceeds of employee-initiated equity issues and all other variables are
as in equation (2). By adding the proceeds of employee-initiated issues to free cash flow and
cash reductions, this definition allows us to obtain a lower bound for the prevalence and
magnitude of actively financed payouts. 15
Columns 5 and 6 show that a remarkable 36% of all payersor 17% of all public firms set
a payout level that forces them to initiate a debt or an equity issue to fund their payout. Column 8
shows that the aggregate magnitude of active payout gaps represents on average 26% of the
aggregate capital annually paid out by public U.S. firms.
Figure 4 shows that both the percentage of firms with an active payout gap and the aggregate
dollar magnitude of these gaps are highly pro-cyclical, declining markedly around the three
recessionary periods in our sample: the early 1990s recession, the early 2000s recession, and the
Great Recession. This suggests that in recessionary years, when it is hardest and costliest to raise
capital (e.g., Campello, Graham, and Harvey, 2010; Erel, Julio, Kim, and Weisbach, 2012), firms
are reluctant to set payout levels that they need to finance by initiating security issues. Yet they
seem to have no problem doing so when the economy appears to be performing well: In 2007,

15

Indeed, it follows from equation (1) that whenever TPit > FCFit + CRit + EEit, then AIit > 0 (where as in Section 2
AIit captures capital that is actively raised through either net debt or firm-initiated equity issues).

20

right before the onset of the Great Recession, close to 40% of all payers had active payout gaps;
they had to raise a combined $190 billion through firm-initiated issues to close their gaps.
We can combine Tables 2 and 4 to determine average ranges for the prevalence and dollar
magnitude of both financed and actively financed payouts. The summary table below does this:
Firm counts
% of payout payers

Dollar magnitudes
% of aggregate total payouts

Lower bound
(1)

Upper bound
(2)

Lower bound
(3)

Upper bound
(4)

Financed payouts

46.0%

82.4%

30.1%

40.6%

Actively financed payouts

36.1%

42.1%

26.1%

31.8%

Summary table. This table combines the results of Table 2 (Columns 2 and 4; Panels A and B) and Table 4
(Columns 2, 4, 6, and 8; Panel A) to provide lower and upper bounds for the prevalence and dollar magnitude of
financed and actively financed payouts. All figures are annual figures averaged over all years in our sample period.

The widest range between the lower and upper bounds obtained in Sections 2 and 3 is in the
percentage of payers that finance their payouts: Between 46% and 82% of payers can be said to
use the proceeds of security issues to finance their payouts, depending on whether we assume
that these proceeds are the source of capital of last or first resort for funding payouts,
respectively. This wide range reflects the fact that among the 82% of payers that simultaneously
issue securities there is a large number of firms with only employee-initiated equity issues that
do not have a payout gap. If we focus on the dollar magnitude of financed payouts, we no longer
find the same wide range between the corresponding lower and upper bounds: Firms finance
between 30% and 41% of aggregate total payouts. This narrower range is consistent with the
notion that the employee-initiated issues that are responsible for the wide range in the percentage
of payers with financed payouts tend to be small in dollar magnitude.
On the other hand, if we focus on actively financed payouts, we find narrow ranges both
when examining firm counts and dollar magnitudes: Between 36% and 42% of payers use the
proceeds of firm-initiated security issues to actively finance their payouts; as a result, between
21

26% and 32% of aggregate total payouts are actively financed. These narrow ranges indicate that
most payers that simultaneously initiate securities issues do so because they need the proceeds of
these issues to close their payout-funding gaps.
Figure 5 combines Figures 3 and 4 to examine how both the percentage of payers with
actively financed payouts and the dollar magnitude of their financed payouts have changed over
time. The figure confirms that actively financed payout gaps are strongly pro-cyclical, consistent
with our prior interpretation that it is primarily in expansionary years when firms set payout
levels that they need to finance by actively raising capital. This finding is consistent with the
evidence in Bliss, Cheng, and Denis (2014) that exogenous shocks to the supply of credit lead
firms to reduce their payouts as a substitute form of financing. Indeed, our results show that
reducing payouts allows firms not just to conserve internal capital, but also to reduce their
external capital needs.
3.2 Are payout gaps the result of timing mismatches between free cash flow and payouts?
The large prevalence and dollar magnitude of both financed and actively financed payouts
raise an important question: Are the gaps on which we base our lower bounds for financed
payouts the result of time mismatches between free cash flow and payouts? This concern is
particularly relevant given that it has long been known that firms tend to smooth their payouts
relative to their free cash flow (Lintner, 1956). In particular, if firms set their payout equal to
their average free cash flow, our analysis will identify a payout gap every year that a firm has
below average free cash flow. Crucially, if time mismatches were driving the gaps we have
identified in Section 3.1, the gaps should all but disappear if we measure firms sources and uses
of cash over longer horizons.
We test this hypothesis in Panel B of Table 4, where we define payout-funding gaps over
four-year intervals:

22

3
3
PGit4 year min max TPit j FCFit j CRit j , 0 , TPit j
j 0

j 0
j 0

(4)

and analogously for active payout gaps. Interestingly, both payout gaps and active payout gaps
are somewhat more prevalent and of similar aggregate dollar magnitude when we define them
over four-year intervals than when we define them annually in Panel A.16 Therefore, we find no
support for the notion that our findings are driven by time mismatches between firms free cash
flow and payouts.
4. Why do firms choose to finance their payouts?
The summary table in Section 3 indicates that, in the average sample year, between 36% and
42% of payers engage in active payout financing, relying on the proceeds of firm-initiated
security issues to fund their payouts. These actively financed payouts are not just widespread but
they are also economically important, representing between 26% and 32% of the aggregate
capital paid out by public U.S. firms.17
Our goal in this section is to shed light on the motives driving the unexpectedly large
magnitude of financed payouts. This magnitude runs counter to the common notion that payouts
are first and foremost a vehicle to return free cash flow to investors (e.g., Grullon, Michaely, and
Swaminathan, 2002)whether it is to ensure that the cash is not misspent or to signal the
payers quality.
Section 4.1 starts by briefly discussing why actively financed payouts are costly; these costs
are well-known and underscore the surprising nature of our findings. Section 4.2 then turns to
examining the potential benefits of financing payouts and discusses the extent to which these
benefits are likely to explain the striking prevalence of financed payouts observed in the data.
16

This finding is in fact not surprising given that we expect firms that smooth their payout relative to their free cash
flow to use their cash holdings to do so. Our baseline definitions of payout gap (equations (2) and (3)) add cash
reductions to free cash flow, and so Panel A does not identify such temporal smoothing behavior as having a gap.
17
An additional 4% to 9% of aggregate payouts are financed through employee-initiated equity issues, which play a
perhaps less surprising but nevertheless significant role in the funding of payouts.

23

4.1 Costs of actively financing payouts

Our discussion of the costs of actively financed payouts focuses on three key costs: the
wedge between a firms internal and external costs of funds, the notion that actively financed
payouts can crowd out profitable investments, and tax-related costs.
4.1.1 Wedge between the cost of internal and external capital
Perhaps the most notable reason why actively financing payouts is costly stems from the fact
that (virtually) all firms face a wedge between their internal and external costs of funds. Indeed,
any small transaction cost of raising external funds implies that external capital is more costly
than internal capital (see, e.g., Kaplan and Zingales, 1997). In fact, information asymmetry
between firms and investors implies that the wedge between a firms cost of internal and external
capital can be much higher than is implied by pure transaction costs (e.g., Myers and Majluf,
1984). In particular, informational frictions can imply that the cost of new debt and equity may
differ substantially from the opportunity cost of internal finance generated through cash flow and
retained earnings (Fazzari, Hubbard, and Petersen, 1988, p. 142). It is thus not surprising that
Miller and Rock (1985) point out that it would be uneconomic as well as pointless for firms to
pay dividends and raise capital simultaneously (as quoted by Easterbrook, 1984, p. 650).
4.1.2 Underinvestment
To the extent that firms ability to raise capital is not unlimited, using actively raised funds to
make payouts can imply that the firms have to pass up profitable investment opportunities. In
line with this notion, Bens, Nagar, and Wong (2002) show that firms that shift resources away
from real investments towards the repurchase of their own shares appear to experience a
subsequent decline in performance. In recent work, Asker, Farre-Mensa, and Ljungqvist (2014)
show that public U.S. firms invest less and are less responsive to investment opportunities than
their privately held counterparts. At the same time, they show that public firms have higher

24

payouts, a finding that they note is consistent with public firms inefficiently prioritizing high
payouts over profitable investments due to short-termist pressures.
4.1.3 Taxes
Our analysis in Tables 2 and 3 indicates that firms finance both their dividends and their
share repurchases. But because dividends are taxed, if a firm raises $1 of capital from investors
and immediately pays it out as a dividend, its shareholders will receive $(1 ), where is the
dividend tax. This creates an incremental cost for those firms that actively finance their
dividends that needs to be added to the costs discussed above.
4.2 Benefits of actively financing payouts
Why do well over a third of all payers choose to finance their payouts, despite the seemingly
large costs associated with doing so? In this section, we discuss four potential benefits of
financing payouts: adjusting a firms capital structure, mitigating the agency costs of free cash
flow, signaling the firms quality, and increasing earnings-per-share.18 Our results suggest that
three of these four benefits play an important role in explaining the widespread payout-financing
behavior we observe in the data (the exception is signaling).
While our discussion of the potential benefits of financing payouts shows that this behavior is
consistent with some classic payout theories, financed payouts have generally received little
attention in the more recent literature (e.g., see the surveys by DeAngelo et al. (2008) and FarreMensa et al. (2014)).19 The striking prevalence of financed payouts suggests that the benefits of
financing payouts might be more important than it is often assumedwhile, at the same time,
the costs associated with financing payouts might be less important.

18

Our discussion of the benefits of actively financed payouts does not aim to be exhaustive. A fifth potential benefit
that we do not discuss is market timing, which might prompt firms to raise debt when they believe their equity is
undervalued and use the issuance proceeds to repurchase their shares.
19
The only exception are leveraged recapitalizations, but these are rare (Denis and McKeon, 2012) and thus do not
drive our findings.

25

4.2.1 Capital structure adjustments

A key potential benefit of actively financed payouts, specifically in the case of debt-financed
payouts, is that they allow firms to increase their leverage. These capital structure adjustments
can be motivated by at least three reasons. First, as pointed out by Jensen (1986), debt creation,
without retention of the proceeds of the issue, enables managers to effectively bond their promise
to pay out future cash flows. () Issuing large amounts of debt to buy back stock also sets up the
required organizational incentives to motivate managers and to help them overcome normal
organizational resistance to retrenchment which the payout of free cash flow often requires (p.
324). Second, as noted by Easterbrook (1984), financing projects out of retained earningsif
unanticipated by bondholderstransfers wealth from shareholders to debtholders by decreasing
the risk of the already-issued debt (p. 653). Debt-financed payouts can reverse these leverage
decreases, while still preserving the cash the firm needs to fund its investments. Third, debtfinanced payouts can also be motivated by a desire to increase leverage in order to optimize the
tax-efficiency of a firms capital structure without depleting its cash reserves (as would happen if
the firm paid out its internal funds). Consistent with this argument, Heider and Ljungqvist (2014)
show that firms affected by state corporate income tax rises increase their debt holdings without
changing their asset base, implying that firms swap debt for equity by paying out the debt
issuance proceeds.20
Our analysis in Table 3 shows that the vast majority of actively-financed payouts are
financed through debt issues rather than via firm-initiated equity issues: In our average sample
year, 90% of firms that simultaneously pay out and actively raise capital do so by issuing net
debt. In terms of dollar magnitudes, the difference is even more stark: 95% of the capital that is

20

In addition to allowing firms to adjust their capital structure without using internal funds that the firms may need
to invest or may wish to conserve for precautionary reasons, there is an additional reason why some firms prefer to
finance their payouts by issuing debt. This is the case of firms whose cash is overseas and would be subject to
repatriation taxes if it were brought back to the U.S. to fund the firms payouts.

26

paid out and actively raised during the same year by the same firms is raised via net debt issues.
This is consistent with the notion that capital structure adjustments motivated by agency, riskrebalancing, or tax considerations play a key role in explaining the striking prevalence of
actively financed payouts we observe in the data.
4.2.2 Mitigating agency costs of free cash flow
Recall that Table 3, Panel B, shows that as much as 19% of the proceeds of firm-initiated
equity issues are paid out by the same issuing firms during the same year. Needless to say, these
equity-financed payouts have no effect on a firms capital structure, and so they cannot be
motivated by the capital structure adjustments discussed in the prior section. How can we explain
this behavior?
Easterbrook (1984) notes that payouts can mitigate agency problems by compelling firms to
raise new money in order to carry out their activities (p. 655), thereby subjecting managerial
decisions to the scrutiny of the capital markets. Boards can exercise this monitoring by setting
payouts at a level such that firms pay out all their operating cash flow, with the only exception of
those funds required for routine, maintenance-type investments. By doing so, they force
managers to raise capital to fund any new investment projects (even those that could have been
funded using the firms internal funds had these funds not been paid out), thus allowing investors
to decide whether the projects should be funded. This monitoring strategy naturally results in
equity-financed payouts. In fact, to the extent that the monitoring can also be exercised by debt
investors, this strategy can also motivate debt-financed payouts.
To what extent are the actively financed payouts we observe in the data driven by firms with
unusually high levels of investment (and thus unusually low levels of free cash flow) that are
likely undertaking new investment projects? To answer this question, we need to define
unusually high levels of investment by establishing a benchmark for the expected level of
routine investment. In Table 5, we use two different approaches to capture a firms expected
27

investment. This allows us to examine how the prevalence and economic magnitude of active
payout gaps (our lower bound for actively financed payouts) change if we assume that no firm
invests more than its expected level.
First, in Columns 1 through 4, we follow Daniel, Denis, and Naveen (2010) and measure a
firms expected investment as the median ratio of (CAPEX + acquisitions)/ lagged assets of the
firms industry in that year multiplied by the firms lagged assets. We can then define a firms
industry-counterfactual active payout gap, APGitIndI , analogously as in equation (3) but
substituting the firms free cash flow by the free cash flow it would have if it invested no more
than the median firm in its industry.21 Second, Columns 5 through 8 report an analogous analysis
but in this case our counterfactual gap measure, APGitLagI , assumes that no firm invests more
than it did in the previous year. (Analogously as with APGitIndI , we do allow firms to invest less
than they did in the previous year.) Finally, for ease of comparison, Columns 9 through 12 of
Table 5 reproduce our active payout gap baseline results from Table 4, which use a firms actual
level of investment.
A comparison of the average percentage of payers with an industry-counterfactual gap
(Column 2) to the actual percentage (Column 10) reveals that 61% of payers with an active gap
still have a gap when assuming that they invest no more than their industry median (22.1/36.1 =
61.3). If instead we compare Columns 6 and 10, the percentage of firms that still have a gap
when we assume that no firm invests more than it did the prior year is somewhat larger, at 71%.
In terms of dollar magnitudes, the aggregate sum of counterfactual gaps represent 60% (Column
4; industry-median counterfactual) and 69% (Column 8; prior-year counterfactual) of the actual
dollar magnitude of active payout gaps (Column 12).

21

If a firm invests less than the median firm in its industry, no substitution is made and active payout gap is defined
using the firms actual investment rather than the higher investment level of its industry peers. This ensures that a
firms counterfactual gap is always smaller than its actual gap.

28

Taken together, the results in Table 5 indicate that approximately 30% to 40% of the active
payout gaps identified in Table 4 are instances of firms with unusually high investment; these
firms would have no gap if they had just routine investment levels. Conversely, had these firms
paid out no capital, they would now need to raise less (or no) capital to fund their investment
surge. However, their boards chose to set payout levels that ensured that they had to raise capital
for any non-routine investment projects rather than being able to use their internal funds to fund
these projects. This allows the boards to mitigate the agency costs of free cash flow.
We can perform an analogous analysis to the one in Table 5 but assuming that no firm
invests less than the median firm in its industry or than it did in the previous year. Table IA.1 in
the Internet Appendix shows the results of this analysis. In this case, the counterfactual
prevalence of active payout gaps increases by between 12 and 14 percentage points, respectively,
relative to its actual value in Table 4. Although these results are only suggestive, they indicate
that between 12% and 14% of payers might be underinvesting in order to avoid having to
actively raise capital to fund their payouts. Importantly, this does not include those payers whose
payout level does force them to actively raise capital and by doing so are shifting away resources
from profitable investments (see Section 4.1.2).
4.2.3 Signaling
Firms might also benefit from actively financing their payouts due to signaling motives. In a
number of payout models, payouts (typically dividends) are used to signal a firms quality. Our
goal here is not to discuss all these models; rather, we will examine the extent to which actively
financed payouts arise as a natural prediction of two of the most well-known payout signaling
models: Bhattacharya (1979) and Miller and Rock (1985).
In Bhattacharyas (1979) model, a firm should be able to meet its dividend commitment
without recourse to extra, unanticipated new financing; however, in states of the world in
which the firm suffers a negative cash flow shock, the dividend is still paid, and it is assumed
29

that making up the shortfall results in [financing] costs to current shareholders (p.262).
Therefore, in Bhattacharyas model, financed dividends can arise in the (low probability) event
that a firm is hit by a negative cash flow shock. (To the extent that dividends represent a stronger
commitment than share repurchases (Jagannathan et al., 2000), this model has a harder time
explaining financed repurchases. Indeed, any announced repurchase plans could be cancelled in
the face of a negative shock with little signaling cost.)
Table 6 examines the extent to which actively financed dividends can be explained as
instances of firms that decide to maintain their dividends in the face of negative cash flow
shocks, as in Bhattacharyas model. To do so, we start by defining a firms active dividend gap,
ADGit, similarly as we define its active payout gap, substituting the firms total payout in
equation (3) by its dividend payout. This allows us to identify those firms for which the sum of
their free cash flow, cash reduction, and proceeds of employee-initiated equity issues are not
enough to fund their dividend, and thus they have to actively raise capital to close their dividend
gapin addition to having to close any additional gap created by share repurchases.
Columns 1 through 4 show the prevalence and dollar magnitude of active dividend gaps
throughout our sample period. In addition, Columns 5 through 8 explore how the prevalence and
magnitude of active dividend gaps change if we assume that no firm is less profitable than the
median firm in its industry, following an analogous approach to that used in Table 5. Similarly,
Columns 9 through 12 examine counterfactual active dividend gaps under the assumption that no
firm is less profitable than it was in the previous year.
Two results stand out from Table 6. First, a comparison of Columns 1 through 4 in Table 6 to
our baseline active payout gap results in Table 4, Panel A, reveals that active dividend gaps are
substantially less prevalent than active payout gaps: In our average sample year, less than 10% of
all public firms have an active dividend gap (Column 1 in Table 6) while over 17% of firms have
an active payout gap (Column 5 in Table 4, Panel A). In terms of dollar magnitudes, the
30

aggregate sum of active dividend gaps represents 11% of aggregate payouts (Column 3 in Table
6), while the aggregate sum of active payout gaps represents 26% of aggregate payouts (Column
8 in Table 4, Panel A). Therefore, and consistent with our evidence in Tables 2 and 3, it is not the
case that firms only finance their dividends. Instead, much of payout financing activity
corresponds to share repurchases and thus cannot be explained by Bhattacharyas signaling
modelor, more generally, by any theory that relies on the notion that cutting dividends is
costly due to signaling or other considerations.
Second, a comparison of the dividend gaps in Columns 1 through 4 to the counterfactual gaps
in Columns 5 through 12 indicates that 78% of firms with a gap would still have a gap if they
had been at least as profitable as their industry peers or as they themselves were in the previous
year. This suggests that most of the dividend gaps we observe in the data do not conform to the
prediction of Bhattacharyas model that dividend gaps are the result of firms deciding to
maintain their dividend after suffering a negative cash flow shock.
Actively financed payoutsparticularly over four-year intervals as we found in Table 4,
Panel Bare not a natural prediction of Miller and Rocks (1985) signaling model either.
Indeed, these authors write that because a signaling equilibrium is fully revealing, the model
here presented provides no support, of course, for a policy of sustaining dividends in the face of
earnings disasters (p. 1045). All in all, we find little support for the notion that an important
driving force behind actively financed payouts is firms desire to signal their quality to the
capital markets or, more generally, to avoid dividend cuts at all costs.
4.2.4 Increasing earnings-per-share
There is ample evidence in the literature that (at least some) public-firm managers are
concerned about meeting or beating earnings forecasts (see, e.g., Graham, Harvey, and Rajgopal
(2005) for survey evidence, and Bartov, Givoly, and Hayn (2002) and Bhojraj, Hribar, Picconi,
and McInnis (2009) for empirical evidence). Analysts forecasts are typically expressed in terms
31

of earnings-per-share (EPS), and the evidence in Bens, Nagar, Skinner, and Wong (2003)
indicates that one way how firms try to increase EPS is by repurchasing shares.22
To what extent are financed repurchases motivated by firms desire to increase EPS?
(Clearly, this motivation plays no role in the case of financed dividends.) Table 3, Panel C,
shows that firms pay out the vast majority of the proceeds of their employee-initiated equity
issues via share repurchases, particularly in recent years. This finding is consistent with the
notion that firms are concerned about the dilutive effect of these issues and aim to undo this
dilution by repurchasing shares. Importantly, though, the fact that stock options are exercised at
below a firms stock market price implies that using the proceeds of option exercises to
repurchase shares is generally not enough to undo the dilutive effect of the exercises. As a result,
firms need to fund part of their repurchases by issuing debt or by using internal funds if they
want to repurchase as many shares as options have been exercised.
Almeida, Fos, and Kronlund (2013) find that firms raise their debt levels in order to finance
so-called EPS-motivated repurchases. This suggests that the desire to increase EPS may play a
significant role in explaining the striking prevalence of debt-financed repurchases that we
observe in the data.
5. Conclusions
This paper is the first to systematically study the extent to which industrial public firms in the
U.S. rely on the proceeds of security issues to fund their payouts. Our analysis indicates that 42%
of payout payers simultaneously raise funds in the capital markets. At least 30% of aggregate
payouts are not funded with free cash flow but with security issues, and over a third of the
proceeds of firm-initiated debt and equity issues are paid out in the same year. We show that
financed payouts are pro-cyclical, take the form of both dividends and share repurchases, and are
persistent in timein fact, our findings remain virtually unchanged if we measure firms sources
22

This is not to say that all share repurchases are EPS increasing; see Bens et al. (2003) for details.

32

and uses of funds over four-intervals. Our results thus paint a very different picture from the
commonly held view that payouts are first and foremost a vehicle to return free cash flow to
investors (e.g., Grullon, Michaely, and Swaminathan, 2002), except in the rare instances when
firms carry out a leveraged recapitalization (Denis and McKeon, 2012).
While, at first glance, financed payouts may appear uneconomic as well as pointless
(Miller and Rock, 1985) and have received little attention in the literature (see the recent surveys
by DeAngelo et al. (2008) and Farre-Mensa et al. (2014)), we show that they are in fact
consistent with some classic payout theories. In particular, we find most support for the notion
that firms use financed payouts to minimize the agency costs of free cash flow (Easterbrook,
1984) and to gradually adjust their capital structures. By contrast, it appears unlikely that payout
financing is motivated by signaling considerations or, more generally, by a desire to avoid costly
dividend cutsexcept during recessions. The pervasiveness, economic magnitude, and
persistence of financed payouts suggest that the benefits associated with this behavior might be
more important than it is often assumed.
Our findings suggest that firms payout and issuance decisions are intrinsically related, and
thus much can be gained by studying them jointly as interdependent elements of the financial
ecosystem. For instance, how does the source of payout funding affect investors reaction to
payout announcements? And, conversely, how does payout financing behavior affect firms
access and cost of capital? We leave the exploration of these and related questions for future
research.

33

References
Allen, Franklin, and Michaely, Roni, 2003, Payout policy, in Handbook of the Economics of
Finance, eds. George Constantinides, Milton Harris, and Ren Stulz, Amsterdam: Elsevier.
Almeida, Heitor, Vyacheslav Fos, and Mathias Kronlund, 2013, The real effects of share
repurchases, Unpublished working paper, University of Illinois at Urbana-Champaign.
Andrade, Gregor, and Steven N. Kaplan, 1998, How costly is financial (not economic) distress?
Evidence from highly leveraged transactions that became distressed, Journal of Finance 53,
1443-1493.
Asker, John, Joan Farre-Mensa, and Alexander Ljungqvist, 2014, Corporate investment and
stock market listing: A puzzle?, Review of Financial Studies, forthcoming.
Bartov, Eli, Dan Givoly, and Carla Hayn, 2002, The rewards to meeting or beating earnings
expectations, Journal of Accounting and Economics 33, 173-204.
Bens, Daniel A., Venky Nagar, Douglas J. Skinner, and M. H. Wong, 2003, Employee stock
options, EPS dilution, and stock repurchases, Journal of Accounting and Economics 36, 5190.
Bens, Daniel A., Venky Nagar, and M. H. Wong, 2002, Real investment implications of
employee stock option exercises, Journal of Accounting Research 40, 359-393.
Bhattacharya, Sudipto, 1979, Imperfect information, dividend policy, and the bird in the hand
fallacy, Bell Journal of Economics 10, 259-270.
Bhojraj, Sanjeev, Paul Hrbar, Marc Picconi, and John McInnis, 2009, Making sense of cents: An
examination of firms that marginally miss or beat analyst forecasts, Journal of Finance 64,
2361-2388.
Bliss, Barbara, Yingmei Cheng, and David Denis, 2014, Corporate payout, cash retention, and
the supply of credit: Evidence from the 2008-09 Credit Crisis, Journal of Financial
Economics, forthcoming.
Brav, Alon, John Graham, Campbell Harvey, and Roni Michaely, 2005, Payout policy in the 21st
century, Journal of Financial Economics 77, 483-527
Campello, Murillo, John R. Graham, and Campbell R. Harvey, 2010, The real effects of financial
constraints: Evidence from a financial crisis, Journal of Financial Economics 97, 470-487.
Daniel, Naveen D., David J. Denis, and Lalitha Naveen, 2010, Sources of financial flexibility:
Evidence from cash flow shortfalls, Unpublished working paper, University of Pittsburgh.
DeAngelo, Harry, Linda DeAngelo, and Douglas J. Skinner, 2004, Are dividends disappearing?
Dividend concentration and the consolidation of earnings, Journal of Financial Economics
72, 425-456.
DeAngelo, Harry, Linda DeAngelo, and Douglas J. Skinner, 2008, Corporate payout policy,
Foundations and Trends in Finance 3, 95-287
DeAngelo, Harry, Linda DeAngelo, and Ren Stulz, 2006, Dividend policy and the
earned/contributed capital mix: A test of the life-cycle theory, Journal of Financial
Economics 81, 227-254
34

Denis, David J., and Diane K. Denis, 1993, Managerial discretion, organizational structure, and
corporate performance: A study of leveraged recapitalizations, Journal of Accounting and
Economics 16, 209-236.
Denis, David J., and Stephen B. McKeon, 2012, Debt financing and financial flexibility:
Evidence from pro-active leverage increases, Review of Financial Studies 25, 1897-1929.
Easterbrook, Frank H., 1984, Two agency-cost explanations of dividends, American Economic
Review 74, 650-659.
Erel, Isil, Brandon Julio, Woojin Kim, and Michael S. Weisbach, 2012, Macroeconomic
conditions and capital raising, Review of Financial Studies 25, 341-376.
Fama, Eugene F., and Kenneth R. French, 2001, Disappearing dividends: Changing firm
characteristics or lower propensity to pay? Journal of Financial Economics 60, 3-43.
Farre-Mensa, Joan, Roni Michaely, and Martin Schmalz, 2014, Payout policy, in Annual Review
of Financial Economics Vol. 6, eds. Andrew W. Lo and Robert C. Merton, Palo Alto: Annual
Reviews.
Fazzari, Steven R., Glenn Hubbard, and Bruce C. Petersen, 1988, Financing constraints and
corporate investment, Brookings Papers on Economic Activity 1, 141-195.
Foley, C. Fritz, Jay C. Hartzell, Sheridan Titman, and Garry Twite, 2007, Why Do Firms Hold So
Much Cash? A tax-based explanation, Journal of Financial Economics 86, 579-607.

Graham, John R., Campbell R. Harvey, and Shiva Rajgopal, 2005, The economic implications of
corporate financial reporting, Journal of Accounting and Economics 40, 3-73
Grullon, Gustavo, Roni Michaely, and Bhaskaran Swaminathan, 2002, Are dividend changes a
sign of firm maturity?, Journal of Business 75, 387-424.
Grullon, Gustavo, Bradley Paye, Shane Underwood, and James Weston, 2011, Has the
propensity to pay out declined?, Journal of Financial and Quantitative Analysis 46, 1-24.
Heider, Florian, and Alexander Ljungqvist, 2014, As certain as debt and taxes: Estimating the tax
sensitivity of leverage from state tax changes, Journal of Financial Economics, forthcoming.
Jagannathan, Murali, Clifford P. Stephens, and Michael S. Weisbach, 2000, Financial flexibility
and the choice between dividends and stock repurchases, Journal of Financial Economics 57,
355-384.
Jensen, Michael C., 1986, Agency costs of free cash flow, corporate finance and takeovers,
American Economic Review 76, 323-329.
Kaplan, Steven N., and Luigi Zingales, 1997, Do investment-cash flow sensitivities provide
useful measures of financing constraints?, Quarterly Journal of Economics 112, 169-215.
Kim, Woojin, and Michael S. Weisbach, 2008, Motivations for public equity offers: An
international perspective, Journal of Financial Economics 87, 281-307.
Leary, Mark T., and Roni Michaely, 2011, Determinants of dividend smoothing: Empirical
evidence, Review of Financial Studies 24, 3197-3249.
Lintner, John, 1956, Distribution of incomes of corporations among dividends, retained earnings,
and taxes, American Economic Review 46, 97-113
35

McKeon, Stephen, 2013, Firm-initiated versus investor-initiated equity issues, Unpublished

working paper, University of Oregon.
Miller, Merton H., and Kevin Rock, 1985, Dividend policy under asymmetric information,
Journal of Finance 40, 1031-1051.
Modigliani, F., Miller, M., 1963. Corporate income taxes and the cost of capital: a correction.
American Economic Review 53, 433-443.
Myers, Stewart C., and Nicholas S. Majluf, 1984, Corporate financing and investment decisions
when firms have information that investors do not have, Journal of Financial Economics 13,
187-221.
Ross, Stephen A., Randolph W. Westerfield, and Jeffrey Jaffe, 2013, Corporate Finance, 10th
ed., New York: McGraw-Hill/Irwin.
Weld, William C., 2008, Recycled Equity, Unpublished working paper, Cornell University.
Wruck, Karen H., 1994, Financial policy, internal control, and performance: Sealed Air
Corporations leveraged special dividend, Journal of Financial Economics 36, 157-192.

36

Figure 1. Aggregate payout activity.

For each year from 1989 to 2012, the top graph shows the percentage of all public U.S. firms that are payout payers
(i.e., pay dividends or repurchase shares), pay a dividend, or repurchase shares during that year. The bottom graph
shows the aggregate total payout (the sum of dividends and share repurchases) paid by all public U.S. firms each
year, as well as the aggregate dividends and share repurchases. Dollar magnitudes are in billions of dollars of 2012
purchasing power.
Firm counts (% of all public firms)
80%

60%

40%

20%

0%
1989

1991

1993

1995

1997

Payout payers

1999

2001

2003

2005

Dividend payers

2007

2009

2011

Share repurchasers

Dollar magnitudes (billions of 2012 $)

900
800
700
600
500
400
300
200
100
0
1989

1991

1993

1995

1997

Total payout

1999

2001

Dividends

2003

2005

2007

2009

2011

Share repurchases

37

Figure 2. Aggregate capital raising activity.

For each year from 1989 to 2012, the top graph shows the percentage of firms with net debt issues (debt issues net of
debt repurchases), firm-initiated equity issues, and employee-initiated equity issues. The bottom graph shows the
aggregate dollar amount raised via net debt issues, firm-initiated equity issues, and employee-initiated equity issues
by all public U.S. firms each year. Dollar magnitudes are in billions of dollars of 2012 purchasing power.
Firm counts (% of all public firms)
100%
80%
60%
40%
20%
0%
1989

1991

1993

1995

Net debt issuers

1997

1999

2001

2003

Firm-initiated equity issuers

2005

2007

2009

2011

Employee-initiated equity issuers

Dollar magnitudes (billions of 2012 $)

500
450
400
350
300
250
200
150
100
50
0
1989

1991

1993

Net debt issues

1995

1997

1999

2001

Firm-initiated equity issues

2003

2005

2007

2009

2011

Employee-initiated equity issues

38

Figure 3. Simultaneous payouts and firm-initiated security issues.

For each year from 1989 to 2012, the solid line in the top graph plots the percentage of all public firms that pay out
capital and initiate security issues during the same year; the dotted line plots the percentage of all payout payers that
initiate security issues during the same year; and the dashed line plots the percentage of all firms initiating security
issues that pay out capital during the same year. The bottom graph plots, for each year t, the aggregate magnitude of
the simultaneous payouts and firm-initiated security issues of all public U.S. firms (i.e., the sum of min{AIit, TPit}
across all public firms, where AI captures the proceeds of firm-initiated security issues and TP is total payout).
Dollar magnitudes are in billions of dollars of 2012 purchasing power.
Simultaneous payouts and firm-initiated issues (firm counts)
80%

60%

40%

20%

0%
1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

% of all public firms that pay out capital and initiate securities issues
% of all payout payers that initiate securities issues
% of all firms initiating securities issues that pay out capital

Simultaneous payouts and firm-initiated issues (dollar magnitude in billions of 2012 $)

250

200

150

100

50

0
1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

Aggregate sum of simultaneous payouts and firm-initiated securities issues

39

Figure 4. Can firms fund their payouts without actively raising capital?
For each year from 1989 to 2012, the solid line in the top graph plots the percentage of all public firms that have an
active payout gap, i.e., firms with APGit min{max{TPit (FCFit + CRit + EEit), 0}, TPit} > $100,000; these firms
require the proceeds of firm-initiated security issues to fund their payouts. (TP is total payout; FCF is free cash flow,
the sum of operating cash flow and investment cash flow; CR is cash reduction; and EE captures the proceeds of
employee-initiated equity issues.) The dotted line in the top graph plots the percentage of all payout payers that have
an active payout gap. The bottom graph plots, for each year t, the aggregate magnitude of the combined active
payout gaps of all public U.S. firms (i.e., the sum of APGit across all public firms). Dollar magnitudes are in billions
of dollars of 2012 purchasing power.
Active payout gap (firm counts)
60%

40%

20%

0%
1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

2009

2011

2009

2011

% of all public firms that have an active payout gap

% of all payout payers that have an active payout gap

Active payout gap (dollar magnitude in billions of 2012 $)

200

150

100

50

0
1989

1991

1993

1995

1997

1999

2001

2003

2005

2007

Aggregate sum of active payout gaps

40

Figure 5. Actively financed payouts over time.

For each year from 1989 to 2012, the top graph plots the upper and lower bounds of the percentage of payers with
actively financed payouts; the bottom graph plots the upper and lower bounds of the aggregate magnitude of the
combined actively financed payouts paid by all public U.S. firms. (Dollar magnitudes are in billions of dollars of
2012 purchasing power.) As discussed in Section 3.2, we define the upper bound of a firms actively financed
payout as the amount of capital that the firm simultaneously raises and pays out (see Figure 3); the lower bound is
defined as the firms active payout gap (see Figure 4).
Actively financed payouts (firm counts - % of all payers)
60%

40%

20%

0%
1989

1991

1993

1995

1997

1999

2001

Actively financed payouts - upper bound

2003

2005

2007

2009

2011

Actively financed payouts - lower bound

Actively financed payouts (dollar magnitude in billions of 2012 $)

250
200
150
100
50
0
1989

1991

1993

1995

1997

1999

Actively financed payouts - upper bound

2001

2003

2005

2007

2009

2011

Actively financed payouts - lower bound

41

Table 1. Aggregate payout out and capital raising activity.

Panel A shows the percentage of all public U.S. firms that pay out capital (columns 1 through 3), and the percentage
of firms that raise capital (columns 4 through 8). Specifically, for each column, we compute the relevant annual
percentage of firms that pay out or raise capital each year in our sample period, and then report these annual
percentages averaged over four-year intervals to conserve space. Analogously, Panel B shows the aggregate annual
dollar amount (in billions of dollars of 2012 purchasing power) paid out or raised by all public U.S. firms, also
averaged over four-year intervals to conserve space. Financial firms and utilities are excluded throughout the paper.

Panel A. Firm counts.

Annual percentage of all public firms that
pay out capital via
Annual
figures
averaged
over

No.
public
firms

raise capital via

dividends
or share
repurchases

dividends

share
repurchases

any
security
issue

any firminitiated
issue

net debt

firminitiated
equity

employee
-initiated
equity

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

1989-1992

3,626

45.6%

36.3%

24.5%

66.6%

45.7%

36.7%

13.8%

41.1%

1993-1996

4,314

42.6%

32.3%

23.5%

78.0%

51.6%

40.3%

18.1%

54.9%

1997-2000

4,537

47.0%

27.1%

35.7%

83.1%

52.8%

42.7%

16.7%

64.0%

2001-2004

3,602

45.8%

26.5%

32.7%

83.5%

37.7%

27.8%

14.3%

72.5%

2005-2008

3,109

52.9%

32.1%

40.3%

88.0%

41.5%

33.5%

11.9%

77.7%

2009-2012

2,627

57.1%

34.3%

42.7%

79.9%

37.2%

28.4%

12.1%

67.2%

all years

3,636

48.5%

31.4%

33.2%

79.8%

44.4%

34.9%

14.5%

62.9%

Panel B. Dollar magnitudes.

Aggregate annual dollar amount (in billions of 2012 dollars) that is
paid out via

raised via

dividends
or share
repurchases

dividends

share
repurchases

any
security
issue

any firminitiated
issue

net debt

firminitiated
equity

employeeinitiated
equity

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

1989-1992

159.6

107.8

51.8

206.9

189.7

155.9

33.7

17.2

1993-1996

191.4

115.8

75.6

266.9

239.4

193.0

46.4

27.4

1997-2000

315.9

128.7

187.2

502.8

445.9

374.4

71.4

56.9

2001-2004

308.7

134.9

173.8

352.4

292.6

245.6

47.0

59.8

2005-2008

628.3

194.9

433.4

443.8

362.1

321.4

40.7

81.7

2009-2012

518.5

218.0

300.5

329.7

270.9

238.0

32.9

58.8

all years

353.7

150.0

203.7

350.4

300.1

254.7

45.4

50.3

Annual
figures
averaged
over

42

Table 2. Simultaneous payouts and security issues.

This table examines the extent to which firms simultaneously pay out and raise capital during the same fiscal year.
In Panel A, we consider all security issues; specifically, SI is the sum of the proceeds of positive net debt issues (i.e.,
max{debt issued debt repurchased, 0}), firm-initiated equity issues, and employee-initiated equity issues. In Panel
B, we focus only on instances in which firms actively raise capital by considering only firm-initiated security issues;
AI is thus defined as the sum of the proceeds of positive net debt issues and firm-initiated equity issues. On the
payout side, Columns 1 through 5 examine total payout (TP) without distinguishing between dividends and share
repurchases; Columns 6 and 7 focus on dividends only (Div); and Columns 8 and 9 focus on share repurchases (with
Rep denoting the dollar amount of shares repurchased).

Panel A. All security issues.

Firm counts

Dollar magnitudes

% public
firms
that raise
& pay
out
capital

%
payout
payers
that also
raise
capital

% firms
raising
capital
that also
pay out
capital

(1)

(2)

1989-1992

32.9%

1993-1996

33.7%

1997-2000

Aggregate sum of
min{SI, TP} over

Aggregate sum of
min{SI, Div} over

Aggregate sum of
min{SI, Rep} over

aggregate
sum of
TP

aggregate
sum of SI

aggregate
sum of
Div

aggregate
sum of SI

aggregate
sum of
Rep

aggregate
sum of SI

(3)

(4)

(5)

(6)

(7)

(8)

(9)

72.2%

49.5%

51.4%

39.7%

56.6%

29.8%

62.5%

15.6%

79.2%

43.3%

41.8%

29.9%

48.4%

21.3%

49.6%

13.7%

39.4%

83.8%

47.4%

48.5%

30.9%

62.3%

16.2%

53.9%

20.4%

2001-2004

39.3%

85.7%

47.1%

36.3%

32.2%

50.8%

19.9%

40.4%

20.3%

2005-2008

47.6%

90.2%

54.2%

33.9%

48.5%

51.9%

23.3%

39.6%

38.9%

2009-2012

47.5%

83.1%

59.4%

31.8%

50.0%

44.3%

29.6%

40.8%

36.9%

all years

40.1%

82.4%

50.1%

40.6%

38.5%

52.4%

23.3%

47.8%

24.3%

Annual
figures
averaged
over

Panel B. Only firm-initiated security issues (net debt and firm-initiated equity).
Firm counts

Dollar magnitudes

% public
firms
that
actively
raise &
pay out
capital

%
payout
payers
that also
actively
raise
capital

% firms
actively
raising
capital
that also
pay out
capital

aggregate
sum of
TP

aggregate
sum of AI

aggregate
sum of
Div

aggregate
sum of AI

aggregate
sum of
Rep

aggregate
sum of AI

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

1989-1992

21.1%

46.3%

46.2%

45.7%

38.5%

50.5%

29.0%

53.2%

14.6%

1993-1996

20.6%

48.4%

40.0%

34.3%

27.4%

39.1%

19.2%

39.0%

12.0%

1997-2000

23.8%

50.7%

45.1%

38.9%

27.9%

51.1%

14.9%

41.3%

17.6%

2001-2004

15.5%

33.7%

40.9%

24.9%

26.6%

35.8%

16.8%

24.1%

14.6%

2005-2008

20.7%

39.1%

49.8%

24.1%

42.0%

36.1%

19.8%

26.7%

32.2%

2009-2012

19.8%

34.1%

52.2%

23.1%

44.3%

32.5%

26.4%

28.7%

32.2%

all years

20.2%

42.1%

45.7%

31.8%

34.5%

40.9%

21.0%

35.5%

20.5%

Annual
figures
averaged
over

Aggregate sum of
min{AI, TP} over

Aggregate sum of
min{AI, Div} over

Aggregate sum of
min{AI, Rep} over

43

Table 3. Simultaneous payouts and security issues the role of debt and equity.
This table examines which securities firms issue when they simultaneously pay out and raise capital during the same
year. Panel A focuses on net debt issues, with ND denoting the proceeds of positive net debt issues (i.e., max{debt
issued debt repurchased, 0}). Panel B examines firm-initiated equity issues, with FE denoting the proceeds of
firm-initiated equity issues, while Panel C does the same for employee-initiated equity issues, with EE denoting the
proceeds of these issues. Columns 1 through 5 focus on total payout (TP) without distinguishing between dividends
and share repurchases, columns 6 and 7 focus on dividends (Div), and columns 8 and 9 on share repurchases (Rep).

Panel A. Net debt issues.

Firm counts

Dollar magnitudes

% public
firms
that
issue net
debt &
pay out
capital

%
payout
payers
that also
issue net
debt

% firms
issuing
net debt
that also
pay out
capital

(1)

(2)

1989-1992

18.6%

1993-1996

Aggregate sum of
min{ND, TP} over

Aggregate sum of
min{ND, Div} over

Aggregate sum of
min{ND, Rep} over

aggregate
sum of
TP

aggregate
sum of
ND

aggregate
sum of
Div

aggregate
sum of
ND

aggregate
sum of
Rep

aggregate
sum of
ND

(3)

(4)

(5)

(6)

(7)

(8)

(9)

40.7%

50.5%

39.5%

41.0%

44.2%

31.4%

46.8%

15.5%

18.3%

43.0%

45.5%

31.6%

31.6%

36.7%

22.5%

36.2%

13.8%

1997-2000

21.5%

45.7%

50.4%

38.1%

32.4%

50.7%

17.5%

40.2%

20.3%

2001-2004

13.4%

29.2%

48.1%

22.6%

28.9%

33.1%

18.7%

21.8%

15.8%

2005-2008

19.3%

36.4%

57.4%

23.7%

46.9%

35.8%

22.3%

26.3%

36.0%

2009-2012

17.9%

30.8%

61.8%

22.7%

50.4%

31.8%

30.3%

28.4%

36.2%

all years

18.2%

37.6%

52.3%

29.7%

38.5%

38.7%

23.8%

33.3%

22.9%

Annual
figures
averaged
over

Panel B. Firm-initiated equity issues.

% public
firms
that
issue
firm-init.
equity &
pay out
capital

Firm counts
%
payout
payers
that also
issue
firminitiated
equity

% firms
issuing
firminitiated
equity
that also
pay out
capital

(1)

(2)

1989-1992

4.3%

1993-1996

4.2%

1997-2000

Dollar magnitudes
Aggregate sum of
min{FE, TP} over

Aggregate sum of
min{FE, Div} over

Aggregate sum of
min{FE, Rep} over

aggregate
sum of
TP

aggregate
sum of FE

aggregate
sum of
Div

aggregate
sum of FE

aggregate
sum of
Rep

aggregate
sum of FE

(3)

(4)

(5)

(6)

(7)

(8)

(9)

9.4%

30.7%

8.2%

40.6%

8.7%

28.6%

10.0%

18.7%

9.8%

22.9%

3.6%

14.2%

3.9%

9.7%

3.4%

5.9%

4.2%

9.0%

25.4%

2.2%

10.1%

2.5%

4.2%

2.3%

6.8%

2001-2004

3.3%

7.3%

23.4%

3.1%

18.2%

3.3%

8.7%

3.2%

10.6%

2005-2008

2.6%

4.9%

22.1%

1.3%

18.2%

2.7%

12.1%

0.6%

6.3%

2009-2012

2.9%

5.2%

24.0%

0.7%

11.0%

1.2%

8.7%

0.4%

5.2%

all years

3.6%

7.6%

24.7%

3.2%

18.7%

3.7%

12.0%

3.3%

9.0%

Annual
figures
averaged
over

44

Panel C. Employee-initiated equity issues.

Firm counts

Dollar magnitudes

% public
firms that
issue
emp.-init.
equity &
pay out
capital

% payout
payers that
also issue
employeeinitiated
equity

% firms
issuing
employeeinitiated
equity
that also
pay out
capital

(1)

(2)

1989-1992

22.9%

1993-1996
1997-2000

Aggregate sum of
min{EE, TP} over

Aggregate sum of
min{EE, Div} over

Aggregate sum of
min{EE, Rep} over

aggregate
sum of
TP

aggregate
sum of
EE

aggregate
sum of
Div

aggregate
sum of
EE

aggregate
sum of
Rep

aggregate
sum of
EE

(3)

(4)

(5)

(6)

(7)

(8)

(9)

50.4%

56.1%

9.1%

83.5%

11.7%

74.1%

16.5%

48.2%

25.8%

60.7%

47.2%

11.0%

76.6%

15.4%

65.6%

16.5%

43.6%

31.6%

67.4%

49.5%

12.1%

68.1%

18.7%

43.2%

16.8%

56.1%

2001-2004

35.7%

77.9%

49.3%

13.9%

72.7%

20.6%

46.5%

19.2%

56.2%

2005-2008

43.8%

83.0%

56.7%

11.1%

85.9%

23.2%

55.4%

15.0%

79.7%

2009-2012

42.3%

74.1%

62.9%

9.8%

85.9%

17.3%

64.3%

15.2%

74.6%

all years

33.7%

68.9%

53.6%

11.2%

78.8%

17.8%

58.2%

16.6%

59.7%

Annual
figures
averaged
over

45

Table 4. Can firms fund their payouts without raising capital?

This table examines whether firms with simultaneous payouts and security issues could have funded their payouts
without the proceeds of these issues. To do so, Columns 1 and 2 examine the prevalence of firms with a payout gap,
i.e., firms with PGit min{max{TPit (FCFit + CRit), 0}, TPit} > $100,000 (denoted PGit >> 0), while columns 3
and 4 examine the dollar magnitude of these gaps relative to total payouts. Columns 5 and 6 examine the prevalence
of firms with an active payout gap, i.e., firms with APGit min{max{TPit (FCFit + CRit + EEit), 0}, TPit} >> 0,
while columns 7 and 8 examine the dollar magnitude of these active gaps relative to total payouts. (TP is total
payout; FCF is free cash flow, the sum of operating cash flow and investment cash flow; CR is cash reduction; and
EE captures the proceeds of employee-initiated equity issues.) In Panel A, PG and APG are defined annually and,
similarly as in Tables 1 through 3, we report annual counts or magnitudes averaged over four years to conserve
space. In contrast, in Panel B, we define gaps (and also total payouts) over four-year intervals, as follows:
3
3
3

PGitt 3 min max TPit j FCFit j CRit j , 0 , TPit j 0 (and analogously for active gaps).
j 0
j 0
j 0

Payout gaps (PG)

(i.e., firms that cannot fund their payout
without raising capital)

Active payout gaps (APG)

(i.e., firms that cannot fund their payout
without actively raising capital)

Firm counts

Firm counts

Dollar magnitudes

Dollar magnitudes

% of
public
firms with
a payout
gap

% of
payout
payers
with a
payout
gap

Average
of the
ratio
PG /TP
across all
firms with
a gap

Aggregate
sum of PG
over
aggregate
sum of TP
paid by all
public firms

% of
public
firms with
an active
payout
gap

% of
payout
payers
with an
active
payout
gap

Average
of the
ratio
APG /TP
across all
firms with
an active
gap

Aggregate
sum of APG
over
aggregate
sum of TP
paid by all
public firms

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Panel A. Gap defined annually (we report annual figures averaged over four-year intervals)
1989-1992

20.3%

44.6%

80.7%

41.0%

18.3%

40.1%

83.5%

38.0%

1993-1996

21.4%

50.4%

81.3%

32.7%

18.6%

43.7%

85.0%

29.8%

1997-2000

26.1%

55.6%

80.0%

40.5%

21.9%

46.5%

84.3%

35.6%

2001-2004

17.9%

39.1%

71.1%

22.2%

12.2%

26.7%

81.2%

18.2%

2005-2008

25.9%

49.0%

62.3%

24.6%

17.7%

33.5%

74.1%

19.6%

2009-2012

21.6%

37.2%

61.5%

19.6%

15.1%

25.9%

73.6%

15.3%

average of
all years

22.2%

46.0%

72.8%

30.1%

17.3%

36.1%

80.3%

26.1%

Panel B. Gap defined over four-year intervals (sources and uses of capital aggregated over four years)
1989-1992

36.2%

52.6%

78.3%

41.0%

32.8%

47.6%

80.0%

36.3%

1993-1996

40.8%

62.3%

82.4%

33.2%

36.6%

55.9%

83.4%

28.0%

1997-2000

52.7%

71.6%

82.4%

43.2%

46.4%

63.0%

83.9%

37.0%

2001-2004

33.0%

47.4%

77.4%

19.9%

25.1%

36.1%

79.6%

15.1%

2005-2008

46.3%

62.1%

66.0%

28.2%

36.6%

49.1%

68.6%

20.6%

2009-2012

35.7%

47.4%

64.3%

20.5%

27.5%

36.5%

71.7%

14.5%

average of all
four-year intervals

40.8%

57.2%

75.1%

31.0%

34.2%

48.0%

77.9%

25.3%

46

Table 5. The role of investment in explaining active payout gaps.

This table examines the extent to which the active payout gaps identified in Table 4 are the result of firms with unusually high levels of investment. Specifically,
the panel replicates the analysis shown in Columns 5 through 8 of Table 4, Panel A, but using two modified definitions of active payout gap. In Columns 1
through 4, we assume that no firm invests more than the median firm in its industry, thus defining a firms modified active payout gap as follows: APGitIndI
min{max{TPit (OCFit + max{ICFit, Industry median ICFit}+ CRit + EEit), 0}, TPit}, where in Industry median ICFit we substitute CAPEX and acquisitions
(both of which enter the definition of investment cash flow, ICFit, with a negative sign) by the median ratio of (CAPEX + acquisitions)/lagged assets of the firms
industry in that year multiplied by the firms lagged assets. (TP is total payout; OCF is operating cash flow; CR is cash reduction; and EE captures the proceeds
of employee-initiated equity issues. Industry is defined at the 2-digit SIC level.) In Columns 5 through 8, we do the same but in this case we assume that no firm
invests more than it did in the previous year; our modified definition of active payout gap is then: APGitLagI min{max{TPit (OCFit + max{ICFit, Lagged
ICFit}+ CRit + EEit), 0}, TPit}, where in Lagged ICFit we substitute CAPEX and acquisitions by their lagged values. For ease of comparison, Columns 9 through
12 reproduce the baseline active payout gaps results we show in Columns 5 through 8 of Table 4, Panel A, which use a firms actual level of investment.
Active payout gaps
We assume that no firm invests more than the
median firm in its industry
Counts

Dollar magnitudes
Average
across all
Aggregate
firms
sum of
with a
APGIndI
modified
over
active
aggregate
gap of the
sum of
ratio
TP
IndI
APG
over TP

% of
public
firms
with an
active
payout
gap

% of
payout
payers
with an
active
payout
gap

(1)

(2)

(3)

1989-1992

11.3%

24.8%

1993-1996

12.1%

28.5%

1997-2000

14.5%

2001-2004

6.8%

2005-2008

10.3%

2009-2012
all years

Annual
figures
averaged
over

We assume that no firm invests more than it did

in the previous year
Counts

Dollar magnitudes
Average
across all
Aggregate
firms
sum of
with a
APGLagI
modified
over
active
aggregate
gap of the
sum of
ratio
TP
LagI
APG
over TP

Actual level of investment

Counts

Dollar magnitudes

% of
public
firms
with an
active
payout
gap

% of
payout
payers
with an
active
payout
gap

Average
across all
firms
with an
active
gap of the
ratio APG
over TP

Aggregate
sum of
APG over
aggregate
sum of
TP

(9)

(10)

(11)

(12)

28.4%

18.3%

40.1%

83.5%

38.0%

81.7%

20.4%

18.6%

43.7%

85.0%

29.8%

34.7%

81.0%

24.6%

21.9%

46.5%

84.3%

35.6%

17.8%

75.5%

11.8%

12.2%

26.7%

81.2%

18.2%

22.6%

68.0%

13.5%

17.7%

33.5%

74.1%

19.6%

9.6%

16.5%

67.1%

9.5%

15.1%

25.9%

73.6%

15.3%

12.2%

25.5%

75.4%

18.0%

17.3%

36.1%

80.3%

26.1%

% of
public
firms
with an
active
payout
gap

% of
payout
payers
with an
active
payout
gap

(4)

(5)

(6)

(7)

(8)

77.8%

21.9%

14.0%

30.6%

79.4%

80.5%

19.4%

13.2%

31.0%

30.8%

78.8%

22.2%

16.3%

14.8%

75.7%

9.6%

8.2%

19.5%

65.6%

12.4%

12.0%

8.3%

14.2%

66.0%

7.9%

10.6%

22.1%

74.1%

15.5%

47

Table 6. The role of negative cash flow shocks in explaining active dividend gaps.
This table examines the prevalence and economic magnitude of active dividend gaps and the extent to which these are the result of firms deciding to maintain
their dividends in the face of negative cash flow shocks. First, in Columns 1 through 4, we define a firms dividend gap, ADGit, analogously to the way we define
its active payout gap: ADGit min{max{Divit (FCFit + CRit + EEit), 0}, Divit}. (Div is dividend payout; FCF is free cash flow; CR is cash reduction; EE
captures the proceeds of employee-initiated equity issues; and TP is total payout.) Next, in Columns 5 through 8, we use a modified version of a firms active
dividend gap that assumes that no firm is less profitable than the median firm in its industry: ADGitIndO min{max{Divit (max{OCFit, Industry median OCFit}
+ ICFit + CRit + EEit), 0}, Divit}, where Industry median OCFit is defined as the median ratio of operating cash flow/lagged assets of the firms industry in that
year multiplied by the firms lagged assets. (Industry is defined at the 2-digit SIC level.) Similarly, in Columns 9 through 12, we assume that no firm is less
profitable than it was in the previous year; our modified definition of active dividend gap is then: ADGitLagO min{max{Divit (max{OCFit, OCFit-1} + ICFit +
CRit + EEit), 0}, Divit}.
Active dividend gaps

Actual level of profitability

Counts

We assume that no firm is less profitable than the

median firm in its industry

Dollar magnitudes

Counts

Dollar magnitudes

% of
public
firms
with an
active
dividend
gap

% of
dividend
payers
with an
active
dividend
gap

Aggregate
sum of
ADG over
aggregate
sum of TP

Aggregate
sum of
ADG over
aggregate
sum of
Div

% of
public
firms
with an
active
dividend
gap

% of
dividend
payers
with an
active
dividend
gap

Aggregate
sum of
ADGIndO
over
aggregate
sum of TP

Aggregate
sum of
ADGIndO
over
aggregate
sum of
Div

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

1989-1992

13.1%

36.1%

22.5%

33.4%

10.3%

28.4%

19.8%

1993-1996

12.8%

39.8%

16.1%

26.1%

10.3%

32.0%

1997-2000

10.9%

40.1%

13.2%

32.5%

8.9%

2001-2004

5.9%

22.5%

7.5%

17.2%

2005-2008

7.9%

24.7%

4.3%

14.3%

2009-2012

6.7%

19.3%

4.4%

all years

9.6%

30.4%

11.4%

Annual
figures
averaged
over

We assume that no firm is less profitable than it

was in the previous year
Counts

Dollar magnitudes

% of
public
firms
with an
active
dividend
gap

% of
dividend
payers
with an
active
dividend
gap

Aggregate
sum of
ADGLagO
over
aggregate
sum of TP

Aggregate
sum of
ADGLagO
over
aggregate
sum of
Div

(9)

(10)

(11)

(12)

29.4%

10.3%

28.2%

18.0%

26.7%

13.7%

22.3%

10.4%

32.2%

13.6%

22.1%

32.7%

12.1%

29.9%

8.8%

32.4%

10.9%

26.8%

4.2%

15.9%

6.9%

15.6%

4.3%

16.2%

6.6%

15.0%

5.9%

18.5%

3.8%

12.6%

6.3%

19.5%

3.8%

12.6%

9.6%

5.1%

14.4%

4.0%

8.7%

5.1%

14.5%

3.7%

8.0%

22.2%

7.4%

23.6%

10.1%

19.8%

7.5%

23.8%

9.4%

18.6%

48

INTERNET APPENDIX

(NOT INTENDED FOR PUBLICATION)

Table IA.1. Avoiding active payout gaps by underinvesting?

This table examines whether the prevalence and economic magnitude of active payout gaps identified in Table 4 would be greater if firms invested in line with
their industry or what they invested in the previous year. Specifically, the panel replicates the analysis shown in Columns 5 through 8 of Table 4, Panel A, but
using two modified definitions of active payout gap. In Columns 1 through 4, we assume that no firm invests less than the median firm in its industry, thus
defining a firms modified active payout gap as follows: APGitIndI min{max{TPit (OCFit + min{ICFit, Industry median ICFit}+ CRit + EEit), 0}, TPit}, where
in Industry median ICFit we substitute CAPEX and acquisitions (both of which enter the definition of investment cash flow, ICFit, with a negative sign) by the
median ratio of (CAPEX + acquisitions)/lagged assets of the firms industry in that year multiplied by the firms lagged assets. (TP is total payout; OCF is
operating cash flow; CR is cash reduction; and EE captures the proceeds of employee-initiated equity issues. Industry is defined at the 2-digit SIC level.) In
Columns 5 through 8, we do the same but in this case we assume that no firm invests more than it did in the previous year; our modified definition of active
payout gap is then: APGitLagI min{max{TPit (OCFit + min{ICFit, Lagged ICFit}+ CRit + EEit), 0}, TPit}, where in Lagged ICFit we substitute CAPEX and
acquisitions by their lagged values. For ease of comparison, Columns 9 through 12 reproduce the baseline active payout gaps results we show in Columns 5
through 8 of Table 4, Panel A, which use a firms actual level of investment when defining its active payout gap.
Active payout gaps
We assume that no firm invests less than the
median firm in its industry
Counts

Dollar magnitudes
Average
across all Aggregate
firms
sum of
with an
APGIndI+
active
over
gap of the aggregate
ratio
sum of
APGIndI+
TP
over TP

% of
public
firms
with an
active
payout
gap

% of
payout
payers
with an
active
payout
gap

(1)

(2)

(3)

1989-1992

22.9%

50.1%

1993-1996

23.8%

1997-2000

28.2%

2001-2004

We assume that no firm invests less than it did

last year
Counts

Dollar magnitudes
Average
across all Aggregate
firms
sum of
with an
APGLagI+
active
over
gap of the aggregate
ratio
sum of
APGLagI+
TP
over TP

Actual level of investment

Counts

Dollar magnitudes

% of
public
firms
with an
active
payout
gap

% of
payout
payers
with an
active
payout
gap

Average
across all
firms
with an
active
gap of the
ratio APG
over TP

Aggregate
sum of
APG over
aggregate
sum of
TP

(9)

(10)

(11)

(12)

45.4%

18.3%

40.1%

83.5%

38.0%

85.5%

35.4%

18.6%

43.7%

85.0%

29.8%

86.1%

42.1%

21.9%

46.5%

84.3%

35.6%

42.1%

81.6%

26.4%

12.2%

26.7%

81.2%

18.2%

25.9%

48.8%

75.8%

26.4%

17.7%

33.5%

74.1%

19.6%

20.2%

23.8%

41.3%

75.3%

23.1%

15.1%

25.9%

73.6%

15.3%

31.9%

24.1%

50.0%

81.5%

33.1%

17.3%

36.1%

80.3%

26.1%

% of
public
firms
with an
active
payout
gap

% of
payout
payers
with an
active
payout
gap

(4)

(5)

(6)

(7)

(8)

82.2%

42.4%

24.2%

53.1%

84.8%

55.8%

84.2%

37.5%

23.2%

54.5%

60.0%

83.5%

42.9%

28.2%

59.9%

17.1%

37.4%

78.8%

22.9%

19.3%

2005-2008

24.9%

47.1%

70.1%

25.7%

2009-2012

21.2%

36.5%

70.8%

all years

23.0%

47.8%

78.3%

Annual
figures
averaged
over