CEO Marital Status

CEO Marital Status and Corporate Cash Holdings
Ahmed Elnahas 1
Md Noman Hossain 2
Siamak Javadi 3
Abstract
We examine the effect of CEO marital status on corporate cash holdings. Consistent with the
classical agency framework, we find that firms with single CEOs hold more cash compared to
otherwise similar firms with married CEOs and that the excess cash held by single CEOs is
significantly discounted by shareholders. Our findings survive a battery of tests to ease
endogeneity and selection bias, confirming that results are not simply reflecting innate
heterogeneity in preferences. Overall, our findings indicate that a variable outside the common
firm- and macro-level determinants, CEO marital status, can significantly influence corporate
policies.
Keywords: Cash Holdings; Marital Status; CEO; Firm Value
JEL Classification: G30, G32, G35

We would like to express our gratitude to the anonymous referees for their thoughtful comments and suggestions.
Further, we are indebted to Hua Cheng (Discussant), Gautam Hazarika, Suin Lee, Yu Liu, Nam Nguyen, Mahdi
Rastad, Bruno Arthur, Incheol Kim, Levent Kutlu, Armando Lopez-Velasco, Andre Mollick, Monika Rabarison,
Jean-Baptiste Tondji, Salvador Contreras, and other seminar participants at the 2020 Southern Finance Association,
the 2021 Southwest Finance Association, and the 2021 economics and finance seminar series at the University of
Texas Rio Grande Valley. All remaining errors are ours. The data that support the findings of this study are available
from the corresponding author upon reasonable request.
1
Robert C. Vackar College of Business and Entrepreneurship, University of Texas Rio Grande Valley, 1201 W
University Dr, Edinburg, TX 78539. Voice: (956) 665-3827; E-mail: [email protected]
2
College of Business, Central Washington University, 400 E University Way, Ellensburg, WA 98926.Voice: (509)
963-1409; E-mail: [email protected]
3
Robert C. Vackar College of Business and Entrepreneurship, University of Texas Rio Grande Valley, 1 W
University Blvd, Brownsville, TX 78520. Phone: (956) 882-8858; E-mail: [email protected]
Electronic copy available at: https://ssrn.com/abstract=4085811

CEO Marital Status and Corporate Cash Holdings
Abstract
We examine the effect of CEO marital status on corporate cash holdings. Consistent with the
classical agency framework, we find that firms with single CEOs hold more cash compared to
otherwise similar firms with married CEOs and that the excess cash held by single CEOs is
significantly discounted by shareholders. Our findings survive a battery of tests to ease
endogeneity and selection bias, confirming that results are not simply reflecting innate
heterogeneity in preferences. Overall, our findings indicate that a variable outside the common
firm- and macro-level determinants, CEO marital status, can significantly influence corporate
policies.
Keywords: Cash Holdings; Marital Status; CEO; Firm Value
JEL Classification: G30, G32, G35

1. Introduction
There is no doubt that a CEO is the single most influential individual in a modern corporation.
Consequently, financial economists have paid significant attention to studying how CEOs'
personal traits affect their firms’ policies and decision-making processes. For example,
researchers investigate the impact of CEOs’ gender (Adhikari, 2018), overconfidence (Aktas,
Louca & Petmezas, 2019; and Malmendier & Tate, 2005), age (Andreou, Louca & Petrou, 2017),
military affiliation (Benmelech & Frydman, 2015), political ideology (Di Giuli & Kostovetsky,
2014; Elnahas & Kim, 2017; and Hossain et al., 2022), and personal debt and risk-taking
(Cronqvist, Makhija, & Yonker, 2012; and Cain & McKeon, 2016). Recently, researchers have
studied the impact of a CEO’s marital status on corporate decision-making. For example, Hilary,
Huang, and Xu (2017) find that firms headed by single CEOs display a higher degree of earnings
management than those headed by married CEOs. Roussanov and Savor (2014) show that firms
headed by single CEOs exhibit higher stock return volatility, have more aggressive investment
policies, and are less likely to respond to changes in idiosyncratic risk. Lu, Ray, and Teo (2016)
conclude that hedge funds run by single managers take more risks than those run by married
managers. Similarly, Belenzon, Patacconi, and Zarutskie (2016) find that firms run by married
couples behave more conservatively than other family-owned firms. Other studies also document
the effect of marital status on risk aversion in terms of asset allocation, saving and portfolio
choices, life insurance policies, and accounting policies (Riley Jr and Chow (1992); Roszkowski,
Snelbecker, and Leimberg (1993); Hallahan, Faff, and McKenzie (2004); Love (2009); Hong &
Ríos-Rull (2012)). In this paper, we extend this literature by investigating the effect of the marital
status of CEOs on corporate cash holdings.
The recent attention that financial economists have paid to the impact of CEO marital status
on corporate policies is not surprising. Social scientists have always considered marriage as the
3

building block of communities (Slater, 1963). In their report titled “why marriage matters”, a
group of renowned family scholars documents that marital status affects several aspects of human
life that include family (e.g., quality relationships and biosocial effects), economic (e.g., poverty
rates and wealth creation), physical health and longevity (e.g., lower rates of injury, illness,
disability, and life expectancy), and mental health and emotional well-being (e.g., rates of crime,
suicide, and mental illness) (Wilcox et al. (2005)). According to the behavioral consistency
theory, the personal traits of a CEO, such as marital status, not only affect her/his personal
behavior and choices but also the policies and decisions of the firm she/he leads.
Cash holding is one of the corporate policies that has received growing academic attention in
the literature (Opler et al.1999; Harford 1999; Dittmar et al. 2003; Pinkowitz et al. 2006; Dittmar
and Mahrt-Smith 2007; Foley et al. 2007; Harford et al. 2008; Bates et al. 2009; Duchin 2010;
Nikolov and Whited 2014; Gao et al. 2013). Understanding corporate cash holding policy is
important because cash holdings act not only as a safety net in the presence of costly external
financing but also may impact investors’ concerns about managerial self-dealing. Indeed, the two
predominant theories of corporate cash holdings, the precautionary motive and the agency
framework, make mixed predictions about the possible impact of CEO marital status on cash
holdings.
Cash provides a safety cushion against bankruptcy risks and ameliorates adverse cash flow
shocks. Consistent with this view, Bates, Kahle, and Stulz (2009) find that the increase in
corporate cash holdings is positively correlated with the increase in firm-specific risks. Similarly,
Adhikari (2018) provides evidence that firms with more risk-averse managers hold more cash.
Prior literature has characterized single individuals as less risk-averse than married individuals.
Roussanov and Savor (2014) show that single CEOs are less risk-averse compared to married

CEOs. Similarly, Lu, Ray, and Teo (2016) conclude that hedge funds run by single managers
take more risk than those run by married managers. It follows that a single CEO with a lower
degree of risk aversion is expected to hold less cash than a married CEO. We call this our risk
aversion hypothesis.
In addition to having low risk aversion, single individuals have always been characterized as
more self-centered by social scientists. DePaulo and Morris (2005) document that single
individuals have been consistently viewed as more self-centered and envious than married
individuals. Similarly, Wilcox et al. (2005) claim that marriage turns individuals’ attention away
from dangerous, antisocial, and self-centered activities. Further, Campbell & Lee (1992) link
singlehood to social isolation and view marriage as an attachment to wider communities and
increased investment in social ties. By definition, a self-centered CEO who is more willing to
seek her/his interest at the expense of those of others (including shareholders) would engage
more in self-dealing, which aggravates the agency problem in firms headed by single CEOs.4
However, how would this aggravated agency problem affect the cash holdings of firms
headed by single CEOs? The agency framework provides opposing predictions regarding the
association between marital status and cash holdings. On the one hand, researchers show that
poorly governed firms with escalated agency issues dissipate more cash (Dittmar and Mahrt-
Smith (2007)), particularly on acquisitions (Harford, Mansi, and Maxwell, (2008).)
Consequently, this argument predicts that firms headed by single CEOs dissipate more cash, i.e.,
4
Self-centeredness is defined by Dambrun and Ricard (2011) as “the self takes on a central point of reference with
regard to many psychological activities (i.e., conation, motivation, attention, cognition, affect/emotion, and
behavior). The exaggerated importance given to the self emerges mainly from self-centeredness and refers to the
increased degree with which the individual considers that his own condition is more important than that of others
and this takes unquestionable priority.” Further, they argue that Self-centered psychological functioning includes
characteristics such as biased self-interest, egoism, egocentrism, and egotism.

hold less cash —a prediction similar to our risk-aversion hypothesis. We call this our poor
governance hypothesis.
On the other hand, the classical agency framework pioneered by Jensen (1986) suggests that
corporate insiders generally have a personal incentive to retain cash rather than increase
dividends and that they rely on cash and liquid assets to finance their risky projects, empire-
building and self-dealing, allowing them to avoid the market scrutiny necessary for raising
external finance. This classical agency framework predicts that firms headed by single CEOs,
due to escalated agency issues, hold more cash. The positive association between agency
problems and cash holdings is consistent with the work of Dittmar et al. (2003), Kalcheva and
Lins (2007), and Pinkowitz et al. (2006). We refer to this as our classical agency problem
hypothesis.
The impact of CEO marital status on corporate cash holding is ultimately an empirical
question that depends on how marital status affects a CEO’s risk aversion and agency problem.
Using a sample of 19,864 firm-year observations between 1993 and 2008, we find that firms with
single CEOs hold about 1.3 percentage points more cash compared to firms with married CEOs,
after controlling for the known determinants of cash holdings (Opler et al. (1999)), other CEO
characteristics such as tenure, age, duality, pay-performance sensitivity (Delta), risk-taking
incentives (Vega), CEO ownerships, and the inclusion of industry and year fixed effects. This
effect is economically significant. For the average firm in our sample that has a total asset size
of about $4.3 billion, the above figure is equivalent to holding about $56 Million more cash by
firms with single CEOs.5
5
Our sample ends in 2008, because CEOs marital status data is available between 1993 and 2008.

Next, we address the endogeneity concerns. Specifically, (i) culture, religion, or managers'
early life experience, among other traits, might be jointly correlated with CEO marital status and
cash holdings, leading to an omitted variable bias in our baseline results. For instance, corporate
cash holdings could be affected by the level of CEO overconfidence (Chen, Ho, and Yeh (2020));
(ii) the association between CEO marital status and cash holdings may be spurious and attributable
to the innate heterogeneity in preferences rather than marital status; and (iii) there is a concern that
firms with single CEOs are fundamentally different from those with married CEO, suggesting that
our analysis may suffer from selection bias. We address these concerns in several ways. First, we
employ a Difference-in-Differences (DiD) estimation around CEO turnover events and find that
firms that replace a married CEO with a single CEO experience an increase in their cash holdings
while those that replace a single CEO with a married one experience a decrease in their cash
holdings. Second, to address the concern that our results are driven by CEO innate heterogeneity
rather than by their marital status, we use the instrumental variables (IV) regression approach used
by Roussanov & Savor (2014) and find that our baseline regression results continue to hold. Third,
we use propensity score matching (PSM) to address the selection bias concern. There are
significant differences between firms run by single CEOs and those run by married CEOs. Thus,
we match firms with single CEOs (the treatment) and firms with married CEOs (control groups)
on the firm and CEO characteristics as well as year and industry to mitigate selection bias concerns
(Rosenbaum & Rubin (1983), Roberts & Whited (2013)). We then run our baseline regression
using the matched sample and find that our baseline results continue to hold. We also test our DiD
estimate using the matched sample and find consistent results to our baseline findings. We also
find similar results when we estimate our baseline model while controlling for CEO

overconfidence.6 Overall, while it is impossible to completely refute endogeneity concerns, it is
unlikely that our results are driven by endogeneity.
Our baseline results are more consistent with the classical agency problem hypothesis rather
than the risk aversion and poor governance hypotheses. However, we note that as an alternative
interpretation of our results, hoarding cash may be explained by precautionary motive rather than
our classical agency problem hypothesis, as our interpretation suggests. We conduct a battery of
tests to rule out precautionary motives as an alternative explanation of our results. First, we
investigate the association between CEO marital status and cash holdings for subsamples of
financially constrained and unconstrained firms. While cash-holding policies in financially
constrained firms can be greatly driven by precautionary motives (Almeida et al. (2004); Han &
Qiu (2007), Denis & Sibilkov (2009)), such motives are not expected to play a significant role in
financially unconstrained firms. Consequently, finding higher levels of cash holdings for
financially unconstrained firms with single CEOs would refute the possibility that the cash
holdings of single CEOs are driven by precautionary motives. Using several conventional
measures of financial constraint as well as different proxies for cash holding, we find robust higher
levels of cash holdings for unconstrained firms with single CEOs compared to otherwise similar
firms with married CEOs. Next, we conduct a test using subsamples of firms with high and low
firm-specific risk. Following the same intuition, the precautionary motive is not expected to play
a vital role in holding excess cash in firms with low levels of risk. Using several measures of firm-
specific risk, we find robust higher levels of cash holdings for low-risk firms with single CEOs
compared to otherwise similar firms with married CEOs, which is inconsistent with the
precautionary explanation.
6
The results of the test that controls for CEO overconfidence are similar to our baseline results. These results are un-
tabulated and are available upon request.

In the next step, we provide more corroborating evidence consistent with the classical agency
problem hypothesis which also hints at the possible channels for the positive association between
cash holdings and single CEOs. First, since tax avoidance lowers the firm’s overall tax burden and
thus increases the firm’s net income, managers benefit from tax avoidance activities in the form of
higher compensation. Consequently, self-centered single CEOs may attempt to pump their
compensations by engaging in aggressive tax avoidance practices. Using two proxies of corporate
tax avoidance (Wilson (2009), Frank, Lynch, & Rego (2009)), we find that firms run by single
CEOs are more likely to engage in corporate tax avoidance than those run by married CEOs.
Second, using various measures of dividend payout, we find that firms with single CEOs pay out
significantly lower dividends compared to otherwise similar firms with married CEOs. Third, we
find that single CEOs rely more on cash holdings compared to married CEOs to fund the risky
investments documented in Roussanov and Savor (2014) and that external financing in general is
significantly lower for firms with single CEOs. The findings of these tests are generally consistent
with the predictions of our classical agency problem hypothesis.
Lastly, we explore the possible impact of the excess cash held by single CEOs on firm value.
Following the valuation models of Dittmar and Mahrt-Smith (2007) and Pinkowitz and
Williamson (2007), we find robust evidence consistent with our classical agency problem
hypothesis that shareholders heavily discount the additional cash that single CEOs hold.
This study makes several contributions to the existing literature. First, this study adds
additional evidence to the small but growing finance literature on behavioral consistency theory.
Prior studies provide evidence of behavioral consistency in corporate policies of firms, such as
CEO overconfidence (Malmendier & Tate (2005)), personal tax aggressiveness (Chyz (2013)),
CEO political ideology (Hutton et al. (2014)), and pilot CEOs (Sunder, Sunder, & Zhang (2017)).

In contrast, we examine whether CEOs’ personal traits i.e., marital status affect firms’ cash
holdings. It is worth noting that in almost all prior research that investigates CEO marital status,
risk aversion and agency framework do not necessarily have opposing predictions. For example,
Roussanov and Savor (2014) investigate the association between marital status and aggressive
investment policies. Nicolosi and Yore (2015) examine the association between marital status and
deal-making as well as risk-taking behavior, while Hilary, Huang, and Xu (2017) analyze the
association between marital status and earnings management. Across all these studies, both the
risk aversion and the agency frameworks mostly have similar predictions regarding the decision-
making process of single CEOs. However, due to the competing predictions of the risk aversion
hypothesis and the classical agency problem hypothesis of the cash holdings literature, cash
holding provides an appropriate setting to disentangle risk-taking motives and agency issues. The
closest paper to ours is a concurrent study by Al Mamun et al (2021). While it is reassuring that our
main results and theirs are similar in that single CEOs hold more cash than married CEOs, our
paper is different from their paper in several important ways. First, our paper rules out the
possibility that excess cash holdings by single CEOs are driven by precautionary motives. Second,
our paper provides some insights into the possible channels by presenting empirical evidence on
the engagement of single CEOs in tax avoidance, dividend cuts, and risky investment policies.
Further, our tests employ a wider set of measures of cash holding, measures of dividends payout
policy, as well as varying cash valuation models including those of Dittmar and Mahrt-Smith
(2007), and Pinkowitz and Williamson (2007), among others.
Further, this study contributes to the corporate cash holdings literature. Most studies in this
literature assume that cash holding is a function of firm-level (Opler et al. (1999); Bates et al.
10

(2009)) or macro-level factors (Duong et al. (2020); Phan et al. (2019); Javadi et al. (2021)). We,
however, provide robust evidence that CEO personal traits also affect cash holdings.
The paper proceeds as follows. Section 2 describes the data sources, sample construction, and
summary statistics. Section 3 presents the empirical results and the steps we took to address
endogeneity. We explore the precautionary as well as the agency explanations in Section 4. We
present additional tests regarding possible channels, risky investments, and valuation of excess
cash holdings in section 5, and conclude in Section 6.
2. Data, sample selection, and summary statistics
2.1. Data
We use the data on CEO's marital status provided by Roussanov & Savor (2014).7 They collected
the names, biographical information, and compensation of all CEOs covered by the Compustat
Executive Compensation (ExecuComp) database which covers firms in the S&P 1500 index from
1993 to 2008. To identify the CEOs' marital and family status, they use several public sources such
as the Marquis Who’s Who in Finance and Industry, the Notable Names Database, the U.S.
Securities and Exchange Commission’s insider filings, and various media mentions.8 Following
their study, we define the CEO’s marital status as Single which is an indicator variable that equals
1 if a CEO is unmarried and 0 otherwise.
We then merge this database with Compustat to construct the cash-holding measures and firm-
level control variables. Following prior studies, we exclude financial firms (SIC codes between
6000 and 6999) and utilities (SIC codes between 4900 and 4999) since these industries are highly
7
We thank Roussanov & Savor (2014) for sharing their CEOs marital status data, which is available at
http://dx.doi.org/10.1287/mnsc.2014.1926
8
We request readers to go through Roussanov & Savor (2014) for more details about the data collection process,
biases, and accuracy of the CEOs’ marital status.
11

regulated.9 We also exclude firms with a missing value of total assets and require firms to have
positive cash holding and sales. Our final sample includes 19,864 firm-year observations of 2,305
unique firms.
2.2. Measures of Cash holding
Following Bates, Kahle, & Stulz (2009), we define firms’ cash holdings, Cash, as the ratio of cash
plus marketable securities to the book value of assets. We also construct an alternative measure of
cash holdings for robustness check, NetCash which is the ratio of cash to net assets, where net
assets are defined as the book value of assets minus cash following Dittmar et al. (2003). We also
construct FCash which is the one-year forwarded value of Cash (Casht+1) and FNetCash
(NetCasht+1) which is the one-year forwarded value of NetCash. We use these two alternative
measures in our estimation to address the cash-flow identity issue (Kalcheva & Lins (2007). We
report the descriptive statistics of the key variables used in our analyses in Table 1.
[Please insert Table 1 here]
Statistics in Table 1 show that the mean value of Single is 0.173, indicating that 17.3% of the
firm-year observations have single CEOs, which is consistent with statistics reported by
Roussanov & Savor (2014), and Hilary et al. (2017). The mean value of Cash is 0.144 indicating
that, on average, firms hold 14.4% of their total assets in cash and marketable securities, which is
consistent with the statistics reported by Opler et al. (1999), Bates et al. (2009), and Nguyen et al.
(2018). Similarly, the mean value of NetCash implies that, on average, firms hold 18.6% of their
net assets in net cash, which is consistent with the statistics reported by Dittmar et al. (2003).
9
See more at Dittmar et al. (2003), and Faulkender & Wang (2006).
12

Further, the average firm in our sample has financed 21.3% of its assets by debt and about half of
our sample firms pay dividends. We report Pearson correlation coefficients in Table 2.
Table 2 shows that the correlation between Single and Cash is 0.13, indicating that single CEOs
hold more cash than married CEOs. The positive correlations between Single and R&D,
Acquisition, and Industry_sigma are consistent with the higher risk-taking propensity of single
CEOs. The negative relation between Single and Ln(Age) indicates that the older the CEO, the
higher the likelihood of her/him being married.10 The negative correlation between Cash and
Ln(Age) indicates that older CEOs are less likely to hold cash as compared to younger CEOs. The
positive correlation between Cash and Ln(Tenure) indicates that entrenched managers hold more
cash as CEO entrenchment increases with their tenure (Coles, Daniel, & Naveen (2014), Jiang &
Lie (2016)).
3. Analysis and Results
3.1. Baseline Results
To test the relation between CEO marital status and cash holdings, we estimate the following
regression model:
𝐶𝑎𝑠ℎ𝑖𝑗𝑡 = 𝛽0 + 𝛽1 𝑆𝑖𝑛𝑔𝑙𝑒𝑖𝑗𝑡 + 𝛾𝑖𝑗𝑡−1 + 𝐼𝑗 + 𝜏𝑡 + Ɛ𝑖𝑗𝑡 (1)
where 𝐶𝑎𝑠ℎ𝑖𝑗𝑡 is our measure of cash holdings for firm i that belongs to industry j in year t as
defined in Section 2.2. Single is an indicator variable that equals 1 if a CEO is unmarried and zero
otherwise. The coefficient of Single (β1) measures the effect of the CEO’s marital status on the
10
Roussanov & Savor (2014) mention that “According to U.S. census data, 70% of men in the 35-59 age range were
married in 2000.” Thus, it is logical that we find only 17.3 % of the CEOs are Single in our sample period and the
correlation between Single and CEO age is negative.
13

firms’ cash holdings decisions. A positive sign on β1 is consistent with our classical agency
problem hypothesis whereas a negative sign would be consistent with our risk aversion and poor
governance hypotheses. Following the prior literature (Opler et al. (1999), Dittmar et al. (2003),
Almeida, Campello, & Weisbach (2004), Dittmar & Mahrt-Smith (2007), Acharya, Almeida, &
Campello (2007), Bates et al. (2009), Nguyen, Phan, & Sun (2018), and Javadi et al. (2021)), we
control for a vector of firm- and CEO-level control variables (𝛾𝑖𝑗𝑡−1). For instance, prior studies
argue that there are economies of scale to holding cash, thus we control for firm size (Ln(assets)).
The value of cash is more important for firms with better investment opportunities and financially
constrained firms, so we control for the ratio of market-to-book value (MB). We control for
Leverage because firms with debt constraints may use cash to reduce leverage, whereas the
hedging argument of Acharya et al. (2007) predicts a positive relationship between cash holdings
and leverage. We also control for R&D because firms with higher R&D expenditures have a higher
cost of financial distress and may require more cash holdings. We control for Cash_Flow because
firms with higher cash flow accumulate more cash. We control for NWC because prior research
shows a negative relationship between net working capital and cash holdings.
Prior studies demonstrate that if capital expenditures are used as a proxy for the cost of
financial distress or investment opportunities then they could be positively related to cash, whereas
if firms temporarily increase capital expenditures due to a productivity shock, or if capital
expenditures create assets that can be subsequently used as collateral, then they could lower cash
holdings. Consequently, we control for capital expenditures (CAPX). We control for Dividend
because dividend-paying firms are less risky and less financially constrained and have greater
access to capital markets reducing the precautionary motive for cash holdings. We control for
acquisition expenditure (Acquisition) due to the expected effect of acquisition expenditures on cash
14

outflows. Prior studies suggest that firms with greater cash flow risk hold more cash due to
precautionary motives, thus we control for cash flow volatility (Industry_sigma).
With regard to the CEO-level controls, prior studies suggest that CEO tenure and age are
related to career concerns (Gibbons & murphy (1992)) and that older CEOs are more risk-averse
due to career concerns (Chevalier & Ellison (1999), Ali & Zhang (2015), Andreou, Louca, &
Petrou (2017)). Similarly, Duru, Iyengar, & Zampelli (2016), and DeBoskey, Luo, & Zhou (2019)
argue that CEO tenure and CEO-chairman duality affect firms’ policies (i.e., earnings
announcement tone and firm performance.) Thus, we control for CEO tenure (Ln(Tenure)), CEO
age (Ln(Age)), and Duality. We control for CEO risk-taking incentive (Ln(Vega)) since prior
research finds a positive relation between Vega and cash holdings and a negative relation between
Vega and the value of cash to shareholders (Tong (2010), Liu & Mauer (2011), Aktas et al. (2019)).
Following Opler et al. (1999), we control for CEO ownership (CEO_Own) which is the percentage
of shares outstanding owned by a CEO. We also control for CEO pay-performance sensitivity
(Ln(Delta)) where Delta and Vega are measured following Guay (1999), and Core and Guay
(2002). All the control variables are in the lagged form to address the possible simultaneity
concerns (Kalcheva & Lins (2007)). In all our models, we include a set of dummy variables to
capture year (𝜏𝑡 ) and industry fixed effects (𝐼𝑗 ) to control for time-varying unobservable industry
characteristics, and clustered robust standard errors at the firm level are used. This model accounts
for correlation among unobservable time-invariant industry effects and explanatory variables,
unobservable aggregate time trends, and within-firm serial correlation. Table 3 presents the results
of our baseline regression model of the association between CEO marital status and cash holdings.
15

The dependent variables in models (1) and (2) in Table 3 are Cash, and NetCash, respectively.
The coefficient estimates of Single in models (1) and (2) are positive and statistically significant
indicating that firms headed by single CEOs hold more cash than those run by married CEOs. To
mitigate concerns related to simultaneity, we also use the one-year forwarded values, FCash, and
FNetCash in models (3) and (4), respectively.11 Similar to the models that use contemporaneous
values of cash measures, the coefficient estimates of Single are positive and statistically significant
in models (3) and (4). Specifically, on average, firms run by single CEOs hold about 1.0 to 1.8
percentage points more cash than those run by married CEOs. Taking the average of the
coefficients on Single, our results indicate that firms with single CEOs hold about 1.3 percentage
points more cash relative to firms with married CEOs. This impact is economically significant.
For the average firm in our sample that holds 14.4% of its total assets ($4.3 Billion) in cash, this
figure is equivalent to a 9% rise in cash holdings and translates into about $56 Million more cash
held by firms with single CEOs.12
The signs of the coefficients of other control variables are similar to prior studies. For instance,
we find a negative relationship between firm size (Ln(assets)) and cash holdings and a positive
relationship between firms’ growth opportunities (MB) and cash holdings. Similarly, high
dividend-paying firms hold less cash, whereas firms with higher cash flow volatility hold more
cash. Further, consistent with the agency problems framework, we find a positive and significant
effect of CEO tenure on cash holdings.
11
We note that for these two specifications, unlike the specification in equation (1), all the explanatory variables are
in time t since the dependent variable is shifted one time-step forward (t+1). In this setup all the explanatory variables
are essentially lagged.
12
It is worth noting that the average firm size in our sample is relatively smaller than that reported in Roussanov and
Savor (2014), possibly due to the exclusion of financial firms in our sample.
16

We note that the prediction that single CEOs would hoard more cash has ample theoretical
support in other fields, outside of financial economics. Psychology literature directly links marital
status and hoarding behavior. This literature indicates that several factors drive hoarding behavior
such as anxiety avoidance, fear of losing information, and feelings of pleasure associated with
possessions (Bubrick et al. (2004); Grisham et al. (2006)). Canale & Klontz (2013) note that
individuals hoard to shield themselves from potential future loss and avoid future uncertainty, and
argue that hoarding cash is similar to hoarding possessions. Grisham et al. (2006) document that
individuals who hoard are typically single and lack personal connections with others, leading to
stronger attachments to their possessions which become a part of their identity. Thus, the
psychology literature predicts that firms with single CEOs would hoard more cash than firms with
married CEOs.
3.2. Endogeneity issues
Our baseline results show a positive and statistically significant relationship between Single and
cash holdings. However, CEOs' marital status and cash holdings could be jointly correlated with
unobservable variables, such as CEO social, economic, physical, and demographic characteristics,
implying that our findings could be subject to the omitted variable bias. For instance, Xu,
Hudspeth, & Bartkowski (2005), and Uecker (2014) find a relation between the timing of marriage
and individual religious tradition, religious service attendance, and religious commitment. Other
scholars focus on marriage timing and economic factors (Oppenheimer (1988), Oppenheimer,
Kalmijn, & Lim (1997), Sweeney & Cancian (2004)) and marriage timing and cultural factors
(Lesthaeghe & Surkyn (1988), Jennings, Axinn, & Ghimire (2012)). Thus, the relationship
between single CEOs and cash holdings may be attributable to the innate heterogeneity in
preferences rather than marital status. Also, there is another concern that firms with single CEOs
are different from those with married CEOs, giving rise to selection bias. Lastly, Chen, Ho, and
17

Yeh (2020) show that corporate cash holding is affected by the level of CEO overconfidence. So,
our results could be affected by omitted CEO overconfidence. We address these concerns in the
following sections using a Difference-in-Differences (DiD) estimate, an instrumental variables
(IV) regression, a propensity score matching (PSM) approach, and a test that controls for CEO
overconfidence.
3.2.1. Corporate cash holdings around CEO turnover: A DiD test
To establish a stronger causal relationship between CEO marital status and cash holding, we
employ a DiD regression around the CEO turnover events. For this analysis, we first restrict the
sample to outgoing CEOs who had held their position for at least three years before their departure
and to incoming CEOs who held their position for at least 3 years after taking the position. This
setup assures that CEOs could imprint their personal preferences in corporate policies. Then, to
mitigate the effect of other confounding events, we restrict our DiD estimation window to the -3,
+3 years surrounding CEO turnover events. Table 4 presents the results of our DiD estimation
tests.
The dependent variable in models (1) and (2) is Cash, while the dependent variable in models
(3) and (4) is NetCash. After is a dummy variable equal to 1 for the post-turnover period and 0 for
the pre-turnover period. Treated is a dummy variable equal to 1 if a firm replaces a married CEO
with a single CEO, and 0 otherwise. Our variable of interest is the difference-in-difference
coefficient (After*Treated). Our classical agency problem hypothesis predicts a positive
coefficient on this interaction term since a change from a married CEO to a single CEO should
cause an increase in cash holding. All models include firm and CEO-level characteristics, and
industry and year fixed effects, and we cluster the standard error at the firm level. Focusing on the
18

first two columns, the coefficients of the interaction term After*Treated are positive and
statistically significant in both models indicating that firms that replaced a married CEO with a
single CEO experience an increase in cash holdings.
Further, our framework also predicts the opposite effect, a reduction in cash holdings, when a
married CEO replaces a single CEO. Thus, in models (3) and (4), Treated is equal to 1 if a firm
replaces a single CEO with a married CEO. The coefficients on the interaction term in both models
are significantly negative. We note that these results are not the mirror image of the results in the
first two columns, because a firm could replace a married CEO with another married CEO, or a
single CEO could be succeeded by another single CEO.
Further, in models (5)-(8) of Table 4, we repeat our DiD analysis surrounding the CEO
turnover events using a propensity score matched sample.13 Consistent with the results for the
overall sample, the coefficient of the interaction term After*Treated is positive and statistically
significant in models (5) and (6), and significantly negative in models (7) and (8). These results
suggest that in a matched sample, firms replacing a married CEO with a single CEO experience
an increase in cash holdings while cash holdings decrease when a married CEO succeeds a single
CEO. Together, these results provide strong empirical support for the association between CEO
marital status and corporate cash holding and enhance the causal interpretation of our results.
3.2.2. Instrumental variables approach
To mitigate the concern of innate heterogeneity, we employ an instrumental variables approach.
Following Roussanov & Savor (2014), we examine the cost of getting married by differentiating
between firms located in states that adopt community property and those that adopt equitable
division divorce laws. In the community property system, a divorce would result in an equal
13
We discuss our matching procedure in detail in Section 3.2.3.
19

division of the properties of a married couple, whereas the division of property is determined by a
judge based on multiple factors in the equitable division system. Since, on average, CEOs are
expected to be wealthier than their spouses, marriage will be costlier for them under community
property laws. Therefore, a CEO is more likely to be single if he/she resides in a state with a
community property law. Further, it is unlikely that this law can affect corporate cash holdings
other than through its impact on CEO’s marital status. Thus, following Roussanov & Savor (2014),
we employ the instrument, Community, which is an indicator variable equal to 1 if a firm is located
in a state that has adopted the community property system, and zero otherwise.14 Table 5 presents
the results of our instrumental variables regression tests.
Model 1 in Table 5 reports the results of the first stage estimate. Even though the community
property system has been adopted regardless of the economic conditions of the state, we control
for state-level real per capita income and Coincident Economic Activity Index (CEAI) to address
the possibility that state-level economic conditions might predict the individuals’ marital status.15
In the first stage (model (1)), we find that the coefficient of community is positive and statistically
significant (i.e., 𝛽1= 0.126 with a t-stat = 1.99), indicating that CEOs who are residing in a
community property state are more likely to be single.
Our second stage test results are reported in model (2) in Table 5 where the dependent variable
is Cash and the explanatory variable is the predicted values from the first stage regression
Singlepred. The coefficient on the instrumented Single, Singlepred, is positive and statistically
14
In the United States, nine states have adopted the community property system: Arizona, California, Idaho, Louisiana,
Nevada, New Mexico, Texas, Washington, and Wisconsin. Puerto Rico is also a community property jurisdiction.
15
Roussanov & Savor (2014) argue that the community property system is not related to the state characteristics such
as rich vs poor states, large vs small states or political preference of the populations, such as liberal vs conservative.
In addition, they argue that as the community property system was adopted well before the sample period (1993-
2008), the potential endogeneity concern of political economy consideration is likely to be muted for these states.
20

significant. Overall, our findings indicate that CEOs’ marital status is also an important
determinant of corporate cash holdings. Moreover, the instrumental variable approach potentially
mitigates the effect of the CEO's innate heterogeneity that could potentially bias our estimates. It
also reduces the possible effect of measurement error bias that could arise due to the classification
of CEOs' marital status into “single” versus “married.” 16
3.2.3. Propensity Score Matching (PSM)
In this section, we first examine the differences in the characteristics of firms with single CEOs
and those with married CEOs.
Statistics in Panel A of Table 6 show that the differences in cash holdings (both Cash and
NetCash) are positive and statistically significant which indicates that single CEOs have higher
cash holdings compared to married CEOs. Additional descriptive statistics demonstrate that firms
run by single CEOs have a higher investment in R&D and Acquisition and have higher
Industry_sigma, which are consistent with the risk-taking preference by single CEOs documented
by Roussanov & Savor (2014). Further, these descriptive statistics also reveal that Single CEOs
are, on average, younger and have less wealth and share ownership compared to married CEOs,
which is consistent with the literature that relates wealth to success in the market for marriage
(Korenman & Neumark (1991), Chun & Lee (2001), Roussanov & Savor (2014), Nicolosi and
Yore (2015)). Moreover, firms run by single CEOs are on average smaller, have high growth
potential (MB), and have lower cash flows compared to those run by married CEOs. These
16
We note that the coefficient estimated using the IV approach is substantially larger than those reported in Table 3
which could be due to the IV model identifying the local average treatment effect of the endogenous variable on the
outcome variable (Imbens & Angrist, 1994). Moreover, Jiang (2017) also argues that “there is no reason to expect that
the causal effects in close to 85% of all the cases studied by researchers should be predominantly higher than the
simple correlational effect”. Thus, while the results of this IV estimation alleviates the endogeneity concerns, we are
hesitant to draw conclusion on the magnitude of the effect.
21

significant differences between single-run and married-run firms may raise selection bias concerns
with our baseline results. Although we control for firm and CEO characteristics in our regression
models, we alleviate this concern by applying the propensity score matching (PSM) approach. In
this section and onward, we run our analyses using the PSM sample.
Specifically, we employ PSM and match firms based on their characteristics and their CEO
attributes to mitigate selection bias (Rosenbaum & Rubin (1983), Roberts & Whited (2013)). We
also include year and industry fixed effects in this matching process. In the first step of our PSM,
we estimate a logistic regression of Single on multiple firms and CEO characteristics. Using the
estimated coefficients from the logistic model in the first step, we then compute the propensity
scores (i.e., the predicted likelihood of Single = 1) for all firms in our sample. We then match firms
with Single CEOs (treatment groups) with those with married CEOs (control groups) using the
nearest neighbor-matching technique where one married CEO firm-year enters into the sample
only once. Panel B of Table 6 shows the differences in the characteristics of treatment and control
for the matched sample. There are no significant differences between the treatment and control
groups, indicating that the matching was conducted successfully. In the next step, we re-estimate
our baseline regression models using the PSM sample and report the results in Table 7.
The dependent variables in Table 7 are Cash, NetCash, FCash, and FNetCash, in turn. Models
(1-4) include firm characteristics, while models (5-8) include both firm and CEO characteristics.
All models include industry and year fixed effects, and we cluster standard error at the firm level.
Consistent with our baseline results (in Table 3), we find a positive and statistically significant
association between Single and cash holdings in all regression models for the PSM sample,
alleviating concerns that our findings may suffer from selection bias.
22

4. The precautionary vs. the classical agency hypotheses
Our finding that single CEOs hold more cash than married CEOs is consistent with our classical
agency problem hypothesis. However, CEOs may hold excess cash for precautionary motives,
giving rise to the concern that our results could have an alternative interpretation based on the
precautionary motive framework rather than the agency problem. In this section, we attempt to
rule out this possibility.
4.1. constrained vs. unconstrained firms
Prior studies document that single CEOs are less risk-averse. For instance, firms run by single
CEOs experience higher stock return volatility, pursue more aggressive investment policies, are
less likely to respond to changes in idiosyncratic risk, and behave less conservatively (Roussanov
and Savor (2014), Lu, Ray, and Teo (2016), Belenzon, Patacconi, and Zarutskie (2016)). Because
single CEOs are less risk-averse, their cash holdings policies are less likely to be motivated by
precautionary motives. However, any rational managers might hold more cash when facing
financial constraints. Han & Qiu (2007), and Denis & Sibilkov (2009) argue that financially
constrained firms hold more cash compared to unconstrained firms due to precautionary motives.
They show that higher future cash flow volatility and constraints on external financing induce
financially constrained firms to hold excess cash. Similarly, Almedia et al. (2004) show that
financially constrained firms save more cash out of their cash flow. Consistent with this argument,
financial constraints may encourage managers to increase cash holdings due to precautionary
motives regardless of their marital status. Thus, to formally rule out the possibility that our results
are driven by the precautionary motives associated with financial constraints, we conduct a cross-
sectional test for subsamples of financially constrained and unconstrained firms. While cash-
holding policies in financially constrained firms can be greatly shaped by precautionary motives,
23

the same is not true for financially unconstrained firms. In other words, finding higher cash
holdings for unconstrained firms with single CEOs would refute the possibility that the
precautionary motive is the main driver of our results.
Specifically, we use five different proxies to identify financially constrained firms: Dividends,
annual payout, bond rating, commercial paper rating, and KZ index. The financially constrained
[FC] (Financially unconstrained [Non-FC]) subgroup includes firms not paying (do paying)
dividends, below (above) median values of Annual Payout, below (above) S&P long-term credit
ratings investment grade, below (above) S&P short-term credit ratings investment grade, and
above (below) medium scores on the KZ index. Table 8 presents the results for this cross-section
test using our full sample, reported in the left panels, as well as the PSM sample, reported in the
right panels.
The dependent variables in Table 8 are Cash and NetCash, in turn. All regression models
include firm and CEO characteristics, year and industry fixed effects, and we cluster the standard
errors at the firm level. Focusing on the left panels, models 1 to 4, across all measures of financial
constraints, we report a positive and significant coefficient for Single only in the subsamples of
financially unconstrained firms. This finding cannot be explained by the precautionary motives
which are expected to have a negligent effect on cash holdings policies of financially unconstrained
firms. In contrast, the coefficient estimates of Single are not statistically significant in the
financially constrained subsamples. This result indicates that in the presence of financial
constraints, there is no difference in firms’ cash holdings due to CEO’s marital status. It seems
that, regardless of their marital status, managers in financially constrained firms are facing a more
binding cash-holding policy that restricts their ability to implant their preferences into firms’ cash-
24

holding policies.17 We find similar results in the right panels, where we use the PSM sample. For
example, focusing on the results based on the KZ index, we find that firms with single CEOs hold
about 2 percentage points more cash relative to otherwise comparable firms with married CEOs
only when the firms are unconstrained. These results are consistent with our earlier findings and
further alleviate the concerns about our findings being spurious. Overall, these results suggest that
the potential reason that Single CEOs hold more cash in these unconstrained firms cannot be
attributed to precautionary motives (which may affect the constrained sample), but rather they are
more consistent with our classical agency problem hypothesis. In the next section, we further
explore this conjecture.
4.2. Firm-specific risk
Similar to financially constrained firms, managers of firms with high specific risk might hold
excess cash for precautionary motives. Consequently, an alternative way to rule out the
precautionary hypothesis is to investigate our results in subgroups of firms with high vs. low levels
of firm-specific risk. If the precautionary motives are the reason behind the excess cash holdings
of single CEOs, we should observe stronger results in the high firm-specific risk subsample. To
formally test this conjecture, we use four proxies to capture firm-specific risk; Idiosyncratic
Volatility which is the standard deviation of the residual from a regression of monthly stock returns
on the CRSP value-weighted market portfolio return, Total Volatility which is the annualized
standard deviation of monthly stock returns over the previous year, Industry_sigma which is the
average of the standard deviation of the ratio of cash flow to book value of assets over the last 10
17
It is worth noting that our un-tabulated statistics show that financially unconstrained firms (regardless of the marital
status of the CEO) have lower average cash holdings ratio than financially constrained firms. On average, financially
constrained firms have, a cash holding ratio of 19-23%, while financially unconstrained firms have, a cash holding
ratio of 8-11% (depending on a CEO’s marital status). These statistics are consistent with Han & Qiu (2007), and
Denis & Sibilkov (2009).
25

years of firms in the same 2-digit SIC code industries, and Leverage which is the ratio of total debt
divided by the book value of total assets. We report the results of this test in Table 9.
Panels A to D in Table 9 report results for Idiosyncratic Volatility, Total Volatility,
Industry_sigma, and Leverage, respectively. Models (1) and (2) report the results for high firm-
specific risk, while models (3) and (4) report the results for low firm-specific risk. Consistent with
the results of the financial constraints test (Table 8), the coefficient estimates of Single are positive
and statistically significant only in the models of low firm-specific risk. This finding is robust
across all four measures of firm-specific risk. Since the precautionary motives are not expected to
play a vital role in determining cash holdings of the low-risk firms, these results provide strong
evidence against the precautionary hypothesis as an alternative explanation of our results.
5. Potential channels and funding of risky investments
5.1. Tax avoidance and Dividend policy
In this section, we investigate a few self-centered behaviors that may explain the mechanism for
the positive association between single CEOs and corporate cash holdings. Self-centered CEOs
might attempt to pump their performance-based compensation by aggressively engaging in tax
avoidance. The traditional view suggests that corporate tax avoidance can increase shareholder
value by transferring wealth from the government to shareholders. Managers, particularly single
CEOs who suffer from higher personal income tax levels, also benefit from corporate tax
avoidance activities in the form of higher compensation because tax avoidance potentially reduces
the overall tax burden and increases the firm’s net income. Consequently, the notion that single
CEOs are more self-centered would predict a positive association between single CEO and
corporate tax avoidance, possibly enabling them to receive higher compensation. Prior studies
26

document a positive association between corporate cash holdings and tax avoidance (Foley et al.
(2007), Wang (2015), Khuong et al. (2019)). Thus, we investigate the association between CEO
marital status and tax avoidance and present the results in Panel A of Table 10.
The dependent variables in Panel A of Table 10 are measures of corporate tax avoidance.
Following prior studies, we employ two different proxies for corporate tax avoidance. Our first
proxy is DTAX which is the discretionary permanent book-tax difference for firm i in year t (Frank
et al. (2009)). Following Wilson (2009), our second proxy is BTD which is the total book-tax
difference defined as book income less taxable income scaled by lagged assets. The coefficient of
Single is positive and statistically significant in all models, indicating that firms run by single
CEOs are more likely to avoid taxes than those run by married CEOs. In sum, this set of results
suggests that tax avoidance could be a potential mechanism for the higher levels of cash holdings
among firms led by single CEOs.
Next, we investigate corporate payout policy as another mechanism through which the agency
problem in firms headed by single CEOs may manifest itself. Prior studies argue that asset
liquidity, such as cash and/or cash equivalent assets is a key determinant of a firm’s dividends
policy (Holder, Langrehr, & Hexter (1998), Aivazian, Booth, & Cleary (2003), Ferreira & Vilela
(2004)). Ozkan & Ozkan (2004) suggest that there might be a positive relationship between cash
holdings and dividends. Further, Jiang & Lie (2016) argue that self-interested managers are
reluctant to disburse excess cash allowing cash levels to stockpile unless the firms face external
pressure. Motivated by these studies, we argue that the self-centered single CEOs are expected to
be more reluctant to pay dividends than married CEOs, predicting a negative relationship between
Single and measures of dividend payment. To test this conjecture, we use multiple measures of
27

dividend payment, including dividends dummy, dividends to assets ratio, dividends to cash flow
ratio, dividend yield, natural log of dividends, and dividends per share. Panels B1 and B2 in Table
10 report the result of this test. We find that the coefficients of Single in all models, whether we
use the full or the PSM sample, are significantly negative, indicating that firms run by single CEOs
are less likely to pay dividends and on average, they pay lower dividends compared to otherwise
similar firms run by married CEOs. This set of results also provides evidence that higher levels of
cash held by single CEOs could be explained, at least partly, by the lower amount of dividend
payout. We note that this result is also more consistent with our classical agency problem
hypothesis.
Overall, these tests indicate that single CEOs engage in more tax avoidance and pay less
dividends compared to married CEOs, possibly to enrich themselves in the form of higher
compensation and entrench themselves through holding larger piles of cash, potentially indicating
elevated agency issues for firms headed by single CEOs. Tax avoidance and lower dividend
payouts can then be viewed as two potential underlying channels that may explain the positive
association between Single and cash holdings we report in this paper. Further, these results are
more consistent with our classical agency problem, rather than the precautionary hypothesis.
5.2. Funding aggressive investment policies
Our main result that firms with single CEOs tend to hold more cash seems to contradict the main
result of Roussanov and Savor (2014) that firms with single CEOs tend to be riskier and pursue
more aggressive investment policies. Also, their main result raises the question of how single
CEOs fund these risky projects. In this section, we provide evidence that reconciles this seeming
contradiction.
28

The agency problems framework (Jensen (1986)) suggests that corporate insiders generally
have a personal incentive to retain cash rather than increase dividends and that they rely on cash
and liquid assets to finance their risky projects, allowing them to avoid the market scrutiny
necessary for raising external finance. This framework predicts that compared to married CEOs,
single CEOs should raise a significantly lower amount of external finance and would rely more on
cash holdings to fund future risky and aggressive investment policies. In this section, we provide
empirical evidence consistent with these predictions.
We follow Roussanov and Savor (2014) and Guay (1999) and use R&D as a proxy for risky
investment. Roussanov and Savor (2014) argue that R&D investment is a proxy for firm risk-
taking behavior while Guay (1999) uses R&D as a measure for CEO risk-taking. We use a lead-
lag regression setup in panel A of Table 11 and compare the link between cash holdings in year t-
1 and R&D investment in year t between firms with single and married CEOs. We find that the
association between Casht-1 and R&Dt is twice as strong for firms with single CEOs (β=0.127 with
t-stat= 2.86 for the single subset vs. β=0.063 with t-stat= 3.9 for the married subset) after
controlling for the lagged value of R&D as well as a set of firm and CEO characteristics. This
result continues to hold when we control for future values of cash holdings. Removing the lagged
value of R&D or including more lags of cash does not alter this result.18 These results also reconcile
the seeming contradiction between our findings and those of Roussanov and Savor (2014). While
they show single CEOs are less risk averse and increase the riskiness of firms by pursuing
aggressive investment policies, we show that they hoard cash and rely on this internal source to
fund these projects and thereby avoid market scrutiny.
18
We have not reported these results for brevity.
29

In panel B of Table 11, we examine whether single CEOs raise significantly less external
finance to avoid market scrutiny as predicted by the agency framework. Our measure of external
finance is calculated as the sale of common and preferred stocks minus the purchase of common
and preferred stocks plus the issuance of long-term debt minus the reduction of long-term debt
(Leary and Roberts (2014)). We regress the natural log of this variable on the single dummy along
with a set of variables controlling for firm and CEO characteristics. Whether we estimate the
regression using the full sample or the PSM sample, the coefficient on the single dummy is
negative and statistically significant. Specifically, firms with single CEOs raise about 20% less
external finance than otherwise similar firms with married CEOs (β=-0.198 with t-stat= 3.20).
The stronger association between cash holdings and future R&D expenditure for firms with
single CEOs combined with significantly lower external financing by these firms and lower
dividend payout are all consistent with our classical agency problem hypothesis.
5.3. Single CEO’s cash holdings and firm value
A natural question that follows immediately from the above discussion is how would the excess
cash held by single CEOs, presumably due to the agency problem, affect firm value? Prior studies
show that shareholders discount the value of excess cash holdings when there are concerns about
agency problems. For instance, Harford (1999) argues that cash-rich firms are more likely to make
diversifying acquisitions that are value-destroying and experience abnormal declines in operating
performance. In a similar vein, Dittmar and Mahrt-Smith (2007) document that cash in poorly
governed firms is discounted heavily. Martínez-Sola, García-Teruel, & Martínez-Solano (2013)
find an optimal level of cash holdings and argue that deviating from that optimal level of cash
holdings decreases the firm value. Based on these arguments, we examine whether and how cash
holdings by single CEOs affect firm value. Our classical agency problem hypothesis predicts that
the value of cash held by single CEOs should be discounted by shareholders.

30

To examine the above conjecture, we employed various cash valuation models proposed by
prior literature. First, we employ the valuation of the excess cash model following Dittmar and
Mahrt-Smith (2007), where excess cash is measured as the residual from the normal level of cash
model based on Opler, Pinkowitz, Stulz, and Williamson (1999).19 Next, we employ the valuation
models of change in, and level of, cash following Pinkowitz and Williamson (2007).20 Lastly, we
run a model that captures the association between our baseline measure of cash, Cash, and Tobin’s
Q. We run each of these valuation models of cash separately for subgroups of firms with single
and married CEOs and report the results in Table 12.
The test that follows Dittmar and Mahrt-Smith’s (2007) valuation model of excess cash is
reported in Panel A of Table 12. For the PSM sample, the excess cash held by single CEOs is
valued at around 16 percent less as compared with cash held by married CEOs (β=1.191 with t-
stat= 3.73 for the single subset vs. β=1.385 with t-stat= 3.87 for the married subset.) A similar, but
more economically significant, results are reported when using Pinkowitz and Williamson’s
(2007) change in, and level, of cash valuation models which are reported in Panels B and C of
Table 12. In Panel B, the PSM coefficient estimates indicate that a lagged change in cash by single
CEOs is valued at around 30 percent less as compared to a similar change in cash by married CEOs
(β=2.195 with t-stat= 6.00 for the single subset vs. β=2.852 with t-stat= 8.45 for the married
subset.) Similarly, In Panel C, the PSM coefficient estimates indicate that cash held by single
CEOs is valued at around 30 percent less than cash held by married CEOs (β=1.733 with t-stat=
19
The specification used in this test is similar to model (2) in Dittmar and Mahrt-Smith (2007), PP 616. Due to the
use of excess cash, this model suffers less from the endogeneity concern inherent in the other cash valuation models.
20
The specification of change in the value of cash is similar to model (1), while the specification of level of cash is
similar to model (2) in Pinkowitz and Williamson (2007), PP 76.
31

5.09 for the single subset vs. β=2.270 with t-stat= 7.44 for the married subset.) This finding is
robust to the use of Tobin’s Q as an alternative measure of firm value in Panel D of Table 12.
These findings are consistent whether we use the matched sample or the full sample. The
coefficient on cash for the single subset is significantly lower (by about 50%) than that for the
married subset, indicating that the cash held by single CEOs is discounted by the shareholders.
These findings are consistent with our agency hypothesis and suggest that firms headed by single
CEOs hold excess cash at the expense of shareholder value.
6. Conclusion
This study examines the relationship between CEO marital status and corporate cash holdings.
Our empirical analysis provides evidence that firms headed by single CEOs hold more cash than
those by married CEOs. To address the concern that our results are driven by CEO innate
heterogeneity, rather than their marital status, we first employ a Difference-in-Differences (DiD)
estimation around the CEO turnover events and find that firms that replaced a married CEO with
a single CEO experience an increase in cash holdings. We also document the opposite effect, a
decrease in cash holdings, when a single CEO is replaced with a married one. Second, we use the
instrumental variables (IV) regression approach and find that our baseline regression results
continue to hold. Third, we use the propensity score matching (PSM) approach to address the
selection bias concern. We also test our DiD estimate for the matched sample and find consistent
results to our baseline findings. Further, we control for CEO overconfidence which has been shown
in previous research to affect corporate cash holdings. Together, these results are more consistent
with and suggestive of a causal relationship between CEO marital status and corporate cash
holdings.
32

We conduct a battery of additional tests to investigate the precautionary vis-à-vis the agency
hypotheses. To investigate the precautionary explanation of our results, we first test the association
between CEO marital status and cash holdings for financially constrained and unconstrained firms
and find robust higher levels of cash holdings for unconstrained firms with single CEOs compared
to otherwise similar firms with married CEOs. Next, we test the association between CEO marital
status and cash holdings for subgroups of firms with high vs. low levels of firm-specific risk. The
results of this test show that firms with single CEOs hold more cash than those with married CEOs
in the subsample of firms with low firm-specific risk. Together, these tests refute the possibility
that single CEOs hold cash for precautionary purposes. Consistent with the agency hypothesis, we
find that firms with single CEOs engage in more corporate tax avoidance activities, pay smaller
dividends, and hold more cash to fund future aggressive investment policies (Roussanov and Savor
(2014)). Lastly, our cash valuation models indicate that excess cash held by single CEOs is
discounted heavily by shareholders as compared to that held by married CEOs.
Overall, we find a robust relationship between CEO marital status and corporate cash holdings
and argue that CEOs' marital status plays a crucial role in shaping their corporate policies. Our
results contribute to the growing literature that provides evidence that corporate policies are
influenced by factors outside of macroeconomic conditions or firm characteristics.
33

References
Acharya, V. V., Almeida, H., & Campello, M. (2007). Is cash negative debt? A hedging perspective on
corporate financial policies. Journal of Financial Intermediation, 16(4), 515-554.
Adhikari, B. K. (2018). Female executives and corporate cash holdings. Applied Economics
Letters, 25(13), 958-963.
Aivazian, V., Booth, L., & Cleary, S. (2003). Do emerging market firms follow different dividend policies
from US firms? Journal of Financial Research, 26(3), 371-387.
Aktas, N., Louca, C., & Petmezas, D. (2019). CEO overconfidence and the value of corporate cash
holdings. Journal of Corporate Finance, 54, 85-106.
Ali, A., & Zhang, W. (2015). CEO tenure and earnings management. Journal of Accounting and
Economics, 59(1), 60-79.
Al Mamun, Md, Boubaker S., Ghafoor, A., & Suleman, M. T. (2021) Is Marriage a Turning Point? Evidence
from Cash Hoardings Behavior of Married CEOs (November 10, 2021). Available at
SSRN: https://ssrn.com/abstract=3960753 or http://dx.doi.org/10.2139/ssrn.3960753
Almeida, H., Campello, M., & Weisbach, M. S. (2004). The cash flow sensitivity of cash. The Journal of
Finance, 59(4), 1777-1804.
Andreou, P. C., Louca, C., & Petrou, A. P. (2017). CEO age and stock price crash risk. Review of
Finance, 21(3), 1287-1325.
Bates, T. W., Kahle, K. M., & Stulz, R. M. (2009). Why do US firms hold so much more cash than they
used to? The journal of finance, 64(5), 1985-2021.
Belenzon, S., Patacconi, A., & Zarutskie, R. (2016). Married to the firm? A large‐scale investigation of the
social context of ownership. Strategic Management Journal, 37(13), 2611-2638.
Benmelech, E., & Frydman, C. (2015). Military CEOs. Journal of Financial Economics, 117(1), 43-59.
Bubrick, J., Neziroglu, F., & Yaryura-Tobias, J. (2004). Overcoming compulsive hoarding: Why you save
and how you can stop. New Harbinger Publications.
Campbell, K. E., & Lee, B. A. (1992). Sources of personal neighbor networks: social integration, need, or
time?. Social forces, 70(4), 1077-1100.
Campbell, T.C., Gallmeyer, M., Johnson, S.A., Rutherford, J., Stanley, B.W., 2011. CEO optimism and
forced turnover. J. Financ. Econ. 101, 695-712.
Cain, M. D., & McKeon, S. B. (2016). CEO personal risk-taking and corporate policies. Journal of Financial
and Quantitative Analysis, 51(1), 139-164.
Canale, A., & Klontz, B. (2013). Hoarding disorder: It’s more than just an obsession-implications for
financial therapists and planners. Journal of Financial Therapy Volume, 4(2).
Chen, Y. R., Ho, K. Y., & Yeh, C. W. (2020). CEO overconfidence and corporate cash holdings. Journal
of Corporate Finance, 62, 101577.
Chevalier, J., & Ellison, G. (1999). Career concerns of mutual fund managers. The Quarterly Journal of
Economics, 114(2), 389-432.
Chun, H., & Lee, I. (2001). Why do married men earn more: productivity or marriage selection?. Economic
Inquiry, 39(2), 307-319.
Chyz, J. A. (2013). Personally tax aggressive executives and corporate tax sheltering. Journal of Accounting
and Economics, 56(2-3), 311-328.
Coles, J. L., Daniel, N. D., & Naveen, L. (2014). Co-opted boards. The Review of Financial Studies, 27(6),
1751-1796.
Core, J., & Guay, W. (2002). Estimating the value of employee stock option portfolios and their sensitivities
to price and volatility. Journal of Accounting Research, 40(3), 613-630.
Cronqvist, H., Makhija, A. K., & Yonker, S. E. (2012). Behavioral consistency in corporate finance: CEO
personal and corporate leverage. Journal of financial economics, 103(1), 20-40.
34

Dambrun, M., & Ricard, M. (2011). Self-Centeredness and Selflessness: A Theory of Self-Based
Psychological Functioning and Its Consequences for Happiness. Review of General Psychology, 15(2),
138–157.
DeBoskey, D. G., Luo, Y., & Zhou, L. (2019). CEO power, board oversight, and earnings announcement
tone. Review of Quantitative Finance and Accounting, 52(2), 657-680.
Denis, D. J., & Sibilkov, V. (2009). Financial constraints, investment, and the value of cash holdings. The
Review of Financial Studies, 23(1), 247-269.
DePaulo, B. M., & Morris, W. L. (2005). Singles in society and in science. Psychological Inquiry, 16(2-3),
57-83.
Dittmar, A., Mahrt-Smith, J., & Servaes, H. (2003). International corporate governance and corporate cash
holdings. Journal of Financial and Quantitative analysis, 38(1), 111-133.
Dittmar, A., & Mahrt-Smith, J. (2007). Corporate governance and the value of cash holdings. Journal of
financial economics, 83(3), 599-634.
Di Giuli, A., & Kostovetsky, L. (2014). Are red or blue companies more likely to go green? Politics and
corporate social responsibility. Journal of Financial Economics, 111(1), 158-180.
Duchin, R. (2010). Cash holdings and corporate diversification. The Journal of Finance, 65(3), 955-992.
Duong, H. N., Nguyen, J. H., Nguyen, M., & Rhee, S. G. (2020). Navigating through economic policy
uncertainty: The role of corporate cash holdings. Journal of Corporate Finance, 62, 101607.
Duru, A., Iyengar, R. J., & Zampelli, E. M. (2016). The dynamic relationship between CEO duality and
firm performance: The moderating role of board independence. Journal of Business Research, 69(10),
4269-4277.
Elnahas, A. M., & Kim, D. (2017). CEO political ideology and mergers and acquisitions decisions.
Journal of Corporate Finance, 45, 162-175.
Faulkender, M., & Wang, R. (2006). Corporate financial policy and the value of cash. The Journal of
Finance, 61(4), 1957-1990.
Ferreira, M. A., & Vilela, A. S. (2004). Why do firms hold cash? Evidence from EMU countries. European
Financial Management, 10(2), 295-319.
Foley, C. F., Hartzell, J. C., Titman, S., & Twite, G. (2007). Why do firms hold so much cash? A tax-based
explanation. Journal of financial economics, 86(3), 579-607.
Frank, M. M., Lynch, L. J., & Rego, S. O. (2009). Tax reporting aggressiveness and its relation to aggressive
financial reporting. The Accounting Review, 84(2), 467-496.
Gao, H., Harford, J., & Li, K. (2013). Determinants of corporate cash policy: Insights from private
firms. Journal of Financial Economics, 109(3), 623-639.
Gibbons, R., & Murphy, K. J. (1992). Optimal incentive contracts in the presence of career concerns:
Theory and evidence. Journal of Political Economy, 100(3), 468-505.
Grisham, J. R., Frost, R. O., Steketee, G., Kim, H. J., & Hood, S. (2006). Age of onset of compulsive
hoarding. Journal of anxiety disorders, 20(5), 675-686.
Guay, W. R. (1999). The sensitivity of CEO wealth to equity risk: an analysis of the magnitude and
determinants. Journal of Financial Economics, 53(1), 43-71.
Hallahan, T. A., Faff, R. W., & McKenzie, M. D. (2004). An empirical investigation of personal financial
risk tolerance. Financial Services Review-greenwich-, 13(1), 57-78.
Han, S., & Qiu, J. (2007). Corporate precautionary cash holdings. Journal of Corporate Finance, 13(1), 43-
57.
Harford, J. (1999). Corporate cash reserves and acquisitions. The Journal of Finance, 54(6), 1969-1997.
Harford, J., Mansi, S. A., & Maxwell, W. F. (2008). Corporate governance and firm cash holdings in the
US. Journal of financial economics, 87(3), 535-555.
Hilary, G., Huang, S., & Xu, Y. (2017). Marital status and earnings management. European Accounting
Review, 26(1), 153-158.
35

Hirshleifer, D., Low, A., Teoh, S.H, 2012. Are overconfident CEOs better innovators? J. Finan. 67, 1457-
1498.
Holder, M. E., Langrehr, F. W., & Hexter, J. L. (1998). Dividend Policy Determinants: An Investigation of
the Influences of. Financial management, 27(3), 73-82.
Hong, J. H., & Rios-Rull, J. V. (2012). Life insurance and household consumption. American Economic
Review, 102(7), 3701-30.
Hossain, M. N., Elnahas, A., Gao, L., & Kim J.B. (2022). CEO political ideology and voluntary forward-
looking disclosure. Available at SSRN 3637608.
Hutton, I., Jiang, D., & Kumar, A. (2014). Corporate policies of Republican managers. Journal of Financial
and Quantitative Analysis, 49(5-6), 1279-1310.
Imbens, G., J. Angrist. 1994. Identification and estimation of local average treatment effects. Econometrica,
62, pp. 467-475
Javadi, S., Mollagholamali, M., Nejadmalayeri, A., Al-Thaqeb, S. 2021 “Corporate Cash Holdings, Agency
Problems and Economic Policy Uncertainty,” International Review of Financial Analysis, Vol. 77,
101859
Jensen, M. C. (1986). Agency costs of free cash flow, corporate finance, and takeovers. The American
economic review, 76(2), 323-329.
Jennings, E. A., Axinn, W. G., & Ghimire, D. J. (2012). The effect of parents’ attitudes on sons’ marriage
timing. American sociological review, 77(6), 923-945.
Jiang, W., 2017. Have instrumental variables brought us closer to the truth? Review of Corporate Finance
Studies, 6 (2) pp. 127-140
Jiang, Z., & Lie, E. (2016). Cash holding adjustments and managerial entrenchment. Journal of Corporate
Finance, 36, 190-205.
Kalcheva, I., & Lins, K. V. (2007). International evidence on cash holdings and expected managerial agency
problems. The Review of Financial Studies, 20(4), 1087-1112.
Kaplan, S. N., & Zingales, L. (1997). Do investment-cash flow sensitivities provide useful measures of
financing constraints?. The quarterly journal of economics, 112(1), 169-215.
Khuong, N. V., Ha, N. T. T., Minh, M. T. H., & Thu, P. A. (2019). Does corporate tax avoidance explain
cash holdings? The case of Vietnam. Economics & Sociology, 12(2), 79-93.
Korenman, S., & Neumark, D. (1991). Does marriage really make men more productive?. Journal of
Human Resources, 282-307.
Leary, M. T., & Roberts, M. R. (2014). Do peer firms affect corporate financial policy?. The Journal of
Finance, 69(1), 139-178.
Lesthaeghe, R., & Surkyn, J. (1988). Cultural dynamics and economic theories of fertility
change. Population and development review, 1-45.
Liu, Y., & Mauer, D. C. (2011). Corporate cash holdings and CEO compensation incentives. Journal of
financial economics, 102(1), 183-198.
Love, D. A. (2009). The effects of marital status and children on savings and portfolio choice. The Review
of Financial Studies, 23(1), 385-432.
Lu, Y., Ray, S., & Teo, M. (2016). Limited attention, marital events and hedge funds. Journal of Financial
Economics, 122(3), 607-624.
Malmendier, U., & Tate, G. (2005). CEO overconfidence and corporate investment. The journal of
finance, 60(6), 2661-2700.
Martínez-Sola, C., García-Teruel, P. J., & Martínez-Solano, P. (2013). Corporate cash holding and firm
value. Applied Economics, 45(2), 161-170.
Nguyen, H. T., Phan, H. V., & Sun, L. S. (2018). Shareholder litigation rights and corporate cash holdings:
Evidence from universal demand laws. Journal of Corporate Finance, 52, 192-213.
36

Nicolosi, G., & Yore, A. S. (2015). “I do”: does marital status affect how much CEOs “do”?. Financial
Review, 50(1), 57-88.
Nikolov, B., & Whited, T. M. (2014). Agency conflicts and cash: Estimates from a dynamic model. The
Journal of Finance, 69(5), 1883-1921.
Opler, T., Pinkowitz, L., Stulz, R., & Williamson, R. (1999). The determinants and implications of
corporate cash holdings. Journal of financial economics, 52(1), 3-46.
Oppenheimer, V. K. (1988). A theory of marriage timing. American journal of sociology, 94(3), 563-591.
Oppenheimer, V. K., Kalmijn, M., & Lim, N. (1997). Men’s career development and marriage timing
during a period of rising inequality. Demography, 34(3), 311-330.
Ozkan, A., & Ozkan, N. (2004). Corporate cash holdings: An empirical investigation of UK
companies. Journal of banking & finance, 28(9), 2103-2134.
Phan, H. V., Nguyen, N. H., Nguyen, H. T., & Hegde, S. (2019). Policy uncertainty and firm cash
holdings. Journal of Business Research, 95, 71-82.
Pinkowitz, L., Stulz, R., & Williamson, R. (2006). Does the contribution of corporate cash holdings and
dividends to firm value depend on governance? A cross‐country analysis. The Journal of
Finance, 61(6), 2725-2751.
Pinkowitz, L., & Williamson, R. (2007). What is the market value of a dollar of corporate cash?. Journal of
Applied Corporate Finance, 19(3), 74-81.
Riley Jr, W. B., & Chow, K. V. (1992). Asset allocation and individual risk aversion. Financial Analysts
Journal, 48(6), 32-37.
Roberts, M. R., & Whited, T. M. (2013). Endogeneity in empirical corporate finance1. In Handbook of the
Economics of Finance (Vol. 2, pp. 493-572). Elsevier.
Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies
for causal effects. Biometrika, 70(1), 41-55.
Roussanov, N., & Savor, P. (2014). Marriage and managers' attitudes to risk. Management Science, 60(10),
2496-2508.
Roszkowski, M. J., Snelbecker, G. E., Leimberg, S. R., & Leimberg, S. R. (1993). The tools and techniques
of financial planning.
Slater, P. E. (1963). On social regression. American sociological review, 339-364.
Sunder, J., Sunder, S. V., & Zhang, J. (2017). Pilot CEOs and corporate innovation. Journal of Financial
Economics, 123(1), 209-224.
Sweeney, M. M., & Cancian, M. (2004). The changing importance of white women's economic prospects
for assortative mating. Journal of Marriage and Family, 66(4), 1015-1028.
Tong, Z. (2010). CEO risk incentives and corporate cash holdings. Journal of Business Finance &
Accounting, 37(9‐10), 1248-1280.
Uecker, J. E. (2014). Religion and early marriage in the United States: Evidence from the Add Health
Study. Journal for the scientific study of religion, 53(2), 392-415.
Wang, L. (2015). Tax enforcement, corporate tax aggressiveness, and cash holdings. China Finance Review
International.
Wilson, R. J. (2009). An examination of corporate tax shelter participants. The accounting review, 84(3),
969-999.
Wilcox, W.B., Doherty, W.J., Fisher, H., Galston, W.A., Glenn, N.D., Gottman, J., Lerman, R.,
Mahoney,A., Markey, B., Markman, H.J., Nock, S., Popenoe, D., Rodriguez, G.G., Stanley, S.M.,
Waite, L.J., Wallerstein, J.(2005). Why Marriage Matters, Twenty Six Conclusions from the Social
Sciences, 2nd edition, ISBN: 978-1-931764-10-7
Xu, X., Hudspeth, C. D., & Bartkowski, J. P. (2005). The timing of first marriage: Are there religious
variations? Journal of Family Issues, 26(5), 584-618.
37

Appendix. Variable definition
Variable Definition
CEO Marital Status
Single An indicator variable that equals 1 if a CEO is unmarried, and zero otherwise.
[Roussanov and Savor, 2014].
Cash holdings measures
Cash The ratio of cash plus marketable securities to the book value of assets. (che/at]
NetCash The ratio of cash to net assets, where net assets are defined as the book value of
assets minus cash. [che/(at-ch)]
Firm Characteristics
Ln(assets) The natural logarithm of total assets (at).
MB The ratio of market to book value of equity. [(at-ceq+me)/at]
Leverage The ratio of total debt divided by the book value of total assets. [(dltt+dlc) / at]
LTM The ratio of long-term debt scaled by lagged assets. [(dltt/att-1)]
R&D Expenditures on research and development scaled by total sales. [xrd/sale]
Cash_Flow The ratio of operating income before depreciation minus interest, taxes, and
common dividends, all divided by the book value of assets. [(oibdp-xint-txt-
dvc)/at]
NWC The ratio of net working capital, measured as working capital minus cash and
marketable securities, divided by the book value of assets. [(wcap-che)/at]
CAPX The ratio of capital expenditures to the book value of assets. [capx/at]
INTANG The ratio of intangible assets for a firm in a given year, scaled by lagged assets.
[(intan/at t-1)]
NOL An indicator variable equals one if loss carry forward for a firm in a given year is
positive, zero otherwise. [tlcf]
ESUB Equity income in earnings for a firm in a given year, scaled by lagged assets.
[esub/at t-1]
Dividend An indicator variable equals 1 if a firm pays a common dividend, and 0 otherwise.
Dividend to Assets is defined as the ratio of dividends declared on common shares (DVC) over total
assets (at)
Dividend to Cash The ratio of dividends declared on common shares (dvc) to the summation on
Flow income before extraordinary items (ib) plus depreciation and amortization (dp).
Dividend Yield The ratio of dividends declared on common shares to the market value of equity
(prcc_f*csho) in the model.
Log of Dividend The natural logarithm of one plus the amount of dividend declared on common
shares.
Dividend Per is the ratio of dividends declared on common shares to common shares outstanding
Share (CSHO).
Annual Payout The ratio of total payout (dividend plus stock repurchases) to operating income.
[(dvc+prstkc)/oibdp]
Acquisition The ratio of corporate acquisition expenditures to the book value of assets.
[(aqc/at)]
Industry_sigma The average of the standard deviation of the ratio of cash flow to book value of
assets over the last 10 years of firms in the same 2-digit SIC code industries.
InstOwn The percentage of shares owned by institutional investors.
ΔNol Change in loss carry forward scaled by lagged total assets. [Δtlcf/at]
Total Volatility The annualized standard deviation of monthly stock returns over the previous year.
Idiosyncratic The standard deviation of the residual from a regression of monthly stock returns
Volatility on the CRSP value-weighted market portfolio return.
38

Appendix. Variable definition. Cont’d
FC (Non-FC) Financially constrained (Financially unconstrained) subgroup is identified as non-
dividend (dividend) payers, below (above) median values of Annual Payout, below
(above) S&P long-term credit ratings investment grade, below (above) S&P short-
term credit ratings investment grade, and above (below) medium scores on the KZ
index, in turn.
Bond Rating An indicator variable that equals 1 if a firm is above Standard and Poor’s (S&P)
long-term credit ratings investment grade, and 0 otherwise.
CP Rating An indicator variable equals 1 if a firm is above S&P short-term credit ratings
investment grade, and 0 otherwise.
KZ index Kaplan & Zingales (1997) index.
MV which is the sum of the market value of equity and the book value of short-term
debt and long-term debt over total assets
CEO Characteristics
Ln(Tenure) The natural logarithm of CEO tenure, where tenure is defined as the length of a
CEO's tenure with her/his current firm (measured as fiscal year minus year joined as
CEO).
Ln(Age) The natural logarithm of the age of a CEO.
Duality An indicator variable that equals 1 if a CEO is also the chairman, and zero otherwise.
Ln(Delta) The natural logarithm of the expected dollar changes in CEO wealth for a 1% change
in stock price computed as in Core and Guay (2002).
Ln(Vega) The natural logarithm of the expected dollar changes in CEO wealth for a 1% change
in stock return volatility computed as in Guay (1999).
CEO_Own The percentage of shares outstanding owned by a CEO. [Shrown_excl_opts /
(csho*1000)]
Overconfidence An indicator variable equals 1 if a CEO holds vested options with average
moneyness greater than 67 percent, and 0 otherwise. Starting in the first year
when a CEO displays this behavior. Option moneyness is calculated as
follows: first, we calculate the realizable value per option as the total
realizable value of the exercisable options divided by the number of
exercisable options [Value_Per_option = (OPT_UNEX_EXER_EST_VAL /
OPT_UNEX_EXER_NUM)]. Second, we compute the estimate of the
average exercise price of the options by subtracting the per-option realizable
value from the stock price at the fiscal year-end [avg_exercise_price =
(prcc_f - Value_Per_option)]. Lastly, the average percent moneyness of an
option equals the per-option realizable value divided by the estimated average
exercise price [avg_pctg_moneyness_opt = (Value_Per_option /
avg_exercise_price)]. [Malmendier and Tate, 2005; Campbell et al., 2011;
Hirshleifer et al., 2012]
Tax Avoidance Measures
BTD Following Wilson (2009), the total book-tax difference which equals book income
less taxable income scaled by lagged assets. Book income is pretax income in year
t. Taxable income is calculated by summing the current federal tax expense and
current foreign tax expense and divided by the statutory tax rate (STR) and then
subtracting the change in NOL carryforwards in year t. If the current federal tax
expense is missing, the total current tax expense is calculated by subtracting deferred
taxes, state income taxes, and other income taxes from the total income taxes in year
t. 𝐵𝑇𝐷 = (𝑃𝐼 − (𝑇𝑋𝐹𝐸𝐷 + 𝑇𝑋𝐹𝑂)/0.35) − 𝛥𝑇𝐿𝐶𝐹)/𝐴𝑇𝑡 − 1
39

Appendix. Variable definition. Cont’d
DTAX The discretionary permanent book-tax difference for firm i in year t (Frank et al.
(2009)). DTAXi,t is the residual from the following regression estimated by two-digit
SIC code and fiscal year:
𝑃𝐸𝑅𝑀𝐷𝐼𝐹𝐹𝑖, 𝑡 = 𝛽0 + 𝛽1 𝐼𝑁𝑇𝐴𝑁𝐺𝑖, 𝑡 + 𝛽2 𝑈𝑁𝐶𝑂𝑁𝑖, 𝑡 + 𝛽3 𝑀𝐼𝑖, 𝑡 +
𝛽4 𝐶𝑆𝑇𝐸𝑖, 𝑡 + 𝛽5 ∆𝑁𝑂𝐿𝑖, 𝑡 + 𝛽6 𝐿𝐴𝐺𝑃𝐸𝑅𝑀𝑖, 𝑡 + 𝜀𝑖, 𝑡;
Where:
𝑃𝐸𝑅𝑀𝐷𝐼𝐹𝐹𝑖, 𝑡 = 𝐵𝐼𝑖, 𝑡 − [(𝐶𝐹𝑇𝐸𝑖, 𝑡 + 𝐶𝐹𝑂𝑅𝑖, 𝑡)/𝑆𝑇𝑅𝑖, 𝑡] − (𝐷𝑇𝐸𝑖, 𝑡/𝑆𝑇𝑅𝑖, 𝑡);
BI is pre-tax book income (PI), CFTE is the current federal tax expense (TXFED),
CFOR is the current foreign tax expense (TXFO), DTE is the deferred tax expense
(TXDI), STR is the statutory tax rate (0.35), INTANG is the goodwill and other
intangibles (INTAN), UNCON is the income (loss) reported under the equity
method (ESUB), MI is the income (loss) attributable to minority interest (MII),
CSTE is the current state income tax expense (TXS), ΔNOL is the change in net
operating loss carryforwards (TLCF), LAGPERM is the one-year lagged
PERMDIFF. To handle the missing value problems in calculating DTAX, we follow
the similar methodology in Frank et al. (2009). If minority interest (MII), current
foreign tax expense (TXFO), income from unconsolidated entities (ESUB), or
current state tax expense (TXS) is missing on Compustat, we set MII, CFOR,
UNCON, or CSTE, respectively, to zero. If the current federal tax expense (TXFED)
is missing on Compustat, we set the value of CFTE to total tax expense (TXT) less
current foreign tax expense (TXFO) less current state tax expense (TXS) less
deferred tax expense (TXDI). If goodwill and other intangibles (INTANG) is
missing on Compustat, we set the value for INTANG to 0. If INTANG equals “C”,
then we set the value of INTANG to that for goodwill (GDWL).
State-Level Variables
Ln(IncomeState) The logarithm of real per-capita income for each state. [available at:
https://www.bea.gov/]
CEAI The in-state Coincident Economic Activity Index was constructed by the Federal
Reserve bank of Philadelphia (CEAI). This index provides a summary of the current
economic conditions in a single statistic by combining four state-level variables: (i)
nonfarm payroll employment, (ii) average hours worked in manufacturing by
production workers, (iii) the unemployment rate, and (iv) wage and salary
disbursements deflated by the consumer price index (CPI) of U.S. city average. For
each state, the trend of this index is set to the trend of its gross domestic product
(GDP), thus long-term growth in the state’s index matches long-term growth in its
GDP. Available at: https://www.philadelphiafed.org/
40

Table 1
Summary statistics
This table reports descriptive statistics for measures of CEO marital status, cash holdings, and control
variables for our sample covering the period 1993-2008. Single is an indicator variable that equals 1 if a
CEO is unmarried, and 0 otherwise. Cash is the ratio of cash plus marketable securities to the book value
of assets. NetCash is the ratio of cash to net assets, where net assets are defined as the book value of
assets minus cash. All other variables are defined in the Appendix. Firms and CEO characteristics
variables are lagged, and all continuous variables are winsorized at the 1st and 99th percentiles.
variable No. Mean Std Dev 25th Perc Median 75th Perc
CEO Marital Status
Single 19,864 0.173 0.378 0.000 0.000 0.000
Cash Holding
Cash 19,864 0.144 0.171 0.022 0.071 0.209
NetCash 19,864 0.186 0.257 0.022 0.075 0.246
Firm Characteristics
Ln(assets) 19,864 6.994 1.559 5.860 6.840 7.974
MB 19,864 2.201 1.553 1.278 1.698 2.488
Cash_Flow 19,864 0.085 0.084 0.055 0.089 0.126
NWC 19,864 0.083 0.149 -0.018 0.071 0.178
R&D 19,864 0.053 0.136 0.000 0.002 0.049
Capex 19,864 0.064 0.057 0.026 0.048 0.082
Leverage 19,864 0.213 0.177 0.050 0.198 0.327
Dividend 19,864 0.498 0.500 0.000 0.000 1.000
Acquisition 19,864 0.028 0.062 0.000 0.000 0.023
Industry_sigma 19,864 0.140 0.130 0.061 0.098 0.148
Tobin’s Q 19,864 2.130 1.469 1.253 1.654 2.410
CEO Characteristics
Ln(Tenure) 17,210 1.774 0.786 1.099 1.792 2.303
Ln(Age) 16,944 4.001 0.139 3.912 4.007 4.094
Duality 17,210 0.630 0.483 0.000 1.000 1.000
Ln(Delta) 17,125 5.393 1.613 4.409 5.387 6.413
Ln(Vega) 17,038 3.634 1.707 2.688 3.796 4.796
CEO_Own 16,558 0.029 0.064 0.001 0.004 0.018
41

Table 2
Pearson correlation
This table reports Pearson correlation coefficients of key-dependent and independent variables. Single is an indicator variable that equals 1 if a
CEO is unmarried, and 0 otherwise. Cash is the ratio of cash plus marketable securities to the book value of assets. All other variables
are defined in the Appendix. Firms and CEO characteristics variables are lagged, and all continuous variables are winsorized at the
1st and 99th percentiles.
Industry_sigma
Cash_Flow
Acquisition
Ln(Tenure)
CEO_Own
Ln(assets)
Ln(Delta)
Ln(Vega)
Leverage
Dividend
Ln(Age)
Duality
Capex
Single
NWC
R&D
Cash
MB
Single 1.00
Cash 0.13 1.00
Ln(assets) -0.20 -0.33 1.00
MB 0.02 0.39 -0.16 1.00
Leverage -0.06 -0.38 0.29 -0.27 1.00
R&D 0.07 0.53 -0.21 0.27 -0.13 1.00
Cash_Flow -0.03 -0.18 0.05 0.22 -0.17 -0.45 1.00
NWC 0.03 -0.18 -0.28 -0.11 -0.15 -0.17 0.12 1.00
Capex -0.02 -0.16 -0.05 0.06 0.03 -0.10 0.25 -0.14 1.00
Dividend -0.15 -0.35 0.36 -0.14 0.09 -0.25 0.06 0.06 -0.01 1.00
Acquisition 0.02 -0.13 0.00 -0.04 0.14 -0.03 0.02 0.01 -0.13 -0.01 1.00
Industry_sigma 0.09 0.28 -0.08 0.17 -0.18 0.22 -0.05 -0.15 -0.16 -0.20 0.05 1.00
Ln(Tenure) -0.10 0.04 -0.06 0.04 -0.05 0.01 0.05 0.06 0.04 -0.02 0.00 -0.01 1.00
Ln(Age) -0.11 -0.16 0.14 -0.12 0.06 -0.13 0.03 0.09 -0.04 0.22 -0.01 -0.09 0.32 1.00
Duality -0.14 -0.14 0.20 -0.03 0.08 -0.11 0.04 -0.02 0.00 0.16 0.00 -0.09 0.26 0.28 1.00
Ln(Delta) -0.15 0.04 0.42 0.33 -0.06 -0.03 0.23 -0.16 0.03 0.04 0.03 0.06 0.32 0.11 0.22 1.00
Ln(Vega) -0.09 -0.05 0.53 0.05 0.09 0.03 0.07 -0.21 -0.09 0.08 0.03 0.11 -0.03 -0.02 0.11 0.42 1.00
CEO_Own -0.04 0.08 -0.19 0.07 -0.09 -0.06 0.05 0.06 0.06 -0.06 -0.03 -0.04 0.30 0.11 0.12 0.43 -0.31 1.00
42

Table 3
CEO marital status and cash holding: A Baseline analysis
This table presents tests of the association between CEO marital status and a firm’s cash holding. Single
is an indicator variable that equals 1 if a CEO is unmarried, and 0 otherwise. Cash is the ratio of cash plus
marketable securities to the book value of assets. NetCash is the ratio of cash to net assets, where net
assets are defined as the book value of assets minus cash. FCash and FNetCash is the one-year forwarded
value of Cash and NetCash. All other independent variables are defined in the Appendix. Firms and CEO
characteristics variables are lagged, and all continuous variables are winsorized at the 1st and 99th
percentiles. All models include year and industry fixed effects. T-statistics are computed using robust
standard errors clustered at the firm level and are reported in parentheses. ***, ** and * denote significance
at the 1%, 5%, and 10% levels, respectively.
Cash NetCash FCash FNetCash
(1) (2) (3) (4)
Single 0.010* 0.018** 0.010* 0.017**
(1.90) (2.13) (1.78) (2.02)
Ln(assets) -0.021*** -0.036*** -0.022*** -0.037***
(-9.82) (-11.38) (-9.73) (-11.54)
*** *** ***
MB 0.020 0.029 0.020 0.030***
(10.39) (10.22) (9.95) (9.80)
Leverage -0.191*** -0.246*** -0.191*** -0.245***
(-14.06) (-11.46) (-13.79) (-11.29)
R&D 0.351*** 0.477*** 0.366*** 0.491***
(13.90) (11.44) (13.96) (11.53)
** *** **
Cash_Flow -0.078 -0.206 -0.075 -0.200***
(-2.55) (-3.90) (-2.35) (-3.67)
NWC -0.257*** -0.387*** -0.262*** -0.394***
(-14.57) (-13.60) (-14.64) (-13.67)
Capex -0.450*** -0.653*** -0.484*** -0.704***
(-11.86) (-11.59) (-12.15) (-11.90)
*** *** ***
Dividend -0.034 -0.044 -0.033 -0.042***
(-7.79) (-7.12) (-7.38) (-6.71)
Acquisition -0.270*** -0.396*** -0.282*** -0.411***
(-16.22) (-15.43) (-16.41) (-15.68)
Industry_sigma 0.027 0.052** 0.021 0.043*
(1.61) (2.00) (1.26) (1.65)
** ** **
Ln(Tenure) 0.007 0.009 0.006 0.007*
(2.51) (2.18) (2.09) (1.78)
Ln(Age) -0.034** -0.052** -0.032** -0.051**
(-2.18) (-2.22) (-2.01) (-2.11)
Duality -0.006* -0.008 -0.005 -0.006
(-1.65) (-1.38) (-1.35) (-1.14)
Ln(Delta) 0.002 0.002 0.002 0.004
(0.71) (0.77) (0.86) (1.17)
Ln(Vega) -0.000 0.000 -0.001 -0.000
(-0.17) (0.14) (-0.32) (-0.07)
CEO_Own 0.113** 0.146* 0.109** 0.128
(2.24) (1.86) (2.10) (1.62)
Year & Ind. FE Yes Yes Yes Yes
Observations 16,178 16,178 14,941 14,941
Adjusted R2 0.564 0.517 0.570 0.525
43

Table 4
CEO marital status and corporate cash holding. A DID test
This table presents estimates from the Difference-in-Difference (DID) regressions of the change in cash holdings around CEO
turnover events (-3, +3). For models 1, 2, 5, and 6, Treated is an indicator variable that equals 1 if a firm replaces a married CEO
with a single CEO, and 0 otherwise. For models 3, 4, 7, and 7, Treated is an indicator variable that equals 1 if a firm replaces a
single CEO with a married CEO, and 0 otherwise. After is an indicator variable that equals 1 (0) for the three years after (before)
the CEO turnover year. All other independent variables are defined in the Appendix. Models (1)-(4) use the full sample, while
models (5)-(8) use the PSM sample. Firms and CEO characteristics variables are lagged, and all continuous variables are winsorized
at the 1st and 99th percentiles. All models include year and industry fixed effects. T-statistics are reported in parentheses and
computed using robust standard errors clustered at the firm level. ***, ** and * denote significance at the 1%, 5%, and 10% levels,
respectively.
Full sample PSM sample
Treated = single Treated = married Treated = single Treated = married
replacing married replacing single replacing married replacing single
(1) (2) (3) (4) (5) (6) (7) (8)
Cash NetCash Cash NetCash Cash NetCash Cash NetCash
After*Treated 0.218*** 0.533*** -0.168** -0.407*** 0.193*** 0.474*** -0.166*** -0.538***
(5.63) (6.67) (-2.51) (-2.67) (3.58) (3.40) (-2.67) (-7.32)
Treated -0.213*** -0.526*** 0.176*** 0.433*** -0.185*** -0.460*** 0.163*** 0.547***
(-5.60) (-6.61) (2.81) (2.92) (-3.42) (-3.28) (3.15) (9.20)
After -0.279*** -0.632*** -0.147** -0.309** -0.251*** -0.547*** -0.114*** -0.127***
(-8.19) (-8.11) (-2.41) (-2.10) (-7.16) (-4.21) (-2.61) (-2.76)
Ln(assets) -0.003 -0.009* -0.003 -0.009* -0.015** -0.025** -0.015** -0.024**
(-0.85) (-1.76) (-0.88) (-1.78) (-2.36) (-2.58) (-2.39) (-2.55)
MB 0.023*** 0.028*** 0.023*** 0.029*** 0.018*** 0.023*** 0.018*** 0.024***
(6.96) (6.74) (7.03) (6.91) (3.22) (3.00) (3.30) (3.20)
Leverage -0.164*** -0.217*** -0.160*** -0.207*** -0.212*** -0.300*** -0.208*** -0.287***
(-6.96) (-6.59) (-6.74) (-6.38) (-5.62) (-5.45) (-5.48) (-5.32)
R&D 0.431*** 0.646*** 0.426*** 0.635*** 0.447*** 0.743*** 0.440*** 0.722***
(7.64) (8.07) (7.58) (8.19) (6.06) (6.03) (5.99) (6.03)
Cash_Flow -0.015 -0.133 -0.025 -0.158 0.123 0.083 0.117 0.062
(-0.23) (-1.42) (-0.38) (-1.65) (1.18) (0.52) (1.11) (0.38)
NWC -0.113*** -0.178*** -0.114*** -0.180*** -0.167*** -0.273*** -0.168*** -0.273***
(-3.66) (-4.07) (-3.66) (-4.06) (-3.21) (-3.39) (-3.19) (-3.31)
Capex -0.295*** -0.338*** -0.291*** -0.329*** -0.292*** -0.345** -0.301*** -0.366**
(-3.84) (-3.39) (-3.77) (-3.27) (-2.66) (-2.25) (-2.71) (-2.35)
Dividend -0.050*** -0.063*** -0.050*** -0.063*** -0.056*** -0.073*** -0.057*** -0.073***
(-5.17) (-5.04) (-5.26) (-5.09) (-4.38) (-4.08) (-4.59) (-4.27)
Acquisition -0.176*** -0.223*** -0.176*** -0.225*** -0.209*** -0.296*** -0.214*** -0.304***
(-5.02) (-4.28) (-5.03) (-4.32) (-3.80) (-3.30) (-3.88) (-3.37)
Industry_sigma 0.071* 0.114* 0.076* 0.127** 0.052 0.106 0.055 0.123
(1.82) (1.94) (1.95) (2.18) (0.71) (0.96) (0.75) (1.10)
Ln(Tenure) 0.007 0.010 0.007 0.010* 0.014 0.022* 0.015 0.024*
(1.50) (1.64) (1.48) (1.65) (1.56) (1.71) (1.57) (1.81)
Ln(Age) -0.010 -0.023 -0.011 -0.024 -0.015 -0.009 -0.014 -0.013
(-0.33) (-0.51) (-0.33) (-0.53) (-0.27) (-0.11) (-0.25) (-0.15)
Duality 0.002 0.004 0.003 0.005 0.014 0.017 0.014 0.018
(0.33) (0.43) (0.39) (0.55) (1.19) (1.01) (1.23) (1.04)
Ln(Delta) -0.009** -0.013** -0.009** -0.013** -0.009 -0.012 -0.009 -0.012
(-2.25) (-2.21) (-2.21) (-2.13) (-1.41) (-1.14) (-1.37) (-1.16)
Ln(Vega) -0.003 -0.004 -0.003 -0.004 0.005 0.002 0.005 0.001
(-0.86) (-0.92) (-0.93) (-1.07) (1.12) (0.30) (1.06) (0.20)
CEO_Own 0.118 0.146 0.118 0.146 -0.065 -0.243 -0.066 -0.227
(0.95) (0.76) (0.95) (0.76) (-0.32) (-0.72) (-0.33) (-0.69)
Year & Ind. FE Yes Yes Yes Yes Yes Yes Yes Yes
Observations 2,870 2,870 2,870 2,870 912 912 912 912
Adjusted R2 0.590 0.572 0.590 0.572 0.605 0.597 0.605 0.600
44

Table 5
CEO marital status and corporate cash holding: Instrumental variable regression
This table presents tests of the association between CEO marital status and cash holding using the instrumental variable approach
where model (1) presents the results of first-stage regression and model (2) presents the second-stage regression. Single is an
indicator variable that equals 1 if a CEO is unmarried, and 0 otherwise. The IV, Community is an indicator variable that equals
1 if the ﬁrm is headquartered in a community property state and 0 otherwise. All other independent variables are defined in the
Appendix. Firms and CEO characteristics variables are lagged, and all continuous variables are winsorized at the 1 st and 99th
percentiles. All models include year and industry fixed effects. T-statistics are computed using robust standard errors clustered
by the state of headquarters and are reported in parentheses. ***, ** and * denote significance at the 1%, 5%, and 10% levels,
respectively.
Single (1st stage) Cash (2nd stage)
(1) (2)
Community 0.126**
(1.99)
Singlepred 0.285**
(2.35)
Ln(assets) -0.217*** -0.009
(-6.82) (-1.64)
MB -0.018 0.020***
(-1.17) (11.36)
Leverage 0.149 -0.192***
(0.74) (-7.87)
R&D -0.136 0.356***
(-0.57) (8.39)
Cash_Flow 0.324 -0.079**
(1.25) (-2.02)
NWC 0.071 -0.249***
(0.31) (-6.96)
Capex -0.302 -0.432***
(-0.55) (-8.49)
Dividend -0.159*** -0.021***
(-2.60) (-3.16)
Acquisition 0.368* -0.290***
(1.75) (-6.77)
Industry_sigma -0.093 0.023
(-0.36) (1.02)
Ln(Tenure) -0.187*** 0.017***
(-3.63) (3.40)
Ln(Age) -0.469** -0.010
(-2.25) (-0.34)
Duality -0.100* 0.003
(-1.88) (0.53)
Ln(Delta) -0.039 0.004
(-1.28) (1.12)
Ln(Vega) -0.015 0.000
(-0.89) (0.10)
CEO_Own -0.050 0.122*
(-0.07) (1.86)
CEAI -0.008 -0.001
(-1.43) (-1.10)
Ln(IncomeState) 0.139 0.060
(0.42) (1.31)
Year & Ind. FE Yes Yes
Observations 15,511 15,511
Adjusted R2 0.117 0.012
Weak Identification Test: Cragg-Donald Wald F Statistic 81.36
Weak Identification Test: Kleibergen-Paap rk F Statistic 30.23
45

Table 6
CEO marital status and cash holding. Propensity score matching (PSM) Diagnostic
This table presents the results of the univariate analysis in Panel A and differences in characteristics of
Treatment and Control for propensity score-matched sample in panel B where Treatment denotes single
CEOs and Control refers to the matching sample of married CEOs. Single is an indicator variable that
equals 1 if a CEO is unmarried and 0 otherwise. Married is an indicator variable that equals 1 if a CEO
is married and 0 otherwise. Cash is the ratio of cash plus marketable securities to the book value of assets.
NetCash is the ratio of cash to net assets, where net assets are defined as the book value of assets minus
cash. All other variables are defined in Appendix A. ***, **, and * denote significance at the 1%, 5%, and
10% levels, respectively.
Panel A. Univariate test
variable Single Married Difference
Cash 0.193 0.134 0.059***
NetCash 0.260 0.171 0.090***
Ln(assets) 6.309 7.137 -0.828***
MB 2.300 2.181 0.119***
Cash_Flow 0.079 0.087 -0.008***
NWC 0.095 0.081 0.014***
R&D 0.077 0.048 0.029***
CAPX 0.062 0.065 -0.003***
Leverage 0.188 0.218 -0.030***
Dividend 0.334 0.531 -0.197***
Acquisition 0.031 0.028 0.003***
Industry_sigma 0.167 0.135 0.032***
Ln(Tenure) 1.602 1.810 -0.208***
Ln(Age) 3.969 4.008 -0.039***
Duality 0.481 0.661 -0.180***
Ln(Delta) 4.875 5.500 -0.625***
Ln(Vega) 3.283 3.707 -0.424***
CEO_Own 0.023 0.030 -0.007***
Panel B. Differences in Characteristics of Treatment and Control for PSM sample.
Variable Treatment Control Difference T-test
Ln(assets) 6.423 6.409 0.015 0.420
MB 2.261 2.223 0.038 0.880
Leverage 0.191 0.192 -0.001 -0.200
R&D 0.076 0.078 -0.002 -0.450
Cash_Flow 0.079 0.078 0.001 0.360
NWC 0.090 0.094 -0.004 -0.900
Capex 0.060 0.061 0.000 -0.180
Acquisition 0.031 0.031 0.000 0.000
Industry_sigma 0.172 0.170 0.002 0.490
Ln(Tenure) 1.614 1.639 -0.025 -1.250
Ln(Age) 3.968 3.966 0.002 0.440
Duality 0.482 0.489 -0.007 -0.510
Ln(Delta) 4.968 4.958 0.011 0.270
Ln(Vega) 3.324 3.302 0.022 0.530
CEO_Own 0.023 0.023 0.000 -0.060
46

Table 7
CEO marital status and cash holding. PSM sample
This table presents tests of the association between CEO marital status and the firm’s cash holding for
the propensity score-matched sample. Single is an indicator variable that equals 1 if a CEO is unmarried,
and 0 otherwise. Cash is the ratio of cash plus marketable securities to the book value of assets. NetCash
is the ratio of cash to net assets, where net assets are defined as the book value of assets minus cash.
FCash and FNetCash is the one-year forwarded value of Cash and NetCash. All other independent
variables are defined in the Appendix. Firms and CEO characteristics variables are lagged, and all
continuous variables are winsorized at the 1st and 99th percentiles. All models include year and industry
fixed effects. T-statistics are computed using robust standard errors clustered at the firm level and are
reported in parentheses. ***, ** and * denote significance at the 1%, 5%, and 10% levels, respectively.
Cash NetCash FCash FNetCash Cash NetCash FCash FNetCash
(1) (2) (3) (4) (5) (6) (7) (8)
Single 0.013** 0.020** 0.012* 0.019** 0.014** 0.021** 0.012** 0.019**
(2.37) (2.32) (1.91) (1.96) (2.41) (2.36) (1.98) (2.03)
Ln(assets) -0.029*** -0.050*** -0.028*** -0.048*** -0.031*** -0.054*** -0.030*** -0.051***
(-10.87) (-11.56) (-9.81) (-10.60) (-10.12) (-10.70) (-8.64) (-9.12)
MB 0.017*** 0.025*** 0.018*** 0.027*** 0.014*** 0.022*** 0.017*** 0.026***
(8.09) (7.61) (6.64) (6.33) (6.18) (5.90) (5.57) (5.57)
Leverage -0.221*** -0.308*** -0.219*** -0.297*** -0.215*** -0.299*** -0.207*** -0.279***
(-11.88) (-10.39) (-11.11) (-9.39) (-11.57) (-10.17) (-10.59) (-9.01)
R&D 0.321*** 0.410*** 0.316*** 0.417*** 0.314*** 0.401*** 0.312*** 0.413***
(11.14) (8.01) (9.53) (7.31) (10.77) (7.77) (9.32) (7.18)
* ** **
Cash_Flow -0.019 -0.135 -0.051 -0.170 -0.033 -0.155 -0.063 -0.188**
(-0.47) (-1.85) (-1.15) (-2.11) (-0.81) (-2.10) (-1.42) (-2.29)
NWC -0.322*** -0.513*** -0.317*** -0.493*** -0.323*** -0.515*** -0.318*** -0.493***
(-11.79) (-10.91) (-10.90) (-9.82) (-11.94) (-11.02) (-11.20) (-10.10)
Capex -0.492*** -0.745*** -0.484*** -0.737*** -0.503*** -0.761*** -0.501*** -0.765***
(-9.61) (-9.54) (-8.61) (-8.77) (-9.61) (-9.52) (-8.66) (-8.88)
*** *** *** *** *** *** ***
Dividend -0.033 -0.043 -0.032 -0.040 -0.030 -0.039 -0.032 -0.040***
(-5.24) (-4.50) (-4.69) (-3.88) (-4.92) (-4.21) (-4.71) (-3.91)
Acquisition -0.322*** -0.482*** -0.342*** -0.516*** -0.332*** -0.497*** -0.343*** -0.518***
(-12.24) (-11.54) (-11.59) (-11.51) (-12.27) (-11.57) (-11.42) (-11.35)
Industry_sigma 0.002 0.029 0.009 0.047 0.001 0.029 0.009 0.045
(0.07) (0.69) (0.31) (0.95) (0.06) (0.67) (0.28) (0.88)
Ln(Tenure) 0.001 -0.001 0.004 0.003
(0.35) (-0.12) (0.73) (0.35)
Ln(Age) -0.032 -0.051 -0.027 -0.044
(-1.30) (-1.36) (-0.97) (-0.99)
Duality -0.009 -0.012 -0.004 -0.006
(-1.53) (-1.40) (-0.58) (-0.57)
Ln(Delta) 0.007** 0.011** 0.004 0.008
(2.15) (2.23) (1.17) (1.28)
Ln(Vega) -0.001 -0.003 0.000 -0.003
(-0.52) (-0.83) (0.08) (-0.74)
CEO_Own 0.026 0.008 0.086 0.091
(0.31) (0.06) (0.82) (0.55)
Observations 5,526 5,526 4,287 4,287 5,526 5,526 4,155 4,155
Adjusted R2 0.570 0.513 0.574 0.525 0.571 0.515 0.582 0.532
47

Table 8
Ruling out the precautionary motive. Financially constrained vs. Unconstrained firms
This table presents tests of the association between CEO marital status and the firm’s cash holding for subsamples of financially
constrained and unconstrained firms for the full sample [models (1)-(4)] and the PSM sample [models (5)-(8)]. We use five
measures of financial constraints, Panel A uses Dividends which is an indicator variable that equals 1 if a firm pays a common
dividend, and 0 otherwise, Panel B uses Annual Payout which is the ratio of total payout (dividend plus stock repurchases) to
operating income, Panel C uses Bond Rating which is an indicator variable that equals 1 if a firm is above Standard and Poor’s
(S&P) long-term credit ratings investment grade, and 0 otherwise, Panel D uses S&P short-term credit ratings (Commercial
Paper) which is an indicator variable equals 1 if a firm is above S&P short-term credit ratings investment grade, and 0 otherwise,
and Panel E uses KZ index calculated following Kaplan & Zingales (1997). All other independent variables are defined in the
Appendix. Firms and CEO characteristics variables are lagged, and all continuous variables are winsorized at the 1st and 99th
percentiles. All models include year and industry fixed effects. T-statistics are computed using robust standard errors clustered
at the firm level and are reported in parentheses. ***, ** and * denote significance at the 1%, 5%, and 10% levels, respectively.
Panel A. Dividend
Full sample PSM sample
Non-FC FC Non-FC FC
(1) (2) (3) (4) (5) (6) (7) (8)
Single 0.017** 0.027** 0.003 0.008 0.019** 0.027** 0.008 0.013
(2.20) (2.19) (0.52) (0.79) (2.39) (2.25) (1.05) (1.13)
Observations 8,219 8,219 7,959 7,959 2,048 2,048 3,478 3,478
Adjusted R2 0.436 0.393 0.555 0.501 0.477 0.427 0.557 0.498
Panel B. Annual Payout
Non-FC FC Non-FC FC
(1) (2) (3) (4) (5) (6) (7) (8)
Single 0.022*** 0.037*** 0.000 0.002 0.024*** 0.038*** 0.005 0.009
(2.92) (3.12) (0.05) (0.19) (3.22) (3.15) (0.79) (0.76)
Observations 8,408 8,408 7,770 7,770 2,365 2,365 3,161 3,161
Adjusted R2 0.530 0.475 0.591 0.539 0.528 0.464 0.599 0.537
Panel C. Bond Rating
Non-FC FC Non-FC FC
Single 0.015* 0.027* 0.005 0.008 0.018** 0.025 0.008 0.013
(1.74) (1.92) (0.94) (1.00) (1.99) (1.63) (1.37) (1.41)
Observations 7,557 7,557 8,621 8,621 2,071 2,071 3,455 3,455
Adjusted R2 0.623 0.571 0.538 0.499 0.627 0.562 0.543 0.492
Panel D. CP Rating
Non-FC FC Non-FC FC
Single 0.019** 0.034** 0.004 0.007 0.024** 0.035** 0.007 0.011
(2.06) (2.32) (0.83) (0.93) (2.47) (2.18) (1.15) (1.32)
Observations 6,435 6,435 9,743 9,743 1,925 1,925 3,601 3,601
Adjusted R2 0.634 0.578 0.520 0.485 0.619 0.545 0.537 0.493
Panel E. KZ Index
Non-FC FC Non-FC FC
Single 0.020*** 0.036*** -0.003 -0.005 0.022*** 0.031** -0.004 -0.002
(2.71) (2.88) (-0.68) (-0.70) (2.95) (2.39) (-0.60) (-0.27)
Observations 8,155 8,155 8,023 8,023 2,766 2,766 2,760 2,760
Adjusted R2 0.581 0.520 0.590 0.562 0.580 0.503 0.594 0.564
Controls Yes Yes Yes Yes Yes Yes Yes Yes
48

Table 9
Ruling out the precautionary motive. Subsample analysis based on firm-specific risk
This table presents tests of the association between CEO marital status and the firm’s cash holding for
subsamples of firms with high vs low specific risk. To measure firm-specific risk, we use Idiosyncratic
Volatility, Total Volatility, Industry_Sigma, and Leverage in Panels A, B, C, and D, respectively. All
other independent variables are defined in the Appendix. Firms and CEO characteristics variables are
lagged, and all continuous variables are winsorized at the 1st and 99th percentiles. All models include
year and industry fixed effects. T-statistics are computed using robust standard errors clustered at the
firm level and are reported in parentheses. ***, ** and * denote significance at the 1%, 5%, and 10% levels,
respectively.
Panel A. Idiosyncratic Volatility
High Low
Cash NetCash Cash NetCash
(1) (2) (3) (4)
Single 0.003 0.007 0.017** 0.030***
(0.55) (0.71) (2.42) (2.65)
Observations 8,177 8,177 8,001 8,001
Adjusted R2 0.575 0.524 0.499 0.448
Panel B. Total Volatility
High Low
(1) (2) (3) (4)
**
Single 0.005 0.009 0.014 0.026**
(0.73) (0.88) (2.10) (2.37)
Observations 8,056 8,056 8,122 8,122
Adjusted R2 0.580 0.529 0.494 0.444
Panel C. Industry_sigma
High Low
(1) (2) (3) (4)
*
Single 0.006 0.010 0.014 0.023**
(0.88) (0.98) (1.89) (1.97)
Observations 8,344 8,344 7,834 7,834
Adjusted R2 0.560 0.509 0.423 0.389
Panel D. Leverage
High Low
(1) (2) (3) (4)
**
Single 0.001 -0.000 0.015 0.028**
(0.26) (-0.03) (2.06) (2.38)
Observations 8,190 8,190 7,988 7,988
Adjusted R2 0.518 0.507 0.544 0.489
Controls Yes Yes Yes Yes
49

Table 10
CEO marital status and cash holding: The channels
This table presents tests of the channels through which single CEOs can build larger stacks of cash. Panel A reports
tests of the association between CEO marital status and corporate tax avoidance. DTAX is the discretionary
permanent book-tax difference for firm i in year t, BTD is the total book-tax difference, which equals book income
less taxable income scaled by lagged assets. Panel B presents tests of the association between CEO marital status
and alternative measures of dividend policy. Dividend is an indicator variable that equals 1 if a firm pays a common
dividend, and 0 otherwise. Dividend to Assets is defined as the ratio of dividends declared on common shares over
total assets. Dividend to Cash Flow is the ratio of dividends declared on common shares to the summation on income
before extraordinary items plus depreciation and amortization. Dividend Yield is the ratio of dividends declared on
common shares to the market value of equity in the model. Log of Dividend which is the natural logarithm of one
plus the amount of dividend declared on common shares. Dividend Per Share is the ratio of dividends declared on
common shares to common shares outstanding. All other independent variables are defined in the Appendix. All
continuous variables are winsorized at the 1st and 99th percentiles. All models include Year and Industry fixed
effects. T-statistics are computed using robust standard errors clustered by the firm level and are reported in
parentheses. ***, ** and * denote significance at the 1%, 5%, and 10% levels, respectively.
Panel A. Marital status and tax avoidance
Full Sample PSM Sample
DTAX BTD DTAX BTD
(1) (2) (3) (4)
Single 0.021** 0.009** 0.029** 0.010*
(1.99) (1.98) (2.16) (1.82)
Observations 10,196 10,213 3,076 3,080
Adjusted R2 0.124 0.702 0.137 0.730
Panel B1. Marital status and dividends (Full sample)
Dividend to Dividend to Dividend Log of Dividend Per
Dividend
Assets Cash Flow Yield Dividend Share
(1) (2) (3) (4) (5) (6)
Single -0.270** -0.001** -0.010* -0.001** -0.174*** -0.024
(-2.13) (-2.38) (-1.90) (-2.36) (-2.79) (-1.59)
Year & Ind. FE Yes Yes Yes Yes Yes Yes
Observations 16,064 16,143 16,143 16,140 16,143 16,140
Pseudo/Adj. R2 0.314 0.306 0.231 0.305 0.540 0.397
Panel B2. Marital status and dividends (PSM sample)
Dividend to Dividend to Dividend Log of Dividend
Dividend
Assets Cash Flow Yield Dividend Per Share
(1) (2) (3) (4) (5) (6)
Single -0.326** -0.002*** -0.011** -0.001*** -0.220*** -0.031**
(-2.50) (-2.89) (-2.11) (-2.73) (-3.79) (-2.21)
Year & Ind. FE Yes Yes Yes Yes Yes Yes
Observations 5,465 5,517 5,517 5,516 5,517 5,516
Pseudo/Adj. R2 0.308 0.214 0.183 0.263 0.412 0.295
50

Table 11
Funding future investments and external financing
This table presents tests of the association between lagged values of cash and future R&D t expenditure as a proxy for risky
investments in panel A. Panel B presents tests of the association between CEO marital status and external financing measured
by Ln(ExternalFinancing) which is the natural logarithm of the sum of net new equity and debt issues, calculated as the sale of
common and preferred stock minus the purchase of common and preferred stock plus the issuance of long-term debt minus the
reduction of long-term debt. All other independent variables are defined in the Appendix. All continuous variables are
winsorized at the 1st and 99th percentiles. All models include Year and Industry fixed effects. t-statistics are computed using
robust standard errors clustered by the firm level and are reported in parentheses. ***, ** and * denote significance at the 1%, 5%,
and 10% levels, respectively.
Panel A: Cash and future risky investment
Single CEOs Married CEOs
(1) (2) (3) (4)
Casht+1 0.031 0.019
(1.30) (1.64)
Casht -0.014 -0.009 -0.010 -0.028
(-0.44) (-0.22) (-0.76) (-1.58)
Casht-1 0.127*** 0.102** 0.063*** 0.066***
(2.86) (2.07) (3.94) (3.77)
Casht-2 -0.041 -0.032 -0.001 -0.002
(-1.46) (-0.98) (-0.14) (-0.17)
Ln(assets)t -0.001 0.000 -0.001 -0.001
(-0.36) (0.06) (-1.25) (-1.19)
MBt -0.004** -0.004** 0.000 0.000
(-2.14) (-2.09) (0.06) (0.20)
Leveraget -0.014 -0.011 0.001 0.003
(-1.50) (-1.03) (0.15) (0.57)
R&D t-1 0.801*** 0.783*** 0.845*** 0.852***
(18.96) (14.78) (43.21) (41.96)
Cash_Flowt -0.012 0.008 0.008 0.015
(-0.30) (0.15) (0.57) (1.02)
NWCt -0.008 -0.009 0.005 0.010**
(-0.66) (-0.63) (1.25) (2.38)
Capext 0.008 0.003 -0.000 0.001
(0.24) (0.07) (-0.03) (0.05)
Dividendt -0.003 -0.002 -0.003*** -0.002***
(-1.17) (-0.60) (-3.76) (-2.86)
Acquisitiont -0.050* -0.053 -0.041*** -0.037***
(-1.77) (-1.50) (-3.99) (-3.50)
Industry_sigmat 0.008 0.008 0.010** 0.010**
(0.68) (0.53) (2.26) (2.09)
Ln(Tenure)t 0.001 0.002 0.000 -0.000
(0.29) (0.76) (0.71) (-0.17)
Ln(Age)t 0.004 -0.002 -0.001 -0.001
(0.36) (-0.22) (-0.49) (-0.23)
Dualityt -0.007** -0.010*** -0.001 -0.001
(-2.35) (-3.06) (-1.13) (-0.98)
Ln(Delta)t 0.001 0.001 0.000 0.000
(0.47) (0.52) (0.61) (0.62)
Ln(Vega)t 0.002 0.002 0.001*** 0.001***
(1.43) (1.36) (3.58) (3.17)
CEO_Ownt -0.008 -0.015 -0.022*** -0.022**
(-0.16) (-0.25) (-2.65) (-2.52)
Observations 2,334 1,874 11,526 9,933
Adjusted R2 0.839 0.834 0.895 0.896
51

Table 11
Funding future investments and external financing. Cont’d
Panel B: Single CEOs and External Financing
Full Sample PSM Sample
Single -0.127** -0.198***
(-2.30) (-3.20)
Size 0.786*** 0.189***
(35.90) (8.20)
MB 0.115*** 0.876***
(6.15) (22.20)
Profitability -0.795*** -0.951***
(-3.57) (-3.37)
Mature -0.472*** -0.486***
(-6.07) (-4.39)
Tangibility 0.615*** 0.704***
(4.00) (3.04)
Ln(Tenure) -0.031 -0.038
(-0.99) (-0.78)
Ln(Age) -0.323** -0.093
(-1.97) (-0.38)
Duality 0.068 0.036
(1.54) (0.55)
Ln(Delta) 0.116*** 0.089**
(4.24) (2.01)
Ln(Vega) 0.064*** 0.038
(3.12) (1.14)
CEO_Own -2.946*** -2.039**
(-5.33) (-2.20)
Year & Ind. FE Yes Yes
Observations 7,635 2,840
Adjusted R2 0.465 0.409
52

Table 12
Cash holdings and firm value: the effect of marital status
This table presents tests of the effect of CEO marital status on the association between cash holding and firm value.
Panel A presents the results of the value of excess cash models following Dittmar and Mahrt-Smith (2007). Panels
B (C) presents the results of the change in (level of) cash following Pinkowitz and Williamson (2007). The
dependent variable in panels A:C is MVi,t which is the sum of the market value of equity and the book value of
short-term debt and long-term debt over total assets. Panel D presents results for the value of cash models when
firm value is measured by Tobin’s Q which is the ratio of the market value of equity plus total assets less the book
value of equity all divided by total assets. Single is an indicator variable that equals 1 if a CEO is unmarried and 0
otherwise. All models include the same set of control variables, or the control variables from the original models
replicated in panels A:C (coefficients are dropped for brevity). All continuous variables are winsorized at the 1 st
and 99th percentiles. All models include year and two-digit industry fixed effects. T-statistics are computed using
robust standard errors clustered by the firm and are reported in parentheses. ***, ** and * denote significance at the
1%, 5%, and 10% levels, respectively.
Panel A. Value of excess cash (Dittmar and Mahrt-Smith (2007))
Dependent variable = MVi,t
Single Married PSM
(1) Full (2) PSM (3) Full (4) PSM
Excess_Cashi,t 0.964*** 1.191*** 1.834*** 1.385***
(2.94) (3.73) (8.16) (3.87)
Observations 3,039 2,476 15,353 2,549
Adjusted R2 0.375 0.380 0.366 0.338
Panel B. Change value of cash (Pinkowitz and Williamson (2007))
Single Married PSM
ΔCashi,t-1 2.640*** 2.195*** 2.772*** 2.852***
(7.84) (6.00) (12.79) (8.45)
ΔCashi,t+1 2.441*** 1.938*** 2.412*** 2.046***
(8.59) (6.50) (13.78) (5.96)
Observations 3,095 2,478 15,582 2,573
Adjusted R2 0.445 0.431 0.399 0.368
Panel C. Level value of cash (Pinkowitz and Williamson (2007))
Single Married PSM
Cashi,t 1.815*** 1.733*** 3.054*** 2.270***
(5.51) (5.09) (13.26) (7.44)
Observations 3,095 2,478 15,582 2,573
Adjusted R2 0.418 0.420 0.416 0.355
Panel D. Firm value is measured by Tobin’s Q
Dependent variable = Tobin’s Q
Single Married PSM
Cash 0.344* 0.344* 0.785*** 0.673***
(1.77) (1.77) (7.27) (3.26)
Observations 2,763 2,763 13,415 2,763
Adjusted R2 0.601 0.601 0.676 0.609
53

CEO Marital Status

Uploaded by

Copyright:

Available Formats

CEO Marital Status

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CEO Marital Status

Uploaded by

Copyright:

Available Formats

CEO Marital Status and Corporate Cash Holdings

Keywords: Cash Holdings; Marital Status; CEO; Firm Value

JEL Classification: G30, G32, G35

Electronic copy available at: https://ssrn.com/abstract=4085811

Keywords: Cash Holdings; Marital Status; CEO; Firm Value

JEL Classification: G30, G32, G35

Electronic copy available at: https://ssrn.com/abstract=4085811

status of CEOs on corporate cash holdings.

Electronic copy available at: https://ssrn.com/abstract=4085811

Electronic copy available at: https://ssrn.com/abstract=4085811

Smith (2007)), particularly on acquisitions (Harford, Mansi, and Maxwell, (2008).)

Electronic copy available at: https://ssrn.com/abstract=4085811

characteristics such as tenure, age, duality, pay-performance sensitivity (Delta), risk-taking

firms with single CEOs.5

Electronic copy available at: https://ssrn.com/abstract=4085811

Electronic copy available at: https://ssrn.com/abstract=4085811

unlikely that our results are driven by endogeneity.

financially constrained and unconstrained firms. While cash-holding policies in financially

Electronic copy available at: https://ssrn.com/abstract=4085811

with the predictions of our classical agency problem hypothesis.

Electronic copy available at: https://ssrn.com/abstract=4085811

(2007), and Pinkowitz and Williamson (2007), among others.

Electronic copy available at: https://ssrn.com/abstract=4085811

endogeneity. We explore the precautionary as well as the agency explanations in Section 4. We

cash holdings in section 5, and conclude in Section 6.

2. Data, sample selection, and summary statistics

1 if a CEO is unmarried and 0 otherwise.

Electronic copy available at: https://ssrn.com/abstract=4085811

2.2. Measures of Cash holding

[Please insert Table 1 here]

Electronic copy available at: https://ssrn.com/abstract=4085811

[Please insert Table 2 here]

3. Analysis and Results

3.1. Baseline Results

𝐶𝑎𝑠ℎ𝑖𝑗𝑡 = 𝛽0 + 𝛽1 𝑆𝑖𝑛𝑔𝑙𝑒𝑖𝑗𝑡 + 𝛾𝑖𝑗𝑡−1 + 𝐼𝑗 + 𝜏𝑡 + Ɛ𝑖𝑗𝑡 (1)

Electronic copy available at: https://ssrn.com/abstract=4085811

if firms temporarily increase capital expenditures due to a productivity shock, or if capital

Electronic copy available at: https://ssrn.com/abstract=4085811

precautionary motives, thus we control for cash flow volatility (Industry_sigma).

[Please insert Table 3 here]

Electronic copy available at: https://ssrn.com/abstract=4085811

held by firms with single CEOs.12

effect of CEO tenure on cash holdings.

Electronic copy available at: https://ssrn.com/abstract=4085811

3.2. Endogeneity issues

Electronic copy available at: https://ssrn.com/abstract=4085811

following sections using a Difference-in-Differences (DiD) estimate, an instrumental variables

3.2.1. Corporate cash holdings around CEO turnover: A DiD test

[Please insert Table 4 here]

coefficient (After*Treated). Our classical agency problem hypothesis predicts a positive

Electronic copy available at: https://ssrn.com/abstract=4085811

single CEO experience an increase in cash holdings.

single CEO could be succeeded by another single CEO.

3.2.2. Instrumental variables approach

To mitigate the concern of innate heterogeneity, we employ an instrumental variables approach.

Electronic copy available at: https://ssrn.com/abstract=4085811

the results of our instrumental variables regression tests.

[Please insert Table 5 here]

community property state are more likely to be single.

Electronic copy available at: https://ssrn.com/abstract=4085811